WO2018052175A1 - Real-time demand bidding method and system for energy management in discrete manufacturing facility - Google Patents

Abstract

The present invention provides a real-time demand bidding method and system for energy management in a discrete manufacturing facility. To this end, the present invention relates to a real-time demand bidding method for energy management of a discrete manufacturing system in which each machine processes an independent operation at a predetermined time interval, the method comprising the steps of: allowing the discrete manufacturing system to calculate a first maximum profit and a first optimum machine operation schedule indicating the operating or non-operating of each machine at a time interval in a state in which there is no real-time demand bidding event from a utility company supplying power; allowing the discrete manufacturing system to calculate a second maximum profit and a second optimum machine operation schedule for event attendance when a real-time demand bidding event is notified from the utility company; allowing the discrete manufacturing system to attend the real-time demand bidding event to bid for an optimum load reduction amount when the second maximum profit is greater than the first maximum profit; and allowing the discrete manufacturing system to operate each machine by the second optimum machine operation schedule when the bid is awarded from the utility company.

Description

Real-Time Demand Bidding Method and System for Energy Management in Discrete Manufacturing Facilities

The present invention is to provide a real-time demand bidding method and system for energy management in discrete manufacturing facilities.

Demand Bidding (DB) is the most efficient Demand Response (DR) method of smoothing the demand curve in the Smart Grid. DR induces a change in demand-side consumption, preventing grid instability and large voltage fluctuations caused by unbalanced power generation and consumption. DR can be classified into incentive-based DR and price-based DR. Incentive-based DR provides incentives for consumers by reducing the load during peak periods of the power system. Price-based DR (time-based DR) allows consumers to change their energy usage patterns in response to electricity prices that change over time.

Most of DR's previous research has focused on price-based DR in residential or commercial areas. However, 42.6% of the world's electricity is consumed by industry, so industrial sector DR is particularly important in reducing peak demand on power grids. The preceding paper cited barriers to the implementation of DR in industrial facilities, which classify them into two aspects. First, when modeling DR problems in the industrial sector, the machine-to-machine interdependence of the production line must be taken into account. For example, if the raw material feeding machine does not supply raw material, the machine cannot start operation. Modeling machine interdependence increases the complexity of mathematical models when compared to DR in residential or commercial areas. Second, there are significant changes in industrial processes and loads among other industrial applications. Therefore, it is difficult to design a standard model for industrial DR. Industrial DRs that exist today have application specificities, with specific models reported for industrial refrigeration systems, meat industry, cement industry, and food industry. All of these methods are manual DR, or manual decisions in response to DR events, which optimize the load reduction rules, actively adjust production and energy plans, or limit the scope of real-time load shedding.

In 2013, the refinery's automatic DR method was proposed based on the one-day real-time price (DA-RTP), where the interdependence of process constraints was considered. However, the production model does not take into account intermediate storage and is not practical in practice. In 2014, an automatic DR method was introduced that classified processing tasks into non-schedulable tasks and schedule-schedulable tasks. Here, the DR method was evaluated using oxygen production based on DA-RTP as an example. One limitation is that the mathematical model is tailored to continuous processes and does not take into account the order of operation of the machine. The current problem with DA-RTP is that if electricity prices fall within an hour, many consumers suddenly consume energy, leading to a new peak on the demand side. To overcome this problem, a cooperative DR approach has been proposed, including penalties mechanisms, to prevent non-cooperative behavior of consumers, which has been applied to industrial freezers.

There are three limitations to the current approach. First, most of the work is tailored to continuous processes, and there is no universal automated DR model for discrete manufacturing. Second, most industrial DR approaches consider day-ahead DR, but real-time DR has more potential to balance power grids due to production constraints and uncertainty in forecasting. Third, while all previous reports on industrial DR focusing on price-based DR are US statistics, it shows that incentive-based DR contributes to 93% of peak load, while price-based DR is only 7%. Using a comprehensive model based on psychology or economics, the researchers found that consumer responses to price-based DR and consumer responses to incentive-based DR differ significantly, and that incentive-based DR has a greater impact on consumer behavior. While there has been much research on relatively price-based DR, few studies on incentive-based DR. Therefore, extensive investigation of incentive-based DR with application for industrial consumers is required.

Demand Bidding (DB) is an important form of incentive-based DR. There are two types of demand in the wholesale and retail electricity markets as shown in FIG.

As shown in FIG. 1, an independent system operator (ISO) 40 mediates the supply of energy demand between the utility company 10 and the power generation company 30 in the wholesale electricity market. Demand bidding in the wholesale electricity market refers to a process in which the utility company 10 proposes demand bidding to the wholesale market to secure energy. In the retail market, demand bidding refers to incentive-based DR algorithms based on day-ahead prices, where utility companies offer incentives to large consumers (20) to suggest capacity reduction bids and execute load reductions. do. These two types of demand bids differ considerably in terms of purpose and scope of application.

The present invention was devised to solve the above problems, and an object of the present invention is to provide a method and system that enables a power consumer to quickly participate in a demand bid by being notified of a demand bid event in real time.

