A multi-energy virtual power plant economic and ancillary service optimization method

Abstract

The present application relates to a kind of multi-purpose virtual power plant economy and auxiliary service optimization method, belong to virtual power plant dispatching technical field, including the following steps: considering renewable energy uncertainty, construct virtual power plant economy and auxiliary service joint optimization model;Using the hybrid optimization strategy of fusion improved particle swarm optimization algorithm and snow melting optimizer is solved to joint optimization model, obtains optimal scheduling scheme.The present application realizes the collaborative balance between minimum operating cost and auxiliary service maximization;Optimization method has the advantages of few parameters, convergence fast, high precision, significantly enhances the solving performance of model and the optimization quality of scheduling scheme.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of virtual power plant dispatching, and particularly relates to a multi-energy virtual power plant economic and auxiliary service optimization method. BACKGROUND

[0002] In recent years, China has vigorously promoted the green and low-carbon transformation of energy, and renewable energy development has achieved new breakthroughs, entering a new stage of large-scale and high-quality leap-forward development. Domestic renewable energy project construction momentum is good, and will still maintain a high level of production in the future.

[0003] However, renewable energy itself has natural intermittency, volatility and randomness. Taking photovoltaic as an example, its output time is highly concentrated, and is limited by distributed layout and power grid dispatching mechanism, causing the problem of "producing, not sending, not using" increasingly prominent. The concentrated output of new energy reshapes the traditional load curve, forming a "deep valley" in the net load of the power grid in some periods, and the demand for load shifting, peak shaving and valley filling is extremely urgent. These problems will be further aggravated after large-scale renewable energy access, which is easy to cause power frequency fluctuation and voltage instability, seriously affecting power quality and system safety.

[0004] To cope with the above challenges, virtual power plant (VPP) emerges as a new type of aggregation and coordination mechanism. VPP integrates distributed power sources, controllable loads and energy storage devices, realizes unified monitoring and optimal dispatching, and improves the controllability and market adaptability of renewable energy. Virtual power plant is becoming an important support to ensure the safe operation of the power grid and improve the utilization rate of new energy.

[0005] At the same time, with the continuous advancement of new-type power system construction, China's power auxiliary service market is also accelerating improvement. The system's demand for frequency modulation, backup, rapid ramping and other regulation resources has significantly increased. With the characteristics of resource aggregation and rapid response, virtual power plant plays an increasingly important role in the auxiliary service market. Its ability to coordinate renewable energy output and grid demand is becoming one of the key ways to promote the green, safe and efficient development of energy systems. However, the current virtual power plant does not consider the problem of power auxiliary service market in the dispatching of resources.

[0006] In addition, the virtual power plant coordinated dispatching model has the characteristics of high nonlinearity, strong coupling and multiple constraints. Traditional analytical methods are difficult to solve directly, and linearization processing is easy to lose the essential structure of the problem. Therefore, intelligent optimization algorithms are needed to search the global variable space to obtain the optimal dispatching strategy. However, in the process of global search, the algorithm is easy to fall into local maximum solution. SUMMARY

[0007] The present application aims to provide a multi-energy virtual power plant economic and ancillary service optimization method to solve the problem that the existing virtual power plant resource scheduling algorithm does not consider the power ancillary service market and the algorithm is prone to fall into local maximum solution.

[0008] In order to achieve the above-mentioned purpose, the technical scheme of the present application is as follows:

[0009] The present application relates to a multi-energy virtual power plant economic and ancillary service optimization method, which comprises the following steps:

[0010] S1. Considering the uncertainty of renewable energy, a virtual power plant economic and ancillary service joint optimization model is constructed;

[0011] S2. A hybrid optimization strategy combining improved particle swarm optimization algorithm and snowmelt optimizer is used to solve the joint optimization model to obtain the optimal scheduling scheme.

[0012] Preferably, the S1 of constructing a virtual power plant economic and ancillary service joint optimization model comprises: in the typical daily scheduling scenario, a variety of resources in the virtual power plant are uniformly modeled, and an auxiliary service participation mechanism and a benefit evaluation model are constructed.

[0013] Preferably, the S1 of uniformly modeling a variety of resources in the virtual power plant comprises: establishing a minimum virtual power plant operation cost objective function and establishing virtual power plant constraint conditions.

