A cost-benefit analysis-based approach to aggregate adjustable power domains in virtual power plants

Abstract

The present invention discloses a method for aggregating adjustable power domains of virtual power plants based on cost-benefit analysis, constructing a heterogeneous distributed energy model that takes into account risk reserve capacity; constructing a virtual generator model that characterizes the adjustable power domain of the virtual power plant, and solving its parameters using the high-dimensional convex polyhedron interior approximation method; considering the uncertainty of distribution network dispatching instructions and distributed energy power, and with the goal of maximizing comprehensive economic benefits, constructing a two-stage robust optimization aggregation model for the adjustable power domain; using piecewise linearization and strong duality theory methods to simplify the model, and using the CC&G algorithm and Gurobi solver to solve it. By comparing and analyzing the comprehensive economic benefits of the adjustable power domain of the virtual power plant obtained by the reliability-based method and the method proposed by the present invention in conventional scenarios and extreme scenarios, it can be found that the method proposed by the present invention can obtain an adjustable power domain of the virtual power plant that takes into account both reliability and economy.

Description

Technical field:

[0001] The present invention relates to the technical field of virtual power plant applications, and in particular to a method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis. Background technology:

[0002] Since the Industrial Revolution, global energy crises, climate change, and environmental pollution have become increasingly severe. A "high-efficiency, low-carbon, and clean" development model has become the inevitable path for technological transformation in the global energy system. Renewable energy generation, represented by wind and solar power, has garnered increasing attention. As installed capacity of renewable energy generation continues to increase, grid connection methods have gradually shifted from localized to multi-regional centralized and distributed. This has led to a significant increase in random fluctuations on both the power supply and load sides, and a trend toward a diversification of daily operating scenarios, posing numerous challenges to the safe and stable operation of the power system. To promote the efficient use of renewable energy, flexible resources such as energy storage, diesel generators, temperature-controlled loads, and electric vehicles have emerged on the distribution and consumption sides. The large number, small capacity, and uneven distribution of these distributed energy resources make individual unit integration expensive, and they are often invisible to system operators, making management difficult. Virtual power plants, leveraging advanced information and communication technologies and a higher-level software architecture, offer a new approach to coordinated operation of distributed energy resources on the distribution side.

[0003] However, existing global optimization and dispatching of distribution networks based on detailed physical models of virtual power plants (VPPs) present numerous challenges. First, optimizing the model for all distributed energy resources within a VPP makes global optimization difficult due to the sheer number of variables. Second, VPPs face the burden of maintaining and managing detailed VPP models, often resulting in erroneous or missing model information. Furthermore, in the future, VPPs will be reluctant to disclose their internal model information due to competitive power markets, leaving VPPs with limited access to their bid data, hindering the ability to perform optimized dispatch calculations based on detailed models. Consequently, VPPs need to effectively encapsulate the dispatching characteristics of their internal distributed energy resources to obtain aggregated parameters that represent their overall performance.

[0004] Regarding the optimization aggregation methods of the adjustable power domain of virtual power plants, their aggregation accuracy and efficiency have attracted widespread attention from scholars at home and abroad, such as the half-plane approximation method and the ring-band approximation method. In fact, considering the obvious random volatility of distributed energy, the aggregation parameters of the virtual power plant should be a set of random rather than deterministic model parameters. Directly using the parameters obtained by the above methods will have a significant impact on the safe and stable operation of the power system. To address this issue, reliability-based methods limit the adjustable power domain of virtual power plants to an artificially specified reliability level in the form of constraints, lacking the coordination between reliability and economy. Further research is needed on the aggregation methods of the adjustable power domain of virtual power plants based on cost-benefit analysis. Summary of the invention:

[0005] The purpose of the present invention is to provide a method for aggregating the adjustable power domain of a virtual power plant based on cost-benefit analysis, and to design an internal approximate solution method for a high-dimensional convex polyhedron of the adjustable power domain of a virtual power plant from the perspective of geometric space. On this basis, the uncertainty of the distribution network dispatching instructions and the uncertainty of the distributed energy power are considered, and a two-stage robust optimization aggregation model of the adjustable power domain of a virtual power plant is established with the goal of optimizing the comprehensive economic benefits, thereby ensuring the reliability and economy of the solution.

[0006] To achieve the above objectives, the present invention provides a method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis, comprising:

[0007] Construct a heterogeneous distributed energy model that takes risk-based reserve capacity into account;

[0008] Constructing a virtual generator model that represents the adjustable power domain of the virtual power plant, and optimizing and solving the unknown parameters of the virtual generator model using a high-dimensional convex polyhedron interior approximation method;

[0009] Considering the uncertainty of distribution network dispatching instructions and distributed energy power, a two-stage robust optimization aggregation model with adjustable power domain is constructed with the goal of maximizing comprehensive economic benefits.

[0010] The two-stage robust optimization aggregation model is simplified using piecewise linearization and strong duality theory methods, and is solved using the CC&G algorithm and Gurobi solver.