Another object of the present invention is to provide a method and system that can maximize the production benefit of a discrete manufacturing system participating in a demand bid event.

To this end, the real-time demand bidding method according to the present invention is a real-time demand bidding method for energy management of a discrete manufacturing system in which each machine processes an independent operation at a predetermined time interval, a real-time demand bidding event from a utility company supplying power Calculating, by the discrete manufacturing system, a first optimal machine operating schedule indicating whether each machine is operating at a first maximum benefit and a time interval in a non-existent state, and receiving a real-time demand bidding event from the utility company. The system calculates a second maximum benefit and a second optimal machine operation schedule for event participation, and if the second maximum benefit is greater than the first maximum benefit, the discrete manufacturing system participates in a real time demand bidding event to optimize load. Bidding with a reduction amount; If the bid is approved from the discrete manufacturing system includes the steps of operating each machine to the second optimum machine operation schedule.

In addition, the discrete manufacturing system according to the present invention includes a plurality of machines for processing independent operations at regular time intervals to output a result, a plurality of buffers for storing raw materials, intermediate materials or final output from each machine, and A discrete manufacturing system comprising a control device for controlling independent operation, wherein the control device indicates the operation of each machine at a first maximum profit and time interval in the absence of a real-time demand bid event from a utility company supplying power. Calculate a first optimal machine operation schedule, and when the real-time demand bidding event is notified from the utility company, calculate a second maximum benefit and a second optimal machine operation schedule for event participation, such that the second maximum benefit is the first maximum; If it is greater than the profit, bid with the optimal load reduction for the real-time demand bidding event, If a bid is approved from the utility company, the machines are operated according to the second optimal machine operation schedule.

In addition, the real-time automatic demand bidding method for energy management of the discrete manufacturing system according to the present invention, the utility company supplying power to the demand bidding event including the event time, the minimum amount of energy reduction and the incentive rate at a time interval of 1 hour or less Notifying a discrete manufacturing system; calculating, by the discrete manufacturing system, a maximum benefit for determining participation in a demand bid event in view of the event time, minimum amount of energy reduction, and an incentive rate, wherein the maximum benefit is a demand bid Participating in the demand bid event if the discrete manufacturing system is greater than the maximum benefit calculated before the event.

In addition, the automatic demand bidding system for energy management of the discrete manufacturing system according to the present invention is a utility company for notifying the demand bidding event including the event time, the minimum amount of energy reduction and the incentive rate at a time interval of 1 hour or less, and the event Calculating the maximum profit for determining participation in the demand bid event taking into account time, minimum amount of energy reduction, and incentive ratio, and if the maximum benefit is greater than the maximum benefit calculated before the demand bid event, the discrete participation in the demand bid event of the utility company Manufacturing system.

As described above, by the real-time automatic demand bidding method of the discrete manufacturing system according to the present invention, the utility company can prevent the uncertainty due to the prediction error and ensure the stability of the power supply in real time.

In addition, the consumer of the discrete manufacturing system can automatically demand demand only when the maximum profit is secured while adjusting the energy use schedule in real time, thereby maximizing the production profit of the discrete manufacturing system.

1 shows energy demand in the electricity wholesale and retail markets.

2 illustrates the interaction between a utility company and a consumer in a DA-DB.

Figure 3 illustrates the interaction between the utility company and the consumer in the RT-DB according to the present invention.

4 is a view showing the configuration of a discrete manufacturing system according to the present invention.

5 is a flowchart illustrating an automatic demand bidding process for energy management of a discrete manufacturing system according to the present invention.

6 is a view showing the configuration of a lithium-ion battery module presented as a case study of the present invention.

7 is a view showing the configuration of an assembly system of a lithium-ion battery module presented as a case study of the present invention.

8 illustrates a machine operation schedule when there is no RT-DB event.

9 is a graph showing electricity prices and energy consumption in the absence of an RT-DB event.

10 illustrates a machine operation schedule when there is an RT-DB event.

11 is a graph showing electricity prices and energy consumption when the incentive rate of RT-DB events is 0.54 $ / kWh.

12 is a graph showing electricity price and energy consumption when the incentive rate of RT-DB event is 0.45 $ / kWh.

13 is a graph showing profit according to incentive ratio.

14 is a graph showing load reduction according to an incentive ratio.

15 is a graph showing the yield according to the incentive ratio.

16 is a diagram showing a power consumption pattern of the machine in the discrete manufacturing system according to the present invention.

The present invention proposes a DB program meaning incentive-based DR in the retail market. The DB program according to the present invention is different from price-based DR because it is an incentive-based DR. Incentive-based DR may be a reward-based program or price-based DR may be a penalty-based program. Compensation-based programs are particularly attractive to industrial consumers because they can reduce costs and provide compensation due to reduced load. Recently, DA-DB was released by PG & E and Southern California Edison, but the DB for executing incentive-based DR is relatively new, so little has been reported about the database. DA-DB-related reports focus on hotel energy management and residential energy management.