[0014] Preferably, the expression of the minimum virtual power plant operation cost objective function is:

[0015] (1),

[0016] (2),

[0017] (3),

[0018] Among them, represents the minimum virtual power plant operation cost, t represents the t th scheduling period, T represents the set of scheduling periods, , , , , , The energy cost of the power operator, wind turbine, photovoltaic module, fuel cell system, micro gas turbine and energy storage device in the scheduling period t , , ,​ , , are the power injections of the power operator, wind turbine, photovoltaic component, fuel cell system, micro gas turbine and energy storage device at dispatch time period t ; is the flexible load adjustment variable, representing the power adjustment amount achieved by the postponable or reducible load within the dispatch time period t ; , , , are the operating states of the wind turbine, photovoltaic component, fuel cell system, micro gas turbine at dispatch time period t, with "1" representing operation and "0" representing shutdown; and represent the operating states of the wind turbine, photovoltaic component, fuel cell system, micro gas turbine at dispatch time period t and dispatch time period t-1, with "1" representing operation and "0" representing shutdown, represents the operating cost of the energy storage device, and represent the operating states of the energy storage device, also taking 0 / 1, for the start-stop cost calculation of the energy storage unit; , are the start-stop costs of the i th distributed power generation unit and the j th energy storage device; , are the numbers of distributed power generation units and energy storage devices; , are the resistance and current of the b th branch at dispatch time period t ; , , are the original loss, final loss and new loss after optimization of the system; , are the unit loss cost and the power difference before and after optimization; is the power sold to the grid, is the power selling price, represents the flexible load adjustment cost, represents the total number of grid branches in the power system, i.e. the number of transmission lines participating in line power loss calculation in the entire distribution network.

[0019] Preferably, the virtual power plant constraint conditions include:

[0020] a power balance constraint, the expression of which is:

[0021] (4),

[0022] wherein, , Pdisch and Pcharge are the discharging and charging power of the energy storage device in the i-th dispatching period, t Ploss is the line loss power in the i-th dispatching period, Pload is the load demand in the i-th dispatching period; t t

[0023] The output power constraint of the wind turbine is expressed as:

[0024] (5),

[0025] wherein, , Pmax and Pmin are the maximum and minimum power of the output of the wind turbine, respectively;

[0026] The output power constraint of the photovoltaic module is expressed as:

[0027] (6),

[0028] wherein, , Pmax and Pmin are the maximum and minimum power of the output of the photovoltaic module, respectively;

[0029] The output limit of the micro gas turbine is expressed as:

[0030] (7),

[0031] wherein, , Pmax and Pmin are the maximum and minimum power of the output of the micro gas turbine, respectively;

[0032] The output limit of the fuel cell system is expressed as:

[0033] (8),

[0034] wherein, , Pmax and Pmin are the maximum and minimum power of the output of the fuel cell system, respectively;

[0035] The limit of the electricity purchase from the power operator is expressed as:

[0036] (9),

[0037] wherein, , Pmax and Pmin are the maximum and minimum power of the electricity purchase from the power operator, respectively;​​​

[0038] The adjustable load output limit is expressed as follows:

[0039] (10)

[0040] in, , These are the maximum and minimum power of the adjustable load of the energy storage device, respectively.

[0041] The expression for flexible load regulation limit is:

[0042] (11),

[0043] in, This is the upper limit for flexible load adjustment;

[0044] The charging and discharging limits of energy storage devices are expressed as follows:

[0045] (12),

[0046] (13)

[0047] (14)

[0048] (15)

[0049] (16)

[0050] in, , For energy storage devices in the first t The charging and discharging power during each scheduling period, , Energy storage devices in the t The maximum charging and discharging power during each scheduling period , It is a binary state variable. For the first t Energy storage capacity for each dispatch period This represents the stored energy level at the previous moment; Energy storage charging efficiency, Energy storage discharge efficiency;

[0051] The power sales constraint is expressed as follows:

[0052] (17)

[0053] (18)

[0054] (19),

[0055] wherein, represents the electricity sales, is the remaining electricity, is the actual load, is a binary variable, M is a constant.

[0056] Preferably, the construction of auxiliary service participation mechanism in S1 and the benefit evaluation model specifically includes:

[0057] The construction service response power allocation constraint is expressed as:

[0058] (20),

[0059] wherein, is the resource r The power generation of the first t dispatching period, , respectively, the reserved frequency modulation and standby service response power, is the maximum output capacity of the resource r ;

[0060] The construction of supply and demand matching constraints is expressed as:

[0061] (21),

[0062] (22),

[0063] wherein, , respectively, the frequency modulation, standby demand;

[0064] The construction service benefit objective function is expressed as:

[0065] (23),

[0066] wherein, , respectively, the market price of the first t dispatching period frequency modulation, standby service, is the total auxiliary service benefit of the virtual power plant in the entire dispatching cycle.