[0011] Furthermore, the heterogeneous distributed energy model that takes into account risk reserve capacity includes an energy storage device model, a temperature control load model, an electric vehicle model, a diesel generator model, and a wind turbine model; the cost function of each model includes the power cost and risk reserve cost within the dispatchable period of each model, and the constraints of the cost function of each model include power constraints and reserve constraints.

[0012] Furthermore, the power constraints of the energy storage device model and the temperature control load model include upper / lower power constraints, upward / downward climbing constraints, and upper / lower energy constraints; the power constraints of the electric vehicle include upper / lower power constraints and upper / lower energy constraints; the power constraints of the diesel generator set model include upper / lower power constraints and upward / downward climbing constraints; the power constraints of the wind turbine set model include upper / lower power constraints; the reserve constraints include upward reserve constraints and downward reserve constraints.

[0013] Furthermore, a virtual generator model is used to approximate the actual adjustable power domain of the virtual power plant. From the perspective of geometric space, the unknown parameters of the virtual generator model are optimized using the high-dimensional convex polyhedron interior approximation method. The objective function of the virtual generator model is:

[0014]

[0015] The constraints are:

[0016]

[0017] in, is the lower limit of the power required by the virtual generator; is the upper limit of the power required by the virtual generator; is the upper limit of the virtual generator's pending upward ramp; is the upper limit of the virtual generator's downward ramp-down requirement; It is the actual adjustable power domain of the virtual power plant; represents the virtual electric field approximation of the adjustable power domain, Approximate adjustable power domain for virtual power plants with only upper / lower power constraints; Approximate adjustable power domain for virtual power plant with only up / down ramp constraints; The actual adjustable power domain of the virtual power plant only contains the upper / lower limit constraints of distributed energy power and the upper / lower limit constraints of distributed energy energy; It is the actual adjustable power domain of the virtual power plant that only contains the up / down ramp constraints of distributed energy resources.

[0018] Furthermore, from the perspective of geometric space, a virtual power plant with only upper / lower power constraints can approximate an adjustable power domain. It is equivalent to a polyhedron with rectangular faces. The "lower left" vertex and "upper right" vertex of the polyhedron are both located in the actual adjustable power domain of the virtual power plant. within the equivalent polyhedron.

[0019] Furthermore, from the perspective of geometric space, the virtual power plant with only up / down ramp constraints can approximate the adjustable power domain Equivalent to the enclosed part of the parallel plane group, the enclosed part of the parallel plane group is located in the actual adjustable power domain of the virtual power plant within the enclosed portion of the equivalent parallel plane group.

[0020] Furthermore, the two-stage robust optimization aggregation model of the adjustable power domain includes a main problem and a sub-problem; the main problem comprehensively considers the peak-shaving capacity benefit, risk backup cost, and risk penalty cost, and coordinates the allocation of the peak-shaving capacity and backup capacity of each distributed energy in the virtual power plant; the sub-problem arranges the power generation plan of each distributed energy in the virtual power plant with the goal of optimizing the net power generation benefit of the virtual power plant under the worst scenario of the peak-shaving instruction of the distribution network. The constraints include the power balance constraint between the virtual power plant and the distributed energy, the power constraint of each distributed energy, and the power constraint of the virtual power plant as a whole, and the result is fed back to the main problem to correct the adjustable power domain of the virtual power plant.

[0021] Furthermore, the piecewise linearization method is used to solve the power cost functions of the temperature-controlled load, electric vehicles, and distributed energy of diesel generator sets in the sub-problems, and the power shortage expected value function and power surplus expected value function of the virtual power plant in the main problem; the strong duality theory is used to transform the double-layer linear optimization problem of the sub-problem into a single-layer linear optimization problem for solution; the CC&G algorithm and Gurobi solver are used to solve the main and sub-problems in a double-layer iterative manner.

[0022] Furthermore, the CC&G algorithm and Gurobi solver are used to perform a two-level iterative solution to the main and subproblems, including passing the optimal solution of the virtual power plant's adjustable power domain and risk reserve capacity decision variables in the main problem into the subproblem, so that the subproblem can find the worst-case scenario for the distribution network dispatching instructions; using the worst-case scenario for the distribution network dispatching instructions obtained from the subproblem to generate new constraints and add them to the main problem, updating the optimal solution of the main problem; setting a certain convergence accuracy to obtain results that meet the requirements. Description of the drawings:

[0023] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for use in the implementation. Obviously, the drawings described below are only some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 It is a flow chart of a method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis provided by the present invention;

[0025] Figure 2 It is a schematic diagram of an optimization model of a virtual power plant adjustable power domain aggregation method based on cost-benefit analysis provided by the present invention;

[0026] Figure 3 It is a schematic diagram of the piecewise linearization of a curve provided by the present invention;

[0027] Figure 4 This is a schematic diagram of the peak and valley time-of-use electricity prices for general industrial and commercial use in the Beijing power grid according to an embodiment of the present invention;

