CN113360289B

CN113360289B - Edge calculation-based distributed collaborative optimization method for multi-region comprehensive energy system

Info

Publication number: CN113360289B
Application number: CN202110774332.5A
Authority: CN
Inventors: 赵浩然; 王梦雪; 田航
Original assignee: Shandong University
Current assignee: Shandong University
Priority date: 2021-07-08
Filing date: 2021-07-08
Publication date: 2022-06-10
Anticipated expiration: 2041-07-08
Also published as: CN113360289A

Abstract

The utility model provides a distributed collaborative optimization scheduling method of an edge-computing multi-region comprehensive energy system, wherein each edge computing unit obtains coupling information and sends the coupling information to a cloud computing center according to the acquired parameter data of the edge computing unit and with the minimum economic cost as an optimization target; the cloud computing center updates the consensus information and the Lagrangian multiplier according to the received coupling information and sends the consensus information and the Lagrangian multiplier to each edge computing unit; when the residual error is smaller than a preset value, taking the integer variable obtained by calculation of each edge calculation unit as an initial value, and combining a nested sub-algorithm to obtain an optimized control variable of each edge calculation unit; the method and the device protect the information privacy security of the RIES, greatly reduce the data transmission, storage and processing pressure of cloud computing through edge computing, and solve the problem that mass data is difficult to process in a centralized mode in IES.

Description

Edge calculation-based distributed collaborative optimization method for multi-region comprehensive energy system

Technical Field

The disclosure relates to the technical field of intelligent scheduling of integrated energy systems, in particular to a distributed collaborative optimization method of a multi-region integrated energy system based on edge computing.

Background

The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.

An Integrated Energy System (IES) refers to an Energy supply System based on advanced power electronic control technology and advanced communication acquisition technology, and has the operating characteristics of high flexibility, high toughness and the like. The comprehensive energy system is divided into three types of structural layers: trans-Regional level Integrated Energy systems, Regional Integrated Energy Systems (RIES), and campus level Integrated Energy systems. Remote high-capacity equipment such as a power transmission network, a natural gas backbone network and the like are main component equipment of a trans-regional comprehensive energy system; the medium-voltage distribution network, the gas distribution network and the heat distribution network form a regional comprehensive energy system; the park level comprehensive energy system is positioned in the user side energy management in a small range so as to realize the functions of an intelligent power utilization system, distributed heat and water supply, demand side management and the like.

A Multi-Regional Integrated Energy System (MRIES) refers to an Integrated Energy System composed of multiple Regional-level Integrated Energy systems, and the range is between Regional-level Integrated Energy and trans-Regional Integrated Energy systems.

Edge computing refers to an open platform integrating network, computing, storage and application core capabilities at one side close to physical equipment or a data source, and provides nearest-end services nearby. The application program is initiated at the edge side, so that a faster network service response is generated, and the basic requirements of the industry in the aspects of real-time business, application intelligence, safety, privacy protection and the like are met. The edge computation is between the physical entity and the industrial connection, or on top of the physical entity. And the cloud computing still can access the historical data of the edge computing. An Edge Computing Unit (ECU) is an energy data management Unit corresponding to a cloud Computing center in the IES, and can migrate part of tasks of system Computing to local to complete, so as to solve the problem of insufficient cloud Computing resources.

Distributed optimization refers to the task that distributed optimization effectively achieves optimization through cooperative coordination among multiple agents, and can be used for solving the large-scale complex optimization problem that many centralized algorithms are insufficient.

Energy Hubs (EHs) are important components of IES, accommodating the input of various forms of Energy and diverse load types. Optimizing the type and capacity of the equipment configuring the energy hub is the basis for ensuring the safe and economic operation of the energy hub.

The inventor finds that the current stage MRIES distributed collaborative optimization has the following problems:

(1) the decoupling method of the system according to different energy networks weakens the function of the RIES in energy management and allocation, and complicates the increase of the variety and the number of the coupled equipment;

(2) the distributed optimization part with the EH as the main body does not consider the upper and lower energy network constraints, and the part cannot directly obtain the global optimal solution through information interaction between the main bodies;

(3) the corresponding distributed algorithm considering the EH variable condition model is lacked to support the large-scale expansion of the multi-subject IES.

