Virtual power plant multi-energy flow distributed cooperative regulation and control method

Abstract

The invention relates to a virtual power plant multi-energy flow distributed cooperative regulation and control method, and belongs to the technical field of telecommunication. According to the method, uncertain factors and time delay of multi-energy flow response regulation are considered, loads are brought into the range of energy flow, the adaptive capacity of a model to user behavior uncertainty is enhanced, time delay of natural gas flow response scheduling is also considered, and therefore the response efficiency of a system under the uncertain condition is improved. Risk income comprehensive evaluation is introduced into distributed energy flow optimization, the influence of scheduling increment on the user demand satisfaction probability and the virtual power plant income is considered, and a traditional ADMM optimization framework is expanded. According to the method, the penalty factor in the ADMM method is improved, the penalty factor is dynamically adjusted through comprehensive evaluation based on regulation and control risk benefits, the importance of optimization problems and constraint conditions is balanced, the convergence speed of the algorithm is increased, and the regulation and control cost is reduced.

Description

Technical Field

[0001] The present invention relates to a distributed coordinated control method for multiple energy flows in a virtual power plant, and belongs to the technical field of electrical communications. Background Art

[0002] As an intelligent energy system that integrates multiple energy resources (such as electricity, heat, cooling, etc.), the multi-energy flow system of a virtual power plant achieves efficient utilization and flexible scheduling of energy through advanced monitoring, control and optimization technologies. It is a vital energy structure to promote the sustainable development of the energy system. However, the multi-energy flow system of a virtual power plant is large in scale, and multiple devices and multiple stations in the system need to be uniformly scheduled and coordinated. The multi-energy flow system faces the challenges of information asymmetry and difficult data integration. At the same time, due to the involvement of multiple energy resources and equipment in the multi-energy flow system of a virtual power plant, the coupling relationship and interactive influence between different energy sources have greatly increased the complexity of scheduling the multi-energy flow system of a virtual power plant. In order to achieve the coordinated utilization and complementary integration of multiple energy flows and ensure the intelligent operation of the multi-energy flow system of a virtual power plant, it is necessary to coordinate and regulate the energy resources of the multi-energy flow system of a virtual power plant. Therefore, it is urgent to design a distributed coordinated regulation method for multi-energy flows of a virtual power plant, which can achieve coordinated scheduling and complementary utilization of the multi-energy flow system of a virtual power plant by distributedly managing and collaboratively controlling various energy resources, equipment and loads in the system.

[0003] At present, there have been some studies on the coordinated regulation of multiple energy flows:

[0004] The patent "Cross-regional multi-energy system collaborative control method based on distributed optimization" proposes a cross-regional multi-energy system collaborative control method based on distributed optimization. First, a cross-regional multi-energy system hierarchical collaborative control framework is constructed; secondly, a cross-regional multi-energy system "cross-regional level-regional level" collaborative scheduling model and a "regional level-park level" collaborative scheduling model are constructed; finally, the optimization problem of the cross-regional multi-energy system hierarchical collaborative control model is decomposed into three layers of optimization sub-problems for solution, thereby realizing the cross-regional multi-energy system collaborative scheduling under strong information barriers.

[0005] The patent "A multi-energy complementary microgrid system and method based on reverse coordinated regulation" proposes a multi-energy complementary microgrid system and method based on reverse coordinated regulation. First, the electric energy is utilized through the wind power output module, the photovoltaic power output module and the tidal power output module. Secondly, the real-time power generated by the conversion of each energy source is viewed through the power display module. Then, the electric energy is distributed through the power distribution module. Finally, the mutual complementation between the energy sources is achieved through the power complementary module, thereby ensuring the normal power supply of the power grid and improving the stability of the power grid.

[0006] However, the existing distributed collaborative control method for multiple energy flows in virtual power plants still has the following two shortcomings: First, the unified modeling of multiple energy flow resources in virtual power plants is performed without considering the uncertainty of user energy use, which leads to inaccurate unified modeling of multiple energy flow resources in virtual power plants and the risk of low collaborative control efficiency. Second, the comprehensive evaluation of risk and benefit is not considered, and the distributed collaborative control of multiple energy flows in virtual power plants is only carried out to minimize the dispatch cost, resulting in low safety and poor reliability of distributed collaborative control of multiple energy flows in virtual power plants. Therefore, it is urgent to design a distributed collaborative control method for multiple energy flows in virtual power plants. Summary of the invention

[0007] In order to solve the above technical problems, the purpose of the present invention is to provide a method for distributed coordinated control of multiple energy flows in a virtual power plant.

