Multi-source cooperative power balance optimization model construction method, system and device

Abstract

The invention discloses a multi-source cooperative power balance optimization model construction method, system and device, and relates to the technical field of power system optimization. The method comprises the following steps: firstly, acquiring historical and real-time operation data of thermal power, wind power, photovoltaic and energy storage systems; on the basis of the data, an optimization model with the purpose of minimizing the total operation cost of the system is constructed, objective functions cover thermal power generation and start-stop cost, new energy subsidy and energy abandoning penalty cost and energy storage operation loss cost, and constraint conditions include power balance, operation of all power sources and power grid safety constraints; and finally, an improved intelligent optimization algorithm is adopted to solve the model, the algorithm dynamically divides an exploration subgroup for executing global exploration and a development subgroup for executing local development according to the evolutionary state of an objective function value in each iteration, and different position updating strategies are adopted. According to the invention, multi-source power collaborative optimization can be realized, the stability, the economy and the new energy consumption capability of the power system are improved, and the real-time scheduling requirement is met.

Description

Technical Field

[0001] This document relates to the field of power balance optimization technology, and in particular to a method, system and device for constructing a multi-source collaborative power balance optimization model. Background Technology

[0002] With the increasing penetration of new energy sources in the power system, the collaborative optimization of multi-source power systems has become crucial for ensuring stable grid operation and improving energy efficiency. However, current multi-source power balance optimization still faces many technical bottlenecks: On the one hand, the output of new energy sources is highly uncertain due to natural conditions, easily leading to wind and solar power curtailment. Traditional optimization models struggle to accurately quantify their interference with power balance and do not fully consider the correlation between energy storage system operating losses and lifespan decay, resulting in biased energy storage cost accounting and affecting overall dispatch economics. On the other hand, existing intelligent optimization algorithms often suffer from insufficient global exploration capabilities or excessively rapid local convergence due to a single population search strategy when solving multi-source collaborative models, making it difficult to efficiently obtain the optimal dispatch plan that satisfies multiple constraints such as power balance, power source operation, and grid security.

[0003] Power systems need to balance carbon emission reduction and green energy consumption, but existing models lack a collaborative mechanism for carbon trading and green certificates, failing to effectively combine the carbon emission reduction benefits of new energy sources with the optimization of dispatch costs. This further limits the environmental and economic improvement of multi-source power systems. Furthermore, traditional models do not adequately integrate and utilize multi-source data, making it difficult to adapt to the different operating characteristics of various types of equipment, including thermal power, new energy, and energy storage. This results in poor practicality and real-time performance of dispatch schemes, failing to meet the dynamic balance requirements of complex power systems. Summary of the Invention

[0004] According to embodiments of the present invention, a method, system, and apparatus for constructing a multi-source collaborative power balance optimization model are provided. The aim is to solve the technical problems in multi-source power systems, such as the problem of energy curtailment caused by the uncertainty of new energy output, the deviation in the quantification of energy storage operation loss costs, the low efficiency of the single-consistency solution of the population search strategy of traditional intelligent optimization algorithms, and the lack of a collaborative interaction mechanism between carbon trading and green certificates, which makes it impossible to take into account both dispatch economy and carbon emission reduction targets.

[0005] According to an embodiment of the present invention, a method for constructing a multi-source collaborative power balance optimization model is provided, comprising: S1. Obtain historical and real-time operating data of thermal power generation, wind power generation, photovoltaic power generation and energy storage systems in the power system; S2. Based on the data obtained in S1, construct an optimization model with the objective function of minimizing the total system operating cost. The total system operating cost specifically includes: the generation and start-up / shutdown costs of thermal power, the subsidy and curtailment penalty costs of new energy sources, and the operating loss costs of energy storage. The set of constraints for the optimization model includes: power balance constraints, operating constraints of each power source, and grid security constraints. S3. The improved intelligent optimization algorithm is used to solve the model to obtain a power generation and energy storage scheduling plan that satisfies the set of constraints. In each iteration, the improved intelligent optimization algorithm dynamically divides the population into an exploratory subgroup that performs global exploration and a development subgroup that performs local development, based on the evolutionary state of the objective function value, and adopts different position update strategies for different subgroups.