The present invention discloses a real-time DB (RT-DB) program. The RT-DB program according to the present invention reduces the time used for notification one day ago to 10 minutes. This prevents uncertainty due to forecast errors, enabling real-time operation of demand-side energy consumption management and grid stability. Utility companies announce RT-DB during peak loads or grid emergencies, and industrial consumers decide whether to participate in RT-DB events and how much load reduction bids to maximize profits. When the utility company approves the bid, industrial consumers adjust their production and energy plans to reduce the load.

The automatic RT-DB according to the present invention is designed to perform optimal load reduction bidding decisions and active adjustment of production and energy planning. The automated RT-DB according to the present invention was developed based on a general discrete manufacturing production model. Optimization problems are formulated to maximize producer profits while minimizing energy costs. The optimization problem is transformed into a MILE (Mixed Integer Linear Programming) problem, and once the MILP problem is solved, the automatic RT-DB algorithm can adjust the production and energy plans to determine the optimal bid. The algorithm performance is evaluated using a lithium ion battery module manufacturing plant as an example. Simulation results show that the automatic RT-DB algorithm effectively reduces the load during RT-DB events while maximizing the manufacturer's benefits.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

First, we introduce the current DA-DB program and explain how it can be extended to RT-DB program. In general, the DA-DB program is a voluntary DR that provides an opportunity to receive incentives for load reduction during peak load periods.

2 illustrates the interaction between a utility company and a large consumer.

Referring to FIG. 2, for example, when a utility company 10 satisfies at least one of the following conditions at 12 noon the day before, a specific company may notify the consumer 20 of the DA-DB event.

(1) When it is expected that power generation resources of the next day are not enough

(2) When transmission or distribution reliability problem occurs on the following day

(3) The estimated load of the system operator's next day exceeds 43,000 MW.

(4) When the system operator is expected to issue a warning notice or the next day a warning or high level notice.

(5) When forecast temperature of the next day exceeds fixed value

DA-DB event notifications include the duration of a DA-DB event (e.g., 4-6 pm), a minimum load reduction required for participation (e.g., 10kW reduction per hour), an incentive rate (e.g., 1kW load per hour) $ 0.50 for each reduction).

After the DA-DB event announcement, the consumer decides whether to participate in the DA-DB program and how much load reduction bids to make to the utility company. Bidding can be made from 12:00 to 4:00 pm. The consumer receives a bid approval or rejection after 5 pm. The next day, if the grid's expected conditions are met and there is a problem with the stability of the grid, the utility company calls the registered bidder and asks them to reduce the load according to the load reduction bid contract. The consumer will receive a different amount of compensation for the actual load reduction. However, if the load reduction fails, there is no cost penalty. The load reduction at time t is given by Equation 1.

Here, CB (t) is a consumer default value, which is the average energy consumption for the corresponding time during the last 10 days. E (t) is the actual energy consumption. The consumer reward for time t is given by equation (2).

Where I (t) is the incentive ratio.

One of the obvious limitations of DA-DB programs is that their efficiency depends on grid prediction techniques. If an out of forecast load peak or grid emergency occurs, it is too late to generate a DA-DB event. In order to overcome this limitation and realize a DB program in near real time (ie, 10-15 minutes), we propose an RT-DB program according to the present invention.

Figure 3 illustrates the interaction between the utility company and the consumer in the RT-DB program according to the present invention.

Referring to FIG. 3, the RT-DB according to the present invention may reduce the notification time and the determination time to 10 minutes.

The 10 minute time period is sufficient for industrial consumers to determine the optimal load reduction bid in response to the RT-DB program by using the real-time communication capability, the automatic control system and the automatic RT-DB algorithm according to the present invention. In addition, the RT-DB algorithm according to the present invention can be applied to the conventional DA-DB through some modifications.

Discrete Manufacturing System Model

According to the present invention, a mathematical model for constructing a discrete manufacturing assembly facility is constructed. An assembly line is a production system that includes processing and parts merging.

4 shows the structure of a discrete manufacturing assembly system used in an actual manufacturing environment.

4, m _ij is a machine and i is the i th serial production line branch. j represents the jth machine in the jth serial production line branch. Machine m _{ij ,} i> 0 is a component machine. Machine m ₀₁ is an assembly machine. Machine m _ij , i = 0; j> 1 is a further processing machine for the assembled product. B _ij is a buffer for storing intermediate products provided by the machine m _ij .

In this discrete manufacturing assembly system, each machine m _ij is capable of schedule adjustment (ie on or off) on the time axis. Here, the time axis is divided into the same operation time interval. Between two consecutive operations of the machine, the length of the time interval is T _S. The length of the time interval is application specific. For other industrial applications, the length of the time interval can be 10 minutes, 15 minutes, 60 minutes, and the like. In the present invention, it is assumed that T _S = 10 minutes. If the time interval is too short, some machines may not complete the action process. However, if the time interval is too long, it may fail to respond to RT-DB events because the energy usage schedule cannot be adjusted in real time. The total number of time intervals on the schedule linear axis is denoted by S. In the case of T _S = 10 minutes in the 24-hour schedule linear axis, S is expressed by Equation 3 below.