[0008] A virtual power plant multi-energy flow distributed coordinated control method of the present invention, the specific control steps are:

[0009] Collect the dispatch delay and dispatch power information of user-side load resources, natural gas resources, and thermal resources to construct the dispatch utility function of each resource And calculate the total utility function Multi-energy flow modeling is performed based on utility functions and various power constraints;

[0010] Obtain the optimization parameters of the VPP area control in the current time slot, including the total utility function of the control optimization in VPPk The regional coupling equation coefficient matrix A obtained based on the power balance constraint within the VPP station area k and c, the control optimization decision variable matrix of VPPk

[0011] Based on the control optimization parameters, the auxiliary variable z and the equality constraint indicator function h(z) and the inequality constraint indicator function y(z) are introduced to construct the intraday control optimization objective function. When the intraday optimal decision variable satisfies the equality constraint, the equality constraint indicator function h(z) = 0, and the equality constraint term of the optimization objective does not work at this time; when the intraday optimal decision variable does not satisfy the equality constraint, the equality constraint indicator function h(z) = +∞, thereby forcing the optimization algorithm to adjust it to the range that satisfies the inequality constraint to reduce the intraday optimization control cost. The specific equality constraint expression is shown in formula (6). The inequality constraint indicator function is similar to the equality constraint indicator function. When the intraday optimal decision variable satisfies the inequality constraint, the inequality constraint indicator function y(z) = 0, otherwise y(z) = +∞. The specific inequality constraint expression is shown in formula (3) and formula (7). Constructing the augmented Lagrangian function Iterative solution and the construction process of augmented Lagrangian function belong to the existing technology, which will not be described here. In the formula, x is the decision variable for intraday regulation optimization; z is the value of the auxiliary variable of equality constraint and inequality constraint; u is the Lagrangian multiplier, which is used to penalize the constraint conditions, so as to make the optimization process as close as possible to the optimal solution of the original problem while considering the constraint conditions. Determine whether the solution meets the iteration accuracy requirements. If so, output the best solution result and Otherwise, continue to iterate until the iteration accuracy requirement is met, perform multi-energy flow control according to the optimal result, and update the penalty factor ρ k,l (t), and conduct the next time slot scheduling, thereby realizing the distributed coordinated control of multiple energy flows in the virtual power plant.

[0012] Furthermore, the user-side load resource scheduling utility function is:

[0013]

[0014] Among them, γ(t) is the electricity price at time t; Dispatching power for user-side loads; The cost of user-side load dispatching; is the user-side load scheduling delay; α is the scheduling delay sensitivity weight.

[0015] Furthermore, the utility function of natural gas resource scheduling is:

[0016]

[0017] Among them, π grid is the electricity-to-gas conversion efficiency of the gas turbine; Dispatching power for natural gas resources; f k,l (t) is the natural gas consumption of the gas turbine at time t; Scheduling costs for natural gas resources; Delay for natural gas resource scheduling; is the natural gas flow scheduling value at time i, is the maximum natural gas flow rate in a single time slot in the pipeline, and β is the sensitive weight of the average natural gas flow rate;

[0018] The natural gas flow constraints are:

[0019]

[0020] in, is the minimum natural gas flow rate in a single time slot in the pipeline.

[0021] Furthermore, the thermal scheduling utility function is:

[0022]

[0023] Among them, χ heat is the electricity-to-heat conversion efficiency of the gas turbine; dispatching power for thermal resources; The cost of dispatching thermal resources; The delay for thermal resource scheduling.

[0024] Furthermore, the total utility function is:

[0025]

[0026] Substation power balance is an important part of ensuring the stable operation of the power system, which involves the supply and demand relationship, quality and efficiency of power in the power grid. The above needs to meet the following constraints

[0027]

[0028] in, and are the minimum and maximum dispatching power of the substation respectively.

[0029] Furthermore, the multi-energy flow modeling model includes two layers: the edge layer and the terminal layer. The edge layer consists of K VPPs, and the set is expressed as VPP is responsible for communicating and interacting with the substations under its jurisdiction; the terminal layer consists of L substations, and the set is represented as The kth VPP contains l k Each station area includes heterogeneous resources such as user-side load, gas turbine, electric boiler, etc. In this model, the entire optimization period is divided into T time slots, each time slot length is τ, and the time slot set is recorded as

[0030] Furthermore, the iteration accuracy requirement is: ||r c+1 ||2≤ε prime And||s c+1 ||2≤ε dual ; Among them, r c+1 and c+1 are the primal residual and dual residual after the c+1th iteration, ε prime With ε dual are the primal infeasibility and the dual infeasibility respectively.

[0031] Furthermore, the update penalty factor ρ k,l (t) is as follows

[0032]

[0033] in, is the average value of the penalty factor at the previous t-1 moment; δ is the risk-return weight; p k,l (t) is the probability that the dispatch value of the substation l at time t meets the user's demand, reflecting the user's satisfaction with the current dispatch mode. The larger the probability value, the higher the user satisfaction, and vice versa. If the value of the historical average total utility function is smaller, the user satisfaction is lower, indicating that the virtual power plant obtains low benefits and increased risks under the current dispatch mode. It is necessary to increase the penalty factor value to increase the emphasis of the augmented Lagrangian function on the constraints, making the augmented Lagrangian function steeper near the constraints and speeding up the optimal solution exploration speed; otherwise, reduce the penalty factor value and continue to maintain the current state for solving;

[0034]

[0035] Among them, P k ' ,l (i) is the actual demand dispatching power value at time i.

