CN105929687B

CN105929687B - High level central facility optimization

Info

Publication number: CN105929687B
Application number: CN201610107795.5A
Authority: CN
Inventors: 迈克尔·J·文策尔; 罗伯特·D·特尼; 柯克·H·德雷斯; 马修·J·阿斯穆斯
Original assignee: Johnson Controls Technology Co
Current assignee: Johnson Controls Technology Co
Priority date: 2015-02-27
Filing date: 2016-02-26
Publication date: 2020-02-04
Anticipated expiration: 2036-02-26
Also published as: CN105929687A

Abstract

An optimization system for a central facility includes processing circuitry configured to receive load forecast data indicative of building energy usage and utility rate data indicative of prices of one or more resources consumed by equipment of the central facility to service the building energy usage. The optimization system includes a high-level optimization module configured to generate an objective function representing a total financial cost of operating the central facility over an optimization period as a function of the utility rate data and an amount of one or more resources consumed by the central facility device. The high-level optimization module is configured to optimize the objective function over the optimization period to determine an optimal distribution of the building energy load over the plurality of sets of the central facility devices subject to load equality constraints and capacity constraints on the central facility devices.

Description

High level central facility optimization

Cross reference to related patent applications

This application claims the benefit and priority of U.S. patent application No.14/634,609, filed on day 27/2/2015, which in turn claims the benefit and priority of U.S. provisional patent application No.61/987,361, filed on day 1/5/2014. Both U.S. patent application No.14/634,609 and U.S. provisional patent application No.61/987,361 are incorporated herein by reference in their entirety.

Technical Field

The present disclosure relates generally to the operation of a central facility for servicing a building thermal energy load. The present disclosure more particularly relates to systems and methods for distributing building thermal energy loads among a plurality of sub-facilities configured to service the building thermal energy loads.

Background

The central facility may include various types of equipment configured to service the thermal energy load of a building or campus (campus) (i.e., a system of multiple buildings). For example, the central facility may include a heater, a chiller, a heat recovery chiller, a cooling tower, or other type of device configured to provide heating or cooling to a building. The central facility may consume resources (e.g., electricity, water, natural gas, etc.) from the utility to heat or cool a working fluid (e.g., water, glycol, etc.), which is circulated to the building or stored for later use to provide heating or cooling to the building. The fluid conduits typically deliver heated or cooled fluid to an air handler located on the roof (rooop) of the building or to various floors or areas of the building. The air handler pushes the air through a heat exchanger (e.g., a heating coil or a cooling coil) through which the working fluid flows to provide heating or cooling to the air. The working fluid is then returned to the central facility to receive further heating or cooling, and the cycle continues.

High efficiency devices may help reduce the amount of energy consumed by a central facility; however, the effectiveness of such devices is very dependent on the control technique used to distribute the load among the sub-facilities. For example, when the energy price is high, it may be more cost effective to run a heat pump chiller rather than a conventional chiller and a water heater. Determining when and to what extent each of the plurality of sub-facilities should be used to minimize energy costs is difficult and challenging. This optimization is even more complex if power demand costs are taken into account.

Thermal energy storage may be used to store energy for later use. When combined with real-time electricity prices and demand charges, thermal energy storage provides a degree of flexibility that can be used to greatly reduce energy costs by shifting production to low cost times or other times when electrical loads are low so that new peak demand is not set. It is difficult and challenging to integrate thermal energy storage with a central facility having multiple sub-facilities and optimize the use of thermal energy storage in coordination with the multiple sub-facilities to minimize energy costs.

Disclosure of Invention

One embodiment of the present disclosure is an optimization system for a central facility configured to service energy usage loads for a building. The optimization system includes processing circuitry configured to receive load prediction data indicative of building energy loads for a plurality of timescales in an optimization period and utility rate data indicative of prices of one or more resources consumed by equipment of the central facility to provide service to the building energy loads at each of the plurality of timescales. The optimization system also includes a high-level optimization module configured to generate an objective function that represents a total financial cost of operating the central facility over the optimization period as a function of the utility rate data and the amount of one or more resources consumed by the central facility equipment at each of the plurality of timescales. The high-level optimization module is configured to optimize the objective function over the optimization period in accordance with load equality constraints and capacity constraints on the central facility devices to determine an optimal distribution of the building energy load over the plurality of sets of the central facility devices at each of the plurality of time steps.

In some embodiments, the high-level optimization module uses linear programming to generate and optimize the objective function. The objective function may include a cost vector having a cost variable representing a financial cost associated with each of the one or more resources consumed by the central facility device to service the building energy load at each of the plurality of timescales. The objective function may further include a decision matrix containing load variables representing energy load for each of a plurality of groups of the central facility devices at each of the plurality of timescales. The high-level optimization module may be configured to determine an optimal value for a load variable in the decision matrix.

In some embodiments, the central facility includes a plurality of sub-facilities. Each of the plurality of sets of the central facility device may correspond to one of the plurality of sub-facilities. In some embodiments, the plurality of sub-facilities includes at least one of a hot thermal energy storage sub-facility and a cold thermal energy storage sub-facility. The thermal energy storage sub-facility may be configured to store thermal energy generated in one of the plurality of time steps for use in another of the plurality of time steps.

In some embodiments, the high-level optimization module is configured to generate a sub-facility curve for each of the plurality of sub-facilities. Each sub-facility curve may indicate a relationship between resource consumption and load production of one of the plurality of sub-facilities. The high-level optimization module may formulate a sub-facility curve constraint using the sub-facility curve and may optimize the objective function in accordance with the sub-facility curve constraint. In some embodiments, generating the sub-facility curve comprises at least one of: converting the non-linear sub-facility curve to a linear sub-facility curve comprising one or more segmented linear segments; and converting the non-convex facility curve into a convex facility curve. In some embodiments, generating the sub-facility curve comprises: receiving an initial sub-facility curve based on manufacturer data for a group of devices corresponding to the sub-facility; and updating the initial sub-facility curve using experimental data from the central facility.

Another embodiment of the present disclosure is a cascade optimization system for a central facility configured to service energy usage loads for a building. The cascade optimization system includes a central facility controller configured to use dynamic programming to divide an optimization problem of the central facility into a high-level optimization and a low-level optimization, a high-level optimization module configured to perform the high-level optimization, and a low-level optimization module configured to perform the low-level optimization. The high-level optimization includes determining an optimal distribution of building energy loads over a plurality of sets of central facility equipment. The low-level optimization includes determining optimal operating conditions of individual devices within each of the plurality of groups of central facility equipment.

In some embodiments, the optimal distribution of the building energy load determined by the high-level optimization module optimizes financial costs of operating the central facility over an optimization period. The optimal operating conditions determined by the low-level optimization module may optimize the amount of energy consumed by each of the plurality of groups of the central facility devices to achieve the optimal distribution of the building energy usage load determined by the high-level optimization module.

In some embodiments, the low-level optimization module is configured to generate a sub-facility curve for each of a plurality of groups of central facility devices. Each sub-facility curve may indicate a relationship between resource consumption and load production for one of a plurality of sets of the central facility devices. The high-level optimization module may be configured to formulate a sub-facility curve constraint using the sub-facility curve and determine an optimal distribution of the building energy load in accordance with the sub-facility curve constraint.

Another embodiment of the present disclosure is a method for optimizing the cost of a central facility configured to service energy usage loads for a building. The method includes receiving, at processing circuitry of a central facility optimization system, load prediction data indicative of building energy loads for a plurality of timescales in an optimization period and utility rate data indicative of prices of one or more resources consumed by equipment of the central facility to provide service to the building energy loads at each of the plurality of timescales. The method also includes generating, by a high-level optimization module of the central facility optimization system, an objective function that represents a total financial cost of operating the central facility over the optimization period as a function of the utility rate data and an amount of one or more resources consumed by the central facility device at each of the plurality of timescales. The method also includes optimizing, by the high-level optimization module, the objective function over the optimization period in accordance with load equality constraints and capacity constraints on the central facility devices to determine an optimal distribution of the building energy load over the plurality of sets of the central facility devices at each of the plurality of time steps.

In some embodiments, the high-level optimization module uses linear programming to generate and optimize the objective function. The objective function may include a cost vector having a cost variable representing a financial cost associated with each of the one or more resources consumed by the central facility device to service the building energy load at each of the plurality of timescales. The objective function may further include a decision matrix containing load variables representing the energy usage load of each of a plurality of groups of the central facility devices at each of the plurality of timescales. The high-level optimization module may be configured to determine an optimal value for a load variable in the decision matrix.

In some embodiments, the method includes generating the load equation constraint and the capacity constraint using the building energy load and capacity limits for the central facility device. The load equation constraints may ensure that the optimal distribution satisfies the building energy load at each of the plurality of timescales. The capacity constraint may ensure that a plurality of sets of the central facility devices operate within the capacity limit at each of the plurality of time steps.

In some embodiments, the central facility includes a plurality of sub-facilities. Each of the plurality of sets of the central facility devices may correspond to one of the plurality of sub-facilities. In some embodiments, the method includes generating a sub-facility curve for each of the plurality of sub-facilities. Each sub-facility curve may indicate a relationship between resource consumption and load production of one of the plurality of sub-facilities. The method may also include formulating a sub-facility curve constraint using the sub-facility curve; and optimizing the objective function in accordance with the sub-facility curve constraints. In some embodiments, generating the sub-facility curve comprises at least one of: converting the non-linear sub-facility curve to a linear sub-facility curve comprising one or more segmented linear segments; and converting the non-convex facility curve into a convex facility curve.

In some embodiments, the method further comprises using dynamic programming to divide the optimization process into a high-level optimization and a low-level optimization. The high-level optimization may include determining an optimal distribution of the building energy load over the plurality of sets of the central facility equipment. In some embodiments, the optimal distribution of energy load for the building optimizes a financial cost of operating the central facility over the optimization period. The low-level optimization may include determining optimal operating conditions for individual devices within each of a plurality of groups of the central facility apparatus. In some embodiments, the optimal operating conditions optimize the amount of energy consumed by each of the plurality of groups of the central facility devices to achieve an optimal distribution of the building energy usage load.

Those skilled in the art will appreciate that the summary is illustrative only and is not intended to be in any way limiting. Other aspects, inventive features, and advantages of the devices and/or processes described herein, as defined solely by the claims, will become apparent in the detailed description set forth herein and illustrated in the accompanying drawings.

Drawings

FIG. 1 is a schematic diagram of a central facility having a plurality of sub-facilities including a heater sub-facility, a heat recovery chiller sub-facility, a hot thermal energy storage sub-facility, and a cold thermal energy storage sub-facility, according to an exemplary embodiment.

FIG. 2 is a block diagram illustrating a central facility system including a central facility controller that may be used to control the central facility of FIG. 1, according to an example embodiment.

FIG. 3 is a block diagram illustrating a portion of the central facility system of FIG. 2 in greater detail, showing a load/rate prediction module, a high-level optimization module, a low-level optimization module, a building automation system, and a central facility device, according to an example embodiment.

FIG. 4 is a block diagram illustrating the high-level optimization module of FIG. 3 in greater detail, according to an example embodiment.

5A-5B are sub-facility curves illustrating relationships between resource consumption and sub-facility load of a sub-facility, and which may be used by the high level optimization module of FIG. 4 to optimize performance of the central facility of FIG. 1, according to an example embodiment.

FIG. 6 is a non-convex and non-linear sub-facility curve that may be generated from experimental data or by combining plant curves of various devices of a central facility, according to an example embodiment.

FIG. 7 is a linearized sub-facility curve that may be generated from the sub-facility curve of FIG. 6 by converting a non-convex and non-linear sub-facility curve into a segmented linear segment (linear), according to an example embodiment.

FIG. 8 is a graph illustrating a set of sub-facility curves that may be generated by the high-level optimization module of FIG. 3 based on experimental data from the low-level optimization module for a plurality of different environmental conditions, according to an example embodiment.

FIG. 9 is a block diagram of a planning system (planning system) including the high-level optimization module of FIG. 3, according to an example embodiment.

FIG. 10 is a diagram illustrating operation of the planning system of FIG. 9 according to an exemplary embodiment.

FIG. 11 is a flowchart of a process for optimizing central facility costs that may be performed by the central facility controller of FIG. 2 or the planning system of FIG. 9, according to an example embodiment.

Detailed Description

Referring generally to the drawings, a system and method for optimizing a central facility according to an exemplary embodiment is shown. The central facility may include various types of equipment configured to service the thermal energy load of a building or campus (i.e., a system of multiple buildings). For example, the central facility may include heaters, chillers, heat recovery chillers, cooling towers, or other types of equipment configured to provide heating or cooling to a building or campus. The central facility devices may be divided into various groups configured to perform specific functions. Such a group of central facility devices is referred to herein as a sub-facility. For example, the central facility may include a heater sub-facility, a chiller sub-facility, a heat recovery chiller sub-facility, a cold thermal energy storage sub-facility, a hot thermal energy storage sub-facility, and the like. A sub-facility may consume resources (e.g., water, electricity, natural gas, etc.) from one or more utilities to service the energy usage load of a building or campus. Optimizing the central facility may include operating the various sub-facilities in a manner that results in the lowest financial cost of servicing the energy load for the building.

In some embodiments, the central facility optimization is a cascaded optimization process that includes a high-level optimization and a low-level optimization. High-level optimization may determine an optimal distribution of energy usage load among the sub-facilities. For example, the high-level optimization may determine a thermal energy load to be generated by each sub-facility at each time element (time element) in the optimization period. In some embodiments, the high-level optimization includes optimizing a high-level cost function that represents a financial cost of operating the sub-facility as a function of the resources consumed by the sub-facility at each time bin of the optimization period. The low-level optimization may use the optimal load distribution determined by the high-level optimization to determine the optimal operating state of the individual devices within each sub-facility. The optimal operating conditions may include, for example, on/off states and/or operating set points of the individual devices of each sub-facility. The low-level optimization may include optimizing a low-level cost function that represents the energy consumption of the sub-facility as a function of the on/off status and/or operating set points of the individual devices of the sub-facility.

The present disclosure focuses on high-level optimization and describes systems and methods for performing high-level optimization. The high-level optimization module may perform high-level optimization. In various embodiments, the high-level optimization module may be a component of a central facility controller configured for real-time control of a physical facility, or a component of a planning tool configured to optimize a simulation facility (e.g., for planning or design purposes).

In some embodiments, the high-level optimization module performs the high-level optimization using a linear programming framework. Advantageously, linear programming can efficiently handle complex optimization scenarios, and can optimize over longer optimization periods (e.g., days, weeks, years, etc.) with shorter timeframes (e.g., seconds, milliseconds, etc.). In other embodiments, the high-level optimization module may use any of a variety of other optimization frameworks (e.g., quadratic programming, linear fractional programming, non-linear programming, combinatorial algorithms, etc.).

The objective function defining the high-level optimization problem can be represented in a linear programming framework as:

obeying Ax ≦ b, Hx ═ g

Where c is a cost vector, x is a decision matrix, A and b are (respectively) matrices and vectors describing inequality constraints on variables in the decision matrix x, and H and g are (respectively) matrices and vectors describing equality constraints on variables in the decision matrix x. The variables in the decision matrix x may include the sub-facility load and/or resource consumption of the sub-facility allocated to the respective sub-facility at each time element in the optimization period. The high-level optimization module may define a cost vector c and optimization constraints (e.g., matrices a and H and vectors b and g), and solve the optimization problem to determine optimal sub-facility load values for the variables in the decision matrix x.