[0028] Figure 5 2 is a schematic diagram of a temperature curve of a typical day in winter in Beijing according to an embodiment of the present invention;

[0029] Figure 6 is a schematic diagram of a predicted power curve of a wind turbine generator system according to an embodiment of the present invention;

[0030] Figure 7 : is a schematic diagram of the results of optimizing and aggregating the adjustable power domains of a virtual power plant according to an embodiment of the present invention; wherein, Figure 7 (a) is the result of the optimized aggregation of adjustable power domains for Scheme 1; Figure 7 (b) Aggregation results of adjustable power domain optimization for Scheme 5;

[0031] Figure 8 1. It is a schematic diagram of the benefits of optimizing and aggregating adjustable power domains of a virtual power plant under different reliability levels according to an embodiment of the present invention; Specific implementation method:

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0033] It should be understood that the step numbers used herein are only for convenience of description and are not intended to limit the order in which the steps are to be executed.

[0034] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0035] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0036] The term "and / or" refers to and includes any and all possible combinations of one or more of the associated listed items.

[0037] Figure 1 A method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis is proposed in an embodiment of the present invention. The method comprises the following steps:

[0038] S101: Construct a heterogeneous distributed energy model that takes into account risk-based reserve capacity. Specific content includes:

[0039] (1) Energy storage device model:

[0040] The cost function of the energy storage device model is:

[0041]

[0042] in, The power cost of the energy storage equipment, including charging, discharging, operation and maintenance costs; Risk backup cost for energy storage equipment; The charging power of the energy storage device; is the discharge power of the energy storage device; is the operation and maintenance cost coefficient corresponding to the charging power of the energy storage equipment; is the operation and maintenance cost coefficient corresponding to the discharge power of the energy storage equipment; To provide reserve for energy storage equipment; To provide reserve for the upward adjustment of energy storage equipment; Provide a risk reserve cost coefficient for energy storage equipment to reduce reserve; Provides a risk reserve cost coefficient for increased reserve for energy storage equipment.

[0043] The power constraints of the energy storage device model include upper / lower power constraints, upward / downward ramp constraints, and upper / lower energy constraints:

[0044]

[0045] in, The upper limit of the charging power of the energy storage device; The upper limit of the discharge power of the energy storage device; It is the upper limit of the upward climbing of the energy storage equipment; The downward climbing limit of the energy storage device; e ES (t) is the stored energy of the energy storage device; θ ES is the energy consumption coefficient of the energy storage equipment; Charging efficiency of energy storage devices; is the discharge efficiency of the energy storage device; Δt is the time interval; E ESThe lower limit of energy storage for energy storage equipment; The upper limit of energy storage capacity of the energy storage device.

[0046] The reserve constraints of the energy storage device model include upward reserve constraints and downward reserve constraints:

[0047]

[0048] Among them, τ ES is the maximum backup response time of the energy storage device.

[0049] (2) Temperature control load model:

[0050] The cost function of the temperature control load model is:

[0051]

[0052] in, is the power cost of the temperature control load, including electricity purchase income and incentive costs; Risk backup cost for temperature control load; is the power consumption of the temperature control load; The adjustment amount of power consumption relative to planned power in direct load control mode for temperature control load; e TCR is the electricity purchase price for temperature control load; a TCR is the excitation coefficient of the temperature control load; To serve as a standby for lowering the temperature control load; To serve as a backup for increasing the temperature control load; The risk reserve cost coefficient for providing downward reserve for temperature control load; The risk reserve cost coefficient for providing upward reserve for temperature control load.

[0053] The power constraints of the temperature control load model include upper / lower power constraints, upward / downward ramp constraints, and upper / lower energy constraints:

[0054]

[0055] in, The lower limit of power consumption of temperature control load; The upper limit of power consumption of the temperature control load; It is the upper limit of the upward slope of the temperature control load; The upper limit of the downward slope of the temperature control load; e TCR (t) is the real-time temperature of the temperature-controlled load; θ TCR is the energy consumption coefficient of the temperature control load; is the electric-thermal conversion efficiency of the temperature control load; w TCR (t) is the ambient temperature of the temperature-controlled load; Δt is the time interval; E TCRis the lower temperature limit of the temperature control load; It is the upper temperature limit of the temperature control load.

[0056] The reserve constraints of the temperature control load model include upward reserve constraints and downward reserve constraints:

[0057]

[0058] Among them, τ TCR is the maximum backup response time of the energy storage device.

[0059] (3) Electric vehicle model:

[0060] The cost function of the electric vehicle model is:

[0061]

[0062] in, The power cost of electric vehicles, including electricity purchase income and incentive costs; Backup costs for the risks of electric vehicles; Charging power for electric vehicles; The regulation amount of charging power relative to planned power in direct load control mode for electric vehicles; e EV is the electricity purchase price of electric vehicles; a EV is the incentive coefficient of electric vehicles; reserve for electric vehicles’ downgrades; reserve for the increase in electric vehicles; Provide risk reserve cost coefficient for electric vehicles with reduced reserve; Provide risk reserve cost coefficient for electric vehicles with increased reserve.