[0036] A virtual power plant multi-energy flow distributed collaborative control system, including a multi-energy flow modeling subsystem and a distributed control subsystem, which are deployed on edge servers in each area;

[0037] Among them, the multi-energy flow modeling subsystem includes resource information acquisition module, utility function construction module, and multi-energy flow resource modeling module;

[0038] The distributed control subsystem includes an optimization parameter acquisition module, a penalty factor update module and a distributed control optimization module;

[0039] The resource information acquisition module is used to collect information such as the scheduling delay and scheduling power of user-side load resources, natural gas resources, and thermal resources, and transmit the data to the utility function construction module;

[0040] The utility function construction module processes the multi-energy flow scheduling data collected by the resource information acquisition module, constructs each resource scheduling utility function, calculates the total utility function, and transmits the function to the multi-energy flow resource modeling module;

[0041] The multi-energy flow resource modeling module is used to perform multi-energy flow modeling according to the utility function and various power constraints, and the obtained model is uploaded to the distributed control subsystem;

[0042] The optimization parameter acquisition module is used to acquire the uploaded multi-energy flow model and the VPP internal area control optimization parameters, and transmit the parameters to the penalty factor update module;

[0043] The penalty factor updating module is used to update the penalty factor based on the area control optimization parameter and transmit the penalty factor to the distributed control optimization module;

[0044] The distributed control optimization module is used to perform distributed control optimization using the improved ADMM according to the updated penalty factor.

[0045] A virtual power plant multi-energy flow distributed coordinated control device, comprising:

[0046] Modeling module: used to collect the dispatch delay and dispatch power information of user-side load resources, natural gas resources, and thermal resources to build the dispatch utility function of each resource And calculate the total utility function Multi-energy flow modeling is performed based on utility functions and various power constraints;

[0047] Optimization module: used to obtain the optimization parameters of the VPP area control in the current time slot, including the total utility function of the control optimization in VPPk The regional coupling equation coefficient matrix A obtained based on the power balance constraint within the VPP station area k and c, the control optimization decision variable matrix of VPP k

[0048] Solution module: used to construct the intraday regulation optimization objective function based on the regulation optimization parameters, introduce the auxiliary variable z and the equality constraint indicator function h(z) and the inequality constraint indicator function y(z), and construct the augmented Lagrangian function Perform iterative solution; determine whether the solution meets the iterative accuracy requirements, and if so, output the best solution and Otherwise, continue to iterate until the iteration accuracy requirement is met, perform multi-energy flow control according to the best solution result, and update the penalty factor ρ k,l (t), and conduct the next time slot scheduling, thereby realizing the distributed coordinated control of multiple energy flows in the virtual power plant.

[0049] Furthermore, the update penalty factor ρ described in the solution module k,l (t) is as follows

[0050]

[0051] in, is the average value of the penalty factor at the previous t-1 moment; δ is the risk-return weight; p k,l(t) is the probability that the dispatch value of the substation l at time t meets the user's demand, reflecting the user's satisfaction with the current dispatch mode. The larger the probability value, the higher the user satisfaction, and vice versa. If the value of the historical average total utility function is smaller, the user satisfaction is lower, indicating that the virtual power plant obtains low benefits and increased risks under the current dispatch mode. It is necessary to increase the penalty factor value to increase the emphasis of the augmented Lagrangian function on the constraints, making the augmented Lagrangian function steeper near the constraints and speeding up the optimal solution exploration speed; otherwise, reduce the penalty factor value and continue to maintain the current state for solving;

[0052]

[0053] Among them, P k ' ,l (i) is the actual demand dispatching power value at time i.

[0054] By means of the above scheme, the present invention has at least the following advantages:

[0055] (1) This paper proposes a unified modeling method for multi-energy flow resources of virtual power plants that takes into account uncertainty. First, this method takes into account uncertainty factors and the delay of multi-energy flow response regulation, and incorporates load into the category of energy flow. It not only enhances the model's adaptability to user behavior uncertainty, but also takes into account the delay of natural gas flow response scheduling, thereby improving the system's response efficiency under uncertain conditions. Secondly, this method takes into account the maximum flow rate constraint of natural gas flow and the uncertainty of user usage behavior. By introducing the average natural gas flow sensitive weight, the model can better handle the dynamic balance between natural gas supply and demand, ensuring the continuity and safety of supply. Finally, by defining utility functions for multiple energy types such as electricity, natural gas, and heat, these utility functions comprehensively consider uncertainty factors such as load demand and market changes, as well as scheduling delays, ensuring the optimization of virtual power plant multi-energy flow resource allocation, and realizing precise regulation and optimization of various energy flows. Scheduling delay sensitive weights are introduced in the utility function setting, so that the model can dynamically adjust the resource allocation strategy according to delay changes.