Disclosure of Invention

In order to solve the defects of the prior art, the distributed collaborative optimization method of the multi-region comprehensive energy system based on edge calculation is provided, and a set of distributed collaborative scheduling framework suitable for MRIES large-scale extended calculation is provided; in the aspect of a model, an IES partitioning method matched with the framework is designed, an edge computing unit is constructed, the characteristics of distributed computing are met, and a model basis is provided for distributed optimization computing; in the aspect of the algorithm, a distributed computing algorithm supporting large-scale expansion is provided, global optimization is supported, the problem of MILP non-convergence introduced during fine modeling is solved, the convergence speed of the algorithm is increased, and distributed computing is accelerated.

In order to achieve the purpose, the following technical scheme is adopted in the disclosure:

in a first aspect, the present disclosure provides a distributed collaborative optimization system for a multi-region integrated energy system based on edge computing.

An edge computing-based distributed collaborative optimization system for a multi-region integrated energy system, comprising: the cloud computing system comprises a cloud computing center and a plurality of edge computing units, wherein each edge computing unit is in communication connection with the cloud computing center, and the edge computing units are in coupling connection through an electric heating air network;

each edge computing unit performs day-ahead operation optimization according to the acquired self-measured data by taking the minimum economic cost as an optimization target to obtain boundary information and send the boundary information to a cloud computing center;

the cloud computing center updates consensus information and Lagrange multipliers according to the received boundary information and feeds back the consensus information and the Lagrange multipliers to each edge computing unit;

according to a preset distributed nested algorithm, the edge computing unit and the cloud computing center complete multiple iterations required by global optimization to obtain an optimal operation result.

Further, partitioning the whole system according to the relation between the energy hub and the energy network nodes, and completing the partition of each edge computing unit, wherein each edge computing unit completely and unrepeatedly comprises all network nodes; in the edge calculation unit, the upper layer electric energy is converted into cold, heat and electricity through energy coupling equipment inside the energy hub, the electricity and the cold are directly supplied to a local load, and the heat is interacted between the heat accumulator and the lower layer heating power ring network.

Further, the node types according to the energy hub connection mode include upstream and downstream types and the same or same node types:

the upstream and downstream type comprises a connection type and a source type or a library type, and the network nodes connected with the energy hubs and the nodes connected with other energy hubs are positioned at the upstream and the downstream of the network physical layer;

the connection type is the condition that a node connected with an energy hub is not a head-end node or a tail-end node, the processing method is that a virtual node is added in the middle section of an energy transmission line on two sides of the node, the virtual node is contained by edge computing units on two sides, and the information of the virtual node obtained when the interior of each edge computing unit is independently optimized is boundary information which needs to be uploaded to a cloud computing center;

the node with the source type connected with the energy hub is a head end node, the node with the sink type connected with the energy hub is a tail end node, and a virtual node is added in the middle section of a line on one side with an external node during processing.

The two energy hubs are simultaneously connected to the same node or nodes at the same level of the network, and a virtual node is added in the middle section of one line for processing the nodes at the same level so as to convert the nodes into upstream and downstream types;

when the same node processes, the information of the two nodes has a strong coupling relation, and a section of virtual pipeline is respectively added in front of the two units and is converted into the processing of the same node.

Further, distributed adjustment is carried out on the basis of Consensus information among edge computing units by utilizing a Consensus-ADMM algorithm in combination with an improved nesting sub-algorithm;

a nesting sub-algorithm comprising:

taking an integer variable value obtained by Consenssus-ADMM as an initial value, and initializing the integer variable value and the variable;

replacing discrete variables of the original problem with continuous variables, and solving the new problem;

updating the discrete variable and the auxiliary variable according to the obtained value of the continuous variable;

and when the residual error is smaller than a preset value, obtaining the final optimization control variable of each edge calculation unit.

Furthermore, a description penalty term is added to the augmented Lagrange equation of each edge calculation unit, and the description penalty term describes the total loss of the heat supply network supported by all the edge calculation units together.