[0063] The power constraints of the electric vehicle model include upper / lower power constraints and upper / lower energy constraints:

[0064]

[0065] in, The lower limit of charging power for electric vehicles; The upper limit of charging power for electric vehicles; e EV (t) is the stored energy of the electric vehicle; θ EV is the energy consumption coefficient of electric vehicles; is the charging efficiency of the electric vehicle; Δt is the time interval; E EV is the lower energy limit of electric vehicles; The energy limit for electric vehicles; The lower limit of energy required by the owner at the end of charging of the electric vehicle.

[0066] The reserve constraints of the electric vehicle model include upward reserve constraints and downward reserve constraints:

[0067]

[0068] (4) Diesel engine model:

[0069] The cost function of the diesel generator set model is:

[0070]

[0071] in, is the power cost of the diesel generator set, including fuel consumption cost; The risk reserve cost of the diesel generator set; is the generating power of the diesel generator set; a DG 、b DG 、c DG is the coefficient of the quadratic cost function of the diesel generator set; To serve as standby for diesel generator sets; To serve as standby for diesel generator sets; Provide risk reserve cost coefficient for diesel generator sets to be lowered to reserve; Provides risk reserve cost coefficient for diesel generator sets to increase reserve.

[0072] The power constraints of the diesel generator set model include power upper / lower limit constraints and upward / downward ramp constraints:

[0073]

[0074] in, is the lower limit of the diesel generator set's power generation capacity; The upper limit of the power generation capacity of the diesel generator set; It is the upper limit of the upward climbing of the diesel generator set; It is the upper limit of the downward climbing of the diesel generator set.

[0075] The reserve constraints of the diesel generator set model include upward reserve constraints and downward reserve constraints:

[0076]

[0077] Among them, τ DG is the maximum backup response time of the energy storage device.

[0078] (5) Wind turbine model:

[0079] The cost function of the wind turbine model is:

[0080]

[0081] in, is the power cost of the wind turbine, which is approximately zero; is the risk reserve cost of wind turbines; is the power generation capacity of the wind turbine; To serve as standby for wind turbines; To provide standby for wind turbines; Provide risk reserve cost coefficient for wind turbines to provide reserve for downgrade; Provide risk reserve cost coefficient for wind turbines to increase reserve capacity.

[0082] The power constraints of the wind turbine model include upper and lower power constraints:

[0083]

[0084] in, is the power generation capacity of the wind turbine; is the predicted power of the wind turbine; u(t) is the power deviation of the wind turbine that obeys a certain probability distribution.

[0085] The reserve constraints of the wind turbine model include upward reserve constraints and downward reserve constraints:

[0086]

[0087] S102: Construct a virtual generator model that represents the adjustable power domain of the virtual power plant and design a high-dimensional convex polyhedron interior approximation method to solve its parameters. Specific contents include:

[0088] The actual adjustable power domain of the virtual power plant is the Minkowski sum of the actual adjustable power domain of distributed energy sources such as energy storage equipment, temperature control loads, electric vehicles, diesel generator sets, and wind turbines in the virtual power plant:

[0089]

[0090] in, It is the actual adjustable power domain of the virtual power plant; The actual adjustable power domain of the energy storage device; The actual adjustable power range for the temperature-controlled load; A practically adjustable power range for electric vehicles; It is the actual adjustable power range of the diesel generator set; It is the actual adjustable power range of the wind turbine.

[0091] The actual adjustable power domains of distributed energy resources within a virtual power plant have characteristics such as power constraint mutuality, power timing coupling, and differentiated adjustable time periods. The specific content has been described in step S101 above. Therefore, the aggregation operation of the adjustable power domain of the virtual power plant based on Minkowski sum is an NP-hard problem, and the actual adjustable power domain of the virtual power plant is unknown. The virtual generator model is used to approximate the actual adjustable power domain of the virtual power plant:

[0092]

[0093] in, represents the virtual electric field approximation adjustable power domain, p VPP (t) is the power generated by the virtual generator; is the lower limit of the power required by the virtual generator; is the upper limit of the power required by the virtual generator; is the upper limit of the virtual generator's pending upward ramp; It is the upper limit of the virtual generator's pending downward ramp.

[0094] From the perspective of geometric space, a high-dimensional convex polyhedron interior approximation method is designed to optimize the unknown parameters of the virtual generator model:

[0095]

[0096] Specifically, the objective function can be expanded as:

[0097]

[0098] Specifically, the constraints can be expanded as follows:

[0099]

[0100] in, Approximate adjustable power domain for virtual power plants with only upper / lower power constraints; Approximate adjustable power domain for virtual power plant with only up / down ramp constraints; The actual adjustable power domain of the virtual power plant only contains the upper / lower limit constraints of distributed energy power and the upper / lower limit constraints of distributed energy energy; It is the actual adjustable power domain of the virtual power plant that only contains the up / down ramp constraints of distributed energy resources.