Next, the assembly system model will be described. The assembly system model considers two models: the production model and the energy model.

A. Production Model

machine

i) Each machine m _ij has a fixed time to process a part and is called the cycle time CT _ij .

ii) In discrete manufacturing, each machine operates in impulse mode for maximum efficiency. That is, it operates at 100% when needed and 0% when not needed. It can be expressed by Equation 4 by the decision variable x _ij (t) of the machine m _ij in the t-th time interval.

buffer

Each buffer B _ij has a capacity CAP _ij .

Starvation Rule

i) buffers B _ij , i = 0, ..., S; If j = 1, ..., M _i -1 is empty, the machine m _{i (j + 1)} is depleted.

ii) The machines m _i1 , i = 1, ..., S are not depleted.

iii) The assembly machine m ₀₁ is exhausted if at least one of the buffers B _iM , i = 1, ..., S is empty.

iv) The depleted machine stops processing operations and enters a low power state to save energy.

Blockage Rule

i) buffers B _ij , i = 0, ..., S; j = 1, ..., M _i Is full, the machine m _ij is shut off.

ii) The blocked machine enters a low power state to stop the processing of the operation and to prevent the overflow of the buffer.

B. Production Model Constraints

For a given time interval, the machine m _ij processes the quantity n _ij of the various parts, as shown in equation (5).

The buffer storage at the end of the time interval t is given by Equations 6-8.

Where α _ij is the buffer B _ij It is a factor that indicates the number of parts required for the latter machine to produce one part.

Starvation Rule

During the time interval t, the buffer satisfies Equation 9, where N ⁰ is a natural set of zeros.

A time interval (t-1) last in the B _ij (t-1) is empty and without supplying the material to the machine m _{i (j + 1)} machine m _{i (j + 1)} of the exhausted during the time interval t Will be. That is, x _{i (j + 1)} (t) = 0. This rule can be expressed as:

This may be expressed as in Equation 10.

If B _ij (t-1) == 0, then x _{i (j + 1)} (t) = 0. If B _ij (t-1)> 0, then B _ij (t-1) ≧ 1 from equation (9). If B _ij (t-1) ≥1 is _{x i (j + 1) (} t) ≤B ij (t-1) from Equation (10). From which, based on the equation _{4 x i (j + 1)} (t) ∈ {0,1} may determine the value _{x i (j + 1) (} t).

At the end of the time interval (t-1), if B _ij (t-1), (i = 1, ..., S; j = M _i ) is empty and the material cannot be provided to the assembly machine m ₀₁ , The assembly machine m ₀₁ will be depleted for the time interval t. That is, x ₀₁ (t) = 0. This law can be expressed as

This can be expressed as in Equation (11).

Blockage Rule

During the time interval t, the buffer is a condition as in Equation 12.

At the end of the time interval (t-1),

Is full, the machine m _ij is shut off. That is, x _ij (t) = 0. This law can be expressed as

This can be expressed as in Equation 13.

C. Energy Model

The power consumption pattern of a typical machine in a discrete manufacturing facility is shown in FIG. 16.

16 shows a typical power consumption pattern of the machine. Initially, the machine starts operating (①) and enters a low power state (②). If the machine is scheduled to process the workpiece during the first hour, the operation rises (③) and enters the operating state (④). In this operating state, the machine repeatedly processes the workpiece. The average time for processing the workpiece is called the cycle time. During the second time the machine goes down (⑤) and enters a low power state (②) if it is not scheduled to process the workpiece. In the present invention, the first time is referred to as "machine on" and the second time is referred to as "machine off". During the "machine on" time interval, the total electrical consumption of the machine can be measured in Eon. During the "machine off" time interval, the overall consumption of the machine can be measured in Eoff. The right part of FIG. 16 shows equal power consumption, which is the same as the gray part on the left.

If machine m _ij is the "on" schedule for time interval t (x _ij (t) = 1), then the electricity consumption per time interval of machine m _ij is Eon _ij . If the machine m _ij is an "off" schedule for the time interval t (x _ij (t) = 0), it enters the low power state and retains only the basic functions, at which time the electricity consumption per time interval of the machine m _ij is Eoff _ij . The energy consumption of the machine m _ij for the time interval t

to be. In this formula, if x _ij (t) = 1 (machine m _ij "on"), then Eon _ij , and if x _ij (t) = 0 (machine m _ij "off"), then Eoff _ij .

The total energy consumption of all machines in the production line over time interval t

to be. The total electricity cost of the production line over time interval t

to be. Where p (t) is the electrical value for the time interval t. The total electricity cost in the schedule linear axis (T)

to be.

D. Energy Model Constraints

The total energy consumed by the plant during the time interval t is kept below the limit E _max . That is, as shown in equation (14).

Where E _max > 0 is determined by the limitation of the local power transmission line.