[0067] Preferably, the specific steps of S2 using a hybrid optimization strategy combining the improved particle swarm optimization algorithm and the snow melting optimizer to solve the joint optimization model are:

[0068] S2.1. Global search is performed using the improved particle swarm optimization algorithm to obtain a scheduling scheme;

[0069] S2.2. Evaluate the scheduling scheme with the fitness function, determine whether the iteration and convergence conditions are met, when all the iteration and convergence conditions are met, take the scheduling scheme as the optimal scheduling scheme and output, when any one of the iteration and convergence conditions is not met, go to S2.3;

[0070] S2.3. Perform local fine perturbation and convergence control in combination with the hybrid optimization strategy of the snow melting optimizer, and optimize the scheduling scheme.

[0071] Preferably, the specific way of obtaining the scheduling scheme in S2.1 is to perform global search by using the improved particle swarm optimization algorithm, update the speed and position of the particle, and gradually obtain the scheduling scheme. The update formula of the particle speed and position is:

[0072] (25),

[0073] (26),

[0074] wherein, g indicates the number of iterations, and respectively indicate the speed and position of the particle, is the historical optimal position of the particle p , is the current global optimal position, and are random vectors, indicates the Hadamard product, is the linearly decreasing result of the inertia weight, and are learning factors.

[0075] Preferably, the fitness function in S2.2 is:

[0076] (27),

[0077] wherein, indicates the fitness, indicates the operating cost of the virtual power plant, indicates the weight coefficient, is the total auxiliary service revenue of the virtual power plant in the entire scheduling period;

[0078] Determine whether the following convergence conditions are met, that is, one of the following two conditions is met:

[0079] Condition 1. Reach the maximum number of iterations ;

[0080] Condition 2. Continue The intra-generation optimal fitness function value change is less than a set threshold That is,

[0081] (28).

[0082] Preferably, the specific steps of the S2.3 combining the hybrid optimization strategy of the snowmelt optimizer for local fine perturbation and convergence control include:

[0083] S2.3.1. Based on the solution of the scheduling scheme obtained in S2.1, generate a random set in matrix form and the upper and lower limits of the solution, and the expression of the random set in matrix form is:

[0084] (29),

[0085] The calculation formula of the upper and lower limits is:

[0086] (30),

[0087] Wherein, N and n represent the number of population particles and the dimension of the solution, m represents the dimension index in the matrix, m ∈ n , U , L are the upper and lower limits of the value, is a random number in [0,1];

[0088] S2.3.2. Update the position of the particle by introducing Brownian motion, and the update formula of the particle position is:

[0089] (31),

[0090] Wherein, is the position of the p th particle, g is the current iteration number, is a random individual in the elite particle set, is a random number vector generated by Gaussian distribution of Brownian motion, is the multiplication by row, is a randomly selected individual in the elite group in the population, is the centroid of the entire particle position;

[0091] S2.3.3. Gradually converge through the snowmelt process to find the best solution, and the position update formula in this stage is:

[0092] (32),

[0093] (33),

[0094] wherein: is a random number in [-1, 1], P is a snow-melt model, is the maximum number of iterations, is a random number in [0, 1], denotes the i-th individual in the population, the randomly selected individual position or index in the generation.

[0095] Compared with the prior art, the technical scheme provided by the application has the following beneficial effects:

[0096] 1. The multi-energy virtual power plant economic and auxiliary service optimization method relates to a virtual power plant economic and auxiliary service joint optimization model, and the optimal scheduling scheme obtained by solving the model realizes the collaborative balance between the minimization of operating costs and the maximization of auxiliary service revenue.

[0097] 2. The multi-energy virtual power plant economic and auxiliary service optimization method adopts a hybrid optimization strategy combining an improved particle swarm optimization algorithm and a snow-melt optimizer to solve the joint optimization model and obtain an optimal scheduling scheme. The improved particle swarm optimization algorithm is used to realize global exploration of the solution space, and the snow-melt optimizer is used to locally and finely optimize the candidate solution, taking into account the global search capability and local convergence precision. The method has the advantages of few parameters, fast convergence, high precision, etc., and significantly enhances the solution performance of the model and the optimization quality of the scheduling scheme. BRIEF DESCRIPTION OF DRAWINGS

[0098] Figure 1 is a general flowchart of the multi-energy virtual power plant economic and auxiliary service optimization method;

[0099] Figure 2 is a virtual power plant model structure diagram for adjustable load and energy storage participation;

[0100] Figure 3 is a hybrid optimization strategy algorithm flowchart combining an improved particle swarm optimization algorithm and a snow-melt optimizer;

[0101] Figure 4 is a 24-hour load and renewable energy output diagram;

[0102] Figure 5 is a comprehensive energy system scheduling output diagram after optimization;

[0103] Figure 6 is a system scheduling diagram without auxiliary service participation;

[0104] Figure 7 This is a system scheduling diagram for scenarios with no adjustable load. Detailed Implementation

[0105] To further understand the content of this invention, the invention will be described in detail with reference to the embodiments. The following embodiments are used to illustrate the invention, but are not intended to limit the scope of the invention.