[0101] Virtual Power Plant Approximation of Adjustable Power Domain From the perspective of geometric space, it can be equivalent to a polyhedron with rectangular faces. It only needs that the "lower left" vertex and "upper right" vertex of the polyhedron are both located in the actual adjustable power domain of the virtual power plant. The virtual power plant can be approximately adjusted within the equivalent polyhedron of All equivalent polyhedrons are located in the actual adjustable power domain of the virtual power plant Within the equivalent polyhedron of :

[0102]

[0103] in, The “upper right” vertex of the equivalent polyhedron that approximates the adjustable power domain of the virtual power plant; The “lower left” vertex of the equivalent polyhedron that approximates the adjustable power domain of the virtual power plant.

[0104] Virtual Power Plant Approximation of Adjustable Power Domain From the perspective of geometric space, it can be equivalent to the enclosed part of the parallel plane group. It only needs to ensure that the enclosed part of the parallel plane group is located in the actual adjustable power domain of the virtual power plant. The enclosed portion of the equivalent parallel plane group is:

[0105]

[0106] in, is the downward climbing limit of distributed energy i in the virtual power plant; It is the upper limit of the upward climbing of distributed energy i in the virtual power plant.

[0107] S103: Considering the uncertainty of distribution network dispatching instructions and distributed energy power, with the goal of maximizing comprehensive economic benefits, a two-stage robust optimization aggregation model with adjustable power domain is constructed, such as Figure 2 The specific contents include:

[0108] (1) The main problem of the two-stage robust optimization aggregation model of the adjustable power domain of the virtual power plant is to comprehensively consider multiple objectives such as peak-shaving capacity benefits, risk backup costs, and risk penalty costs, coordinate the allocation of peak-shaving capacity and backup capacity of each distributed energy source in the virtual power plant, and use the method proposed in S102 to solve the adjustable power domain of the virtual power plant:

[0109] The objective function of the main problem is:

[0110]

[0111] in, The risk reserve cost of the virtual power plant; Penalize costs for the risks of virtual power plants; is the peak-shaving capacity benefit of the virtual power plant.

[0112] Specifically, the cost of electricity generated by the actual use of reserve capacity during operation is not considered, and only the cost of reserve capacity is considered. The risk reserve cost of the virtual power plant is is the sum of the risk backup costs of all distributed energy resources in the virtual power plant:

[0113]

[0114] Specifically, the expected loss is used to describe the improvement level of the average reliability of the virtual power plant due to the backup, and the risk penalty cost of the virtual power plant is is the sum of the power shortage risk penalty cost and power excess risk penalty cost of the virtual power plant:

[0115]

[0116] Where EENS(t) is the expected power shortage of the virtual power plant; EEAN(t) is the expected power surplus of the virtual power plant; EENS (t) is the power shortage risk penalty coefficient of the virtual power plant; e EEAN (t) is the power excess risk penalty coefficient of the virtual power plant; f(u t ) is the probability distribution of wind turbine power deviation; Reserve capacity for upside risk of virtual power plants; It is the downside risk reserve capacity of the virtual power plant.

[0117] Specifically, the peak-shaving capacity electricity price is considered as the weight coefficient, and the objective function in the method proposed in S102 is used as the peak-shaving capacity benefit:

[0118]

[0119] Among them, r VPP (t) is the peak-shaving capacity electricity price of the virtual power plant.

[0120] The constraints of the main problem are:

[0121]

[0122] in, It is the actual adjustable domain of the virtual power plant represented by the power constraints and reserve constraints of each distributed energy in the virtual power plant.

[0123] (2) The sub-problem of the two-stage robust optimization aggregation model of the virtual power plant's adjustable power domain is to optimize the net power generation efficiency of the virtual power plant under the worst-case scenario of the distribution network's peak load instruction. The power generation plan of each distributed energy source in the virtual power plant is arranged, and the result obtained from the sub-problem is fed back to the main problem to correct the adjustable power domain of the virtual power plant:

[0124] The objective function of the subproblem is:

[0125]

[0126] in, is the total power generation cost of the virtual power plant; The original planned power generation benefits of the virtual power plant; For the upward peak mileage benefit of virtual power plants; It is the downward peak shaving mileage benefit of the virtual power plant.

[0127] Specifically, the total power generation cost of the virtual power plant is the sum of the power costs of each distributed energy source in the virtual power plant:

[0128]

[0129] Specifically, the original planned power generation benefits of the virtual power plant are:

[0130]

[0131] Among them, e VPP (t) is the market price of electricity; is the original planned power generation of the virtual power plant.

[0132] Specifically, the upward peak-shaving mileage benefit of the virtual power plant is:

[0133]

[0134] in, To adjust the peak mileage electricity price upwards; To adjust the peak capacity upwards; It is the per-unit value of the upward peak-shaving mileage and has uncertainty.