[0056] (2) The present invention proposes a multi-energy flow distributed collaborative control method based on comprehensive risk-benefit evaluation. By introducing comprehensive risk-benefit evaluation into distributed energy flow optimization, the impact of scheduling increments on the probability of user demand satisfaction and the benefits of virtual power plants is considered, and the traditional ADMM optimization framework is expanded. By introducing the regional coupling equation coefficient matrix and the control optimization decision variable matrix, the power balance optimization of the substation within the virtual power plant is realized, and the accuracy and efficiency of the control are improved. The constructed augmented Lagrangian function is solved iteratively to meet the accuracy requirements, ensuring the accurate search for the optimal solution. In addition, the penalty factor in the ADMM method is improved. By dynamically adjusting the penalty factor based on the comprehensive evaluation of the control risk and benefit, the importance of the optimization problem and the constraint conditions is balanced, the convergence speed of the algorithm is improved, and the control cost is reduced. This method improves the performance of the multi-energy flow distributed collaborative control optimization strategy, enhances the operating efficiency, system stability and reliability of the distribution network, and ensures power supply security under demand fluctuations and uncertainties.

[0057] (3) The present invention proposes a virtual power plant multi-energy flow distributed collaborative control system, including a multi-energy flow modeling subsystem and a distributed control subsystem, wherein the multi-energy flow modeling subsystem includes a resource information acquisition module, a utility function construction module, and a multi-energy flow resource modeling module; the distributed control subsystem includes an optimization parameter acquisition module, a penalty factor update module, and a distributed control optimization module. The virtual power plant multi-energy flow distributed collaborative control system is equipped with a virtual power plant multi-energy flow resource unified modeling method taking into account uncertainty and a multi-energy flow distributed collaborative control optimization method based on an improved ADMM. The penalty factor update module updates the penalty factor based on the area control optimization parameters to balance the importance of the optimization problem and the constraints. The utility function construction module constructs the scheduling utility function of each resource according to the scheduling delay and scheduling power of the user-side load resources, natural gas resources, and thermal resources, and calculates the total utility function. The modules operate in coordination to realize the distributed collaborative control of multi-energy flows in the virtual power plant, improve the accuracy of multi-energy flow collaborative control, and reduce the control cost.

[0058] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show a certain embodiment of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0060] Figure 1 It is a flow chart of the distributed collaborative control method of multiple energy flows in a virtual power plant of the present invention;

[0061] Figure 2 This is a diagram of a multi-energy flow distributed coordinated control system of a virtual power plant according to the present invention; DETAILED DESCRIPTION

[0062] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0063] The present invention proposes a distributed collaborative control method for multi-energy flows in a virtual power plant. First, a unified modeling method for multi-energy flow resources of a virtual power plant taking into account uncertainty is proposed to model multi-energy flow resources of a virtual power plant. Then, a distributed collaborative control method for multi-energy flows based on a comprehensive risk-benefit evaluation is proposed. The computing tasks of the entire system are distributed to each edge through ADMM, and the penalty factor is dynamically adjusted based on the comprehensive risk-benefit evaluation results of the control. The importance of optimization problems and constraints can be dynamically balanced according to the actual operating requirements and uncertainties of the system, thereby avoiding situations where the VPP has low returns or the probability of user demand not being met is low.

[0064] 1. Unified modeling method for multi-energy flow resources in virtual power plants considering uncertainty

[0065] S1: This paper proposes a unified modeling method for multi-energy flow resources of virtual power plants taking into account uncertainty. It is assumed that the multi-energy flow resource scheduling model of virtual power plants includes two layers: edge layer and terminal layer. The edge layer consists of K VPPs, and the set is represented as The VPP is responsible for communicating and interacting with the substations under its jurisdiction. The terminal layer consists of L substations, and the set is represented as The kth VPP contains l k Each station area includes heterogeneous resources such as user-side load, gas turbine, electric boiler, etc. In this model, the entire optimization period is divided into T time slots, each time slot length is τ, and the time slot set is recorded as

[0066] In the integrated energy system, participants include different types of energy suppliers, and their respective utilities depend not only on their own economic behavior, but also on user behavior, which makes the decision-making problem extremely complicated, and the traditional single-agent optimization theory can no longer effectively solve these problems. In the study of power system optimization scheduling, the utility function is an important tool, which is used to measure the performance of scheduling strategies in different aspects. Scheduling delay refers to the time interval from issuing a scheduling command to responding to the command. In the power system, reducing the scheduling delay means that load changes can be responded to faster, improving the reliability of the system and user satisfaction. The scheduling cost is the total cost required to implement the scheduling strategy. Benefits usually refer to the economic benefits obtained by optimizing scheduling, and the economic feasibility of different scheduling schemes is comprehensively evaluated in combination with the scheduling cost. In order to ensure the optimal allocation of VPP multi-energy flow resources, the present invention sets a resource scheduling utility function, which comprehensively considers the uncertain factors such as equipment operation, load demand and market changes such as scheduling delay, scheduling cost and benefits. However, due to the uncertainty of user usage behavior, the scheduling delay is considered in the utility function to ensure that user needs are effectively met, while improving the operating efficiency and reliability of the overall system.