The high-level optimization module may receive as input a predicted or projected energy usage load for the building or campus at each time bin in an optimization period. The high-level optimization module may use the predicted or projected loads to formulate constraints on the high-level optimization problem (e.g., define matrices a and H and vectors b and g). The high-level optimization module may also receive utility rates (e.g., energy prices, water prices, demand charges, etc.) that define the cost of each resource consumed by the central facility to service the energy load. The utility rates may be time-varying rates (e.g., different rates defined at different times) and may include demand charges for various time periods. The high-level optimization module may use utility rates to define the cost vector c.

The high-level optimization module may receive or generate a sub-facility curve for each sub-facility. The sub-facility curve defines the resource consumption of the sub-facility as a function of the load generated by the sub-facility. The sub-facility curve may be generated by a low-level optimization module or generated by a high-level optimization module based on operational data points received from a low-level optimization module. The high-level optimization module may use the sub-facility curves to constrain the resource consumption of each sub-facility to some value along the corresponding sub-facility curve (e.g., based on the load produced by the sub-facility). For example, the high-level optimization module may use the sub-facility curves to define optimization constraints (e.g., matrices a and H and vectors b and g) for the high-level optimization problem.

In some embodiments, the high-level optimization module is configured to introduce demand charges into the high-level optimization process. Demand charges are additional charges imposed by some utility providers based on the maximum rate of resource consumption during the applicable demand charge period. For example, the demand electricity fee may be provided as a cost c per unit of electric power_demandAnd may be multiplied by the peak power usage max (P) over the demand charge period_elec,k) To determine demand charges. Conventional systems have not been able to introduce demand charges into a linear optimization framework due to the non-linear max () function used to calculate demand charges.

Advantageously, the high-level optimization module of the present disclosure may be configured to introduce demand charges into a linear optimization framework by modifying the decision matrix x, the cost vector c, and/or the a and b vectors describing inequality constraints. For example, the high-level optimization module may optimize the peak power consumption in the optimization period by adding a new decision variable x representing the peak power consumption_peakTo modify the decision matrix x. High-level optimization module available demand charge rate c_demandTo modify the cost vector c such that the demand charges a rate c_demandMultiplied by the peak power consumption x_peak. The high-level optimization module may generate and/or impose constraints to ensure peak power consumption x_peakGreater than or equal to the electrical demand per time step in the demand charging period and greater than or equal to its previous value within the demand charging period.

In some embodiments, the high-level optimization module is configured to introduce a load-change penalty into the high-level optimization process. The load change penalty may represent an increase in cost (e.g., equipment degradation, etc.) due to rapid changes in the load assigned to the sub-facility. The high-level optimization module may introduce a load change penalty by modifying the decision matrix x, the cost vector c, and/or the optimization constraints. For example, the high-level optimization module may modify the decision matrix x by adding a load change variable δ for each sub-facility. The load change variable may represent a sub-facility load change for each sub-facility from one time bin to the next. The high-level optimization module may modify the cost vector c to increase the cost associated with changing the sub-facility load. In some embodiments, the high-level optimization module adds constraints that constrain the load change variable δ to corresponding changes in the sub-facility load. These and other enhancements to the high-level optimization process can be introduced into a linear optimization framework, as described in more detail below.

Referring now to FIG. 1, a diagram of a central facility 10 is shown, according to an exemplary embodiment. Central facility 10 is illustrated as including a plurality of sub-facilities including a heater sub-facility 12, a heat recovery chiller sub-facility 14, a chiller sub-facility 16, a cooling tower sub-facility 18, a hot Thermal Energy Storage (TES) sub-facility 20, and a cold Thermal Energy Storage (TES) sub-facility 22. The sub-facilities 12-22 consume resources (e.g., water, natural gas, electricity, etc.) from a utility to service the thermal energy load (e.g., hot water, cold water, heating, cooling, etc.) of a building or campus. For example, the heater sub-facility 12 may be configured to heat water in a hot water circuit 24, the hot water circuit 24 circulating hot water between the central facility 10 and a building (not shown). The chiller sub-facility 16 may be configured to chill water in a cold water loop 26, the cold water loop 26 circulating cold water between the central facility 10 and the building. The heat recovery chiller sub-installation 14 may be configured to transfer heat from the cold water circuit 26 to the hot water circuit 24 to provide additional heating of the hot water and additional cooling of the cold water. The condensate water circuit 28 may absorb heat from the cold water in the chiller sub-installation 16 and reject the absorbed heat in the cooling tower sub-installation 18, or transfer the absorbed heat to the hot water circuit 24. The hot TES sub-facility 20 and cold TES sub-facility 22 store hot and cold thermal energy, respectively, for subsequent use.

The hot water loop 24 and the cold water loop 26 may deliver heated and/or cooled water to an air handler located on the roof of the building or to various floors or areas of the building. The air handler pushes the air through a heat exchanger (e.g., a heating coil or cooling coil) through which water flows to provide heating or cooling to the air. Heated or cooled air may be delivered to various areas of a building to service the thermal energy load of the building. The water is then returned to the central facility 10 to receive further heating or cooling in the subsystems 12-22.

Although the central facility 10 is illustrated and described as heating and cooling water for circulation to a building, it will be understood that any other type of working fluid (e.g., ethylene glycol, CO2, etc.) may be used instead of or in addition to water to service a thermal energy load. In other embodiments, the central facility 10 may provide heating and/or cooling directly to a building or campus without the need for an intermediate heat transfer fluid. The central facility 10 may be physically separate from the building that the sub-facilities 12-22 provide services, or may be physically integrated with (e.g., located within) the building.

Each sub-facility 12-22 may include various devices configured to facilitate sub-facility functionality. For example, the heater sub-facility 12 is illustrated as including a plurality of heating elements 30 (e.g., boilers, electric heaters, etc.) configured to add heat to the hot water in the hot water circuit 24. The heater sub-assembly 12 is also illustrated as including

several pumps

32 and 34 configured to circulate hot water in the hot water circuit 24 and control the flow rate of the hot water through the various heating elements 30. The heat recovery chiller sub-installation 14 is illustrated as including a plurality of heat recovery heat exchangers 36 (e.g., refrigeration circuits) configured to transfer heat from the cold water circuit 26 to the hot water circuit 24. The heat recovery chiller sub-installation 14 is also illustrated as including a number of

pumps

38 and 40 configured to circulate hot and/or cold water through the heat recovery heat exchangers 36 and to control the flow rate of the water through each heat recovery heat exchanger 36.

Chiller sub-installation 16 is illustrated as including a plurality of chillers 42 configured to remove heat from the cold water in cold water loop 26. The chiller sub-installation 16 is also illustrated as including a number of

pumps

44 and 46 configured to circulate chilled water in the chilled water loop 26 and to control the flow rate of the chilled water through each chiller 42. The cooling tower sub-facility 18 is illustrated as including a plurality of cooling towers 48 configured to remove heat from the condensate in the condensate circuit 28. The cooling tower sub-facility 18 is also illustrated as including a number of pumps 50 configured to circulate the condensed water in the condensed water circuit 28 and to control the flow rate of the condensed water through each cooling tower 48.

The hot TES sub-facility 20 is illustrated as including a hot TES water tank 52 configured to store hot water for later use. The hot TES sub-facility 20 may also include one or more pumps or valves configured to control the flow rate of hot water into and out of the hot TES water tank 52. The cold TES sub-facility 22 is illustrated as including a cold TES water tank 54 configured to store cold water for later use. The cold TES sub-facility 22 may also include one or more pumps or valves configured to control the flow rate of cold water into and out of the cold TES water tank 54. In some embodiments, one or more pumps (e.g., pumps 32, 34, 38, 40, 44, 46, and/or 50) in the central facility 10 or a pipeline in the central facility 10 includes an isolation valve associated therewith. In various embodiments, isolation valves may be integrated with or located upstream or downstream of the pumps to control fluid flow in the central facility 10. In other embodiments, more, fewer, or different types of devices may be included in the central facility 10.

Referring now to FIG. 2, a block diagram of a central facility system 100 is shown, according to an example embodiment. The system 100 is illustrated as including a central facility controller 102, a building automation system 108, and a plurality of sub-facilities 12-22. The sub-facilities 12-22 may be the same as previously described with reference to fig. 1. For example, the sub-facilities 12-22 are illustrated as including a heater sub-facility 12, a heat recovery chiller sub-facility 14, a chiller sub-facility 16, a hot TES sub-facility 20, and a cold TES sub-facility 22.

Each sub-facility 12-22 is illustrated as including a device 60 that may be controlled by the central facility controller 102 and/or the building automation system 108 to optimize performance of the central facility 10. The apparatus 60 may include other devices such as a heating device 30, a chiller 42, a heat recovery heat exchanger 36, a cooling tower 48, thermal

energy storage devices

52, 54, pumps 32, 44, 50,

valves

34, 38, 46, and/or sub-facilities 12-22. The various devices of the plant 60 may be switched on and off to regulate the thermal energy load serviced by each sub-facility 12-22. In some embodiments, the various devices of the plant 60 may be operated at variable capacities (e.g., operating chillers at 10% capacity or 60% capacity) according to the operational set points received from the central facility controller 102.

In some embodiments, one or more of the sub-facilities 12-22 includes a sub-facility level controller configured to control the equipment 60 of the corresponding sub-facility. For example, the central facility controller 102 may determine an on/off configuration and a global operational set point for the device 60. In response to the on/off configuration and the received global operational set point, the sub-facility controller may turn on or off various devices of the plant 60 and implement a particular operational set point (e.g., damper position, blade position, fan speed, pump speed, etc.) to reach or maintain the global operational set point.

A Building Automation System (BAS)108 may be configured to monitor conditions within a controlled building or building area. For example, the BAS108 can receive input from various sensors (e.g., temperature sensors, humidity sensors, airflow sensors, voltage sensors, etc.) distributed throughout the building and can report building conditions to the central facility controller 102. The building condition may include, for example, a temperature of the building or a zone of the building, a power consumption (e.g., electrical load) of the building, a state of one or more actuators configured to affect a controlled state within the building, or other types of information related to the controlled building. The BAS108 can operate the sub-facilities 12-22 to affect monitored conditions within the building and to service the thermal energy load of the building.

The BAS108 can receive control signals from the central facility controller 102 specifying on/off states and/or set points for the equipment 60. The BAS108 may control the equipment 60 (e.g., via actuators, power relays, etc.) according to control signals provided by the central facility controller 102. For example, the BAS108 can use closed loop control to operate the plant 60 to achieve a set point specified by the central facility controller 102. In various embodiments, the BAS108 may be integrated with the central facility controller 102 or may be part of a separate building management system. According to an exemplary embodimentBAS108 is sold by Johnson Controls, IncA brand of building management system.

The central facility controller 102 may use information received from the BAS108 to monitor the status of the controlled building. The central facility controller 102 may be configured to predict building thermal energy loads (e.g., heating loads, cooling loads, etc.) for a plurality of time steps in a prediction window (e.g., using weather forecasts from a weather service). The central facility controller 102 may generate on/off decisions and/or set points for the devices 60 to minimize the cost of energy consumed by the sub-facilities 12-22 to service the predicted heating and/or cooling loads for the duration of the prediction window. The central facility controller 102 may be configured to perform the process 1100 (fig. 11) as well as other processes described herein. According to an exemplary embodiment, the central facility controller 102 is integrated within a single computer (e.g., one server, one chassis, etc.). In various other exemplary embodiments, the central facility controller 102 may be distributed across multiple servers or computers (e.g., may exist in distributed locations). In another exemplary embodiment, the central facility controller 102 may be integrated with a smart building manager that manages multiple building systems and/or integrated with the BAS 108.

The central facility controller 102 is illustrated as including a communication interface 104 and a processing circuit 106. The communication interface 104 may include a wired or wireless interface (e.g., receptacle, antenna, transmitter, receiver, transceiver, wire connection, etc.) for data communication with various systems, devices, or networks. For example, the communication interface 104 may include an ethernet card and port for sending and receiving data via an ethernet-based communication network, and/or a WiFi transceiver for communicating via a wireless communication network. The communication interface 104 may be configured to communicate via a local or wide area network (e.g., the internet, a building WAN, etc.) and may use various communication protocols (e.g., BACnet, IP, LON, etc.).

The communication interface 104 may be a network interface configured to facilitate electronic data communication between the central facility controller 102 and various external systems or devices (e.g., the BAS108, the sub-facilities 12-22, etc.). For example, the central facility controller 102 may receive information from the BAS108 indicating one or more measured conditions of the controlled building (e.g., temperature, humidity, electrical load, etc.) and one or more conditions of the sub-facilities 12-22 (e.g., equipment conditions, power consumption, equipment availability, etc.). The communication interface 104 may receive input from the BAS108 and/or the sub-facilities 12-22 and may provide operating parameters (e.g., on/off decisions, set points, etc.) to the sub-facilities 12-22 via the BAS 108. The operating parameters may cause the sub-facilities 12-22 to activate, disable, or adjust set points for various devices of the equipment 60.

Still referring to fig. 2, the processing circuit 106 is illustrated as including a processor 110 and a memory 112. Processor 110 may be a general or special purpose processor, an Application Specific Integrated Circuit (ASIC), one or more Field Programmable Gate Arrays (FPGAs), a set of processing components, or other suitable processing components. The processor 110 may be configured to execute computer code or instructions stored in the memory 112 or received from other computer readable media (e.g., CDROM, network storage, remote server, etc.).

The memory 112 may include one or more devices (e.g., memory units, memory devices, storage devices, etc.) for storing data and/or computer code for performing and/or facilitating the various processes described in this disclosure. Memory 112 may include Random Access Memory (RAM), Read Only Memory (ROM), hard drive storage, temporary storage, non-volatile memory, flash memory, optical storage, or any other suitable memory for storing software objects and/or computer instructions. Memory 112 may include database components, object code components, script components, or any other type of information structure for supporting the various activities and information structures described in this disclosure. Memory 112 may be communicatively connected to processor 110 via processing circuitry 106 and may include computer code for executing (e.g., by processor 106) one or more processes described herein.

Still referring to FIG. 2, the memory 112 is illustrated as including a building status monitor 134. The central facility controller 102 may receive data via the building status monitor 134 regarding the entire building or building space to be heated or cooled by the central facility 10. In an exemplary embodiment, the building status monitor 134 may include a graphical user interface component configured to provide a graphical user interface to a user for selecting building requirements (e.g., overall temperature parameters, schedule for building selection, selection of different temperature levels for different building zones, etc.).

The central facility controller 102 may determine on/off configurations and operating set points to meet the building requirements received from the building status monitor 134. In some embodiments, the building status monitor 134 receives, collects, stores, and/or transmits cooling load requirements, building temperature set points, household data, weather data, energy usage data, calendar data, and other building parameters. In some embodiments, the building status monitor 134 stores data regarding energy costs, such as pricing information (energy charges, demand charges, etc.) available from the utilities 126.

Still referring to fig. 2, memory 112 is illustrated as including a load/rate prediction module 122. The load/rate prediction module 122 may be configured to predict the thermal energy load of a building or campus at each time step k (e.g., k-1 … n) of an optimization periodLoad/rate prediction module 122 is illustrated as receiving weather forecasts from weather service 124. In some embodiments, load/rate prediction module 122 predicts thermal energy load as a function of weather forecast

In some embodiments, the load/rate prediction module 122 predicts the load using feedback from the BAS108

The feedback from the BAS108 can include eachTypes of sensor inputs (e.g., temperature, flow, humidity, enthalpy, etc.) or other data related to the controlled building (e.g., inputs from HVAC systems, lighting control systems, security systems, water systems, etc.).