[0135] Specifically, the downward peak-shaving mileage benefit of the virtual power plant is:

[0136]

[0137] in, To adjust the peak mileage electricity price downward; To adjust the peak capacity downward; It is the per-unit value of downward peak-shaving mileage and has uncertainty.

[0138] The constraints of the sub-problems include the power balance constraint between the virtual power plant and distributed energy resources, the power constraints of the distributed energy resources themselves, and the power constraints of the virtual power plant as a whole:

[0139]

[0140] Among them, p i(t) is a simplified representation of the power of each distributed energy source in the virtual power plant.

[0141] p i (t)∈Ω i (u i ) (34)

[0142] Among them, Ω i (u i ) is a simplified representation of the actual adjustable power domain of each distributed energy source after taking into account the risk reserve constraint.

[0143]

[0144] S104: Simplify the two-stage robust optimization aggregation model using piecewise linearization and strong duality theory, and solve it using the CC&G algorithm and Gurobi solver. Specific content includes:

[0145] (1) Figure 3 As shown in the figure, the power cost functions of distributed energy resources such as temperature-controlled loads, electric vehicles, and diesel generator sets in the sub-problems of the two-stage robust optimization aggregation model for the adjustable power domain of the virtual power plant are nonlinear concave functions. In order to facilitate the Gurobi solver to solve quickly and directly, they need to be piecewise linearized:

[0146]

[0147] Among them, M is the dependent variable; n is the independent variable; k i is the slope of the i-th segment; (M i ,n i ) is the end point of the i-th segment; l is the number of segments.

[0148] Correspondingly, minM(n) can be transformed into:

[0149]

[0150] (2) The sub-problem of the two-stage robust optimization aggregation model for the adjustable power domain of the virtual power plant is a two-level linear optimization problem. The matrix form of the original problem is expressed as:

[0151]

[0152] In order to facilitate the Gurobi solver to solve quickly and directly, it is necessary to use the strong duality theory to transform it. The matrix form of the dual problem is expressed as:

[0153]

[0154] (3) The CC&G algorithm is used to perform a two-layer iterative optimization solution on the two-stage robust optimization aggregation model of the adjustable power domain of the virtual power plant. The specific steps are as follows:

[0155] 1) Set the lower bound LB of the original problem of the two-stage robust optimization aggregation model of the adjustable power domain of the virtual power plant to -∞, the upper bound UB to +∞, and the total number of iterations num to 0;

[0156] 2) Solve the main problem of the two-stage robust optimization aggregation model of the adjustable power domain of the virtual power plant and obtain the optimal solution And update the lower bound

[0157] 3) The optimal solution of the main problem of the two-stage robust optimization aggregation model of the adjustable power domain of the virtual power plant Substitute into the sub-problem to obtain the worst scenario for peak shaving mileage And update the upper bound

[0158] 4) Set the convergence accuracy ε of the column and constraint generation algorithm. If (UB-LB) / LB≤ε is satisfied, the iteration stops and the optimal solution is finally determined. Otherwise proceed to the next step;

[0159] 5) Based on the sub-problems of the two-stage robust optimization aggregation model of the adjustable power domain of the virtual power plant, the worst peak mileage scenario Add new constraints to the main problem and continue to iterate the solution.

[0160] The following comparative analysis of the comprehensive economic benefits of the adjustable power domain of the virtual power plant obtained by the reliability-based method and the method proposed in the present invention in conventional scenarios and extreme scenarios is conducted through specific examples.