In a second aspect, the present disclosure provides a distributed collaborative optimization method for a multi-region integrated energy system based on edge computing.

A distributed collaborative optimization method of a multi-region comprehensive energy system based on edge computing comprises the following steps:

each edge computing unit performs day-ahead operation optimization according to the acquired self-measured data by taking the minimum economic cost as an optimization target, obtains boundary information and sends the boundary information to a cloud computing center;

Further, distributed adjustment is carried out on the basis of Consensus information among edge computing units by utilizing a Consensus-ADMM algorithm in combination with an improved nesting sub-algorithm; a nesting sub-algorithm comprising:

Furthermore, a description penalty term is added to the extension Lagrange equation of each edge computing unit, and the description penalty term describes the total loss of the heat supply network supported by all the edge computing units together.

Compared with the prior art, the beneficial effect of this disclosure is:

1. the method takes an ECU (electronic control Unit) as a unit, the ECU takes an EH (electronic equipment) as a main body, the geographic range and the structural hierarchy are equivalent to RIES (resource allocation system), and a globally optimal scheduling scheme is finally obtained through in-unit optimization calculation and information interaction with a cloud service center.

2. In the aspect of a model, an IES partitioning method matched with the framework is designed according to different EH connection node positions, ECU regions are reasonably divided, the characteristics of distributed computing are met, and a model basis is provided for distributed optimization computing.

3. In the aspect of the algorithm, an improved consensus-ADMM nested algorithm based on a refined RIES model is provided to be matched with the ECU model, distributed calculation of large-scale expansion is supported, global optimization is supported, the problem that MILP (minimum iterative process) is not converged during refined modeling is solved, and the optimization result obtained through a few iteration times under the condition of large-scale expansion of the ECU is ensured, so that the convergence speed of the algorithm is increased.

4. The distributed collaborative optimization method provided by the disclosure not only protects the information privacy security of the RIES, but also greatly reduces the data transmission, storage and processing pressure of the cloud computing through the edge computing, and solves the problem that the mass data of the IES is difficult to be processed in a centralized manner. The framework and the algorithm supplement each other, and theoretical and method support is provided for the MRIES distributed collaborative optimization to the cross-regional IES distributed collaborative optimization evolution.

Drawings

The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and are not to limit the disclosure.

Fig. 1 is a schematic structural diagram of an ECU provided in the embodiment of the present disclosure.

Fig. 2 is a schematic diagram of information flow of centralized optimization and distributed optimization of an ECU architecture provided by the embodiment of the present disclosure.

Fig. 3 is a schematic diagram of energy transmission network partition types and a method according to an embodiment of the disclosure.

Fig. 4 is a schematic diagram of a general model for energy transmission of a unit heat supply network according to an embodiment of the disclosure.

Fig. 5 is a schematic diagram of the internal energy flow of an energy hub provided by an embodiment of the present disclosure.

Fig. 6 is a schematic diagram of consensus variables between computing units according to an embodiment of the present disclosure.

Fig. 7 is a flow chart of an improved consensus-ADMM algorithm provided by an embodiment of the present disclosure.

Fig. 8 is a schematic diagram for comparing residual convergence conditions of the distributed algorithm provided in the embodiment of the present disclosure.

Fig. 9 is a schematic view of a partition of a multi-region integrated energy system based on an ECU according to an embodiment of the present disclosure.

Detailed Description

The present disclosure is further described with reference to the following drawings and examples.

It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.

It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present disclosure. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.

The embodiments and features of the embodiments in the present disclosure may be combined with each other without conflict.

Example 1:

the embodiment 1 of the present disclosure provides a distributed collaborative optimization method for a multi-region integrated energy system based on edge computing, which includes the following processes:

s1: architecture construction

The ECU provided by the embodiment is an energy data management unit corresponding to a cloud computing center in the IES, and can migrate part of tasks of system computing to local to complete, so as to solve the problem of insufficient cloud computing resources, and the ECU model result is shown in fig. 1.