In some embodiments, the load/rate prediction module 122 receives measured electrical loads and/or previously measured load data from the BAS108 (e.g., via the building status monitor 134). Load/rate prediction module 122 may be a function of a given weather forecast

Type of date (day), time of day (t) and previously measured load data (Y)_k-1) Predicting load

Such a relationship can be represented in the following equation:

in some embodiments, load/rate prediction module 122 predicts load using a deterministic plus stochastic model (deterministic plustochastic model) trained from historical load data

Load/rate prediction module 122 may predict the load using any of a variety of prediction methods

(e.g., linear regression is used for the deterministic portion and AR model is used for the stochastic portion). Load/rate prediction module 122 may predict one or more different types of loads for a building or campus. For example, load/rate prediction module 122 may predict hot water load within each time step k within a prediction window

And cold water load

Load/rate prediction module 122 is illustrated as receiving utility rates from utility 126. The utility rate may indicate a cost or price per unit of resource (e.g., electricity, natural gas, water, etc.) provided by utility 126 at each timescale k within the prediction window. In some embodiments, the utility rate is a time-varying rate. For example, the price of electricity may be higher at certain times of the day or on certain days of the week (e.g., during high demand periods) and lower at other times of the day or on other days of the week (e.g., during low demand periods). The utility rates may define various time periods and a cost per unit of resource during each time period. The utility rate may be an actual rate received from utility 126 or a predicted utility rate estimated by load/rate prediction module 122.

In some embodiments, the utility rates include demand charges for one or more resources provided by the utility 126. The demand charge may define an individual cost imposed by the utility 126 based on a maximum usage (e.g., maximum energy consumption) of a particular resource during the demand charge period. The utility rate may define various demand charge periods and one or more demand charges associated with each demand charge period. In some instances, the demand charge periods may partially or completely overlap with each other and/or with the prediction window. Advantageously, the optimization module 128 may be configured to account for demand charges in the high-level optimization process performed by the high-level optimization module 130. Utility 126 may be defined by: time-varying (e.g., hourly) prices, maximum service levels (e.g., maximum consumption rates allowed by physical infrastructure or contracts), and demand charges in the case of electricity or charges for peak consumption rates over a period of time.

Load/rate prediction module 122 may predict the load

And utility rates are stored in memory 112And/or will predict load

And utility rates are provided to the optimization module 128. The optimization module 128 may use the predicted load

And utility rates to determine the optimal load distribution for the sub-facilities 12-22 and to generate on/off decisions and set points for the equipment 60.

Still referring to fig. 2, the memory 112 is illustrated as including an optimization module 128. The optimization module 128 may perform a cascaded optimization process to optimize the performance of the central facility 10. For example, the optimization module 128 is illustrated as including a high-level optimization module 130 and a low-level optimization module 132. The high-level optimization module 130 may control outer (e.g., sub-facility level) loops of the cascade optimization. The high-level optimization module 130 may determine an optimal thermal energy load distribution among the sub-facilities 12-22 for each time step in the prediction window to optimize (e.g., minimize) the energy costs consumed by the sub-facilities 12-22. The low-level optimization module 132 may control inner (e.g., device-level) loops of the cascade optimization. The low-level optimization module 132 may determine how best to run each sub-facility at the load set point determined by the high-level optimization module 130. For example, the low-level optimization module 132 may determine on/off states and/or operational set points of various devices of the plant 60 to optimize (e.g., minimize) the energy consumption of each sub-facility while meeting the thermal load set points of the sub-facilities. The cascade optimization process will be described in more detail with reference to fig. 3.

Still referring to FIG. 2, the memory 112 is illustrated as including a sub-facility control module 138. The sub-facility control module 138 may store historical data regarding past operating conditions, past operating set points, and instructions for calculating and/or implementing control parameters for the sub-facilities 12-22. The sub-facility control module 138 may also receive, store, and/or transmit data regarding the status of various devices of the equipment 60, such as operating efficiency, equipment degradation, date since last service, life parameters, status levels, or other device-specific data. The sub-facility control module 138 may receive data from the sub-facilities 12-22 and/or the BAS108 via the communication interface 104. The sub-facility control module 138 may also receive and store on/off status and operational set points from the low-level optimization module 132.

Data and processing results from the optimization module 128, the sub-facility control modules 138, or other modules of the central facility controller 102 may be accessed by (or pushed to) the monitoring and reporting application 136. The monitoring and reporting application 136 may be configured to generate a real-time "system health" dashboard that may be viewed and navigated (navigator) by a user (e.g., a central facility engineer). For example, the monitoring and reporting applications 136 may include a web-based (web) monitoring application having a number of Graphical User Interface (GUI) elements (e.g., widgets (widgets), dashboard controls, windows, etc.) for displaying Key Performance Indicators (KPIs) or other information to a user of the GUI. Further, GUI elements may summarize relative energy usage and intensity between various central facilities in different buildings (real or modeled), different parks, and so forth. Other GUI elements or reports may be generated and displayed based on the available data, which allows a user to assess performance across one or more central facilities from one screen. The user interface or report (or underlying data engine) may be configured to aggregate and categorize operating conditions by building, building type, equipment type, and the like. The GUI elements may include charts or histograms that allow a user to visually analyze the operating parameters and power consumption of the devices of the central facility.

Still referring to FIG. 2, the central facility controller 102 may include one or more GUI servers, web services 114, or GUI engines 116 to support monitoring and reporting applications 136. In various embodiments, the application 136, web service 114, and GUI engine 116 may be provided as separate components outside of the central facility controller 102 (e.g., as part of an intelligent building manager). The central facility controller 102 may be configured to maintain a detailed historical database (e.g., a relational database, an XML database, etc.) of relevant data and include computer code modules that continuously, frequently, or occasionally query, aggregate, transform, search, or otherwise process the data maintained in the detailed database. The central facility controller 102 may be configured to provide the results of any such processing to other databases, tables, XML files, or other data structures for further querying, computation, or access by, for example, an external monitoring and reporting application.

The central facility controller 102 is illustrated as including a configuration tool 118. The configuration tool 118 may allow a user (e.g., via a graphical user interface, via prompting for a "wizard," etc.) to define how the central facility controller 102 should react to changing conditions in the central facility subsystem. In an exemplary embodiment, the configuration tool 118 allows a user to establish and store situation response scenarios that can span multiple central facility devices, multiple building systems, and multiple enterprise control applications (e.g., work order management system applications, enterprise resource planning applications, etc.). For example, the configuration tool 118 may provide a user with the ability to combine data (e.g., from subsystems, from event histories) with various conditional logic. In different exemplary embodiments, the conditional logic may range from simple logical operators between conditions (e.g., AND, OR, exclusive OR (XOR), etc.) to pseudo code constructions OR complex programming language functions (allowing more complex interactions, conditional statements, loops, etc.). The configuration tool 118 may present a user interface for building such conditional logic. The user interface may allow a user to graphically define policies and responses. In some embodiments, the user interface may allow a user to select a prestored or preconfigured policy and adjust it or enable it for use with their system.

Referring now to FIG. 3, a block diagram illustrating a portion of the central facility system 100 in greater detail is shown, according to an exemplary embodiment. FIG. 3 illustrates a cascaded optimization process performed by the optimization module 128 to optimize the performance of the central facility 10. In the cascade optimization process, the high-level optimization module 130 performs a sub-facility level optimization that determines an optimal allocation of thermal energy loads among the sub-facilities 12-22 for each time step in the prediction window to minimize the cost of the energy consumed by the sub-facilities 12-22. The low-level optimization module 132 performs a plant-level optimization that determines how best to run each sub-facility at the sub-facility load set-points determined by the high-level optimization module 130. For example, the low-level optimization module 132 may determine on/off states and/or operational set points of the various devices of the plant 60 to optimize energy consumption of each sub-facility while meeting the sub-facility's thermal energy load set points.

One advantage of the cascaded optimization process performed by the optimization module 128 is the optimal use of computation time. For example, the sub-facility level optimization performed by the high level optimization module 130 may use a relatively long time horizon due to the operation of the thermal energy storage. However, plant-level optimization performed by the low-level optimization module 132 may use a much shorter timeframe or no timeframe at all because low-level system dynamics are relatively fast (compared to thermal energy storage dynamics) and low-level control of the plant 60 may be handled by the BAS 108. Such optimized use of computation time allows the optimization module 128 to perform central facility optimization in a short amount of time, which allows for real-time predictive control. For example, the short computation time enables the optimization module 128 to be implemented in a real-time planning tool with interactive feedback.

Another advantage of the cascading optimization performed by the optimization module 128 is that the central facility optimization problem can be divided into two cascading sub-problems. The cascading configuration provides an abstraction layer that allows the high-level optimization module 130 to distribute thermal energy loads among the sub-facilities 12-22 without requiring the high-level optimization module 130 to know or use any details about the particular equipment configuration within each sub-facility. The interconnections between the devices 60 within each sub-facility may be hidden from the high-level optimization module 130 and handled by the low-level optimization module 132. For purposes of sub-facility level optimization performed by the high level optimization module 130, each sub-facility may be defined entirely by one or more sub-facility curves 140.

Still referring to fig. 3, the low-level optimization module 132 may generate and provide the sub-facility curve 140 to the high-level optimization module 130. The sub-facility curve 140 may indicate a rate of utility usage (e.g., electricity usage measured in kW, water usage measured in L/s, etc.) for each sub-facility 12-22 as a function of sub-facility load. Exemplary sub-facility curves are shown and described in more detail with reference to fig. 5A-8. In some embodiments, the low-level optimization module 132 generates the sub-facility curves 140 based on the plant models 120 (e.g., by combining the plant models 120 of the various appliances into a lumped curve for the sub-facility). The low-level optimization module 132 may generate the sub-facility curve 140 by running a low-level optimization process for several different load and weather conditions to generate a plurality of data points. The low-level optimization module 132 may fit a curve to these data points to generate the sub-facility curve 140. In other embodiments, the low-level optimization module 132 provides the data points to the high-level optimization module 132, and the high-level optimization module 132 generates the sub-facility curve using the data points.

High-level optimization module 130 may receive load and rate predictions from load/rate prediction module 122 and sub-facility curves 140 from low-level optimization module 132. The load prediction may be based on weather forecasts from the weather service 124 and/or information from the building automation system 108 (e.g., the current electrical load of the building, measurements from the building, previous load history, set point trajectories, etc.). The utility rate prediction may be based on a utility rate received from utility 126 and/or a utility price from another data source. The high-level optimization module 130 may determine an optimal load distribution (e.g., a sub-facility load for each sub-facility) for each time-step sub-facility 12-22 in the prediction window and provide the sub-facility loads as setpoints to the low-level optimization module 132. In some embodiments, the high-level optimization module 130 determines the sub-facility loads by minimizing the total operating cost of the central facility 10 over a prediction window. In other words, given the predicted load and utility rate information from load/rate prediction module 122, high-level optimization module 130 may distribute the predicted load among sub-facilities 12-22 over an optimization period to minimize operating costs.

In some examples, optimal load distribution may include storing thermal energy using TES sub-facilities 20 and/or 22 during a first time period for use during a later time period. Thermal energy storage may advantageously allow thermal energy to be generated and stored during a first period of relatively low energy prices, and subsequently retrieved and used during a second period of relatively high energy prices. The high-level optimization may differ from the low-level optimization in that the high-level optimization has a longer time constant due to the thermal energy storage provided by the TES sub-facilities 20-22. The high level optimization can be described by the following equation:

wherein the content of the first and second substances,

contains an optimal high-level decision (e.g., optimal load for each sub-facility 12-22) for the entire optimization period, and J_HLIs a high level cost function.

To find optimal high-level decisions

The high-level optimization module 132 may minimize the high-level cost function J_HL. High level cost function J_HLMay be the sum of the economic cost of each utility consumed by each sub-facility 12-22 for the duration of the optimization period. In some embodiments, the high-level cost function JHL can be described using the following equation:

wherein n is_hIs the number of time steps k in the optimization period, n_sIs the number of sub-facilities, t_sIs the duration of the time step, c_jkIs the economic cost of utility j at time step k of the optimization period, and u_jikIs the usage rate of the utility j used by the sub-facility i at time step k.

In some embodiments, the cost function J_HLIncluding additional demand charges such as:

wherein, w_dIs a weighted term, c_demandIs the demand cost and the max () term selects the peak electricity usage during the applicable demand charge period. Thus, the high-level cost function J_HLCan be described by the following equation:

decision vector θ_HLMay be subject to several constraints. For example, the constraints may require: the sub-facilities are operated without exceeding their total capacity, the thermal storage is not filled or discharged too quickly or under-flow/over-flow of the water tanks, and the thermal load of the building or park is met. These constraints result in both equality and inequality constraints on the high-level optimization problem, as described in more detail with reference to FIG. 4.

Still referring to FIG. 3, the low-level optimization module 132 may use the sub-facility loads determined by the high-level optimization module 130 to determine an optimal low-level decision for the plant 60

(e.g., binary on/off decision, flow set point, temperature set point, etc.). A low-level optimization process may be performed for each sub-facility 12-22. The low-level optimization module 132 may be responsible for determining which devices of each sub-facility to use and/or the operational set points of those devices that will achieve the sub-facility load set points while minimizing energy consumption. The low-level optimization can be described using the following equation:

wherein the content of the first and second substances,

including optimal low-level decisions, and J_LLIs a low-level cost function.

To find optimal low-level decisionsThe low-level optimization module 132 may minimize a low-level cost function J_LL. Low level cost function J_LLMay represent the total energy consumption of all devices 60 in the applicable sub-facility. Low level cost function J_LLThe following equation can be used to describe:

where N is the number of devices of the equipment 60 in the sub-facility, ts is the duration of the time step, b_jIs a binary on/off decision (e.g., 0 off, 1 on), and u_jIs a function of the set point theta_LLThe energy source used by device j. Each device may have a continuous variable that may be varied to determine the lowest possible energy consumption for the total input conditions.

The low-level optimization module 132 may minimize a low-level cost function J_LLLow level cost function J_LLSubject to inequality constraints based on the capabilities of the device 60 and equality constraints based on energy and mass balances. In some embodiments, optimal low-level decisions

Subject to a switching constraint that defines a short time range for maintaining the device in an on or off state after a binary on/off switch. The switching constraints may prevent the device from cycling quickly between on and off. In some embodiments, the low-level optimization module 132 performs device-level optimization without regard to system dynamics. The optimization process may be slow enough to safely assume that the plant control has reached its steady state. Thus, the low-level optimization module 132 may determine optimal low-level decisions at one time, rather than over a long time horizon

The low-level optimization module 132 may determine optimal operating conditions (e.g., on or off) of a plurality of devices of the apparatus 60. According to an exemplary embodiment, the on/off combinations may be determined using binary optimization and quadratic compensation. Binary optimization may minimize a cost function representing the power consumption of devices in the applicable sub-facility. In some embodiments, non-exhaustive (i.e., not all potential device combinations are considered) binary optimization is used. Quadratic compensation may be used when considering devices whose power consumption is quadratic (rather than linear). The low-level optimization module 132 may also use non-linear optimization to determine the optimal operating set point for the plant. The non-linear optimization may identify an operational set point that further minimizes the low-level cost function JLL. The low-level optimization module 132 may provide on/off decisions and setpoints to the building automation system 108 for use in controlling the central facility device 60.

In some embodiments, the Low-Level Optimization performed by the Low-Level Optimization module 132 is the same or similar to the Low-Level Optimization process described in U.S. patent application No. xx/yyyyyyyy (attorney docket No.081445-0652), filed on the same day as the present application and entitled "Low Level Central Plant Optimization". The entire disclosure of U.S. patent application No. xx/yyyyyy is incorporated herein by reference.

Referring now to FIG. 4, a block diagram illustrating the high-level optimization module 130 in greater detail is shown, according to an exemplary embodiment. High-level optimization module 130 may receive load and rate predictions from load/rate prediction module 122 and sub-facility curves from low-level optimization module 132. The high-level optimization module 130 may determine an optimal sub-facility load for each sub-facility 12-22 as a function of the load and rate forecast and the sub-facility curves. In some embodiments, the optimal sub-facility load minimizes the economic cost of operating the sub-facilities 12-22 to meet the predicted load of the building or campus. The high-level optimization module 130 may output the optimal sub-facility load to the low-level optimization module 132.