[0161] Example Figure 4The peak-valley time-of-use electricity prices for general industrial and commercial users in the Beijing power grid are shown as the energy market price. The peak-shaving capacity price is set at 15% of the energy market price, the upward peak-shaving mileage price is set at 110% of the energy market price, and the downward peak-shaving mileage price is set at 90% of the energy market price. The expected risk loss penalty price is set at 10 times the maximum risk reserve price of distributed energy resources. There is one energy storage device in total, with an operating and maintenance cost coefficient of 0.15 yuan / kWh for both charging and discharging power, and 0.10 yuan / kWh for both upward and downward risk reserve cost coefficients. The upper and lower limits for charging and discharging power are both 10 kW, and the upper and lower limits for upward and downward ramp rates are both 8 kW / h. The energy storage capacity is 65 kWh, with the upper and lower limits and initial energy storage of 0.20, 0.90, and 0.55 of the energy storage capacity, respectively. The charging and discharging efficiency and energy consumption coefficients are 0.90 and 1.00, respectively. There are 5 temperature control loads in total, and their electricity purchase price and incentive coefficient are 0.55 yuan / kWh and 0.1 yuan / (kWh) respectively. 2 The upward and downward risk standby cost coefficients are both 0.25 yuan / kWh, the upper limit of power consumption is 5 kilowatts, the upper and lower limit of climbing rate are both 4 kilowatts per hour, the electric-thermal conversion efficiency and heat dissipation coefficient are 0.95 and 0.85 respectively, the upper and lower limits of the own temperature and the initial temperature are 16 degrees Celsius, 24 degrees Celsius and 18 degrees Celsius respectively, and the ambient temperature is adopted. Figure 5 The temperature curve of a typical winter day in Beijing is shown below. There are 5 electric vehicles in total, and their electricity purchase price and incentive coefficient are 0.55 yuan / kWh and 0.1 yuan / (kWh) respectively. 2 The upward and downward risk standby cost coefficients are both 0.25 yuan / kWh, the upper limit of charging power is 7 kW, the energy storage capacity is 36 kWh, the upper and lower limits of energy storage, initial energy storage, and departure energy storage are 0.20, 0.90, 0.20, and 0.60 of the energy storage capacity respectively, the charging and discharging efficiency and energy consumption coefficient are 0.90 and 1.00 respectively, and the dispatchable period is 9:00-15:00. There is a total of 1 diesel generator set, and its fuel consumption cost coefficient is 0.006 yuan / (kWh). 2 , 0.18 yuan / kWh, 0.05 yuan, the upward and downward risk standby cost coefficients are 0.20 yuan / kWh and 0.30 yuan / kWh respectively, the upper and lower limits of power are 50 kW and 110 kW respectively, and the upper and lower limits of the ramp rate are both 30 kW per hour. There are 2 wind turbines in total, the upward and downward risk standby cost coefficients are 0.25 yuan / kWh and 0.50 yuan / kWh respectively, and the predicted power is Figure 6 The wind power forecast curve shown assumes a normal distribution for the forecast deviation, with the maximum value calculated as 20% of the forecast power. The maximum backup response time for each distributed energy resource is 15 minutes. The relevant calculations were performed on a computer with an Intel Core i5-7400 processor at 3.00GHz and 8GB of memory, using Python for the solution.

[0162] In order to compare and analyze the reliability and economy of the aggregation method model introduced in the embodiment of the present invention, the following comparison scheme is established:

[0163] Option 1: A reliability-based approach is used to solve the adjustable power domain of the virtual power plant, with an extremely low reliability level (the virtual power plant risk reserve capacity is 0% of the maximum power forecast deviation of the virtual power plant);

[0164] Option 2: A reliability-based approach is used to solve the adjustable power domain of the virtual power plant, with a low reliability level (the virtual power plant risk reserve capacity is 15% of the maximum power forecast deviation of the virtual power plant);

[0165] Option 3: A reliability-based approach is used to solve the adjustable power domain of the virtual power plant, with a high reliability level (the virtual power plant risk reserve capacity is 45% of the maximum power forecast deviation of the virtual power plant);

[0166] Option 4: A reliability-based approach is used to solve the adjustable power domain of the virtual power plant, with an extremely high reliability level (the virtual power plant risk reserve capacity is 75% of the maximum power forecast deviation of the virtual power plant);

[0167] Solution 5: Use the cost-benefit analysis-based method proposed in this paper to solve the adjustable power domain of the virtual power plant.

[0168] like Figure 7 As shown, Scheme 1 achieves the maximum adjustable power domain that a virtual power plant can achieve. Its shape is influenced by the electricity market clearing price: when the peak-shaving capacity price is high during a certain period, the peak-shaving capacity is larger; when the peak-shaving capacity price is low during a certain period, the peak-shaving capacity is smaller. However, to mitigate the economic losses of power generation caused by the uncertainty of distribution network dispatch instructions and the risk penalty losses caused by the uncertainty of wind turbine power, the adjustable power domain of the virtual power plant obtained by Scheme 5 is only about 50% of the adjustable power domain of the virtual power plant obtained by Scheme 1, and its shape has also changed significantly.

[0169] At the same time, Monte Carlo simulation technology was used to generate 1,000 scenarios within the uncertain range of distribution network peak-shaving instructions and wind turbine output power. The average comprehensive economic benefit of this set of scenarios was used as the comprehensive economic benefit under the conventional scenario, and the minimum comprehensive economic benefit was used as the comprehensive economic benefit under the extreme scenario. As shown in Table 1, although Scheme 5 has lower peak-shaving capacity benefits than Scheme 1, it achieves higher power generation economic benefits and lower risk penalty losses, improving the overall economic benefits. Furthermore, Scheme 5 has a more significant risk avoidance advantage in extreme scenarios than in conventional scenarios. The proposed method fully guarantees the reliability and economic efficiency of the optimized aggregation of the adjustable power domain of the virtual power plant.

[0170] Table 1

[0171]

[0172] In addition, if Figure 8 As shown in the figure, through the comparative analysis of the comprehensive economic benefits of Schemes 1-4 and Scheme 5 in different scenarios, it can be seen that the reliability level of the adjustable power domain of the virtual power plant obtained by the reliability-based method is not strictly proportional to the economic benefit level. It will be difficult to achieve the coordination of reliability and economy by relying on artificially defined reliability levels. This further verifies the advantage of the method proposed in the present invention in taking into account both reliability and economy in optimizing the aggregation of the adjustable power domain of the virtual power plant.