The upper layer electric energy is converted into cold, Heat and electricity through energy coupling equipment such as an electric refrigerator (EC), a Heat Pump (HP), Combined Heat and Power Generation (CHP), a Gas Boiler (GB) and the like in the EH, the electricity and the cold are directly supplied to a local load, and the Heat is interacted in a Heat accumulator (HS) and a lower layer heating ring network to support the Heat requirement of the local Heat load together. The ECU supports multi-region expansion, and the computing units are connected with each other through a power grid pipeline, an air grid pipeline and a heat supply network pipeline to jointly form the MRIES.

S2: information interaction process comparison

The method comprises the steps that the ECU acquires equipment and node information through a local sensing layer, boundary information obtained through internal optimization is uploaded to a cloud computing center to be processed in a unified and integrated mode, the cloud computing center feeds back consensus information obtained through processing to the ECUs, the ECUs adjust a local operation scheme according to the consensus information, iteration is conducted for multiple times, and global optimization is finally achieved.

As can be seen from the left side of fig. 2, the centralized computing needs to submit the information of all the devices in the network and the EH to the cloud computing center, which undoubtedly generates a large amount of communication facility construction cost and data transmission, storage and processing cost, deviates from the energy development idea of energy conservation and emission reduction, and the information security and privacy cannot be guaranteed.

As shown in the right side of the figure 2 according to the mode of decoupling the system by the ECU, data transmission is carried out in the RIES, only boundary information of coupling positions between the ECUs needs to be submitted to the cloud computing center, and the construction cost of communication facilities and the data transmission, storage and processing cost are greatly reduced on the basis of ensuring the safety and privacy of information in the ECUs. The model collects data of three types of energy of electric heat gas together for centralized processing, can better play the role of multi-energy cooperation of the ECU, and lays a foundation for the next research work of energy unified management and energy unified scaling.

S3: edge computation element modeling

S3.1: IES partitioning method based on ECU

The premise for the global optimization of the ECUs through local optimization is that each ECU must completely and non-repeatedly contain all the network nodes. In order to adapt to the topological structure of the network and provide a model foundation for a distributed algorithm, the whole system is partitioned according to the relation between EH connection energy network nodes, and the partitioning method is classified and summarized.

(1) Upstream and downstream type

The upstream and downstream type refers to that the network node connected with the EH and other nodes connected with the EH are positioned at the upstream and downstream of the network physical layer. As shown in fig. 3, and includes a connection type and a source/sink type.

The connection type is a case where the node to which the EH is connected is not the head-end node nor the end node. The processing method of the type is that a virtual node is added in the middle section of the energy transmission lines on the two sides of the node respectively, and the virtual node is contained by the ECUs on the two sides. And the information of the virtual node obtained when the interior of the unit is independently optimized is boundary information which needs to be uploaded to the cloud computing center.

The source type means that the EH-connected node is a head-end node, and the sink type means that the EH-connected node is a tail-end node. Compared with a connection type, the situation lacks half of connection relation, and only one virtual node is added in the middle section of the line on one side with the external node during processing.

(2) Same/same node type

The same/same-level node type means that two EHs are simultaneously connected to the same node or same-level node of the network, as shown in fig. 3. Adding a virtual node in the middle section of one line for processing the same level of time, so that the virtual node is converted into an upstream type and a downstream type; when the same node processes, if the information of the two nodes has a strong coupling relation, a section of virtual pipeline needs to be added before the two units respectively, and the two units are converted into the same node processing.

And partitioning the energy distribution network according to the method, wherein the energy consumption of a single ECU is the inflow and the outflow.

S3.2: unit internal power grid partitioning model

And simulating the power flow of the power distribution network in the nth ECU by using a distflow power flow equation, and processing model nonlinearity by using second-order cone programming. The node power balance is as described in formulas (1) - (2), formula (3) is a voltage drop equation, and formula (4) represents the relationship between branch current and power.

In the above formula, Π (j) represents a device set connected with the node j of the EUC, and Ω (j) is a terminal node set of a branch connected with the node j;

and

representing active and reactive flowing into node j; r_ij、X_ij、P_ij,t、Q_ij,tThe resistance and reactance of the line ij and the active power and the reactive power flowing through at the moment t are respectively;

respectively representing that the g-th generator, the pv-th photovoltaic generator and the wt-th fan generate electric power;

and

the active power and the reactive power absorbed by the d-th load at the moment t;

representing the power that node i flows into EH at time t,

and

to represent

And

the transmission line transport capacity limit is shown in equation (5).