[0006] According to an embodiment of the present invention, a system for constructing a multi-source collaborative power balance optimization model is provided, comprising: The data acquisition module is used to acquire historical and real-time operating data of thermal power generation, wind power generation, photovoltaic power generation and energy storage systems in the power system; The model building module, based on the data acquired by the data acquisition module, constructs an optimization model with the objective function of minimizing the total system operating cost. The total system operating cost specifically includes: the generation and start-up / shutdown costs of thermal power, the subsidy and curtailment penalty costs of new energy sources, and the operating loss costs of energy storage. The constraint set of the optimization model includes: power balance constraints, operating constraints of each power source, and grid security constraints. The model solving module uses an improved intelligent optimization algorithm to solve the model in order to obtain a power generation and energy storage scheduling plan that satisfies the set of constraints. In each iteration, the improved intelligent optimization algorithm dynamically divides the population into an exploratory subgroup that performs global exploration and a development subgroup that performs local development, based on the evolutionary state of the objective function value, and adopts different position update strategies for different subgroups.

[0007] According to an embodiment of the present invention, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the program to implement the steps of the multi-source collaborative power balance optimization model construction method described above.

[0008] According to an embodiment of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that, when the program is executed by a processor, it implements the steps of the above-described method for constructing a multi-source collaborative power balance optimization model.

[0009] By constructing an objective function that includes the costs of thermal power generation and start-up / shutdown, renewable energy subsidies and curtailment penalties, and energy storage operation losses, the total operating cost of the system is accurately calculated, reducing cost waste. Simultaneously, by combining carbon trading and green certificate interaction mechanisms, the carbon emission reduction benefits of renewable energy are converted into cost deductions, further optimizing economic returns. A constraint set is constructed using power balance, the operation of each power source, and grid security constraints to ensure real-time balance between supply and demand from multiple power sources, reducing the interference of renewable energy output uncertainty on the grid and avoiding risks such as overcharging, over-discharging, and line overload. The improved intelligent optimization algorithm dynamically divides the system into exploration and development subgroups. This approach addresses the shortcomings of traditional algorithm search strategies, such as insufficient exploration or excessively rapid convergence, and efficiently obtains the optimal power generation and energy storage scheduling plan. Furthermore, it improves upon the SSA algorithm by introducing Cauchy mutation perturbations when optimizing LSTM hyperparameters, preventing premature convergence and enhancing the accuracy of cost prediction by the surrogate model, thus aiding in optimization. Through a carbon trading and green certificate interaction mechanism, the carbon emission reductions from wind and solar renewable energy are converted into virtual carbon credits to offset carbon emission quota shortfalls. Simultaneously, it clarifies that renewable energy output cannot repeatedly apply for green certificates, thus incentivizing renewable energy consumption to reduce wind and solar curtailment and promoting carbon emission reduction in the power system, achieving a synergistic improvement in both economic efficiency and environmental protection. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in one or more embodiments of this specification or in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart of the method for constructing a multi-source collaborative power balance optimization model according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the multi-source collaborative power balance optimization model construction system according to an embodiment of the present invention; Figure 3 This is a detailed flowchart of the multi-source collaborative power balance optimization model construction method according to an embodiment of the present invention; Figure 4 This is a schematic diagram of a power system according to an embodiment of the present invention. Detailed Implementation

[0012] To enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of this document.

[0013] Method Implementation Examples According to embodiments of the present invention, a method for constructing a multi-source collaborative power balance optimization model is provided. Figure 1 This is a flowchart of the multi-source collaborative power balance optimization model construction method according to an embodiment of the present invention. Figure 1 As shown, the multi-source collaborative power balance optimization model construction method of this embodiment of the invention specifically includes: S1. Obtain historical and real-time operating data of thermal power generation, wind power generation, photovoltaic power generation and energy storage systems in the power system; The historical operational data includes power generation, start-up and shutdown records, and fuel consumption data of thermal power units over the past 1-3 years; output time-series data of wind and solar power; and charging and discharging power, SOC change curves, and lifespan degradation records of energy storage systems. Real-time operational data is collected in real time through the power system SCADA system, the new energy power station monitoring platform, and the energy storage BMS, with a sampling frequency of 15 minutes per sampling to ensure data timeliness. After data acquisition, preprocessing is required: missing data is filled using linear interpolation or LSTM prediction models; outlier data is identified using the 3σ criterion and replaced with the average of adjacent time periods, ultimately forming a standardized dataset to provide high-quality data support for subsequent model construction.