[0173] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis, comprising: Construct a heterogeneous distributed energy model that takes risk-based reserve capacity into account; A virtual generator model is used to approximate the actual adjustable power domain of the virtual power plant. From the perspective of geometric space, the unknown parameters of the virtual generator model are optimized using the high-dimensional convex polyhedron interior approximation method. The objective function of the virtual generator model is: ， The constraints are: ， in, is the lower limit of the power required by the virtual generator; is the upper limit of the power required by the virtual generator; is the upper limit of the virtual generator's pending upward ramp; is the upper limit of the virtual generator's downward ramp-down requirement; It is the actual adjustable power domain of the virtual power plant; represents the virtual electric field approximation of the adjustable power domain, Approximate adjustable power domain for virtual power plants with only upper / lower power constraints; Approximate adjustable power domain for virtual power plant with only up / down ramp constraints; The actual adjustable power domain of the virtual power plant only contains the upper / lower limit constraints of distributed energy power and the upper / lower limit constraints of distributed energy energy; The actual adjustable power domain of the virtual power plant containing only the up / down ramp constraints of distributed energy resources; From the perspective of geometric space, the virtual power plant with only upper / lower power constraints can approximate the adjustable power domain. It is equivalent to a polyhedron with rectangular faces. The "lower left" and "upper right" vertices of the polyhedron are both located in the actual adjustable power domain of the virtual power plant. within the equivalent polyhedron; From the perspective of geometric space, the virtual power plant with only up / down ramp constraints approximates the adjustable power domain Equivalent to the enclosed part of the parallel plane group, the enclosed part of the parallel plane group is located in the actual adjustable power domain of the virtual power plant within the enclosed portion of the equivalent parallel plane group; Considering the uncertainty of distribution network dispatching instructions and distributed energy power, a two-stage robust optimization aggregation model with adjustable power domain is constructed with the goal of maximizing comprehensive economic benefits. The two-stage robust optimization aggregation model is simplified by using piecewise linearization and strong duality theory methods, and the Algorithms and The solver solves it.

2. A method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis according to claim 1, characterized in that: The heterogeneous distributed energy model that takes into account risk reserve capacity includes an energy storage device model, a temperature control load model, an electric vehicle model, a diesel generator model, and a wind turbine model; the cost function of each model includes the power cost and risk reserve cost within the dispatchable period of each model, and the constraints of the cost function of each model include power constraints and reserve constraints.

3. The method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis according to claim 2, characterized in that: The power constraints of the energy storage device model and the temperature control load model include upper / lower power constraints, upward / downward climbing constraints, and upper / lower energy constraints; the power constraints of the electric vehicle include upper / lower power constraints and upper / lower energy constraints; the power constraints of the diesel generator set model include upper / lower power constraints and upward / downward climbing constraints; the power constraints of the wind turbine set model include upper / lower power constraints; the reserve constraints include upward reserve constraints and downward reserve constraints.

4. The method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis according to claim 1, characterized in that: The two-stage robust optimization aggregation model of the adjustable power domain includes a main problem and a sub-problem; the main problem comprehensively considers the peak-shaving capacity benefits, risk backup costs, and risk penalty costs, and coordinates the allocation of the peak-shaving capacity and backup capacity of each distributed energy in the virtual power plant; the sub-problem arranges the power generation plan of each distributed energy in the virtual power plant with the goal of optimizing the net power generation benefit of the virtual power plant under the worst scenario of the peak-shaving instruction of the distribution network. The constraints include the power balance constraint between the virtual power plant and the distributed energy, the power constraints of each distributed energy, and the power constraints of the virtual power plant as a whole, and the result is fed back to the main problem to correct the adjustable power domain of the virtual power plant.

5. The method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis according to claim 4, characterized in that: The power cost functions of temperature control load, electric vehicle and diesel generator set distributed energy in the sub-problem and the power shortage expected value function and power surplus expected value function of virtual power plant in the main problem are solved by piecewise linearization method; the double-layer linear optimization problem of the sub-problem is converted into a single-layer linear optimization problem by using strong duality theory; Algorithms and The solver solves the main and subproblems in a two-level iterative manner.

6. A method for aggregating adjustable power domains of a virtual power plant based on cost-benefit analysis according to claim 5, characterized in that: use Algorithms and The solver's two-level iterative solution to the main and subproblems includes passing the optimal solution of the virtual power plant's adjustable power domain and risk reserve capacity decision variables in the main problem into the subproblem, which is used to find the worst-case scenario for the distribution network dispatching instructions in the subproblem; using the worst-case scenario for the distribution network dispatching instructions obtained from the subproblem to generate new constraints and add them to the main problem, updating the optimal solution to the main problem; setting a certain convergence accuracy to obtain results that meet the requirements.

Images