The high-level optimization module 130 is illustrated as including an optimization framework module 142. The optimization framework module 142 may be configured to select and/or establish an optimization framework for use in calculating the optimal sub-facility load. In some embodiments, the optimization framework module 142 uses linear programming as the optimization framework. The linear programming problem has the following form:

obeying Ax ≦ b, Hx ═ g

Where c is a cost vector, x is a decision matrix, A and b are (respectively) matrices and vectors describing inequality constraints on the optimization problem, and H and g are (respectively) matrices and vectors describing equality constraints on the optimization problem.

The following paragraphs describe an exemplary linear optimization framework that may be used by the high-level optimization module 130 to calculate the optimal sub-facility load. Advantageously, the linear programming framework described herein allows the high-level optimization module 130 to determine sub-facility load allocations for long optimization periods, including load change penalties, demand charges, and sub-facility performance curves, within a very short timeframe. However, the linear optimization framework is only one example of an optimization framework that may be used by the high-level optimization module 130 and should not be viewed as limiting. It should be appreciated that in other embodiments, the high-level optimization module 130 may calculate the optimal sub-facility load using any of a variety of other optimization frameworks and/or optimization techniques (e.g., quadratic programming, linear fractional programming, non-linear programming, combinatorial algorithms, etc.).

Still referring to fig. 4, the high-level optimization module 130 is illustrated as including a linear programming module 144. The linear programming module 144 may be configured to formulate and solve a linear optimization problem to calculate the optimal sub-facility load. For example, the linear programming module 144 may determine and set the values of a cost vector c, a matrix and b vectors describing inequality constraints, and an H matrix and g vector describing equality constraints. The linear programming module 144 may determine a minimization cost function c^Tx, the optimal decision matrix x. The optimal decision matrix x may correspond to optimal decisions (for each time step k within the optimization period)

It minimizes the high-level cost function JHL as described with reference to fig. 3.

For a central facility 10 that includes chillers, heat recovery chillers, hot water generators, and thermal energy storage, the facility assets (plant assets) between which the load is to be distributed may include a chiller sub-facility 16, a heat recovery chiller sub-facility 14, a heater sub-facility 12, a hot thermal energy storage sub-facility 20, and a cold thermal energy storage sub-facility 22. The load across each sub-facility 12-22 may be a decision variable in the decision matrix x, which is determined by the high level optimization for each time step k within the optimization period. For example, the linear programming module 144 may formulate the decision matrix x as:

wherein the content of the first and second substances,

and

are n-dimensional vectors that represent the thermal energy load assigned to the chiller sub-facility 16, the heat recovery chiller sub-facility 14, the heater sub-facility 12, the hot TES sub-facility 20, and the cold TES sub-facility 22, respectively, at each of the n time steps within the optimization period.

The linear programming module 144 may formulate a linear programming for simple cases where only energy costs and equipment constraints are considered. The simplified linear plan may then be modified by an inequality constraint module 146, an equality constraint module 148, an under-met load module 150, a ground loop module 152, a heat exchanger module 154, a demand charging module 156, a load change penalty module 158, a tank forced fill module 160, and/or a sub-facility curve module 170 to provide additional enhancements, which will be described in more detail below.

In some embodiments, the linear programming module 144 formulates a simplified linear programming with the assumption that each sub-facility has a particular cost per unit load. For example, the linear programming module 144 may assume that each sub-facility has a constant coefficient of performance (COP) or efficiency for any given time step k. COP may change over time and may have different values for different time steps; however, in the simplest case, the COP of each sub-facility does not change as a function of load. With this assumption, the linear programming module 144 may formulate the cost function c as:

wherein, t_sIs the duration of the time step, n_uIs the total number of resources (e.g., electricity, gas, water, etc.) consumed by the sub-facility, c_jIs the cost per unit of the j-th resource, and u_j,Chiller，u_j,hrChillerAnd u_j,HeaterThe usage of the jth resource by the chiller, heat recovery chiller and

heater sub-facilities

16, 14, 12, respectively, for each of the h time steps within the optimization time period. The first three elements in the formula

A vector representing h sums (i.e., summing all resource usage), one for each time step within the optimization period. The last two elements 0 in this formula_hIs zero to indicate that there is no cost to charge and discharge the thermal energy storage tank (pumping power is ignored).

In some embodiments, the linear programming module 144 uses load and rate predictions to formulate a linear program. For example, the linear programming module 144 may determine u using load prediction_j,Chiller、u_j,hrChillerAnd u_j,HeaterAnd may use rate prediction to determine n_uC of each of the species resources_jThe value of (c). In some embodiments, the linear programming module 144 uses the sub-facility curve to define c as a function of resource usage_j. The linear programming module 144 may use the data from the inequality constraint module 146, the equality constraint module 148,The inputs to the under-met load module 150, the ground loop module 152, the heat exchanger module 154, the demand charging module 156, the load change penalty module 158, the tank fill module 160, and/or the sub-facility curve module 170 to determine and set the values of the various matrices and vectors in the linear program. The

module

146 and 170 may modify the cost vector c, a matrix, b vector, H matrix, and/or g vector to provide additional enhancements and/or functionality to the linear programming. The inputs provided by the

modules

146 and 170 are described in more detail below.

The linear programming module 144 may use any of a variety of linear optimization techniques to solve the linear optimization problem. For example, the linear programming module 144 may use a basis exchange (basis exchange) algorithm (e.g., simplex (simplex) algorithm, crossbar (cross) algorithm, etc.), an interior point (inter point) algorithm (e.g., ellipsoid (ellipsoid) algorithm, projection (projection) algorithm, path-following (path-following) algorithm, etc.), overlay and padding (covering and packing) algorithm, integer programming (integer programming) algorithm (e.g., cutting-place) algorithm, branch-and-bound (branch and bound) algorithm, branch-and-cut (branch and cut) algorithm, branch-pricing (branch and price) algorithm, etc.), or any other type of linear optimization algorithm or technique to solve for linear programming subject to optimization constraints. For embodiments using non-linear optimization, the linear programming module 144 may solve the non-linear optimization problem using any of a variety of non-linear optimization techniques.

Still referring to FIG. 4, the high-level optimization module 130 is illustrated as including an inequality constraint module 146. The inequality constraint module 146 may formulate or define one or more inequality constraints on the optimization problem solved by the linear programming module 144. In some examples, the inequality constraint module 146 defines a decision variable for each time step k within the optimization period corresponding to the load on the chiller, heat recovery chiller and

heater sub-facilities

16, 14, 12, respectively

And

the inequality constraint of (1). For example, each sub-facility 12-16 may have two capacity constraints given by the following formula:

range of

Range of

Wherein the content of the first and second substances,

is the load on the ith sub-facility during time step k, and

is the maximum capacity of the ith sub-facility. The first capacity constraint requires the load on each sub-facility 12-16 at each time step k within the optimization time period

Less than or equal to the maximum capacity of the sub-facility

The second capacity constraint requires the load on each sub-facility 12-16 for each time step k within the optimization periodGreater than or equal to zero.

The inequality constraints for the chiller sub-facilities 16 can be placed in the form of Ax ≦ b by defining the A matrix and the b vector as follows:

wherein [ I ]_h]Representing an h × h identity matrix or an all-one vector of h × 1, [0 ]_h]Or represent an h x h zero matrix orRepresents an all-zero vector of hx 1, and

is the maximum capacity of the chiller sub-installation 16. Similar inequality constraints for the heat recovery chiller sub-facility 14 and the heater sub-facility 12 can be placed in the form of Ax ≦ b by defining the A matrix and the b vector as follows:

wherein the content of the first and second substances,

is the maximum capacity of the heat recovery chiller sub-installation 14, andis the maximum capacity of the heater sub-facility 12.

Inequality constraint module 146 may formulate or define a decision variable corresponding to the load on hot and

cold TES sub-facilities

20, 22 for each time step k within the optimization period

And

the inequality constraint of (1). For example, each sub-facility 20-22 may have two capacity constraints given by the following equation:

range of

Range of

Wherein the content of the first and second substances,is the rate at which the ith TES sub-facility is being discharged at time step k,

is the maximum emission rate of the ith sub-facility, and

is the maximum fill rate of the ith sub-facility.The positive charge value of (A) indicates that the TES sub-facility is discharging, and

indicates that the sub-facility is filling. The first capacity constraint requires the emission rate of each sub-facility 20-22 for each time step k within the optimization period

Less than or equal to the maximum emission rate of the sub-facilityThe second capacity constraint requires a negative emission rate for each sub-facility 20-22 for each time step k within the optimization period

(i.e., fill rate) less than or equal to the maximum fill rate of the sub-facility

The inequality constraints for the hot TES sub-facility 20 can be placed in the form of Ax ≦ b by defining the A matrix and the b vector as follows:

wherein the content of the first and second substances,

is the maximum discharge rate of the hot TES sub-facility 20, and

is the maximum charge rate of the hot TES sub-facility 20. The analogous inequality constraint for a cold TES sub-facility 22 can be placed in the form of Ax ≦ b by defining the A matrix and b vector as follows:

wherein the content of the first and second substances,

is the maximum discharge rate of the cold TES sub-facility 22, andis the maximum charge rate of the cold TES sub-facility 202.

The inequality constraint module 146 may aggregate power usage P for all sub-facilities and buildings/parks_elec,campusAn electrical demand constraint is enforced. The inequality constraint module 146 may require that the total electrical demand be less than or equal to the maximum electrical demand P by defining the A matrix and the b vector as follows_elec,max：

A＝[u_{elec，Chiller}[I_h]，u_{elec，hrChiller}[I_h]，u_{elec，Heater}[I_h]，0_n，0_n]，b＝P_elec，max[I_h]-P_{elec，campus，k}

Wherein u is_elec,Chiller、u_{elec,hrChiller}And u_elec,HeaterThe electricity usage values, P, for the chiller, heat recovery chiller and

heater sub-installations

16, 14 and 12, respectively_{elec,campus,k}Is the electricity usage of the building/campus at time k,and P is_elec,maxIs the maximum total power usage for the central facility 10 and the building/campus.

Inequality constraint module 146 may enforce tank capacity constraints on hot TES sub-facility 20 and cold TES sub-facility 22. Tank capacity constraints may require that each TES tank never fill above its maximum capacity or vent below zero. These physical requirements result in a series of constraints to ensure an initial tank level Q for each TES tank at the beginning of the optimization period₀Plus that all fills during steps 1 through k within the optimization period is less than or equal to the maximum capacity Q of the TES tank_max. Similar constraints may be implemented to ensure an initial tank level Q for each TES tank at the beginning of an optimization period₀All emissions during the subtraction of steps 1 to k in the optimization period are greater than or equal to zero.

The tank capacity constraint for the hot TES sub-facility 20 can be placed in the form of Ax ≦ b by defining the A matrix and the b vector as follows:

wherein Q is_0,HotIs the initial charge level, Q, of the hot TES sub-facility 20 at the beginning of the optimization period_max,HotIs the maximum charge level, Δ, of the hot TES sub-facility 20_hIs the lower triangular matrix of 1, and t_sIs the duration of the time step. Draining the tank is represented in the top row of the a matrix as a positive flow from the tank, while filling the tank is represented in the bottom row of the a matrix as a negative flow from the tank. The analogous inequality constraint for a cold TES sub-facility 22 can be placed in the form of Ax ≦ b by defining the A matrix and b vector as follows:

wherein Q is_0,ColdIs the initial charge level of the cold TES sub-facility 22 at the beginning of the optimization period, and Q_max,ColdBeing cold TES sub-facilities 22A maximum fill level.

Still referring to FIG. 4, the high-level optimization module 130 is illustrated as including an equality constraint module 148. The equality constraint module 148 may formulate or define one or more equality constraints on the optimization problem solved by the linear programming module 144. The equality constraints ensure that the predicted thermal energy load of the building or campus is met at each time step k in the optimization period. The equality constraint module 148 may formulate equality constraints for each type of thermal energy load (e.g., hot water, cold water, etc.) to ensure that the load is satisfied. The equality constraints can be given by the following equation:

range of

Wherein the content of the first and second substances,

is the thermal energy load of type p (e.g., hot water, cold water, etc.) on the ith sub-facility during time step k, n_sIs the total number of sub-facilities that can serve the thermal energy load p, and

is the predicted thermal energy load of type p that must be met at time step k. The predicted thermal energy load may be received as a load prediction from load/rate prediction module 122.

In some embodiments, the predicted thermal energy load comprises a predicted hot water thermal energy load per time step kAnd predicting cold water thermal energy load

Predicting hot water thermal load

May be satisfied by a combination of heat recovery chiller sub-facility 14, heater sub-facility 12, and hot TES sub-facility 20. Preparation ofMeasuring cold water heat load

May be satisfied by a combination of heat recovery chiller sub-facility 14, chiller sub-facility 16, and cold TES sub-facility 22.

The equality constraint can be placed in the form of Hx ═ g by defining the H matrix and g vector as follows:

wherein the content of the first and second substances,

and

k-dimensional vectors for the predicted cold water load and predicted hot water load at each time step k, respectively, and u_{elec,hrChiller}Is the power consumption of the heat recovery chiller sub-facility 14. For a central facility that services one or more additional types of loads, additional rows may be added to the H matrix and g vector to define the equality constraints for each additional load serviced by the central facility.

For this example problem, assuming an optimization period of 72 one hour samples, the linear programming has 360 decision variables and 1224 constraints. However, the linear programming module 144 may solve the linear programming in less than 200 milliseconds using a linear programming framework to determine the optimal sub-facility load value. Advantageously, this allows the high-level optimization module 130 to determine the sub-facility load distribution for a long optimization period within a very short time frame.

Still referring to FIG. 4, the high-level optimization module 130 is illustrated as including an under-met load module 150. In some examples, the central facility apparatus 60 may not have sufficient capacity or reserve storage (reserve storage) to meet the predicted thermal energy load, regardless of how the thermal energy load is distributed among the sub-facilities 12-22. In other words, even if the applicable sub-facility is running at maximum capacity, the high level optimization problem may not satisfy the solutions of all inequality and equality constraints. The under-met load module 150 may be configured to fix the high-level optimization problem to account for this possibility and allow the high-level optimization to find a solution that results in the smallest amount of under-met load.

In some embodiments, the under-met load module 150 modifies the decision variable matrix x by introducing slack variables for each type of thermal energy load. The relaxation variable represents an unsatisfied (e.g., unsatisfied, delayed, etc.) amount of each type of thermal energy load. For example, the under-met load module 150 may modify the decision variable matrix x as follows:

wherein Q is_{ColdUnmet,1…n}And Q_{HotUnmet,1…n}Respectively, are n-dimensional vectors representing the total delayed cold thermal energy load and the total delayed hot thermal energy load at each time step k within the optimization period. In some embodiments, the decision variable Q_{ColdUnmet,1…n}And Q_{HotUnmet,1…n}Representing the total delay load accumulated up to each time step k, rather than the incrementally delayed load at each time step. The total delay load may be used because any delay load may increase the load required during subsequent time steps.

The under-load module 150 may correct the equation constraints to account for any delayed thermal energy load. The revised equality constraints may require the following: the predicted thermal energy load is equal to the total load met by the sub-facilities 12-22 plus any unsatisfied thermal energy loads. The revised equality constraint can be placed in the form of Hx ═ g by defining the H matrix and g vector as follows:

wherein [ D ]_-1]Is the lower diagonal matrix of 1.