Optimization Problems and RT-DB Algorithms

A. Formulation of Optimization Problems

Microeconomic theory, like raw material consumers, increases demand to the point where the profits that electric consumers can derive are equal to the additional costs for this. For example, a producer will not produce a product if it does not make a profit on the sale due to the cost of electricity. In other words, the plant will decide whether or not to produce in a way that maximizes profits and when and how. The optimization objective of the plant used here is to maximize profits.

Profit is gross income minus cost. Suppose the factory produces _n output units (u ₁ , ... u _n ) and uses _m input units (n1, ... n _m ). If the value of the output product is (v ₁ , ... v _n ) and the input cost is (c1, ... c _m ), the profit π that the company gets is equal to (15).

here,

Is the total income of the output product,

Is the total input cost. The production output is priced at the market price. Input costs include all costs associated with production, including raw materials, energy and labor costs.

Assume that the factory participates in the RT-DB program. If there is no RT-DB event, the optimization objective function is shown in Equation 16.

The first term is the amount of imports, the second term is the input cost excluding the cost of electricity, and the third term is the cost of electricity.

This optimization problem follows the constraints of Equations 4-14. These constraints and the optimization objective function of Equation 16 create a MILE (Mixed Integer Linear Programming) problem. Through the MILP solution, the objective (ie, benefit) can be maximized using the decision variable x _ij (t).

When the RT-DB event occurs and the factory participates in the RT-DB event, the optimization objective function becomes as shown in Equation 17.

The first to third terms are the same as in Equation 16, and the fourth term represents an incentive for the RT-DB event. here,

Is the load reduction during RT-DB event T ',

Is the default load,

Is the energy consumption that is scheduled during the RT-DB event.

Since the RT-DB event also specifies a minimum load reduction LR _min , the constraint is not equal to Equations 4-14 and is equal to Equation 18.

The constraints of equations (4) to (14) and (18) and the optimization objective function of equation (17) create a MILE (Mixed Integer Linear Programming) problem. MILP solution reduces load to bid with decision variable x _ij (t) during RT-DB event

Optimize the optimization objective function to maximize the benefit by changing.

B. Automatic RT-DB Algorithm

Based on the mathematical model described above, the present invention proposes an automatic RT-DB algorithm as shown in FIG. The algorithm according to the present invention has been implemented using a Manufacturing Operations Management System (MOMS), which allows industrial consumers to decide whether to participate in RT-DB events and submit optimal reduction capacity bids and adjustments to production and energy plans to utility companies. Make sure The algorithm according to the present invention allows industrial consumers to intelligently interact with utility companies in response to RT-DB events, as well as actively generate optimal operating schedules for machines that maximize profits.

3 described above illustrates the interaction between the utility company 10 and the consumer 20 in the RT-DB event. The consumer 20 of FIG. 3 is discrete when explaining the automatic RT-DB algorithm shown in FIG. It becomes a manufacturing system. In addition, the subject processing the RT-DB event by specifically executing the RT-DB algorithm may be a control device of a utility company 10 and a discrete manufacturing system, a computer or a hardware device, but for convenience of description, the subject is discrete. It is set as a manufacturing system.

In FIG. 5, at the start of the day, the discrete manufacturing system solves the MILP problem (Equations 4-14, 16) in the absence of an RT-DB event, with the largest objective function value π ₁ (maximum benefit) and the set of decision variables S ₁ (24-hour optimal machine operation schedule) is obtained (S10).

The discrete manufacturing system checks whether the RT-DB event is notified from the utility company (S12), and if the RT-DB event is notified from the utility company, the discrete manufacturing system solves the MILP problem (Equations 4-14, 17-18). , The largest objective function value π ₂ (maximum gain), the set of decision variables S ₂ , and the optimum load reduction amount LR (S14).

The discrete manufacturing system then compares the maximum benefit π ₂ obtained for participation in the RT-DB event with the maximum benefit π ₁ obtained without the RT-DB event (S16), if π ₂ is greater than π _1. In order to participate in the bidding to the utility company to submit the optimal load reduction amount LR (S18).

The discrete manufacturing system checks whether the bid is approved from the utility company (S20), and if the bid is approved, operates the production system according to the decision variable set S ₂ (optimal machine operation schedule) (S22).

If an RT-DB event is not notified from the utility company, π ₂ is less than π ₁ , or the bid is not approved, the discrete manufacturing system will ensure that the production system operates according to the decision variable set S ₁ (optimal machine operating schedule). (S24)

Next, the discrete manufacturing system checks whether the day is over (S26), terminates the RT-DB algorithm if the day is over, and checks the notification of the RT-DB event every predetermined time (10 minutes) if the day is not over. .

simulation

The present invention presents a modeling for a discrete manufacturing system that takes part in an RT-DB event assuming a large capacity lithium-ion manufacturing plant.

A. Case Study: Lithium-Ion Battery Manufacturing

Large-capacity lithium-ion battery manufacturing is one example of typical discrete manufacturing, with four manufacturing processes: assembly, implantation, forming, and grading.