[0106] See attached document Figure 2 As shown, the virtual power plant model involved in this embodiment is based on the day-ahead market and uses the local load's predicted electricity demand over 24 hours as a dispatch reference. Figure 4 As shown, local distributed generation (DG, including wind power, photovoltaic power, micro gas turbines, etc.) is prioritized to meet load demand. When DG output is insufficient or costs are high, the system operator (i.e., the energy management system) will communicate with the public grid to purchase electricity to supplement the shortfall. Conversely, if DG output is sufficient, excess electricity will be converted into heat energy for storage or stored in fuel cells (FC) and sold back to the grid at a predetermined profit rate. Simultaneously, to improve the flexibility and economy of system dispatch, flexible load resources (adjustable loads) are introduced. By adjusting some electricity consumption behavior within a permissible time range, dynamic response on the load side and auxiliary regulation of power balance are achieved, thereby further optimizing local energy utilization and reducing electricity purchase costs.

[0107] See attached document Figure 1 As shown, this invention relates to a method for optimizing the economy and ancillary services of a multi-energy virtual power plant, which includes the following steps:

[0108] S1. Initialize the virtual power plant's ancillary service parameters (such as the power output of each device, the SOC of the energy storage device, the input load, the photovoltaic power generation, the wind power generation, the electricity price, etc.). Considering the uncertainty of renewable energy, construct a joint optimization model for the virtual power plant's economy and ancillary services. Specifically, this includes: using a multi-source collaborative modeling module to uniformly model various resources within the virtual power plant under typical intraday scheduling scenarios, that is, to uniformly model distributed resources, flexible loads, and ancillary services; and constructing an ancillary service participation mechanism and revenue evaluation model based on market mechanisms and revenue.

[0109] The unified modeling of various resources within a virtual power plant specifically includes: establishing an objective function to minimize the operating cost of the virtual power plant and establishing constraints for the virtual power plant.

[0110] The objective function for minimizing the operating cost of the virtual power plant is expressed as follows:

[0111] (1),

[0112] (2),

[0113] (3),

[0114] where, denotes the minimization of the operation cost of the virtual power plant, t denotes the t th dispatch period, T denotes the set of dispatch periods, , , , , , and t are the energy costs of the power operator, wind turbine, photovoltaic component, fuel cell system, micro gas turbine and energy storage device in the dispatch period , , , , , are the power injections of the power operator, wind turbine, photovoltaic component, fuel cell system, micro gas turbine and energy storage device in the dispatch period t , is the flexible load adjustment variable, denoting the power adjustment amount achieved by the postponable or reducable load in the dispatch period t , , , , are the operating states of the wind turbine, photovoltaic component, fuel cell system, micro gas turbine, in the dispatch period t, with "1" indicating operation and "0" indicating shutdown; and denote the operating states of the wind turbine, photovoltaic component, fuel cell system, micro gas turbine in the dispatch period t and the dispatch period t-1, with "1" indicating operation and "0" indicating shutdown, denotes the operating cost of the energy storage device, and denote the operating states of the energy storage device, also taking 0 / 1, for the start-stop cost calculation of the energy storage unit; , are the start-stop costs of the i th distributed power generation unit and the j th energy storage device; , are the numbers of distributed power generation units and energy storage devices; , are the resistance and current of the b th branch in the dispatch period t . , , These are the original system loss, the final loss, and the new loss after optimization, respectively. , These represent the unit loss cost and the power difference before and after optimization, respectively. The power of electricity sold to the grid. For electricity sales price, This indicates the cost of flexible load adjustment. This represents the total number of power grid branches in the power system, that is, the number of transmission lines in the entire distribution network that participate in the calculation of line power loss.

[0115] The constraints of the virtual power plant include:

[0116] Power balance constraints: In this stage, only the modeling power constraints of the virtual power plant itself are considered. To ensure the balance of energy supply and demand in the system, the virtual power plant should satisfy the balance relationship between total generation power and total load demand in each dispatch cycle (usually 24 hours). This balance relationship takes into account the output and consumption of resources from distributed generation, grid purchases, energy storage systems, etc. The expression for the power balance constraint is:

[0117] (4),

[0118] in, , These are the energy storage devices. t Discharge and charging power during each scheduling period For the first t Line loss power during each scheduling period No. t Load demand for each scheduling period;

[0119] The output power constraint of a wind turbine generator is expressed as follows:

[0120] (5),

[0121] in, , These are the maximum and minimum power outputs of the wind turbine generator set, respectively.

[0122] The output power constraint of photovoltaic modules is expressed as follows:

[0123] (6),

[0124] in, , These represent the maximum and minimum power output of the photovoltaic modules, respectively.