The unsatisfied load module 150 may modify the cost vector c to associate a cost value with any unsatisfied load. In some embodiments, the under-met load module 150 assigns a relatively higher cost to the under-met load than the costs associated with other types of loads in the decision variable matrix x. Assigning large costs to the under-satisfied load ensures that the optimal solution to the high-level optimization problem uses the under-satisfied load only as the last means (i.e., when the optimization has no solution if the under-satisfied load is not used). Thus, the linear programming module 144 may avoid using the unsatisfied load if any feasible combination of plants is capable of satisfying the predicted thermal energy load. In some embodiments, the unsatisfied load module 150 assigns a cost value to the unsatisfied load that allows the linear programming module 144 to use the unsatisfied load in the optimal solution even if the central facility is able to meet the predicted thermal energy load. For example, the under-met load module 150 may assign a cost value that allows the linear programming module 144 to use the under-met load without regard for the solution that the under-met load would be prohibitively expensive and/or very inefficient.

Still referring to FIG. 4, the high-level optimization module 130 is illustrated as including a sub-facility curve module 170. In the simplest case described with reference to the linear programming module 144, it is assumed that the resource consumption of each sub-facility is a linear function of the thermal energy load produced by the sub-facility. However, this assumption may not hold for some sub-facility devices, let alone for the entire sub-facility. The sub-facility curve module 170 may be configured to correct a high-level optimization problem to account for sub-facilities that have a non-linear relationship between resource consumption and load generation.

Sub-facility curve module 170 is illustrated as including a sub-facility curve updater 172, a sub-facility curve database 174, a sub-facility curve linearizer 176, and a sub-facility curve merger (incorporator) 178. The sub-facility curve updater 172 may be configured to request sub-facility curves for each of the sub-facilities 12-22 from the low-level optimization module 132. Each sub-facility curve may indicate resource consumption (e.g., electricity in kW, water in L/s, etc.) for a particular sub-facility as a function of sub-facility load. Exemplary sub-facility curves are shown and described in more detail with reference to fig. 5A-8.

In some embodiments, the low-level optimization module 132 generates the sub-facility curve by running a low-level optimization process for various combinations of sub-facility loads and weather conditions to generate a plurality of data points. The low-level optimization module 132 may fit a curve to the data points to generate a sub-facility curve and provide the sub-facility curve to the sub-facility curve updater 172. In other embodiments, the low-level optimization module 132 provides the data points to the sub-facility curve updater 172, and the sub-facility curve updater 172 generates sub-facility curves using the data points. The sub-facility curve updater 172 may store sub-facility curves in the sub-facility curve database 174 for use in the high level optimization process.

In some embodiments, the sub-facility curve is generated by combining efficiency curves of the individual devices of the sub-facility. The device efficiency curve may indicate the amount of resource consumption of the device as a function of load. The device efficiency curves may be provided by the device manufacturer or generated using experimental data. In some embodiments, the device efficiency curve is based on an initial efficiency curve provided by the device manufacturer and updated with experimental data. The plant efficiency curve may be stored in the plant model 120. For some devices, the device efficiency curve may indicate that resource consumption is a U-shaped function of load. Thus, when combining multiple plant efficiency curves into a sub-facility curve for the entire sub-facility, the resulting sub-facility curve may be a wave-like curve as shown in fig. 6. These waves are caused by the individual device load rising before it would be more efficient to switch on another device to meet the sub-facility load.

Sub-facility curve linearizer 176 may be configured to convert the sub-facility curve to a convex curve. A convex curve is a curve where a straight line connecting any two points on the curve is always above or along the curve (i.e., not below the curve). Convex curves may be advantageous for use in high-level optimization, because they allow for less computationally expensive optimization processes relative to optimization processes using non-convex functions. Sub-facility curve linearizer 176 may be configured to segment the sub-facility curve into piecewise linear segments that are combined to form a piecewise-defined convex curve. Fig. 6 and 7 show an unmodified sub-facility curve 600 and a linearized sub-facility curve 700 generated by the sub-facility curve linearizer 176, respectively. Sub-facility curve linearizer 176 may store the linearized sub-facility curve in sub-facility curve database 174.

Still referring to FIG. 4, the sub-facility curve module 170 is illustrated as including a sub-facility curve merger 178. The sub-facility curve merger 178 may be configured to correct a high-level optimization problem to introduce the sub-facility curve into the optimization. In some embodiments, the sub-facility curve merger 178 modifies the decision matrix x to include one or more decision vectors that represent resource consumption for each sub-facility. For example, for the chiller sub-facility 16, the sub-facility curve merger 178 may modify the decision matrix x as follows:

wherein u is_{Chiller,elec,1…n}And u_{Chiller,water,1…n}Respectively, are n-dimensional vectors representing the power and water consumption of the chiller sub-facilities 16 at each time step k.

For each sub-facility 12-22, the sub-facility curve merger 178 may add one or more resource consumption vectors to the matrix x. The decision vector added by the sub-facility curve merger 178 for a given sub-facility may represent the resource consumption of each resource (e.g., water, electricity, natural gas, etc.) consumed by that sub-facility at each time step k within the optimization period. For example, if the heater sub-facility 12 consumes natural gas, electricity, and water, the sub-facility curve merger 178 may increase a decision vector u representing the amount of natural gas consumed by the heater sub-facility 12 at each time step_{Heater,gas,1…n}A decision vector u representing the amount of power consumed by the heater sub-facility 12 at each time step_{Heater,elec,1…n}And a decision vector u representing the amount of water consumed by the heater sub-facility at each time step_{Heater,water,1…n}. The sub-facility curve merger 178 can add resource consumption vectors for other sub-facilities in a similar manner.

The sub-facility curve merger 178 may modify the cost vector c to account for the decision matrix xThe resource consumption vector of (2). In some embodiments, the sub-facility curve merger 178 removes (or zeroes) the sub-facility curve directly associated with the sub-facility load (e.g.,etc.) and increase the economic cost associated with the consumption of resources required to generate the sub-facility load. For example, for chiller sub-facility 16, sub-facility curve merger 178 may modify cost vector c as follows:

C＝[… 0_n… C_elec,1…nC_{water，1…n}…]^T

wherein 0n indicates at each time step

Of an n-dimensional zero vector with a direct economic cost of zero, c_elec,1…nIs an n-dimensional vector indicating the cost per unit of electricity at each time step, and c_water,1…nIs an n-dimensional vector indicating the cost per unit of water at each time step. The revised cost vector associates the economic cost with the resources consumed to generate the sub-facility load, rather than with the sub-facility load itself. In some embodiments, c_elec,1…nAnd c_water,1…nIs the utility rate obtained from load/rate prediction module 122.

The sub-facility curve merger 178 can correct the inequality constraints to ensure that the correct amount of each resource is consumed to service the predicted thermal energy load. In some embodiments, the sub-facility curve merger 178 formulates inequality constraints that force the resource usage of each resource in the context map (epigraph) of the corresponding linearized sub-facility curve. For example, the chiller sub-plant 16 may have a linearized sub-plant curve that is a function of the chilled water production of the chiller sub-plant 16 (i.e.,

) Indicating power usage of the chiller sub-facility 16 (i.e., u)_Chiller,elec). Such a linearized sub-facility curve 700 is illustrated in fig. 7. The linearizing sub-facility curve may include connecting a point [ u ]₁，Q₁]Connected to a point u₂，Q₂]First line segment of (1), point [ u ]₂，Q₂]Connected to a point u₃，Q₃]And a second line segment of [ u ], and a point of [ u ]₃，Q₃]Connected to a point u₄，Q₄]The third line segment of (1).

The sub-facility curve merger 178 may formulate an inequality constraint for each segmented segment of the sub-facility curve that will u_Chiller,elecIs constrained to be greater than or equal to the value used for

The line segment of the corresponding value of (a) defines the amount of electricity used. The utility curve constraint for the chiller sub-facility 16 can be placed in the form of Ax ≦ b by defining the A matrix and the b vector as follows:

similar inequality constraints may be formulated for other sub-facility curves. For example, the sub-facility curve merger 178 may use a function of cold water productionDefining the water consumption u of the chiller sub-facilities 16_{Chiller,water},_1…nTo generate the water consumption u for the chiller sub-plant 16_{Chiller,water},_1…nA set of inequality constraints. In some embodiments, the water consumption of the chiller sub-facility 16 is equal to the chilled water production, and the linearized sub-facility curve for water consumption includes the point [ u [ ]₅，Q₅]Connected to a point u₆，Q₆]A single line segment (as shown in fig. 5B). The sub-plant curve constraint for the cold water consumption of the chiller sub-plant 16 can be placed in the form of Ax ≦ b by defining the A matrix and the b vector as follows:

A＝[… [-(u₆-u₅)]I_n… 0_n[(Q₆-Q₅)]I_n…]，b＝[Q₅u₆-Q₆u₅]

the sub-facility curve merger 178 may repeat this process for each sub-facility curve for the chiller sub-facilities 16 and for the other sub-facilities of the central facility 10 to define a set of inequality constraints for each sub-facility curve.

The inequality constraints generated by the sub-facility curve merger 178 ensure that the high level optimization module 130 keeps the resource consumption above all segments of the corresponding sub-facility curve. In most cases, the high-level optimization module 130 has no reason to select resource consumption values that lie above the corresponding sub-facility curve due to the economic cost associated with resource consumption. Thus, the high-level optimization module 130 may expect to select resource consumption values that lie on the corresponding sub-facility curve rather than above it.

The exception to this general rule is the heat recovery chiller sub-facility 14. The equality constraints for the heat recovery chiller sub-facility 14 are specified as follows: the heat recovery chiller sub-facility 14 produces hot water at a rate equal to the chilled water production of the sub-facility plus the electricity usage of the sub-facility. The inequality constraints for the heat recovery chiller sub-facility 14 produced by the sub-facility curve merger 178 allow the high-level optimization module 130 to over-utilize to produce more hot water without increasing the cold water production. This behavior is extremely inefficient and only becomes realistic when the hot water demand is high and cannot be met with more efficient technology. However, this is not the way in which the heat recovery chiller sub-installation 14 actually operates.

To prevent the high-level optimization module 130 from over-utilizing electricity, the sub-facility curve merger 178 may check whether the calculated electricity usage (determined by the optimization algorithm) of the heat recovery chiller sub-facility 14 is above the corresponding sub-facility curve. In some embodiments, this check is performed after each iteration of the optimization algorithm. If the calculated power usage of the heat recovery chiller sub-facility 14 is above the sub-facility curve, the sub-facility curve merger 178 may determine that the high-level optimization module 130 is over-utilizing. In response to determining that the high-level optimization module 130 is over-utilizing electricity, the sub-facility curve merger 178 may constrain the production of the heat recovery chiller sub-facility 14 to its current value and constrain the utilization of electricity by the sub-facility 14 to a corresponding value on the sub-facility curve. The high-level optimization module 130 may then re-run the optimization with the new equation constraints.

Still referring to FIG. 4, the high-level optimization module 130 is illustrated as including a ground loop module 152 and a heat exchanger module 154. In some embodiments, the central facility 10 includes a heat exchanger configured to transfer heat between the hot water circuit 24 and the condensate water circuit 28. In some embodiments, the central facility 10 includes a ground loop that serves as a heat rejection for the chiller sub-facility 16 and/or a heat uptake for the heat recovery chiller sub-facility 14. The ground loop module 152 and the heat exchanger module 154 may be configured to modify optimization issues to account for heat transfer resulting from operation of the heat exchanger and/or the ground loop.

The ground loop module 152 may introduce heat rejection to the ground loop into the optimization problem by varying the power and water usage of the chiller sub-facilities 16. For example, for a load application up to the heat removal capacity of the ground circuit, the chiller sub-facility 16 may use additional power to run the ground circuit pump. The additional power usage may be constant or may vary per unit flow through the ground loop. The water production of the chiller sub-installation 16 may be constant regardless of whether a ground loop is used.

The ground circuit module 152 and the heat exchanger module 154 may in a similar manner introduce heat uptake from the ground circuit and heat transfer between the hot water circuit 24 and the condensate water circuit 28 into the optimization problem. For example, the ground loop module 152 and the heat exchanger module 154 may use heat ingestion from the ground loop and heat transfer between the loops 24 and 28 to correct the load seen by the central facility equipment. The ground loop module 152 can use the ground loop to generate a load that the plant looks like a dummy building load, thereby allowing the heat recovery chiller sub-facility 14 to operate as a heat pump when the building load is not applying sufficient heat to the system. This result may be optimal when the ratio between electricity prices and gas prices is low so that electrically operated circuits and heat exchangers are cheaper than generating heat with natural gas in the heater sub-facility 12.

The heat exchanger module 154 may use heat exchangers to produce what appears to be a pseudo-hot water building load, thereby allowing the heat recovery chiller sub-facility 14 to operate as a conventional chiller. Excess heat from the heat recovery chiller sub-installation 14 may be transferred through a heat exchanger to the condenser loop 28 and ultimately into the atmosphere or the ground. In some embodiments, the heat exchanger module 154 operates the heat exchanger to prevent the condenser circuit from becoming overloaded. For example, the heat exchanger module 154 may limit the total heat rejected to the capacity of the condenser circuit 28 minus the heat generated by a conventional chiller.

The ground loop module 152 and the heat exchanger module 154 may modify the decision matrix x by adding a new decision vector for each type of thermal energy load. The new decision vector may represent the overproduction of each thermal energy load for each time step k within the optimization period. For example, the revised decision matrix may be presented as follows:

wherein the content of the first and second substances,andrespectively, are n-dimensional vectors representing the overproduction rate of the cold and hot thermal energy loads per time step k within the optimization period.

The ground loop module 152 and the heat exchanger module 154 may correct the equation constraints to account for any excess generated thermal energy load. The overproduced thermal energy load may be added to the equality constraint as a slack variable that operates in the opposite direction to the unsatisfied load. The revised equality constraints may require that the predicted thermal energy load plus any overproduction be equal to the total load met by the sub-facilities 12-22 plus any unsatisfied thermal energy load. The revised equality constraint can be placed in the form of Hx ═ g by defining the H matrix and g vector as follows:

wherein [ D ]_-1]Is the lower diagonal matrix of 1. The ground loop module 152 and the heat exchanger module 154 may correct the cost vector c with the additional cost per unit of overproduced pumping power required to operate the ground loop and/or heat exchanger.

Still referring to FIG. 4, the high-level optimization module 130 is illustrated as including a demand charging module 156. As described above, optimization framework module 142 may formulate the optimization problem as:

subject to Ax ≦ b, Hx ═ g

However, such a formula does not take into account demand charges.

Demand charging is an additional charge that certain utility providers implement based on the maximum rate of energy consumption during the applicable demand charging period. For example, the demand charge may be provided in dollars per unit power (e.g., $/kW), and may be multiplied by the peak power usage during the demand charge period (e.g., kW) to calculate the demand charge. In some examples, demand charges may account for more than 15% of the electricity bill. Not including demand charges in the optimization scheme may result in all devices being turned on at the same time (e.g., the most efficient or least cost time). This is optimal from the point of view of consumption costs. However, shifting some load over time may save thousands of dollars in demand costs while only spending a few dollars in consumption costs.

The demand charges module 156 may be configured to revise the optimization problem to account for demand charges. Introducing peak charges into the optimization framework can greatly improve the performance of high-level optimizations. For example, including demand charges in the optimization framework may reduce the total operating cost of the central facility 10 by an additional 5% over the 8-10% cost reduction provided by the other modules of the central facility controller 102. In various embodiments, the savings provided by the demand charging module 156 and/or the central facility controller 102 may generally be greater or less than the exemplary amounts defined herein due to differences in facility configuration and/or energy costs.