The amount of electricity consumed by the ECU is represented by the formula (6) in which

And

respectively representing the set of all the most upstream virtual nodes and the most downstream virtual nodes of the nth computing unit power grid.

S3.3: unit internal gas network node model

A natural gas pipe network model in the ECU is similar to a power distribution network and also comprises node energy balance constraint and pipeline transmission constraint, and the formula (7) represents that the inflow and outflow of a gas distribution node b are balanced.

In the above formula, Γ (b) represents a load set connected by a node b, and Λ (b) is an end node set of a pipeline connected by the node b;

indicating the gas injection quantity of the node b at the time t; g_ab,tRepresenting the momentum of the airflow in the pipeline ab at time t;

the gas consumption at the moment of the d-th gas load t;

the amount of natural gas input to the EH at node b at time t.

The gas distribution network in the model belongs to medium and low pressure, and a compressor model is not considered; similarly, a second-order cone programming is adopted to relax the Weymouth equation of the gas network as shown in the formula (8); the formulas (9) and (10) are the upper and lower limit constraints of the gas transmission capacity and the natural gas pressure of the natural gas pipeline.

(G_ab,t)²+(K_abπ_b,t)²≤(K_abπ_at)² (8)

In the above formula, K_abIs a Weymouth characteristic parameter of the pipeline; pi_a,tIs the pressure value of the node a at the time t.

The natural gas consumption of the ECU is as described in formula (11), wherein

And

respectively representing the set of all the most upstream and most downstream virtual nodes of the air network in the nth ECU.

S3.4: unit internal heat supply network node model

The heat supply network adopts the available heat power H in the formula (12)^avThe method comprises the steps of representing heat flow in a heat supply network pipeline, decoupling variable temperature T with strong coupling property and mass flow q in a heat supply network model, and forming an energy flow model and a flow-temperature basic equation of the heat supply network. The linearized heat loss equation is as shown in equation (13).

H^av＝kq(T-T_rw) (12)

In the above formula, T_e、T_rwAnd T_swRespectively representing the environment temperature, the return water temperature and the water supply temperature in the heat supply network pipeline; k is a proportionality constant; Σ R represents the total thermal resistance of the conduit per kilometer between the heating medium and the surrounding medium; l_uvRepresenting the length of the pipe uv.

The reference direction is set according to the direction shown in figure 4,

represents the heat provided by node u to the EH; t is_u,tThe node heat medium temperature at the u node t moment is shown in the formula (14) in the heat supply network energy flow model.

In the above formula, Ψ (u) represents a set of pipeline nodes to which the distribution network node u is connected;

the available heat power of the pipeline uv at the moment t;

respectively representing the upper and lower limits of the available thermal power flowing in the conduit.

And (3) establishing a heat supply network mixed integer model according to the equation to obtain available heat power distribution, and calculating the transmission temperature and flow of the heat medium in the pipeline by combining a flow-temperature basic equation (15).

The operation cost of the heat supply network is the operation cost of a circulating water pump in the pipeline, the calculation of the power consumption and heat transfer ratio is introduced as the formula (16),

representing the heat supply network operating cost of the nth computing unit.

In the above formula, W is the total number of circulating water pumps in the EUC, one circulating water pump is configured in each pipeline, and p is a pipeline label; EHR_pThe power consumption and heat transfer ratio of the p-th water pump is obtained; c. C_e,tThe price of electricity at the moment t; h_p,tThe heat delivered by the p-th water pump.

S3.5: variable working condition energy hub model

The EH model is an energy conversion main body of the proposed ECU, a standard matrix model is adopted to model the EH, and a piecewise linearization method is used to improve the EH modeling precision.