[0125] The micro gas turbine output limit is expressed as:

[0126] (7),

[0127] wherein, , are the maximum and minimum power of the micro gas turbine output, respectively;

[0128] The fuel cell system output limit is expressed as:

[0129] (8),

[0130] wherein, , are the maximum and minimum power of the fuel cell system output, respectively;

[0131] The power purchase limit from the power operator is expressed as:

[0132] (9),

[0133] wherein, , are the maximum and minimum power of the power purchase from the power operator, respectively;

[0134] The adjustable load output limit is expressed as:

[0135] (10),

[0136] wherein, , are the maximum and minimum power of the adjustable load of the energy storage device, respectively;

[0137] The flexible load adjustment limit is expressed as:

[0138] (11),

[0139] wherein, is the upper limit of the flexible load adjustment;

[0140] The energy storage device charge and discharge limit is expressed as:

[0141] (12),

[0142] (13),

[0143] (14),

[0144] (15),

[0145] (16),

[0146] wherein, , is the charging and discharging power of the energy storage device in the first t dispatching period, , is the maximum charging and discharging power of the energy storage device in the first t dispatching period, , is a binary state variable, is the energy storage energy in the first t dispatching period, is the energy storage energy in the last time; energy storage charging efficiency, energy storage discharging efficiency;

[0147] the power selling constraint, the expression of which is:

[0148] (17),

[0149] (18),

[0150] (19),

[0151] wherein, represents the power selling amount, formula (17) indicates that the power selling cannot be negative, is the remaining energy amount, is the actual load, ensuring that the power selling can only use the “redundant energy amount” and cannot be oversold, is a binary variable, M is a constant, and formula (19) indicates that the power selling and the power purchasing avoid conflicts.

[0152] According to the system auxiliary service market rules, the services that the virtual power plant can participate in mainly include frequency modulation services and standby services. The response power required by the auxiliary service is “reserved” from the maximum output of each resource unit, that is, the device must reserve sufficient capacity for auxiliary services while completing the original scheduling task. The resource units with adjustment capacity (such as fuel cells, micro gas turbines, and battery storages) are mapped as adjustable resources, each of which has a certain frequency modulation or standby response capability. A service participation mapping table is established as follows:

[0153] Table 1 Service participation mapping table

[0154]

[0155] An auxiliary service participation mechanism and a benefit evaluation model are constructed, which specifically include:

[0156] To ensure the feasibility of auxiliary services and the controllability of resources, the service response power allocation constraint needs to be introduced on the basis of the original output power. Therefore, the service response power allocation constraint is constructed, and its expression is:

[0157] (20),

[0158] wherein, is the resource r The generation power of the first t dispatching period, , is the reserved frequency modulation and standby service response power of the first is the maximum output capacity of the resource r

[0159] The supply-demand matching constraint is constructed, and its expression is:

[0160] (21),

[0161] (22),

[0162] wherein, , respectively represent the frequency modulation and standby demand;

[0163] The service revenue objective function is constructed, and its expression is:

[0164] (23),

[0165] wherein, , are the market prices of the frequency modulation and standby service of the first t dispatching period, is the total auxiliary service revenue of the virtual power plant in the entire dispatching period.

[0166] The revenue will be optimized together with the electricity market cost in the total objective function to achieve the "double market revenue maximization" strategy.

[0167] Based on the above steps, the dispatching behavior of the virtual power plant in the electricity market and the auxiliary service market is uniformly modeled, and the "cost-revenue" coordinated optimization objective function is constructed. The target is to minimize the net cost between the total operating cost of the virtual power plant and maximize the auxiliary service market revenue, and the specific form is:

[0168] (24),

[0169] wherein, minJ represents the minimum net cost, α ​To adjust the weight of auxiliary services in the total target to support sensitivity analysis or multi-scenario trade-off decision.

[0170] S2. The virtual power plant cooperative scheduling model constructed by the application has the characteristics of high nonlinearity, strong coupling and multiple constraints, and the traditional analytical method is difficult to directly solve, and the linearization process is easy to lose the essential structure of the problem. Therefore, intelligent optimization algorithms are needed to search the global variable space to achieve the optimal scheduling strategy. However, in the process of global search, the algorithm is easy to fall into local maximum solution; therefore, the application introduces a global-local algorithm combined with an improved particle swarm optimization method and a snowmelt optimizer SAO, the improved particle swarm optimization method initially searches the global solution exploration (multi-direction) to quickly separate the distribution of solutions, and the SAO is locally refined to improve the accuracy of the solution and compensate for the convergence error of the particle. Compared with other optimization algorithms, the algorithm has better balance ability, search efficiency and adaptability in dealing with complex optimization problems, especially for multi-peak and high-dimensional problems.