The demand charging module 156 may consider demand charging by modifying a cost function used by the high-level optimization module 130. The cost function of the modification may be defined as:

obeying Ax ≦ b, Hx ═ g

Wherein, c_demandIs the demand charge (e.g., $/kW) for the applicable demand charge period, and P_elec,kIs the total power consumption of the central facility 10 and the building/campus at time step k. Term max (P)_elec,k) The peak power consumption at any time during the demand charging period is selected. Demand charges c_demandAnd demand charging periods may be defined by utility rate information received from utility 126 and may be provided to high-level optimization module 130 by load/rate prediction module 122.

Introducing demand charges into the optimization framework complicates the optimization problem in two main ways. First, the cost function is no longer linear, since it includes the max () function. Second, consumption item c^Tx calculating the cost over a consumption period defined by a time range, and a demand charge term c_demandmax(P_elec,k) The cost over the demand charge period is calculated. For example, a consumption period may be defined as a period of time beginning at a current time level k and ending at a future time level k + h, where h represents a time range. Demand charge periods may be defined by utility 126 and provided to high-level optimization module 130 along with utility rate information. In some instances, the consumption period and the demand charge period may not be the same. This is due to the fact that the demand is collectedThe potential trade-offs between control decisions that reduce consumption terms at the expense of expense terms (and vice versa) become confusing and complicate the optimization problem.

The demand charges module 156 may modify the optimization problem to introduce demand charges into the linear optimization framework. For example, the demand charging module 156 may modify the decision matrix x by adding a new decision variable xpeak, as follows:

x_nex＝[… u_{Chiller，elec，1…n}… u_{hpChiller，elec，1…n}… u_{Heater，elec，1…n}… x_peak]^T

wherein x is_peakIs the peak power consumption during the demand charge period. The demand charging module 156 may modify the cost vector c as follows:

c_new＝[… c_elec，1…n… c_elec，1…n… c_elec，1…n… c_demand]^T

so that the demand charges c_demandMultiplied by the peak power consumption x_peak。

The demand charge module 156 may formulate and/or apply inequality constraints to ensure that the peak power consumption xpeak is greater than or equal to the maximum electrical demand over the demand charge period. Namely:

range of

The inequality constraints can be represented in a linear optimization framework by defining the a matrix and b vector as follows:

A＝[… [I_h]… [I_h]… [I_h]… -1]，b＝-P_{elec，campus，k}

during the high-level optimization process, the high-level optimization module 130 may select x equal to the maximum electrical demand over the demand charge period_peakTo minimize x_peakThe associated cost.

The demand charging module 156 may apply inequality constraints to ensure charging on demandPeak power consumption decision variable x during a period_peakGreater than or equal to its previous value x_{peak,previous}. The inequality constraints can be represented in a linear optimization framework by defining the a matrix and b vector as follows:

A＝[… －1]，b＝-x_{peak，previous}

advantageously, the modification to the decision variable matrix x, the cost vector c, and the inequality constraints provided by the demand charging module 156 allows the cost function to be written in a linear form as follows:

this linear form, subject to Ax ≦ b, Hx ═ g cost function, may be used in a linear optimization framework.

The cost function written in the foregoing equation has components over different time periods. For example, the consumption item c^Tx is over a consumption period and the demand charge term c_demandx_peakOver a demand charging period. To properly balance between increased demand charges and increased energy consumption costs, the demand charges module 156 may apply a weighting factor to the demand charges term and/or the consumption term. For example, the demand charging module 156 may charge the consumption c^Tx is divided by the duration h of the consumption period (i.e., the time period between the current time and the time range) and multiplied by the amount of time d remaining in the current demand charge period_demandAnd thus the entire cost function is over the demand charge period. The new optimization function can be given as follows:

subject to Ax ≦ b, Hx ═ g

It is equivalent to:

subject to Ax ≦ b, Hx ═ g

The latter form of the new optimization function has the advantage of adjusting only one term of the function instead of several.

Still referring to FIG. 4, the high-level optimization module 130 is illustrated as including a load-change penalty module 158. In some examples, the high-level optimization module 130 determines a solution to the optimization problem that includes significantly changing the load on one or more of the sub-facilities 12-22 within a relatively short timeframe. For example, a lowest cost solution from a resource consumption perspective may involve changing a sub-facility from disconnected to fully loaded and back again to disconnected within only a few time steps. This behavior may result from the high-level optimization module 130 identifying small fluctuations in the economic cost of the resource and operating the central facility 10 accordingly to achieve the minimum economic cost. However, due to various negative effects (e.g., increased equipment degradation) resulting from rapidly changing sub-facility loads, running central facility 10 in this manner may be undesirable, especially when the cost saved is relatively minimal (e.g., cents or dollars).

Load change penalty module 158 may revise the optimization problem to introduce penalties for rapidly changing sub-facility loads. For example, the load change penalty module 158 may modify the decision matrix x by adding a new decision vector for each sub-facility. The new decision vector represents the sub-facility load change for each sub-facility from one time step to the next. For example, the load change penalty module 158 may modify the decision matrix x as follows:

wherein, delta_{Chiller,1…n}、δ_{hrChiller,1…n}And delta_Heater,1…nRespectively, for each time step k, relative to the previous time step k-1

Andthe sub-facility load of (2) varies.

Load change penalty module 158 may modify cost vector c to increase the cost associated with changing the sub-facility load. For example, the load change penalty module 158 may modify the cost vector c as follows:

c＝[… 0_n… 0_n… 0_n… c_{δChiller，1…n}c_{δhrChiller，1…n}c_{δHeater，1…n}]^T

the load change penalty module 158 may add constraints such that each load change variable δ cannot be smaller than the corresponding sub-facility load

A change in (c). For example, the added constraints for the chiller sub-facility 16 may have the form:

wherein the content of the first and second substances,is used at a previous time step

The value of (c). Similar constraints may be added for each sub-facility 12-22.

The constraint condition imposed by the load change penalty module 158 requires that the load change variable δ be greater than or equal to the current value of the corresponding sub-facility load

With previous value of sub-facility load

The magnitude of the difference between. In operation, the high-level optimization module 130 may select a value of the load change variable δ equal to the magnitude of the difference, due to the cost associated with the load change variable. In other words, the high-level optimization module 130 may not choose to make the load change variable δ larger than the actual change in the corresponding sub-facility load because the load change variable is made to be a variableDelta greater than the required variation would be suboptimal.

Still referring to FIG. 4, the high-level optimization module 130 is illustrated as including a tank overfill module 160. The tank overfill module 160 can correct the optimization problem such that the Thermal Energy Storage (TES) tank is forcibly refilled at the end of the optimization period. This feature provides increased robustness in the event of sub-facility failure and/or controller failure by ensuring that the TES tank has sufficient stored thermal energy to satisfy building loads while repairing the failure. For example, the facility operator may use the stored thermal energy to meet building loads while bringing the central facility controller 102 back online.

Tank overfill module 160 can force the TES tank to be filled by increasing the cost of discharging the TES tank. In some embodiments, the tank refill module 160 corrects the cost of discharging the TES tank such that the discharge cost is higher than the other costs in the cost function, but lower than the cost of not meeting the load. This forces the high level optimization module 130 to maximize the revenue (i.e., negative cost) of filling the TES tank.

Referring now to fig. 5A-5B, two

sub-facility curves

500 and 510 are shown, according to an exemplary embodiment. The sub-facility curves 500 and 510 may be used as the sub-facility curve 140, as described with reference to FIG. 3. The sub-facility curves 500 and 510 define resource usage of a sub-facility (e.g., one of the sub-facilities 12-22) as a function of sub-facility load. Each sub-facility curve may be specific to a particular sub-facility and a particular type of resource used by that sub-facility. For example, the sub-facility curve 500 may define the power usage 502 of the chiller sub-facility 16 as a function of the load 504 on the chiller sub-facility 16, while the sub-facility curve 510 may define the water usage 506 of the chiller sub-facility 16 as a function of the load 504 on the chiller sub-facility 16. Each sub-facility 12-22 may have one or more sub-facility curves (e.g., one sub-facility curve for each type of resource consumed by the sub-facility).

In some embodiments, the low-level optimization module 132 generates the

sub-facility curves

500 and 510 based on the plant models 120 (e.g., by combining the plant models 120 for the various appliances into a lumped curve for the sub-facility). The low-level optimization module 132 may generate the

sub-facility curves

500 and 510 by running a low-level optimization process for several different loads and weather conditions to generate a plurality of data points. The low-level optimization module 132 may fit a curve to the data points to generate a sub-facility curve. In other embodiments, the low-level optimization module 132 provides the data points to the high-level optimization module 130, and the high-level optimization module 130 generates the sub-facility curve using the data points.

Referring now to FIG. 6, another sub-facility curve 600 is shown, according to an exemplary embodiment. The sub-facility curve 600 is a function of the chilled water production of the chiller sub-facility 16 (i.e.,

) Define the power usage of the chiller sub-facility 16 (i.e., u)_Chiller,elec). In some embodiments, sub-facility curve 600 is generated by combining efficiency curves of various devices (e.g., various refrigerators, pumps, etc.) of refrigerator sub-facility 16. For example, each chiller in the sub-facility 16 may have an efficiency curve specific to the device that defines the amount of power used by that chiller as a function of the load on that chiller. Many devices operate at lower efficiency at higher loads and have a device efficiency curve that is a function of the U-shaped load. Thus, combining multiple plant efficiency curves to form the sub-facility curve 600 may result in the sub-facility curve 600 having one or more fluctuations 602, as shown in FIG. 6. These fluctuations 602 may be caused by the load of an individual device rising before it would be more efficient to switch on another device to meet the sub-facility load.

Referring now to FIG. 7, a linearized sub-facility curve 700 is illustrated, according to an exemplary embodiment. The sub-facility curve 700 is a function of the chilled water production of the chiller sub-facility 16 (i.e.,

) Define the power usage of the chiller sub-facility 16 (i.e., u)_Chiller,elec). The sub-facility curve 700 may be generated by converting the sub-facility curve 600 into a linearized convex curve. The convex curve is such that: a straight line connecting any two points on the curve is always above or along the curveFollows the curve (i.e., is not below the curve). Convex curves may be advantageous for use in high-level optimization because they allow for less computationally expensive optimization processes relative to optimization processes using non-convex functions.

In some embodiments, sub-facility curve 700 is generated by sub-facility curve linearizer 176, as described with reference to fig. 4. The sub-facility curve 700 may be generated by generating a plurality of linear segments (i.e.,

segments

702, 704, and 706) that approximate the sub-facility curve 600 and combining the linear segments into a piecewise-defined linearized convex curve 700. Linearizing sub-facility curve 700 is illustrated as including point u₁，Q₁]Connected to a point u₂，Q₂]First linear segment 702, point [ u ]₂，Q₂]Connected to a point u₃，Q₃]A second linear segment 704, and a point [ u ]₃，Q₃]Connected to a point u₄，Q₄]Of the third linear segment 706. The end points of the segments 702-706 may be used to form constraints that force the power usage of the chiller sub-facility 16 in the context map of the linearized sub-facility curve 700.

Referring now to FIG. 8, another sub-facility curve 800 is shown in accordance with an exemplary embodiment. The sub-facility curve 800 defines the energy usage of one of the sub-facilities 12-22 as a function of the load on that sub-facility for several different weather conditions. In some embodiments, the sub-facility curve 800 is generated by the sub-facility curve module 170 using experimental data obtained from the low-level optimization module 132. For example, the sub-facility curve updater 172 may request resource usage data from the low-level optimization module 132 for various combinations of load conditions and environmental conditions. In the embodiment shown in FIG. 8, the sub-facility curve updater 172 requests a temperature (e.g., 40)^oF、50^oF、60^oF and 70^oF) And load (e.g., 170 tons, 330 tons, 500 tons, 830 tons, and 1000 tons). The low-level optimization module 132 may perform a low-level optimization process for the requested load and temperature combinations and return an energy usage value for each combination.

The sub-facility curve updater 172 mayThe data points provided by the low-level optimization module 132 are used to find the best piece-wise linear convex function to fit to the data. For example, the sub-facility curve updater 172 may fit the first sub-facility curve 802 to the 70^oF, fitting the second sub-facility curve 804 to 60^oF, fitting a third sub-facility curve 806 to 50^oF, and a fourth sub-facility curve 808 to 40^oData points for F. The sub-facility curve updater 12 may store the generated sub-facility curves 802-808 in the sub-facility curve database 174 for use in the high level optimization algorithm.

In some embodiments, the central facility controller 102 uses the high-level optimization module 130 as part of an operational tool for real-time control of the central facility 10. In this operational tool, the high-level optimization module 130 may receive load and rate predictions from the load/rate prediction module 122 and sub-facility curves (or data that may be used to generate sub-facility curves) from the low-level optimization module 132. When implemented in an operational tool, high level optimization module 130 may determine optimal sub-facility loads for heater sub-facility 12, heat recovery chiller sub-facility 14, chiller sub-facility 16, hot TES sub-facility 20, and/or cold TES sub-facility 22, as described with reference to FIGS. 2-4. In some embodiments, the high level optimization module 130 determines the ground loop and heat exchanger conductivity in addition to the sub-facility loads. When implemented in an operational tool, the high-level optimization module 130 may provide the determined sub-facility loads and/or conductivities to the low-level optimization module 132 for use in determining optimal on/off decisions and/or operational set points for the equipment of each sub-facility.

Referring now to FIG. 9, a block diagram of a planning system 900 is shown, according to an exemplary embodiment. The planning system 900 may be configured to use the optimization module 128 as part of a planning tool 902 to simulate the operation of a central facility over a predetermined period of time (e.g., days, months, weeks, years, etc.) for planning, budgeting, and/or design considerations. When implemented in the planning tool 902, the optimization module 128 may operate in a similar manner as described with reference to fig. 2-4. For example, the optimization module 128 may use building loads and utility rates to determine an optimal sub-facility load distribution to minimize costs over a simulation period. However, the planning tool 902 may not be responsible for real-time control of the building automation system or the central facility.

In the planning tool 902, the high-level optimization module 130 may receive the planned loads and utility rates for the entire simulation session. The projected load and utility rates may be defined by inputs received from the user (e.g., user-defined, user-selected, etc.) via the client device 922 and/or retrieved from the planning information database 926. The high-level optimization module 130 uses the projected loads and utility rates in conjunction with the sub-facility curves from the low-level optimization module 132 to determine an optimal sub-facility load (i.e., an optimal scheduling schedule) for a portion of the simulation period.

A portion of the simulation period during which the high-level optimization module 130 optimizes the sub-facility load may be defined by a prediction window ending at a time horizon. With each iteration of the optimization, the prediction window is translated forward, and the portion of the scheduling schedule that is no longer within the prediction window is accepted (e.g., stored or output as a simulation result). The load and rate predictions may be predefined for the entire simulation and may not be subject to adjustment in each iteration. However, shifting the prediction window forward in time may introduce additional planning information (e.g., planning load and/or utility rates) for the newly added time segment at the end of the prediction window. The new planning information may have no significant impact on the optimal scheduling schedule because only a small portion of the prediction window may change with each iteration.

In some embodiments, the high-level optimization module 130 requests all sub-facility curves used in the simulation from the low-level optimization module 132 at the start of the simulation. Since the projected loads and environmental conditions are known throughout the simulation period, the high-level optimization module 130 may retrieve all relevant sub-facility curves at the start of the simulation. In some embodiments, when the sub-facility curves are provided to the high-level optimization module 130, the low-level optimization module 132 generates a function that maps sub-facility production to equipment-level production and resource usage. These sub-facility-to-device functions may be used to calculate the yield and resource usage of individual devices (e.g., in a post-processing module) based on the results of the simulation.