An energy matrix type (17) of the input and the output of the energy coupling device is established as shown in fig. 5, the meaning represented by a row vector of the energy matrix type corresponds to r in fig. 5 one by one, and the row vector with r being 1x24 represents the energy value of each hour. The normalized matrix Z, which describes the EH energy conversion relationship, multiplied by V, yields a matrix V1, as equations (18) - (19), where the heat output of the EH, together with the thermal energy extracted from the heat grid, supports the thermal load of the area.

V1＝[P^EH G^EH L_cool L_elec (L_heat-H^EH) 0 0]^T (18)

Z×V＝V1 (19)

Replacing the efficiency of a non-linear coupling device with a fixed value efficiency is not accurate and a piecewise linearization approach such as equations (20) - (22) is used to handle the non-linear efficiency function f (x) of the coupling device.

In the above formula, X is a continuous variable; x₀Is the initial value of X; sigma_kAn X value for the kth piecewise linearization of X; f_L(X) is an efficiency equation after segmentation; eta_kIs the efficiency value of the f (x) th segment; i is_kTo ensure F_L(X) and X consecutive auxiliary binary variables;

and _kXrespectively, the upper and lower bounds of the X value of the kth segment of X.

S3.6: edge computing unit objective function

And the ECU model is modeled, an internal power grid and an air grid form an SOCP problem, and a heat supply network and an EH form an MILP problem.

With the minimum economic cost as the optimization target, each computing unitThe objective function is then as shown in equation (23), where c_g,tIs the time-sharing price of natural gas.

S4: improved nested algorithm based on Consensus-ADMM

The ADMM algorithm performs well in a distributed algorithm, and there are various forms of improvement. The standard ADMM ensures that the convergent multi-region expansion form is more complex and does not conform to the calculation form expected by the proposed ECU; GS-ADMM cannot guarantee multi-region expansion convergence for n ≧ 3, and is a serial computing form. The Consensus-ADMM supports multi-region extensions and is a parallel computation, which is then the basis for an improvement of distributed algorithms.

The Consensus-ADMM algorithm is based on the adjustment of Consensus information between ECUs, as shown in FIG. 6, the ECUs are coupled through an electric heating and gas network, the coupling information includes P, Q, I, U of the power line, G and pi of the natural gas pipeline and H in the heat network, and the parameters and the Consensus parameters at the virtual nodes have an equation relationship as shown in formula (24), wherein x is used for simplicity_αβ,tRepresenting the corresponding coupling variable, z_αβ,tAnd the common identification variables correspond to all the parameters.

The Lagrangian extension of Consensus-ADMM, as shown in formula (23), is (25) (-)_nThe output lines are connected with other units for the power grid, the air grid and the heat supply network in the nth ECU; and lambda and rho are Lagrange multipliers and penalty factors corresponding to the coupling variables.

In order to ensure that the MILP formed by the model is converged in the calculation process, a nesting sub-algorithm is added on the basis of the consensus-ADNN algorithm, and a description penalty term (26) is added in a sub-unit augmented Lagrange equation, wherein the description penalty term describes the total loss of the heat supply network supported by all ECUs together, so that the iteration direction of each time is restricted, and the convergence speed is improved.

The algorithm flow chart is shown in fig. 7.

Wherein M represents the number of all the pipes in the whole heat supply network;

representing the available heat loss of the pth pipe, the aforementioned heat network model loss linearization is a prerequisite for improvement to be carried out.

Introduction of a continuous variable C_B∈[0,1]Auxiliary variable z_BAnd step size θ handles the MILP problem, forming the nested sub-algorithm described below.

Step 1: c is to be_B、z_BThe respective discrete variables B and θ make the iteration number k equal to 0, and set the error tolerance δ.

Step 2: combining formulas (1) - (15), (17) - (22), solving (27) to obtain C_BThe original discrete variable B in equation (25) is solved by C_BAnd (4) replacing.

Step 3 updates the discrete variables and auxiliary variables as equation (28).

And 4, step 4: and (4) comparing the error obtained in the step (27) with theta, and jumping out of the iteration process if the error is smaller than theta, otherwise, repeating the step 2 to the step 4.

The addition of the nested sub-algorithm enables the solution to approach a feasible domain infinitely, and the addition of the formula sub-algorithm can greatly accelerate the convergence process when the original problem and the dual error have small difference.