[0171] The application adopts a hybrid optimization strategy combining the improved particle swarm optimization algorithm and the snowmelt optimizer to solve the joint optimization model and obtain the optimal scheduling scheme, and the specific steps are as shown in Figure 3 , including:

[0172] S2.1. Global search is performed by using the improved particle swarm optimization algorithm to obtain the scheduling scheme; the improved particle swarm optimization algorithm is used for global search to update the speed and position of the particle, and the scheduling scheme is gradually obtained, and the update formula of the particle speed and position is:

[0173] The S2.1 uses the improved particle swarm optimization algorithm to perform global search, update the speed and position of the particle, and gradually obtain the scheduling scheme, and the update formula of the particle speed and position is:

[0174] (25),

[0175] (26),

[0176] wherein, g indicates the number of iterations, and respectively indicate the speed and position of the particle, is the historical optimal position of the particle p , is the current global optimal position, and are random vectors, indicates the Hadamard product, is the linearly decreasing result of the inertia weight, and are learning factors.

[0177] S2.2. Evaluate each generation of scheduling scheme using fitness function:

[0178] (27),

[0179] wherein, represents fitness, represents the operating cost of the virtual power plant, represents a weight coefficient for balancing the relationship between operating cost and ancillary service income, is the total ancillary service income of the virtual power plant in the entire scheduling period;

[0180] Determine whether the following convergence condition is met, which is one of the following two conditions:

[0181] Condition 1. Reach the maximum number of iterations ;

[0182] Condition 2. The optimal fitness function value in the continuous generation changes less than a set threshold , that is:

[0183] (28).

[0184] If the convergence condition is met, the current optimal solution is output as the final scheduling scheme; otherwise, enter the local optimization phase, that is, S2.3 combines the hybrid optimization strategy of Snowmelt Optimizer to perform local fine perturbation and convergence control

[0185] S2.3. Combine the hybrid optimization strategy of Snowmelt Optimizer to perform local fine perturbation and convergence control, which includes the following specific steps:

[0186] S2.3.1. Based on the solution of the scheduling scheme obtained in S2.1, generate a random set in matrix form and the upper and lower limits of the solution, and the expression of the random set in matrix form is:

[0187] (29),

[0188] The calculation formula of the upper and lower limits is:

[0189] (30),

[0190] wherein, N and n represent the number of population particles and the dimension of the solution, respectively, m represents the dimension index in the matrix, m ∈ n , used to represent the position of the particle in the m dimension in the population; U , LThese are the upper and lower limits of the value, respectively. A random number within the range [0,1];

[0191] S2.3.2. The particle position is updated by introducing Brownian motion. The formula for updating the particle position is:

[0192] (31),

[0193] in, For the first p The position of each particle. g This represents the current iteration number. A random individual within a collection of elite particles. A vector of random numbers generated from the Gaussian distribution of Brownian motion. To multiply by row, Individuals randomly selected from several elite groups within the representative population. The center of mass of the entire particle position;

[0194] S2.3.3. By gradually converging through the snow melting process, the optimal solution is found. The position update formula for this stage is:

[0195] (32),

[0196] (33),

[0197] In the formula: A random number in the range [-1, 1] P For snow melting models, The maximum number of iterations, A random number within the range [0,1]. Indicates the first The location or index of an individual randomly selected from the generation.

[0198] Example: This example uses a small electronics assembly factory as an example. The factory mainly engages in electronic component welding, module testing, and assembly line assembly, exhibiting a clear daily load fluctuation pattern: low loads in the morning and evening, and high loads at midday, with a maximum daily load of approximately 155 kW. To cope with electricity price fluctuations and achieve energy conservation and emission reduction, the factory deployed a distributed integrated energy system. After optimizing the scheduling scheme using a class-based algorithm, the cost results are shown in Table 2:

[0199] Table 2 Cost results after optimization by various algorithms

[0200]

[0201] Meanwhile, the solution involved in this invention was simulated, and the simulation results are described below:Figure 5 The 24-hour scheduling output curve diagram of the integrated energy system after introducing the flexible load regulation mechanism and the auxiliary service response mechanism is shown. It can be observed from the diagram that the system effectively realizes the balance between supply and demand in high load periods (such as 11, 14, 21 and 22) by mobilizing the charging and discharging of the energy storage system, the response of the flexible load and the calling of the gas turbine and the grid resources. At the same time, the photovoltaic and wind power fluctuates obviously in some time periods (such as 10-13 hours), and the system activates the auxiliary service (such as frequency regulation and standby) in time to provide stable output compensation, thereby improving the safety and stability of system operation. In comparison, Figure 6 The scheduling strategy under the condition that the auxiliary service mechanism is closed is shown. Since the energy storage and fuel cell cannot be used for frequency regulation and standby, the system shows great imbalance in periods of intense wind and light fluctuation, and needs to rely more on grid power purchase or increase the load of the gas turbine, which not only increases the operation cost, but also reduces the adaptability of the system to uncertainty. Figure 7 The system cannot fill the valley through demand side response without introducing the flexible load regulation, resulting in the need to mobilize the gas turbine and energy storage system in the morning and evening peak periods, further increasing the burden of the energy storage system, and some photovoltaic power generation exists surplus and cannot be utilized, affecting the economy of the system.