[0014] S2. Based on the data obtained in S1, construct an optimization model with the objective function of minimizing the total system operating cost. The total system operating cost specifically includes: the generation and start-up / shutdown costs of thermal power, the subsidy and curtailment penalty costs of new energy sources, and the operating loss costs of energy storage. The set of constraints for the optimization model includes: power balance constraints, operating constraints of each power source, and grid security constraints. The operational cost of energy storage in the objective function is quantified by a nonlinear function related to cycle depth and charge / discharge rate, specifically expressed as follows: ; in, denoted as , where is the lifespan cost, DOD is the depth of cycle, CR is the charge / discharge rate, k is the cost coefficient, and α and β are accelerated aging factors obtained by fitting experimental data, with α>1 and β>0.

[0015] The set of constraints specifically includes: The power balance constraint specifically includes: at any scheduling period t, the total generating power of all power sources equals the sum of load demand and network losses, obtained through the following formula: ; in, This represents the actual power generation of the i-th power source at time t. This represents the discharge power of the energy storage system at time t. This represents the charging power of the energy storage system at time t. This represents the total load demand power at time t. This represents the total power loss of the power grid lines at time t; Operating constraints for thermal power units include: output range constraints, ramp-up rate constraints, and minimum start-up / shutdown time constraints. The output range constraint requires that the power generation of thermal power unit g at time t be between its minimum technical output and maximum rated output. The ramp-up rate constraint limits the rate of power change of thermal power unit g between adjacent time periods. The minimum start-up / shutdown time constraint requires that thermal power unit g must operate continuously for at least [duration missing] after startup. During the period, the machine must remain shut down for at least [duration]. Time period; The expression for the output range constraint is: ; in, and These are the minimum technical output and maximum rated output of thermal power unit g, respectively. The start-up and shutdown state of thermal power unit g at time t; The expression for the climbing rate constraint is: ; in, and These are the upward and downward ramp rate limits for thermal power unit g, respectively. The minimum start-stop time constraint expression is: ; New energy power generation constraints include wind power output constraints and photovoltaic power output constraints. The wind power output constraint determines the output range of the wind farm w at time t based on the wind speed-power characteristic curve and prediction uncertainty. The photovoltaic power output constraint determines the output range of the photovoltaic power station v at time t based on the light intensity-power characteristic and temperature correction coefficient. Wind power output constraints are obtained using the following formula: ; in, Let w be the predicted power output of the wind farm at time t. The tolerance for the output prediction error of wind farm w; Photovoltaic output constraints are obtained using the following formula: ; in, The photoelectric conversion efficiency of the photovoltaic power station is v. Let be the actual light intensity at time t. v represents the total installed capacity of the photovoltaic power station. The temperature coefficient of a photovoltaic cell. Let be the actual ambient temperature at time t. Standard test temperature; Energy storage system constraints include: charge and discharge power constraints and state of charge constraints. The charge and discharge power constraints ensure that the charge and discharge power of the energy storage system at time t is within the rated range, and the state of charge constraints ensure that the SOC of the energy storage system is within the safe range to avoid overcharging and over-discharging. The charge / discharge power constraint is obtained using the following formula: ; ; in, and These are the maximum charging and discharging power of the energy storage system, respectively. The state of charge constraints are obtained using the following formula: ; ; in, and These are the minimum and maximum security limits for the State of the Occupancy (SOC), respectively. and These refer to the charging and discharging efficiencies of the energy storage system, respectively. For the duration of the scheduling period, Let t represent the state of charge of the energy storage system at time t.

[0016] S3. An improved intelligent optimization algorithm is used to solve the model to obtain a power generation and energy storage scheduling plan that satisfies the set of constraints. Specifically, in each iteration, the improved intelligent optimization algorithm dynamically divides the population into an exploratory subgroup performing global exploration and a development subgroup performing local development, based on the evolutionary state of the objective function value, and adopts different position update strategies for different subgroups. The dynamic division of the population into the exploratory subgroup performing global exploration and the development subgroup performing local development specifically includes: Calculate the sum and average of the objective function values ​​for all individuals in the current population; Individuals whose objective function values ​​are better than the average value are assigned to the development subgroup, and the remaining individuals are assigned to the exploration subgroup; Specifically, for the exploratory subgroup, an update strategy based on Lévy flight is adopted to enhance global search capabilities; for the development subgroup, a deterministic search strategy based on the gradient information of the current Pareto front solution is adopted to accelerate convergence.