The algorithm optimization can completely meet various load requirements in the computing unit, and can realize global optimization. Comparing the calculated data with the centralized algorithm, as shown in table 1, the total energy cost error is about 0.1%, and the errors of the electricity cost and the gas cost are relatively small, so that the improved distributed algorithm can achieve a better global optimization result.

Table 1: algorithmic error comparison

As shown in fig. 8, the convergence situation of the original residual error and the dual residual error of the proposed improved algorithm and the consensus-ADMM algorithm is compared, the improved algorithm meets the precision requirement after 40 times, while the original algorithm needs to iterate 90 times under the same parameters, so that the improved algorithm has better convergence performance than the original algorithm.

As shown in fig. 9, the ECU-based multi-area integrated energy system partition diagram verifies the rationality of distributed optimization with the ECU as a subsystem by using the improved algorithm provided by the present embodiment, and verifies that the algorithm has better convergence performance compared with the standard form of consensus-ADMM algorithm. The ECU provides a standard unit form for the consensus-ADMM algorithm, and the features of the consensus-ADMM algorithm and the standard unit form are tightly combined, so that a good optimization effect is realized.

As will be appreciated by one skilled in the art, embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.

The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the disclosure. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.

The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.

Claims

1. A distributed collaborative optimization system of a multi-region comprehensive energy system based on edge computing is characterized in that:

the method comprises the following steps: the cloud computing system comprises a cloud computing center and a plurality of edge computing units, wherein each edge computing unit is in communication connection with the cloud computing center, and the edge computing units are in coupling connection through an electric heating and air network;

according to a preset distributed nested algorithm, the edge computing unit and the cloud computing center complete multiple iterations required by global optimization to obtain an optimal operation result;

dividing the whole system into partitions according to the relation between the energy hub and the energy network nodes, and completing the division of each edge computing unit, wherein each edge computing unit completely and unrepeatedly comprises all network nodes;

in the edge calculation unit, the upper layer electric energy is converted into cold, heat and electricity through energy coupling equipment in an energy hub, the electricity and the cold are directly supplied to a local load, and the heat is interacted between the heat accumulator and the lower layer heating power ring network;

the node types comprise upstream and downstream types, the same or same node types according to the energy hub connection mode;

the upstream and downstream type, the network node that the energy hub connects and node that other energy hubs connect divide upstream and downstream of the network physical level, the upstream and downstream type includes the connection type, source type or storehouse type again;

the source type is the node that the energy hub connects is the head end node, the storehouse type is the node that the energy hub connects is the end node, only increase a virtual node in the middle section of one side circuit with external node while dealing with;

the two energy hubs are simultaneously connected to the same node or the same node of the network, and a virtual node is added in the middle section of one line when the same node is processed, so that the same node or the same node is converted into an upstream and downstream type;

2. The distributed collaborative optimization system for multi-regional integrated energy systems based on edge computing of claim 1, wherein:

distributed adjustment is performed on the basis of Consensus information among the edge computing units by using a Consensus-ADMM algorithm in combination with an improved nesting sub-algorithm;

a nesting sub-algorithm comprising:

taking the integer variable value obtained by the Consensus-ADMM as an initial value, and initializing the integer variable value and the variable;

3. The distributed collaborative optimization system for multi-regional integrated energy systems based on edge computing of claim 1, wherein:

and adding a description penalty term into the extension Lagrange equation of each edge computing unit, and describing the penalty term to describe the total loss of the heat supply network supported by all the edge computing units together.

4. A distributed collaborative optimization method of a multi-region comprehensive energy system based on edge calculation is characterized by comprising the following steps:

the method comprises the following steps:

5. The distributed collaborative optimization method of the multi-region integrated energy system based on the edge calculation according to claim 4, wherein:

a nesting sub-algorithm comprising:

6. The distributed collaborative optimization method of the multi-region integrated energy system based on the edge calculation according to claim 4, wherein:

and adding a description penalty term into the augmented Lagrange equation of each edge calculation unit, and describing the penalty term to describe the total loss of the heat supply network supported by all the edge calculation units together.