[0202] From the cost point of view, Figure 5 The scheduling results of various mainstream optimization algorithms in the integrated energy system are listed. After considering the flexible load and auxiliary service, the improved algorithm iPSO+SAO achieves the optimal scheduling result, with the optimal operation cost of 198.45 yuan and the average cost of 199.38 yuan, which is significantly better than other algorithms (such as the average cost of GA is 339.94 yuan, and the average cost of AOA is 232.42 yuan). The result verifies the advantages of the algorithm in search accuracy and global optimization ability, and also embodies the effective support of the flexible strategy and auxiliary service mechanism for improving the economy of the system.

[0203] The above embodiments are described in detail, but the described content is only the preferred embodiments of the present application and cannot be considered as limiting the scope of the present application. Any equivalent changes and improvements made in accordance with the scope of the present application should still fall within the scope of the present application.

Claims

1. A method for economic and ancillary service optimization of a multi-energy virtual power plant, characterized in that: It comprises the following steps: S1. Considering renewable energy uncertainty, constructing a virtual power plant economic and auxiliary service joint optimization model, specifically: in the typical day scheduling scenario, modeling various resources in the virtual power plant, and constructing an auxiliary service participation mechanism and benefit evaluation model, and establishing a minimum virtual power plant operation cost objective function and virtual power plant constraint conditions, the expression of the minimum virtual power plant operation cost objective function is: （1）， （2）， （3）， wherein, represents the minimization of the operation cost of the virtual power plant, t represents the t dispatching time period, T represents the set of dispatching time periods, , , , , , are the energy costs of the power operator, the wind turbine, the photovoltaic component, the fuel cell system, the micro gas turbine and the energy storage device in the dispatching time period t , , , , , are the power injections of the power operator, the wind turbine, the photovoltaic component, the fuel cell system, the micro gas turbine and the energy storage device in the dispatching time period , t is the flexible load adjustment variable, representing the power adjustment amount achieved by the postponable or reducible load in the dispatching time period , t , , , , are the operating states of the wind turbine, the photovoltaic component, the fuel cell system, the micro gas turbine in the dispatching time period t, with "1" representing operation and "0" representing shutdown; and represent the operating states of the wind turbine, the photovoltaic component, the fuel cell system, the micro gas turbine in the dispatching time period t and the dispatching time period t-1, with "1" representing operation and "0" representing shutdown, represents the operating cost of the energy storage device, and represent the operating states of the energy storage device, also taking 0 / 1, for the start-stop cost calculation of the energy storage unit; , are the start-stop costs of the i distributed power generation unit and the j energy storage device; , are the numbers of the distributed power generation unit and the energy storage device; , are the resistance and current of the b branch in the dispatching time period t , , , The system original loss, the final loss and the new loss after optimization, respectively; , The unit loss cost, the power difference before and after optimization, respectively; The power sold to the power grid, The power sold to the power grid, The power sold to the power grid, Indicates the flexible load adjustment cost, Indicates the total number of grid branches in the power system, that is, the number of transmission lines participating in the line power loss calculation in the entire distribution network; S2. A hybrid optimization strategy combining improved particle swarm optimization algorithm and snowmelt optimizer is used to solve the joint optimization model to obtain the optimal scheduling scheme, the specific steps are: S2.

1. Global search is performed using the improved particle swarm optimization algorithm to obtain the scheduling scheme, specifically: the improved particle swarm optimization algorithm is used for global search, the speed and position of the particle are updated, and the scheduling scheme is gradually obtained, the update formula of the particle speed and position is: （25）， （26）， in, g Indicates the number of iterations. and These represent the particle's velocity and position, respectively. For particles p The best historical position This is the current globally optimal position. and For random vectors, This represents the Hadamard product. The result is a linear decrease in inertia weight. and For learning factors; S2.

2. The scheduling scheme is evaluated by the fitness function to determine whether the iteration and convergence conditions are met, when all the iteration and convergence conditions are met, the scheduling scheme is taken as the optimal scheduling scheme and output, when any one of the iteration and convergence conditions is not met, S2.3 is entered; S2.