[0067] S1.1: Power resource scheduling. The utility function of user-side load resource scheduling can be expressed as

[0068]

[0069] Among them, γ(t) is the electricity price at time t; Dispatching power for user-side loads; The cost of user-side load dispatching; is the user-side load scheduling delay; α is the scheduling delay sensitivity weight. The larger the utility function, the greater the benefit obtained by VPP and the better the current resource scheduling mode.

[0070] S1.2: Natural gas resource scheduling. VPP generally uses gas turbines to couple electricity and gas resources together, which improves the flexibility of energy utilization. The high regulation rate of gas turbines enables them to respond quickly to electricity demand, and can also adjust power generation according to the real-time status of natural gas supply. However, when scheduling natural gas resources, the maximum flow rate limit of the pipeline is a key technical parameter that directly affects the natural gas transportation capacity and the operating efficiency of the system. In order to ensure the continuity and safety of natural gas supply, the scheduling utility function needs to consider the maximum flow rate limit of the pipeline flow rate. The natural gas resource scheduling utility function can be expressed as

[0071]

[0072] Among them, π grid is the electricity-to-gas conversion efficiency of the gas turbine; Dispatching power for natural gas resources; fk,l (t) is the natural gas consumption of the gas turbine at time t; Scheduling costs for natural gas resources; Delay for natural gas resource scheduling; is the natural gas flow scheduling value at time i, is the maximum natural gas flow rate in a single time slot of the pipeline, and β is the sensitive weight of the average natural gas flow rate.

[0073] The natural gas flow constraint is:

[0074]

[0075] in, is the minimum natural gas flow rate in a single time slot in the pipeline

[0076] S1.3: Thermal resource scheduling. VPP usually uses electric boilers to couple electricity and heat resources. As the key equipment for coupling electricity and heat resources, electric boilers are responsible for efficiently converting electrical energy into thermal energy. Therefore, the thermal scheduling utility function can be expressed as

[0077]

[0078] Among them, χ heat is the electricity-to-heat conversion efficiency of the gas turbine; dispatching power for thermal resources; The cost of dispatching thermal resources; The delay for thermal resource scheduling.

[0079] S1.4: Based on the utility functions of the above-mentioned multiple types of energy, the total utility function can be expressed as

[0080]

[0081] The power balance of the substation area is an important part of ensuring the stable operation of the power system, which involves the supply and demand relationship, quality and efficiency of power in the power grid. The above needs to meet the following constraints

[0082]

[0083] in, and are the minimum and maximum dispatching power of the substation respectively.

[0084] 2. Distributed collaborative control method of multi-energy flow based on comprehensive risk-benefit evaluation

[0085] The present invention proposes a multi-energy flow distributed collaborative control method based on comprehensive risk-benefit evaluation, and adopts the ADMM method to maximize the VPP multi-energy flow distributed collaborative scheduling utility function. The specific steps are introduced as follows.

[0086] The total utility function of the control optimization in VPPk is: The regional coupling equation coefficient matrix obtained based on the power balance constraint within the VPP is A k and c; the control optimization decision variable matrix of VPPk is

[0087] S2: Distributed collaborative control optimization of multi-energy flows based on improved ADMM.

[0088] S2.1: Based on the control optimization parameters, in order to ensure that the optimization target meets the above-mentioned VPP intra-station control inequality constraints, the auxiliary variable z and the equality constraint indicator function h(z) and the inequality constraint indicator function y(z) are introduced to improve the objective function. When the intra-day optimal decision variable satisfies the equality constraint, the equality constraint indicator function h(z)=0, and the equality constraint term of the optimization target does not work at this time; when the intra-day optimal decision variable does not satisfy the equality constraint, the equality constraint indicator function h(z)=+∞, thereby forcing the optimization algorithm to adjust it to the range that satisfies the inequality constraint to reduce the intra-day optimization control cost. The specific equality constraint expression is shown in formula (6). The inequality constraint indicator function is similar to the equality constraint indicator function. When the intra-day optimal decision variable satisfies the inequality constraint, the inequality constraint indicator function y(z)=0, otherwise y(z)=+∞. The specific inequality constraint expression is shown in formula (3) and formula (7).