Still referring to fig. 9, the planning tool 902 is illustrated as including a communication interface 904 and a processing circuit 906. The communication interface 904 may include a wired or wireless interface (e.g., receptacle, antenna, transmitter, receiver, transceiver, wire connection, etc.) for data communication with various systems, devices, or networks. For example, the communication interface 904 may include an ethernet card and port for sending and receiving data via an ethernet-based communication network, and/or a WiFi transceiver for communicating via a wireless communication network. Communication interface 904 may be configured to communicate via a local or wide area network (e.g., the internet, a building WAN, etc.) and may use various communication protocols (e.g., BACnet, IP, LON, etc.).

The communication interface 904 may be a network interface configured to facilitate electronic data communication between the planning tool 902 and various external systems or devices (e.g., client device 922, results database 928, planning information database 926, etc.). For example, planning tool 902 may receive planned loads and utility rates from client device 922 and/or planning information database 926 via communication interface 904. The planning tool 902 may use the communication interface 904 to output simulation results to the client device 922 and/or to store the results in a results database 928.

Still referring to fig. 9, the processing circuitry 906 is illustrated as including a processor 910 and a memory 912. Processor 910 may be a general or special purpose processor, an Application Specific Integrated Circuit (ASIC), one or more Field Programmable Gate Arrays (FPGAs), a set of processing components, or other suitable processing components. The processor 910 may be configured to execute computer code or instructions stored in the memory 912 or received from other computer-readable media (e.g., CDROM, network storage, remote server, etc.).

The memory 912 may include one or more devices (e.g., memory units, memory devices, storage devices, etc.) for storing data and/or computer code for performing and/or facilitating the various processes described in this disclosure. Memory 912 may include Random Access Memory (RAM), Read Only Memory (ROM), hard drive storage, temporary storage, non-volatile memory, flash memory, optical storage, or any other suitable memory for storing software objects and/or computer instructions. Memory 912 may include database components, object code components, script components, or any other type of information structure for supporting the various behaviors and information structures described in this disclosure. The memory 912 may be communicatively connected to the processor 910 via the processing circuitry 906 and may include computer code for executing (e.g., by the processor 906) one or more processes described herein.

Still referring to fig. 9, the memory 912 is illustrated as including a GUI engine 926, a web service 914, and a configuration tool 918. In an exemplary embodiment, the GUI engine 916 includes a graphical user interface component configured to provide a graphical user interface to a user for selecting or defining planning information (e.g., planned loads, utility rates, environmental conditions, etc.) for simulation. Web services 914 may allow users to interact with planning tool 902 via a web portal and/or from remote systems or devices (e.g., enterprise control applications).

Configuration tool 918 can allow a user (e.g., via a graphical user interface, via prompt-triggered "wizards," etc.) to define various simulation parameters, such as the number and type of sub-facilities, devices within each sub-facility, sub-facility curves, device-specific efficiency curves, simulation durations, durations of prediction windows, durations of each time step, and/or various other types of planning information related to the simulation. Configuration tool 918 can present a user interface for building a simulation. The user interface may allow a user to graphically define simulation parameters. In some embodiments, the user interface allows a user to select pre-stored or pre-constructed simulation facility and/or planning information (e.g., from planning information database 926) and adapt or enable it for use in the simulation.

Still referring to fig. 9, the memory 912 is illustrated as including the optimization module 128. The optimization module 128 may use the projected load and the utility rates to determine an optimal sub-facility load over the prediction window. The operation of the optimization module 128 may be the same or similar to that previously described with reference to fig. 2-4. With each iteration of the optimization process, the optimization module 128 may translate the prediction window forward and apply the optimal sub-facility load for the portion of the simulation period that is no longer within the prediction window. The optimization module 128 may use the new plan information at the end of the prediction window to perform the next iteration of the optimization process. The optimization module 128 may output the applied sub-facility loads to the reporting application 930 for presentation to the client device 922 (e.g., via the user interface 924) or storage in the results database 928.

Still referring to fig. 9, the memory 912 is illustrated as including a reporting application 930. The reporting application 930 may receive the optimized sub-facility load from the optimization module 128 and, in some embodiments, a cost associated with the optimized sub-facility load. The reporting application 930 may include a web-based reporting application having a number of Graphical User Interface (GUI) elements (e.g., widgets, dashboard controls, windows, etc.) for displaying Key Performance Indicators (KPIs) or other information to a user of the GUI. Further, GUI elements may summarize relative energy usage and intensity across various stations, sub-facilities, and the like. Other GUI elements or reports may be generated and displayed based on the available data, which allow the user to evaluate the simulation results. The user interface or report (or underlying data engine) may be configured to aggregate and categorize the sub-facility loads and costs associated therewith, and provide the results to the user via the GUI. The GUI elements may include charts or histograms that allow the user to visually analyze the simulation results. An exemplary output that may be generated by the reporting application 930 is illustrated in FIG. 10.

Referring now to FIG. 10, several graphs 1000 showing the operation of a planning tool 902 are shown, according to an exemplary embodiment. With each iteration of the optimization process, planning tool 902 selects an optimization period (i.e., a portion of the simulation period) over which to perform the optimization. For example, the planning tool 902 may select an optimization period 1002 for use in the first iteration. Once the optimal load distribution 1010 has been determined, the planning tool 902 may select a portion 1018 of the load distribution 1010 to send to the facility schedule 1030. Portion 1018 may be the first b time steps of load distribution 1010. The planning tool 902 may translate the optimization time period 1002 forward in time to arrive at the optimization time period 1004. The amount of prediction window translation may correspond to the duration of time step b.

The planning tool 902 may repeat the optimization process for the optimization time period 1004 to determine the optimal facility load distribution 1012. The planning tool 902 may select a portion 1020 of the facility load allocation 1012 to send to the facility schedule 1030. Portion 1020 may be the first b time steps of load distribution 1012. The planning tool 902 may then translate the prediction window forward in time, thereby resulting in an optimization period 1006. This process may be repeated for each subsequent optimization period (e.g., optimization periods 1006, 1008, etc.) to generate updated load allocations (e.g.,

load allocations

1014, 1016, etc.) and select portions of each load allocation (e.g., portions 1022, 1024) to send to the facility schedule 1030. The facility schedule 1030 includes the first b time steps 1018- > 1024 from each optimization period 1002- > 1008. Once the optimal sub-facility load distribution 1030 has been programmed for the entire simulation period, the results may be sent to the reporting application 930, the results database 928, and/or the client device 922, as described with reference to fig. 9.

Referring now to FIG. 11, a flowchart of a process 1100 for optimizing central facility costs is shown, according to an example embodiment. In various implementations, the process 1100 may be performed by the central facility controller 102 or the planning tool 902. The central facility may include a plurality of sub-facilities (e.g., sub-facilities 12-22) configured to service the energy usage load of the building or campus. The central facility may be an actual facility (e.g., central facility 10) or a simulated central facility comprising a plurality of simulated sub-facilities.

Process 1100 is illustrated as including receiving load forecast data and utility rate data (step 1102). The load forecast data may include a predicted or projected thermal energy load for the building or campus at each time step k (e.g., k 1 … n) of the optimization period. The load forecast data may include a forecast or forecast of one or more different types of loads for a building or campusThe projected value. For example, the load prediction data may include a predicted hot water load for each time step k within a prediction window

And predicting cold water load

In some embodiments, the load forecast data is based on weather forecasts from weather services and/or feedback from buildings or parks. Feedback from the building or campus may include various types of sensor inputs (e.g., temperature, flow, humidity, enthalpy, etc.) or other data related to the building being controlled (e.g., inputs from HVAC systems, lighting control systems, security systems, water systems, etc.). In some embodiments, the load forecast data is generated by load/rate forecast module 122, as described with reference to fig. 2. For example, the load forecast data may be based on measured electrical loads from a building or campus and/or previously measured load data. The load forecast data may be a function of a given weather forecast

Type of date (day), time of day (t) and previously measured load data (Y)_k-1). Such a relationship can be represented in the following equation:

the utility rate data may indicate a cost per unit or price of one or more resources (e.g., electricity, natural gas, water, etc.) consumed by the central facility to service the thermal energy load of the building or campus at each timestep k in the prediction window. In some embodiments, the utility rate is a time-varying rate. For example, the price of electricity may be higher at certain times of day or days of the week (e.g., during high demand periods) and lower at other times of day or days of the week (e.g., during low demand periods). The utility rates may define various time periods and a cost per unit of resource during each time period. The utility rate may be an actual rate (e.g., received from utility 126) or a predicted utility rate (e.g., estimated by load/rate prediction module 122).

In some embodiments, the utility rate includes a demand charge for one or more resources consumed by the central facility. The demand charge may define a separate cost based on a maximum usage (e.g., maximum energy consumption) of a particular resource during the demand charge period. The utility rate may define various demand charge periods and one or more demand charges associated with each demand charge period. In some instances, the demand charge periods may partially or completely overlap with each other and/or with the prediction window. The utility rate data may include a time-varying (e.g., hourly) price, a maximum service level (e.g., a maximum consumption rate allowed by a physical infrastructure or contract), and a demand charge in the case of electricity or a charge for a peak consumption rate over a period of time.

Still referring to FIG. 11, process 1100 is illustrated as including generating an objective function representing a total financial cost of operating the central facility as a function of the utility rate data and the amount of resources consumed by the central facility (step 1104). In some embodiments, the objective function is a high-level cost function J for the central facility_HL. High level cost function J_HLCan represent the sum of the financial costs of each utility consumed by the central facility for the duration of the optimization period. E.g. a high level cost function J_HLThis can be described using the following equation:

wherein n is_hIs the number of time steps k, n, in the optimization period_sIs the number of sub-facilities, t_sIs the duration of the time step, c_jkIs the economic cost of utility j at time step k of the optimization period, and u_jikIs the usage of utility j by sub-facility i at timescale k.

In some embodiments, the objective function is generated using a linear programming framework. For example, step 1104 may include generating an objective function of the form:

subject to Ax ≦ b, Hx ═ g

Where c is a cost vector, x is a decision matrix, A and b are (respectively) matrices and vectors describing inequality constraints on variables in the decision matrix x, and H and g are (respectively) matrices and vectors describing equality constraints on variables in the decision matrix x. In other embodiments, the objective function may be generated using any of a variety of other optimization frameworks (e.g., quadratic programming, linear fractional programming, non-linear programming, combinatorial algorithms, etc.).

In some embodiments, step 1104 includes formulating the decision matrix x. The load across each sub-facility 12-22 may be a decision variable in the decision matrix x. For example, for a central facility that includes a chiller, a heat recovery chiller, a hot water generator, and a thermal energy store, step 1104 may include formulating the decision matrix x as:

wherein the content of the first and second substances,

and

respectively, are n-dimensional vectors representing the thermal energy loads assigned to the chiller sub-facility 16, the heat recovery chiller sub-facility 14, the heater sub-facility 12, the hot TES sub-facility 20, and the cold TES sub-facility 22 at each of the n time steps within the optimization period.

In some embodiments, step 1104 includes generating a decision matrix x to include one or more decision vectors representing resource consumption for each sub-facility. For example, for a central facility that includes a chiller sub-facility, step 1104 may include generating a decision matrix x as follows:

wherein u is_{Chiller,elec,1…n}And u_{Chiller,water,1…n}Respectively, are n-dimensional vectors representing the power consumption and water consumption of the chiller sub-facilities at each time step k.

Step 1104 may include adding one or more resource consumption vectors to the matrix x for each of the sub-facilities 12-22. The decision vector for a given sub-facility added in step 1104 may represent the resource consumption of each resource (e.g., water, electricity, natural gas, etc.) consumed by that sub-facility at each time step k within the optimization period. For example, if the heater sub-facilities consume natural gas, electricity, and water, step 1104 may include increasing a decision vector u representing the amount of natural gas consumed by the heater sub-facilities at each time step_{Heater,gas,1…n}A decision vector u representing the amount of power consumed by the heater sub-facility at each time step_{Heater,elec,1…n}And a decision vector u representing the amount of water consumed by the heater sub-facility at each time step_{Heater,water,1…n}. Step 1104 may include adding resource consumption vectors for other sub-facilities in a similar manner.

In some embodiments, step 1104 includes generating a cost vector c. Generating the cost vector c may include increasing an economic cost associated with resource consumption required to generate the sub-facility load. For example, providing a decision matrix x as above, step 1104 may include generating a cost vector c as follows:

wherein, 0_nIs to indicate at each time step

Of an n-dimensional zero vector with a direct economic cost of zero, c_elec,1…nIs to indicateN-dimensional vector of cost per unit of electricity at each time step, and c_water,1…nIs an n-dimensional vector indicating the cost per unit of water at each timescale. The cost vector associates the economic cost with the resources consumed to generate the sub-facility load, rather than with the sub-facility load itself. In some embodiments, c_elec,1…nAnd c_water,1…nIs the utility rate obtained from the utility rate data received in step 1102.

In some embodiments, step 1104 includes generating a matrix and b vectors describing inequality constraints, and H matrix and g vectors describing equality constraints. Inequality constraints and equality constraints may be generated by inequality constraint module 146 and equality constraint module 148, as described with reference to FIG. 4. For example, step 1104 may include generating inequality constraints that constrain the decision variables in matrix x to be less than or equal to the maximum capacity of the corresponding central facility equipment, and less than or equal to the maximum charge/discharge rate of the thermal energy storage. Step 1104 may include generating inequality constraints that prevent charging thermal energy storage above maximum capacity and/or prevent discharging thermal energy storage below zero. Step 1104 may include generating an equality constraint that ensures that the building energy load is satisfied at each time step in the prediction window.

In some embodiments, step 1104 includes modifying the objective function to account for under-load (e.g., as described with reference to the under-load module 150), to account for heat intake or heat removal by the ground circuit (e.g., as described with reference to the ground circuit module 152), to account for heat exchange between the hot water circuit and the condensate water circuit (e.g., as described with reference to the heat exchanger module 154), to account for sub-facility curves that are not simple linear functions of load (e.g., as described with reference to the sub-facility curve module 170), and/or to force the thermal energy storage tank to be filled at the end of the prediction window (e.g., as described with reference to the tank refill module 160). The modified objective function may include a modified decision matrix x, a cost vector c, a and b vectors describing inequality constraints, and/or H and g vectors describing equality constraints.

Still referring to FIG. 11, process 1100 is illustrated as including modifying the objective function to account for demand charges (step 1106). Step 1106 is an optional step that may be performed by the demand charging module 156 to account for demand charges that may be enforced by the utility provider under certain pricing scenarios. Demand charging is an additional charge implemented by some utility providers based on the maximum energy consumption rate during the applicable demand charging period. For example, the demand charge may be provided in the form of dollars per unit of power (e.g., $/kW), and may be multiplied by the peak power usage during the demand charge period (e.g., kW) to calculate the demand charge.

Consideration of the demand charge may include modifying various components of the objective function, such as the decision matrix x, the cost function c, and/or the a matrix and b vector describing inequality constraints. The modified objective function may be defined as:

subject to Ax ≦ b, Hx ═ g

Wherein, c_demandIs the demand charge of the applicable demand charge period, and P_elec,kIs the total power consumption of the central facility and building/park at time step k. Term max (P)_elec,k) The peak electric power consumption at any time during the demand charge period is selected. Demand charges c_demandAnd the demand charge period may be defined by the utility rate information received in step 1102.