Assembly: As shown in FIG. 6, the battery module includes a plurality of battery cells A, an auxiliary component, that is, an intermediate frame B, a cooling fin C, a compression foam D, and a frame E. It has a hierarchical structure composed of). These components are assembled and coupled into the battery module through continuous operation.

Injection: The cell is filled with electrolyte and sealed.

Formation: The cell is cycled through at least once precisely controlled charging or discharging to activate and transform the working material into usable form.

Ratings: Ratings of battery modules are rated according to their performance based on discharge, resistance and capacity measurements.

This process can be categorized into 10 tasks, and the appropriate equipment is assigned to each task as shown in Table 1.

Task	Task description	Machine
One	Assembly of components A & D		m 11
2	Assembly of components (AD) & A	m 12
3	Assembly of components B & (ADA) & C		m 13
4	Assembly of components E & A	m 21
5	Assembly of components (EA) & C		m 22
6	Assembly of components A & E		m 31
7	Assembly of (EAC) & (BADAC) & (AE)	m 01
8	Filling		m 02
9	Formation		m 03
10	Grading	m 04

Fig. 7 shows the layout of the battery module assembly system together with the power, cycle time and buffer capacity of each machine. In the low power state, each machine has a power of 0.5W. Some machines are arranged in parallel to increase the production capacity of some parts of the process. Table 2 lists raw material and labor costs as well as the market price of the output.

Item	Unit price ($)
Battery cell (A)	20
Intermediate frames (B)	2.5
Cooling fins (C)	0.5
Compression foams (D)	0.6
Frame (E)	3.5
Labor cost for each battery module	26
Large-capacity lithium-ion battery module	121

The other parameters used in this case study are listed in Table 3.

Parameter	Description	Value
E max	Maximum power drawn by the factory	500 kW
p (t)	Hourly electricity price	DA-RTP data

B. Results

Lingo 12.0, a MILP solution, is used to solve the optimization problem according to the present invention. First, a scenario without an RT-DB event will be described.

Scenario 1: No RT-DB Event

At the beginning of the day, the MILP problem represented by Equations 4-14 and 16 should be solved to obtain an optimized 24-hour machine operation schedule, as shown in FIG. In FIG. 8, the horizontal axis represents the schedule time for 24 hours. In this case study, there are 10 time slots, so there are six time slots at each time. The vertical axis represents all the machines in the production line. The gray box here means that the machine is "on" in that time slot. A white box means that the machine is "off". The energy consumption of the production line per hour according to this schedule is shown in FIG. 9. Energy consumption was lower during the peak time (14-17th time). This reduces grid load during peak load times and saves consumer electrical costs. 24 hour production and maximum profits based on this optimized production schedule are listed in the second column of Table 4.

	Scenario 1	Scenario 2	Scenario 3
Profit ($)	2776.86	2780.51	2776.10
Production volume	828 units	786 units	822 units
Load reduction bid (kWh)	-	16th hour: 21217th hour: 35.3	16th hour: 22.517th hour: 12.8
Join the RT-DB event	-	Yes	No

Scenario 2: RT-DB event at 16-17th hour (incentive rate 0.54 $ / kWh)

Now consider the RT-DB event. When the 16th time begins, an RT-DB event occurs. The RT-DB event notification includes three elements: RT-DB event time (16th-17th time), minimum load reduction (10kW) to participate in the RT-DB program, and incentive rate ($ 0.54 / kWh). Even if the MILP problem is solved according to Equations 4-14 and 17-18, a new optimal schedule is obtained during the 16-24th time as in 10. During the RT-DB event (ie 16th-17th hours), most machines will shut down to provide load reduction. The hourly energy consumption of the production line from the new schedule is calculated as in FIG. Red line indicates participation in RT-DB event. During the 16th to 17th hours, the load was significantly reduced compared to the original energy plan indicated by the blue line. The new optimal load reduction bids, output and profits are listed in the third column of Table 4. The profit of the new schedule was $ 2,780.51, which is greater than the original schedule of $ 2,776.86. The plant will have increased profits and will have to attend RT-DB events with a new schedule.

Scenario 3: RT-DB event with incentive rate 0.45 $ / kWh at 16-17th hour

Scenario 3 is identical to scenario 2 except that the incentive rate is low. The optimal operation plan of the machine is obtained identically and the energy consumption per hour is shown in FIG. It can be seen from FIG. 12 that the energy consumption is slightly reduced in the new schedule during the RT-DB event. The last column of Table 4 lists the load reduction bids, new output, and profits. In scenario 3, the industrial consumer should not attend the RT-DB event because the profit of the rescheduled plan ($ 2,776.10) is less than the original plan's profit ($ 2,776.86).

To determine the relationship between incentive rates and industrial consumer responses, additional simulations were performed with incentive rates varying from 0.35-0.7 $ / kWh in 0.05 $ / kWh steps.

FIG. 13 shows the benefit for rebalancing energy for RT-DB event participation as a function of incentive rate. The greater the incentive rate, the greater the profit from attending RT-DB events. As indicated by the red dots, when the incentive rate is below $ 0.51 / kWh, the benefits of participating in the RT-DB are less than the benefits of the original energy plan ($ 2,776.86). Therefore, attending RT-DB events is uneconomical.