[0089] S2.2: Construct the corresponding augmented Lagrangian function Then iterate and solve. The construction process of the augmented Lagrangian function belongs to the existing technology and will not be described here. In the formula, x is the decision variable for intraday regulation optimization; z is the value of the auxiliary variable of the equality constraint and inequality constraint; u is the Lagrangian multiplier, which is used to penalize the constraint conditions, so as to make the optimization process as close as possible to the optimal solution of the original problem while considering the constraint conditions.

[0090] S2.3: Determine whether the solution meets the iteration accuracy requirements, that is, ||r c+1 ||2≤ε prime And||s c+1 ||2≤ε dual Among them, r c+1 and c+1 are the primal residual and dual residual after the c+1th iteration, ε prime With ε dual are the original infeasibility and the dual infeasibility respectively. If the solution meets the iteration accuracy requirements, the best solution is output. and Otherwise, continue iterating until the iteration accuracy requirement is met.

[0091] S2.4: Considering that the fixed penalty factor of traditional ADMM may cause distributed computing to be difficult to converge, the present invention improves the penalty factor of the ADMM method. Based on the comprehensive evaluation of the risk and benefit of regulation, the penalty factor of each time slot in the optimization regulation process is dynamically adjusted to balance the importance of the optimization problem and the constraint conditions, improve the convergence speed of the optimization regulation, and reduce the regulation cost. k,l (t) is updated as follows

[0092]

[0093] in, is the average value of the penalty factor at the previous t-1 moment; δ is the risk-return weight; p k,l (t) is the probability that the dispatch value of the substation l at time t meets the user's demand, reflecting the user's satisfaction with the current dispatch mode. The larger the probability value, the higher the user satisfaction, and vice versa. If the value of the historical average total utility function is smaller, the lower the user satisfaction, indicating that the virtual power plant obtains low benefits and increased risks under the current dispatch mode, it is necessary to increase the penalty factor value to increase the emphasis of the augmented Lagrangian function on the constraints, making the augmented Lagrangian function steeper near the constraints and speeding up the optimal solution exploration. On the contrary, reduce the value of the penalty factor and continue to maintain the current state for solving.

[0094]

[0095] Among them, P k ' ,l (i) is the actual demand dispatching power value at time i.

[0096] Based on the multi-energy flow distributed collaborative control optimization strategy obtained in the above steps, each substation implements this strategy, which not only improves the operating efficiency of the distribution network, but also enhances the stability and reliability of the system, ensuring power supply security under demand fluctuations and uncertainties.

[0097] 3. Method flow

[0098] The specific steps of the present invention are as follows Figure 1 shown.

[0099] Step 1: Collect the dispatch delay and dispatch power of user-side load resources, natural gas resources, and thermal resources.

[0100] Step 2: Construct the resource scheduling utility function according to step 1 And calculate the total utility function

[0101] Step 3: Model the multi-energy flow based on the utility function and various power constraints.

[0102] Step 4: Obtain the optimization parameters of the VPP area control in the current time slot, including the total utility function of the control optimization in VPPk The regional coupling equation coefficient matrix A obtained based on the power balance constraint within the VPP station area k and c, the control optimization decision variable matrix of VPPk

[0103] Step 5: Introduce auxiliary variables z and the equality constraint indicator function h(z) and the inequality constraint indicator function y(z) to construct the intraday regulation optimization objective function.

[0104] Step 6: Construct the augmented Lagrangian function Perform iterative solution.

[0105] Step 7: Determine whether the solution meets the iteration accuracy requirements. If so, output the best solution. and Otherwise, continue iterating until the iteration accuracy requirement is met.

[0106] Step 8: Perform multi-energy flow control based on the best solution and update the penalty factor ρ k,l (t). Perform next time slot scheduling.

[0107] like Figure 2 As shown in the figure, the proposed virtual power plant multi-energy flow distributed collaborative control system includes a multi-energy flow modeling subsystem and a distributed control subsystem, which are deployed on the edge servers of each substation. Among them, the multi-energy flow modeling subsystem includes a resource information acquisition module, a utility function construction module, and a multi-energy flow resource modeling module; the distributed control subsystem includes an optimization parameter acquisition module, a penalty factor update module, and a distributed control optimization module.

[0108] Resource information acquisition module: collects information such as the dispatch delay and dispatch power of user-side load resources, natural gas resources, and thermal resources, and transmits the data to the utility function construction module.

[0109] Utility function construction module: processes the multi-energy flow scheduling data collected by the resource information acquisition module, constructs the utility function of each resource scheduling, calculates the total utility function, and transmits the function to the multi-energy flow resource modeling module.

[0110] Multi-energy flow resource modeling module: Multi-energy flow modeling is performed based on the utility function and various power constraints, and the resulting model is uploaded to the distributed control subsystem.

[0111] Optimization parameter acquisition module: obtains the uploaded multi-energy flow model and VPP internal area control optimization parameters, and transmits the parameters to the penalty factor update module.

[0112] Penalty factor update module: updates the penalty factor based on the area control optimization parameters and transmits the penalty factor to the distributed control optimization module.