3. Local fine perturbation and convergence control are performed by combining the hybrid optimization strategy of the snowmelt optimizer to optimize the scheduling scheme.

2. The multi-energy virtual power plant economy and ancillary service optimization method according to claim 1, characterized in that: The virtual power plant constraint conditions include: The power balance constraint, the expression of which is: （4）， in, , These are the energy storage devices. t Discharge and charging power during each scheduling period For the first t Line loss power during each scheduling period No. t Load demand for each scheduling period; The output power constraint of the wind turbine generator set, the expression of which is: （5）， wherein, , Pmax, Pminare the maximum and minimum power output of the wind turbine, respectively. The output power constraint of the photovoltaic module, the expression of which is: （6）， wherein, , Pmax, Pminare the maximum and minimum power output of the photovoltaic assembly, respectively; The output limit of the micro gas turbine, the expression of which is: （7）， wherein, , Pmax, Pminare the maximum and minimum power of the microturbine output, respectively. The output limit of the fuel cell system, the expression of which is: （8）， wherein, , Pmax, Pminare the maximum and minimum power of the fuel cell system output, respectively. The power purchase limit from the power operator, the expression of which is: （9）， wherein, , Pmax, Pminare the maximum and minimum power to be purchased from the power operator, respectively. The output limit of the adjustable load, the expression of which is: （10）， wherein, , are the maximum and minimum power of the adjustable load of the energy storage device, respectively; The flexible load adjustment limit, the expression of which is: （11）， wherein, is the flexible load regulation upper limit; The charge and discharge limit of the energy storage device, the expression of which is: （12）， （13）， （14）， （15）， （16）， wherein, , is the charging and discharging power of the energy storage device in the first t dispatch period, , is the maximum charging and discharging power of the energy storage device in the first t dispatch period, , is a binary state variable, is the energy storage capacity in the first t dispatch period, is the energy storage capacity in the previous time point; is the energy storage charging efficiency, is the energy storage discharging efficiency; The power selling constraint, the expression of which is: （17）， （18）， （19）， wherein, represents the amount of electricity sold, is the remaining amount of electricity, is the actual load, is a binary variable, M is a constant.

3. The multi-energy virtual power plant economy and ancillary service optimization method according to claim 1, characterized in that: The auxiliary service participation mechanism and benefit evaluation model in S1 specifically includes: The service response power distribution constraint, the expression of which is: （20）， wherein, is the resource r In the first t dispatch period, the generated power, , is the reserved frequency modulation and standby service response power, is the resource r maximum output capacity; The supply and demand matching constraint, the expression of which is: （21）， （22）， wherein , respectively represent frequency modulation, backup demand; The service benefit objective function, the expression of which is: （23）， wherein, , are the market prices of frequency modulation, reserve service in the first t dispatching period, respectively, is the total auxiliary service revenue of the virtual power plant in the entire dispatching period.

4. The multi-energy virtual power plant economy and ancillary service optimization method of claim 1, wherein: The fitness function in S2.2 is: (27)， wherein, represents fitness, represents the operating cost of the virtual power plant, represents a weight coefficient, is the total auxiliary service revenue of the virtual power plant in the entire dispatching cycle; The convergence condition is met when one of the following two conditions is met: Condition 1. Max number of iterations reached ; Condition 2. Continuous Intra-generation optimal fitness function value change is less than a set threshold That is: （28）。 5. The multi-energy virtual power plant economy and ancillary service optimization method of claim 1, wherein: The specific steps of S2.3 combining the hybrid optimization strategy of the snowmelt optimizer for local fine perturbation and convergence control include: S2.3.

1. Based on the solution of the scheduling scheme obtained in S2.1, a matrix form random set and upper and lower limits of the solution are generated, the expression of the matrix form random set is: （29）， The calculation formula of the upper and lower limits is: （30）， where, N and n denote the population particle number and the dimension of the solution, respectively, m denotes the dimension index in the matrix, m ∈ n , U , L are the upper and lower limits of the value, respectively, is a random number within [0, 1]; S2.3.

2. The position of the particle is updated by introducing Brownian motion, the update formula of the particle position is: （31）， in, For the first p The position of each particle. g This represents the current iteration number. A random individual within a collection of elite particles. A vector of random numbers generated from the Gaussian distribution of Brownian motion. To multiply by row, Individuals randomly selected from several elite groups within the representative population. The center of mass of the entire particle; S2.3.

3. Gradually converge through the snowmelt process to find the best solution, the position update formula in this stage is: （32）， （33）， wherein: is a random number in [-1, 1], P is a snowmelt model, is the maximum number of iterations, is a random number in [0, 1], denotes the position or index of a randomly selected individual in the generation.

Images