14 shows the reduction in the load of industrial consumers as a function of incentive ratio. When the incentive rate is below $ 0.51 / kWh, attending the RT-DB event is uneconomical and the consumer must follow the original energy plan, which means there is no load reduction. When the incentive rate exceeds $ 0.51 / kWh, industrial consumers must attend RT-DB events and the load reduction increases until the limit is reached without further reduction capacity for the 16th to 17th hours as a function of the incentive rate. something to do.

15 shows yield as a function of incentive ratio. When the incentive rate is below $ 0.51 / kWh, production is unchanged compared to the original energy plan (828 units) listed in the second column of Table 4, so industrial consumers should not participate in RT-DB events. When the incentive rate exceeds $ 0.51 / kWh, industrial consumers will have to attend RT-DB events, where production will be reduced. If the incentive rate is greater than $ 0.55 / kWh, the readjusted load will be reduced to a minimum during the 16th to 17th hours, which will no longer reduce production.

Table 5 shows the computation time needed to solve optimization problems using Lingo 12.0 software on the PC.

	Solve the optimization problem without theRT-DB event	Solve the optimization problem with an RT-DB event
Computing time (s)	56	Average: 91.4 (Max: 119; Min: 21)

The time required to solve the optimization problem when there was no RT-DB event was 56 seconds, and the average time required to solve the optimization problem when there was an RT-DB event was 91.4 seconds. This shows that the mathematical problem can be solved in a time appropriate to satisfy the real-time needs of the RT-DB program (ie 10 minutes notice).

So far we have discussed the discrete manufacturing production model and the automated RT-DB algorithm. Optimization problems are formulated for the purpose of maximizing manufacturing profits. By solving these optimizations, we run an automated RT-DB algorithm to determine the optimal load reduction bids and generate a coordinated production and energy plan. The performance of the algorithm was verified through a case study of a lithium-ion battery module manufacturing plant. Simulation results show that the algorithm according to the present invention efficiently reduces the load during RT-DB events while maximizing the manufacturer's benefit. The present invention can be considered as a win-win strategy for both electric users and utility companies. It also explained the relationship between incentive rates and consumer demand elasticity, including analysis of changes in output and profits.

On the other hand, the real-time demand bidding method according to the present invention can be implemented in a software program and recorded in a computer-readable predetermined recording medium.

For example, the recording medium may be a hard disk, a flash memory, a RAM, a ROM, or the like as an internal type of each playback device, or an optical disc such as a CD-R or a CD-RW, a compact flash card, a smart media, a memory stick, or a multimedia card as an external type. have.

Functional operations and implementations described in the specification of the present invention may be implemented in digital electronic circuitry, computer software, firmware or hardware, or a combination of one or more of them. Implementations described in the specification of the present invention are one or more modules relating to computer program instructions encoded on a program storage medium of tangible type for controlling or by the operation of one or more computer program products, ie data processing apparatuses. It can be implemented as.

While the drawings of the present invention describe an operational process, it should not be understood that such operations must be performed in the specific order shown or that all illustrated operations must be performed in order to obtain desirable results. In certain cases, multitasking and parallel processing may be advantageous.

The above description is merely illustrative of the present invention, and various modifications may be made by those skilled in the art without departing from the technical spirit of the present invention.

Therefore, the embodiments disclosed in the specification of the present invention are not intended to limit the present invention. The scope of the present invention should be construed by the claims below, and all techniques within the scope equivalent thereto will be construed as being included in the scope of the present invention.

Claims (21)

1. In the real-time demand bidding method for energy management of a discrete manufacturing system in which each machine processes independent operations at regular time intervals,

   Calculating, by the discrete manufacturing system, a first optimal machine operation schedule indicating the operation of each machine at a first maximum benefit and a time interval in the absence of a real-time demand bid event from a utility company supplying power;

   When the real-time demand bid event is notified from the utility company, the discrete manufacturing system calculates a second maximum benefit and a second optimal machine operation schedule for the event participation;

   If the second maximum profit is greater than the first maximum profit, the discrete manufacturing system participates in a real-time demand bidding event and bids at an optimal load reduction amount;

   And if the bid is approved by the utility company, the discrete manufacturing system including operating each machine with the second optimal machine operation schedule.
2. The method of claim 1,

   And calculating the first maximum benefit by the discrete manufacturing system to calculate the first maximum benefit such that an input cost including raw material cost and labor cost is subtracted from the total income, and the electric cost is maximized.
3. The method of claim 2,

   The optimum value of the electricity cost is determined by the electricity price in the time interval and the first optimum machine operation schedule.
4. The method of claim 1,

   The calculating of the second maximum profit by the discrete manufacturing system may include the input cost including the raw material cost and the labor cost minus the electric cost and the incentive according to the amount of energy reduction to the maximum, thereby increasing the second maximum benefit. The method for calculating the.
5. The method of claim 4, wherein