[0113] Distributed control optimization module: Based on the updated penalty factor, the improved ADMM is used to perform distributed control optimization.

[0114] The above system is used to carry out distributed coordinated control of multiple energy flows in the virtual power plant. The specific steps are as follows.

[0115] Step 0: Initialize each module;

[0116] Step 1: The resource information acquisition module collects the dispatch delay and dispatch power of user-side load resources, natural gas resources, and thermal resources, and transmits the data to the utility function construction module.

[0117] Step 2: The utility function building module calculates the utility function of each resource scheduling And calculate the total utility function And transfer the function to the multi-energy flow resource modeling module.

[0118] Step 3: The multi-energy flow resource modeling module performs multi-energy flow modeling according to the utility function and various power constraints, and the obtained model is uploaded to the distributed control subsystem.

[0119] Step 4: The optimization parameter acquisition module obtains the uploaded multi-energy flow model and the VPP internal area control optimization parameters, and transmits the parameters to the penalty factor update module.

[0120] Step 5: The penalty factor update module updates the penalty factor ρ based on the area control optimization parameters k,l (t), and transmit the penalty factor to the distributed control optimization module.

[0121] Step 6: The distributed control optimization module uses the improved ADMM to perform continuous iterations based on the updated penalty factor to achieve distributed control optimization.

[0122] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the technical principles of the present invention, and these improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A distributed collaborative control method for multiple energy flows in a virtual power plant, characterized in that The specific control steps are: Collect the dispatch delay and dispatch power information of user-side load resources, natural gas resources, and thermal resources to construct the dispatch utility function of each resource And calculate the total utility function Multi-energy flow modeling is performed based on utility functions and various power constraints; Obtain the optimization parameters of the VPP area control in the current time slot, including the total utility function of the control optimization in VPPk The regional coupling equation coefficient matrix A obtained based on the power balance constraint within the VPP station area k and c, VPP k The control optimization decision variable matrix Based on the control optimization parameters, the auxiliary variable z and the equality constraint indicator function h(z) and the inequality constraint indicator function y(z) are introduced to construct the intraday control optimization objective function and the augmented Lagrangian function. Perform iterative solution, where x is the decision variable for intraday regulation optimization; z is the value of the auxiliary variable of equality constraint and inequality constraint; u is the Lagrange multiplier; determine whether the solution meets the iteration accuracy requirements, if so, output the best solution and Otherwise, continue to iterate until the iteration accuracy requirement is met, perform multi-energy flow control according to the best solution result, and update the penalty factor ρ k,l (t), and conduct the next time slot scheduling, thereby realizing the distributed coordinated control of multiple energy flows in the virtual power plant.

2. A method for distributed coordinated control of multiple energy flows in a virtual power plant according to claim 1, characterized in that: The user-side load resource scheduling utility function is: Among them, γ(t) is the electricity price at time t; Dispatching power for user-side loads; The load dispatch cost for the user side; is the user-side load scheduling delay; α is the scheduling delay sensitivity weight.

3. A virtual power plant multi-energy flow distributed coordinated control method according to claim 1, characterized in that: The utility function of natural gas resource scheduling is: Among them, π grid is the electricity-to-gas conversion efficiency of the gas turbine; Dispatching power for natural gas resources; f k,l (t) is the natural gas consumption of the gas turbine at time t; Scheduling costs for natural gas resources; Delay for natural gas resource scheduling; is the natural gas flow scheduling value at time i, is the maximum natural gas flow rate in a single time slot in the pipeline, and β is the sensitive weight of the average natural gas flow rate; The natural gas flow constraints are: in, is the minimum natural gas flow rate in a single time slot in the pipeline.

4. A method for distributed coordinated control of multiple energy flows in a virtual power plant according to claim 1, characterized in that: The thermal dispatch utility function is: Among them, χ heat is the electricity-to-heat conversion efficiency of the gas turbine; dispatching power for thermal resources; The cost of dispatching thermal resources; The delay for thermal resource scheduling.

5. A virtual power plant multi-energy flow distributed coordinated control method according to claim 1, characterized in that: The total utility function is: Substation power balance is an important part of ensuring the stable operation of the power system, which involves the supply and demand relationship, quality and efficiency of power in the power grid. The above needs to meet the following constraints in, and are the minimum and maximum dispatching power of the substation respectively.

6. A virtual power plant multi-energy flow distributed coordinated control method according to claim 1, characterized in that: The multi-energy flow modeling model includes two layers: edge layer and terminal layer. The edge layer consists of K VPPs, and the set is expressed as The VPP is responsible for communicating and interacting with the stations within its jurisdiction; The terminal layer consists of L stations, and the set is represented as The kth VPP contains l k Each station area includes heterogeneous resources such as user-side load, gas turbine, electric boiler, etc. In this model, the entire optimization period is divided into T time slots, each time slot length is τ, and the time slot set is recorded as 7. A virtual power plant multi-energy flow distributed coordinated control method according to claim 1, characterized in that: The iteration accuracy requirement is: ||r c+1 ||2≤ε prime And||s c+1 ||2≤ε dual ; Among them, r c+1 and c+1 are the primal residual and dual residual after the c+1th iteration, ε prime With ε dual are the primal infeasibility and the dual infeasibility respectively.