Step 1106 may include adding a new decision variable x_peakAnd the matrix x is modified as follows:

x_new＝[… u_{Chiller，elec，1…n}… u_{hpChiller，elec，1…n}… u_{Heater，elec，1…n}… x_peak]^T

wherein x is_peakIs the peak power consumption during the optimum period. Step 1106 may include revising the cost vector c as follows:

c_new＝[… c_elec，1…n… c_elec，1…n… c_elec，1…n… c_demand]^T

Step 1106 may include generating and/or applying inequality constraints to ensure peak power consumption x_peakGreater than or equal to the maximum electrical demand for each time step in the optimization time period. Namely:

range of

A＝[… [I_h]… [I_h]… [I_h]… -1]，b＝-P_{elec，campus，k}

step 1106 may include generating and/or applying inequality constraints to ensure peak power consumption decision variable x during demand charge periods_peakGreater than or equal to its previous value x_{peak,previous}. The inequality constraints can be represented in a linear optimization framework by defining the a matrix and b vector as follows:

A＝[… -1]，b＝-x_{peak，previous}

advantageously, the modification of the decision variable matrix x, the cost vector c and the inequality constraints in step 1106 may allow writing the objective function in the following linear form:

subject to Ax ≦ b, Hx ═ g

This linear form of the objective function can be used in a linear optimization framework.

In some embodiments, step 1106 includes applying a weighting factor to at least one of the consumption term and the demand charge term of the objective function. For example, the objective function written in the preceding equation has components over different time periods. Xiaoxiao (medicine for eliminating cough and asthma)Consumption c^Tx is over a consumption period and the demand charge term c_demandxpeakOver a demand charging period. To properly balance between increased demand charges and increased energy consumption costs, step 1106 may include applying a weighting factor to the demand charge term and/or the consumption term. For example, step 1106 may include consuming item c^Tx is divided by the duration h of the consumption period (i.e., the time period between the current time and the time range) and multiplied by the amount of time d remaining in the current demand charge period_demandSo that the entire objective function is over the demand charging period. The new optimization function can be given as follows:

subject to Ax ≦ b, Hx ═ g

It is equivalent to:

subject to Ax ≦ b, Hx ═ g

Still referring to FIG. 11, process 1100 is illustrated as including modifying the objective function to account for load change penalties (step 1108). Step 1108 is an optional step that may be performed by load change penalty module 158 to account for the cost of changing the load assigned to each sub-facility. In some instances, the lowest cost solution from a resource consumption perspective may involve changing the sub-facility from disconnected to fully loaded and back again within only a few time steps. However, running a central facility in this manner may be undesirable due to various negative effects of rapidly changing sub-facility loads (e.g., increased equipment degradation), especially when the cost saved is relatively minimal (e.g., cents or dollars).

Step 1108 may include modifying the objective function to introduce a penalty for rapidly changing sub-facility loads. In some embodiments, step 1108 includes modifying the decision matrix x by adding a new decision vector for each sub-facility. The new decision vector represents the sub-facility load change for each sub-facility from one time step to the next. For example, step 1108 may include modifying the decision matrix x as follows:

wherein, delta_{Chiller,1…n}、δ_{hrChiller,1…n}And delta_Heater,1…nRespectively, for each time step k, representing the time step k relative to the previous time step k-1And

of the sub-facility load.

Step 1108 may include modifying the cost vector c to increase the cost associated with changing the sub-facility load. For example, step 1108 may include modifying the cost vector c as follows:

step 1108 may include adding constraints such that each load change variable δ cannot be less than the corresponding sub-facility load

A change in (c). For example, the added constraints for a chiller sub-installation may have the form:

wherein the content of the first and second substances,

at a previous time step

The value of (c). Similar constraints may be added for each sub-facility 12-22. The constraints imposed by step 1108 may require that the load change variable δ be greater than or equal to the current value of the corresponding sub-facility loadWith previous value of sub-facility load

The magnitude of the difference between.

Still referring to FIG. 11, process 1100 is illustrated as including optimizing an objective function over an optimization period subject to a set of constraints to determine an optimal distribution of energy usage loads across a plurality of sets of central facility devices (step 1110). The set of constraints may include inequality constraints and equality constraints formulated in

steps

1104, 1106, and/or 1108. Optimizing the objective function may include determining an optimal decision matrix x that minimizes the cost function cTx. The optimal decision matrix x may correspond to minimizing (for each time step k within the optimization period) the high-level cost function J_HLIs optimized

As described with reference to fig. 3.

Step 1110 may include determining an optimal decision matrix using any of a variety of linear optimization techniques. For example, step 1110 may include solving an optimal decision matrix subject to optimization constraints using a basis exchange algorithm (e.g., a simplex algorithm, a crossbar-cross algorithm, etc.), an interior point algorithm (e.g., an ellipsoid algorithm, a projection algorithm, a path tracking algorithm, etc.), an overlay and fill algorithm, an integer programming algorithm (e.g., a cut facility algorithm, a branch-bounding algorithm, a branch-cutting algorithm, a branch-pricing algorithm, etc.), or any other type of linear optimization algorithm or technique. For embodiments using non-linear optimization, step 1110 may include solving the optimal decision matrix using any of a variety of non-linear optimization techniques. The result of step 1110 may be the optimal energy load distribution over multiple groups of sub-facility devices (i.e., multiple sub-facilities) for each time step k.

Still referring to FIG. 11, process 1100 is illustrated as including using the optimal energy load distribution to determine an optimal operating state for the various devices of the central facility equipment (step 1112). In some embodiments, step 1112 is performed by low-level optimization module 132, as described with reference to fig. 2-4. For example, step 1112 may include using the sub-facility loads determined in step 1110 to determine an optimal low-level decision for the central facility device(e.g., binary on/off decision, flow set point, temperature set point, etc.). In some embodiments, step 1112 is performed for each of a plurality of sub-facilities.

Step 1112 may include determining which devices of each sub-facility to use and/or the operating set points for those devices that will take the sub-facility load set points while minimizing energy consumption. The low-level optimization performed in step 1112 can be described using the following equation:

wherein the content of the first and second substances,

including optimal low-level decisions, and J_LLIs a low-level cost function.

To find optimal low-level decisionsStep 1112 may include minimizing a low-level cost function J_LL. Low level cost function J_LLMay represent the total energy consumption of all devices in the applicable sub-facility. Low level cost function J_LLThis can be described using the following equation:

where N is the number of devices in the sub-facility, t_sIs the duration of the time step, b_jIs a binary on/off decision (e.g., 0 off, 1 on), and u_jIs device j as a function of set point θ_LLThe energy used. Each device may have a continuous variable that may be varied to determine the lowest possible energy consumption for the total input conditions.

In some embodiments, step 1112 includes minimizing a low-level cost function J_LLLow level cost function J_LLSubject to inequality constraints based on the capabilities of the sub-facility devices and equality constraints based on energy and mass balances. In some embodiments, optimal low-level decisions

Subject to a switching constraint that defines a short time range for maintaining the device in an on or off state after a binary on/off switch. The switching constraints may prevent the device from cycling quickly between on and off.

Step 1112 may include determining an optimal operating state (e.g., on or off) for a plurality of devices of the central facility apparatus. According to an exemplary embodiment, the on/off combinations may be determined using binary optimization and quadratic compensation. Binary optimization may minimize a cost function representing the power consumption of devices in the applicable sub-facility. In some embodiments, a non-exhaustive (i.e., not all potential device combinations are considered) binary optimization is used. Quadratic compensation may be used when considering devices whose power consumption is quadratic (rather than linear).

Step 1112 can include utilizing non-linear optimization to determine an optimal operating set point for the plant. Nonlinear optimization may identify further minimization of the low-level cost function J_LLThe operating set point of (c). In some embodiments, step 1112 includes providing the on/off decisions and set points to the building automation system 108 for use in controlling the central facility device 60.

The construction and arrangement of the systems and methods as shown in the various exemplary embodiments are illustrative only. Although only a few embodiments have been described in detail in this disclosure, many variations are possible (e.g., variations in sizes, dimensions, structures, shapes and proportions of the various elements, values of parameters, mounting arrangements, use of materials, orientations, etc.). For example, the positions of elements may be reversed or otherwise varied, and the nature or number of discrete elements or positions may be altered or varied. Accordingly, all such variations are intended to be included within the scope of the present disclosure. The order or sequence of any process or method steps may be varied or re-sequenced according to alternative embodiments. Other substitutions, modifications, changes and omissions may be made in the design, operating conditions and arrangement of the exemplary embodiments without departing from the scope of the present disclosure.

The present disclosure contemplates methods, systems, and program products on a memory or other machine-readable medium for performing various operations. Embodiments of the present disclosure may be implemented using an existing computer processor, or by a special purpose computer processor included for this or other purposes for an appropriate system, or by a hardwired system. Embodiments within the scope of the present disclosure include program products or memory including machine-executable instructions or data structures carried on or having stored thereon. Such machine-readable media can be any available media that can be accessed by a general purpose or special purpose computer or other machine with a processor. By way of example, such machine-readable media can comprise RAM, ROM, EPROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to carry or store desired program code in the form of machine-executable instructions or data structures and which can be accessed by a general purpose or special purpose computer or other machine with a processor. Combinations of the above are also included within the scope of machine-readable media. Machine-executable instructions comprise, for example, instructions and data which cause a general purpose computer, special purpose computer, or special purpose processing machines to perform a certain function or group of functions.

Although the figures may implement a particular order of method steps, the order of steps may differ from that shown in the figures. Two or more steps may likewise be performed simultaneously or partially simultaneously. Such variations will depend on the software and hardware systems selected and design choices. All such variations are within the scope of the present disclosure. Similarly, software implementations can be realized with standard programming techniques with rule based logic and other logic to perform the various connection steps, processing steps, comparison steps and decision steps.

Claims

1. An optimization system of a central facility configured to provide services to building energy loads, the optimization system comprising:

a load/rate prediction module configured to predict the building energy load at a plurality of timesteps in an optimization period using feedback from a building automation system, the feedback from the building automation system including inputs from one or more sensors configured to monitor an environment within a controlled building;

processing circuitry configured to receive utility rate data indicative of prices of one or more resources consumed by equipment of the central facility to service the predicted building energy load at each of the plurality of timescales;

a high-level optimization module configured to generate an objective function representing a total financial cost of operating the central facility over the optimization period as a function of the utility rate data at each of the plurality of timescales and an amount of one or more resources consumed by equipment of the central facility; and

a low-level optimization module configured to generate a sub-facility curve for a sub-facility of the central facility by optimizing resource consumption of the sub-facility for a number of different combinations of thermal energy load and weather conditions, the sub-facility curve indicating a minimum amount of resources consumed by the sub-facility as a function of thermal energy load on the sub-facility;

wherein the high-level optimization module is configured to receive the sub-facility curves from the low-level optimization module, formulate a load equality constraint and optimize the objective function over the optimization period subject to the load equality constraint and a capacity constraint of the equipment of the central facility to determine an optimal distribution of the predicted energy load for the building over the plurality of sets of equipment of the central facility at each of the plurality of time steps, wherein the load equality constraint ensures that the optimal distribution meets the predicted energy load for the building at each of the plurality of time steps.

2. The optimization system of claim 1, wherein the high-level optimization module uses linear programming to generate and optimize the objective function.

3. The optimization system of claim 1, wherein the objective function comprises:

a cost vector including a cost variable representing a financial cost associated with each of the one or more resources consumed by equipment of the central facility to service the building energy load at each of the plurality of timescales; and

a decision matrix containing load variables representing energy load for each of the sets of equipment of the central facility at each of the plurality of timescales, wherein the high-level optimization module is configured to determine optimal values for the load variables in the decision matrix.

4. The optimization system of claim 1, wherein:

the central facility comprises a plurality of sub-facilities; and is

Each of the sets of equipment of the central facility corresponds to one of the plurality of sub-facilities.

5. The optimization system of claim 4, wherein:

the plurality of sub-facilities includes at least one of a hot thermal energy storage sub-facility and a cold thermal energy storage sub-facility; and is

The thermal energy storage sub-facility is configured to store thermal energy generated by one of the plurality of time steps for use in another of the plurality of time steps.

6. The optimization system of claim 4, wherein the high-level optimization module is configured to:

generating a sub-facility curve for each of the plurality of sub-facilities, wherein each sub-facility curve indicates a relationship between resource consumption and load production for one of the plurality of sub-facilities;

formulating a sub-facility curve constraint using the sub-facility curve; and

optimizing the objective function subject to the sub-facility curve constraints.

7. The optimization system of claim 6, wherein generating the sub-facility curve comprises at least one of:

converting a non-linear sub-facility curve to a linear sub-facility curve, the linear sub-facility curve comprising one or more segmented linear segments; and

converting the non-convex facility curve into a convex facility curve.

8. The optimization system of claim 6, wherein generating the sub-facility curve comprises:

receiving an initial sub-facility curve, the initial sub-facility curve based on manufacturer data for a set of devices corresponding to the sub-facility; and

updating the initial sub-facility curve using experimental data from the central facility.

9. A method for optimizing costs in a central facility configured to service building energy loads, the method comprising:

predicting building energy load at a plurality of timesteps in an optimization period using feedback from a building automation system, the feedback from the building automation system including inputs from one or more sensors configured to monitor an environment within a controlled building;

receiving, at processing circuitry of a central facility optimization system, utility rate data indicative of prices of one or more resources consumed by equipment of the central facility to service the predicted building energy load at each of the plurality of timescales;

generating, by a high-level optimization module of the central facility optimization system, an objective function representing a total financial cost of operating the central facility over the optimization period as a function of the utility rate data at each of the plurality of timescales and an amount of one or more resources consumed by equipment of the central facility;

generating, by a low-level optimization module of the central facility optimization system, a sub-facility curve for a sub-facility of the central facility by optimizing resource consumption of the sub-facility for a number of different combinations of thermal energy load and weather conditions, the sub-facility curve indicating a least amount of resources consumed by the sub-facility as a function of thermal energy load on the sub-facility;

receiving, by the high-level optimization module from the low-level optimization module, the sub-facility curves for the sub-facilities;

formulating load equality constraints by the high-level optimization module; and

optimizing, by the high-level optimization module, the objective function over the optimization period subject to load equality and capacity constraints on the equipment of the central facility to determine an optimal distribution of the predicted building energy load over the plurality of sets of equipment of the central facility at each of the plurality of timescales, wherein the load equality constraint ensures that the optimal distribution meets the predicted building energy load at each of the plurality of timescales.

10. The method of claim 9, wherein the high-level optimization module uses linear programming to generate and optimize the objective function.

11. The method of claim 9, wherein the objective function comprises:

a decision matrix containing load variables representing energy usage load for each of the plurality of sets of equipment of the central facility at each of the plurality of timescales, wherein optimizing the objective function comprises determining an optimal value for the load variables in the decision matrix.

12. The method of claim 9, further comprising:

generating the load equality constraint and the capacity constraint using the building energy load and capacity limits of equipment of the central facility;

wherein the capacity constraint ensures that equipment of the plurality of sets of the central facility operates within the capacity limit at each of the plurality of timescales.

13. The method of claim 9, wherein:

the central facility comprises a plurality of sub-facilities; and is

14. The method of claim 13, further comprising:

formulating a sub-facility curve constraint using the sub-facility curve; and

15. The method of claim 14, wherein generating the sub-facility curve comprises at least one of:

converting the non-convex facility curve into a convex facility curve.

16. The method of claim 9, wherein the central facility optimization system uses dynamic programming to divide the method for optimizing costs into a high-level optimization and a low-level optimization;

wherein the high-level optimization comprises determining an optimal allocation of the building energy load over the sets of equipment of the central facility; and is

Wherein the low-level optimization includes determining an optimal operating state of individual devices within each of the plurality of sets of equipment of the central facility.

17. The method of claim 16, wherein the optimal allocation of building energy load optimizes financial costs of operating the central facility over the optimization period; and is

Wherein the optimal operating condition optimizes an amount of energy consumed by each of the plurality of groups of equipment of the central facility to achieve an optimal distribution of the building energy load.