   The optimum value of the electricity cost is determined by the electricity price at a time interval and the second optimum machine operation schedule.
6. The method of claim 4, wherein

   The optimum value (optimum load reduction amount) of the energy reduction amount is a value obtained by subtracting the energy consumption adjusted by the second optimum machine operation schedule from the basic energy usage amount of the discrete manufacturing system.
7. The method of claim 1,

   And the energy consumed by the optimal machine operation schedule is determined by determinants indicative of the energy consumption of the machine and the presence or absence of operation of each machine during the time interval.
8. The method of claim 1,

   If there is no real-time demand bid event from the utility company or if the second maximum profit is less than the first maximum profit or the utility company does not approve the bid, the discrete manufacturing system does not participate in the real-time demand bid event and the first Operating each machine with an optimal machine operation schedule.
9. A plurality of machines for processing independent operations at regular intervals and outputting a result, a plurality of buffers for storing raw materials, intermediates or final products from each machine, and a control device for controlling the independent operation of each machine. In discrete manufacturing systems,

   The control unit calculates a first optimal machine operation schedule indicating the operation of each machine at a first maximum profit and a time interval in the absence of a real time demand bidding event from a utility company supplying power, and a real time demand from the utility company. When a bid event is notified, a second maximum profit and a second optimal machine operation schedule are calculated to participate in the event, and when the second maximum profit is greater than the first maximum profit, the bid is made at an optimal load reduction amount in a real-time demand bid event. If the bid is approved from the utility company,

   Wherein each machine operates according to the second optimal machine operation schedule.
10. The method of claim 9,

    If there is no real time demand bid event from the utility company or the second maximum profit is less than the first maximum profit or the utility company does not approve the bid, the control device does not participate in the real time demand bid event and the first Discrete manufacturing system characterized by operating each machine with an optimal machine operation schedule.
11. The method of claim 9,

    Wherein each machine stops operating and transitions to a low power state when the previously located buffer is empty or the next located buffer is full.
12. The method of claim 9,

    The assembly machine for gathering multiple outputs from among the plurality of machines to create a composite, if any one of the plurality of previously located buffer is empty, stop the operation and switch to a low power state.
13. Notifying the discrete manufacturing system of demand bid events, including event time, minimum amount of energy reduction, and incentive rates, at utility time intervals of one hour or less;

    Calculating, by the discrete manufacturing system, a maximum benefit for determining participation in a demand bid event in view of the event time, the minimum amount of energy reduction, and the incentive rate;

    And if the maximum profit is greater than the maximum profit calculated before the demand bid event, the discrete manufacturing system participating in the demand bid event.
14. The method of claim 13,

    Real-time automatic demand bidding method for energy management of a discrete manufacturing system, characterized in that the maximum profit before the demand bidding event is calculated from the following equation.

    

    Where the first term is the revenue, the second term is the input cost minus the electricity cost, and the third term is the electricity cost.
15. The method of claim 13,

    Real-time automatic demand bidding method for energy management of a discrete manufacturing system, characterized in that the maximum profit for determining the participation in the demand bidding event is calculated from the following equation.

    

    Where the first term is the revenue, the second term is the input cost minus the electricity cost, the third term is the electricity cost, and the fourth term is the incentive for the demand bid event,

    

    Is the amount of energy reduction during demand bid event T ',

    

    Is the basic amount of energy,

    

    Is the energy consumption that is scheduled during the demand bid event.
16. A utility company that notifies demand bidding events, including event time, minimum energy savings, and incentive rates, at intervals of one hour or less,

    Calculating the maximum profit for determining participation in the demand bidding event in consideration of the event time, the minimum amount of energy reduction, and the incentive ratio, and if the maximum benefit is greater than the maximum profit calculated before the demand bidding event, participate in the demand bid event of the utility company An automatic demand bid system for energy management of a discrete manufacturing system comprising a discrete manufacturing system.
17. The method of claim 16,

    The discrete manufacturing system includes a plurality of machines for processing independent operations at regular time intervals and outputting a result, and a plurality of buffers for storing raw materials, intermediate materials or final output from each machine. Automatic demand bidding system for energy management.
18. The method of claim 16,

    The discrete manufacturing system calculates a maximum profit for determining participation in a demand bidding event by maximizing a sum of input costs including raw material costs and labor costs and incentives according to the amount of energy reduction from the total imports. Automatic demand bidding system for energy management of discrete manufacturing system, characterized in that.
19. The method of claim 18,

    The discrete manufacturing system is an automatic demand bidding system for energy management of the discrete manufacturing system, characterized in that for determining the optimum amount of energy reduction so that the maximum benefit is calculated by adjusting the machine operation schedule indicating the operation of the machine.
20. The method of claim 19,

    And the discrete manufacturing system operates according to the adjusted machine operation schedule when a bid is approved from the utility company in response to the demand bidding event of the utility company at the optimal amount of energy reduction.
21. A computer-readable recording medium having recorded thereon a program for executing the method according to any one of claims 1 to 8.

Images