8. A method for distributed coordinated control of multiple energy flows in a virtual power plant according to claim 1, characterized in that: The update penalty factor ρ k,l (t) is as follows in, is the average value of the penalty factor at the previous t-1 moment; δ is the risk-return weight; p k,l (t) is the probability that the dispatch value of the substation l at time t meets the user's demand, reflecting the user's satisfaction with the current dispatch mode. The larger the probability value, the higher the user satisfaction, and vice versa. If the value of the historical average total utility function is smaller, the user satisfaction is lower, indicating that the virtual power plant obtains low benefits and increased risks under the current dispatch mode. It is necessary to increase the penalty factor value to increase the emphasis of the augmented Lagrangian function on the constraints, making the augmented Lagrangian function steeper near the constraints and speeding up the optimal solution exploration speed; otherwise, reduce the penalty factor value and continue to maintain the current state for solving; Among them, P′ k,l (i) is the actual demand dispatching power value at time i.

9. A virtual power plant multi-energy flow distributed coordinated control system, characterized by: It includes a multi-energy flow modeling subsystem and a distributed control subsystem, which are deployed on the edge servers of each area; Among them, the multi-energy flow modeling subsystem includes resource information acquisition module, utility function construction module, and multi-energy flow resource modeling module; The distributed control subsystem includes an optimization parameter acquisition module, a penalty factor update module and a distributed control optimization module; The resource information acquisition module is used to collect information such as the scheduling delay and scheduling power of user-side load resources, natural gas resources, and thermal resources, and transmit the data to the utility function construction module; The utility function construction module processes the multi-energy flow scheduling data collected by the resource information acquisition module, constructs each resource scheduling utility function, calculates the total utility function, and transmits the function to the multi-energy flow resource modeling module; The multi-energy flow resource modeling module is used to perform multi-energy flow modeling according to the utility function and various power constraints, and the obtained model is uploaded to the distributed control subsystem; The optimization parameter acquisition module is used to acquire the uploaded multi-energy flow model and the VPP internal area control optimization parameters, and transmit the parameters to the penalty factor update module; The penalty factor updating module is used to update the penalty factor based on the area control optimization parameter and transmit the penalty factor to the distributed control optimization module; The distributed control optimization module is used to perform distributed control optimization using the improved ADMM according to the updated penalty factor.

10. A virtual power plant multi-energy flow distributed coordinated control device, characterized in that: include: Modeling module: used to collect the dispatch delay and dispatch power information of user-side load resources, natural gas resources, and thermal resources to build the dispatch utility function of each resource And calculate the total utility function Multi-energy flow modeling is performed based on utility functions and various power constraints; Optimization module: used to obtain the optimization parameters of VPP area control in the current time slot, including VPP k The total utility function of the internal regulation optimization The regional coupling equation coefficient matrix A obtained based on the power balance constraint within the VPP station area k and c, VPP k The control optimization decision variable matrix Solution module: used to construct the intraday regulation optimization objective function based on the regulation optimization parameters, introduce the auxiliary variable z and the equality constraint indicator function h(z) and the inequality constraint indicator function y(z), and construct the augmented Lagrangian function Perform iterative solution; determine whether the solution meets the iterative accuracy requirements, and if so, output the best solution and Otherwise, continue to iterate until the iteration accuracy requirement is met, perform multi-energy flow control according to the best solution result, and update the penalty factor ρ k,l (t), and conduct the next time slot scheduling, thereby realizing the distributed coordinated control of multiple energy flows in the virtual power plant.

11. A virtual power plant multi-energy flow distributed coordinated control device according to claim 10, characterized in that: Update the penalty factor ρ as described in the solution module k,l (t) is as follows in, is the average value of the penalty factor at the previous t-1 moment; δ is the risk-return weight; p k,l (t) is the probability that the dispatch value of the substation l at time t meets the user's demand, reflecting the user's satisfaction with the current dispatch mode. The larger the probability value, the higher the user satisfaction, and vice versa. If the value of the historical average total utility function is smaller, the user satisfaction is lower, indicating that the virtual power plant obtains low benefits and increased risks under the current dispatch mode. It is necessary to increase the penalty factor value to increase the emphasis of the augmented Lagrangian function on the constraints, making the augmented Lagrangian function steeper near the constraints and speeding up the optimal solution exploration speed; otherwise, reduce the penalty factor value and continue to maintain the current state for solving; Among them, P′ k,l (i) is the actual demand dispatching power value at time i.

Images