[0017] The population size is set according to the problem complexity, typically 50-100 individuals. When calculating the objective function value, if the scheduling plan corresponding to an individual violates the constraints, a penalty factor is introduced, multiplying the objective function value by 1.5-2 times to reduce the probability of that individual being selected. In the Lévy flight update strategy, the step size follows a Lévy distribution, enhancing the ability of the exploration subgroup to globally traverse the solution space. The Pareto front solution gradient information is calculated through the difference in the objective function of adjacent non-dominated solutions. The development subgroup updates its position along the gradient direction, accelerating convergence to the optimal solution. Furthermore, the Pareto front solution is pruned every 10 iterations to remove redundant solutions, ensuring solution set diversity.

[0018] Before using the improved intelligent optimization algorithm to solve the problem, the step of constructing a proxy model is also included: An agent model is constructed using LSTM, with the planned output and load sequence of each power source as input and the predicted value of the total operating cost of the system as output. The hyperparameters of the LSTM are optimized using an improved SSA, wherein a Cauchy mutation perturbation is introduced into the leader position update to escape local optima; In the optimization process, the predicted value of the objective function of the surrogate model is weighted and fused with the calculated value of the original objective function as a basis for evaluating the fitness of the algorithm iteration, and the weighting coefficient tilts towards the original objective function as the number of iterations increases.

[0019] The LSTM surrogate model has an input layer dimension of (number of scheduling periods × (number of power types + 1)), 2-3 hidden layers with 32-64 neurons per layer, and a 1-dimensional output layer. During training, the Adam optimizer is used with an initial learning rate of 0.001 and mean squared error as the loss function. The training, validation, and test sets are divided in a 7:2:1 ratio to ensure model generalization ability. For weighted fusion, the weights of the surrogate model's predicted values ​​are set as follows: The weights of the original objective function calculations are 1- , The initial value is 0.7, and it is reduced by 0.1 every 5 iterations until it reaches 0.1. This approach utilizes the surrogate model to improve computational speed while ensuring computational accuracy through the original objective function.

[0020] The specific details of introducing the Cauchy mutation perturbation into the leader's position update are as follows: After the leader position of the SSA is updated, a Cauchy mutation operation is applied to the generated candidate leader positions with an adaptive probability. The perturbation step size of the Cauchy mutation operation is controlled by an adaptive mutation strength coefficient. This coefficient is set to a large value in the early stage of the algorithm iteration to promote global exploration, and decreases non-linearly as the iteration progresses, so as to focus on local development in the later stage of the iteration. The value of the adaptive probability is dynamically adjusted according to the diversity index of the current population: when the population diversity is detected to drop below a preset threshold, the adaptive probability is increased to enhance the perturbation and avoid premature convergence of the algorithm; when the population diversity is maintained at a high level, the adaptive probability is decreased to maintain the stability of the convergence process.

[0021] A further embodiment of the multi-source collaborative power balance optimization model construction method of the present invention also includes: a modeling step for the interaction mechanism between carbon trading and green certificates. The objective function includes carbon trading cost and green certificate trading cost terms. Establish an interactive mechanism between carbon trading cost items and green certificate trading cost items. Specifically, the actual carbon emission reductions generated by wind and solar energy output will be converted into virtual carbon credits that can be used to offset the system's carbon emission quota shortfall through a tradable conversion factor. The virtual carbon credits can be directly deducted from the carbon trading cost item, and at the same time, this part of the new energy output will not be used again to apply for green certificates; The set of constraints further includes carbon emission quota constraints and renewable energy quota constraints.

[0022] Figure 3 This is a detailed flowchart of the multi-source collaborative power balance optimization model construction method according to an embodiment of the present invention. Figure 4 This is a schematic diagram of a power system according to an embodiment of the present invention, which can be combined with... Figure 3 and Figure 4 The embodiments of the present invention will be further understood.

[0023] By employing the embodiments of the present invention, the following beneficial effects are achieved: By constructing an objective function that includes the costs of thermal power generation and start-up / shutdown, renewable energy subsidies and curtailment penalties, and energy storage operation losses, the total operating cost of the system is accurately calculated, reducing cost waste. Simultaneously, by combining carbon trading and green certificate interaction mechanisms, the carbon emission reduction benefits of renewable energy are converted into cost deductions, further optimizing economic returns. A constraint set is constructed using power balance, the operation of each power source, and grid security constraints to ensure real-time balance between supply and demand from multiple power sources, reducing the interference of renewable energy output uncertainty on the grid and avoiding risks such as overcharging, over-discharging, and line overload. The improved intelligent optimization algorithm dynamically divides the system into exploration and development subgroups. This approach addresses the shortcomings of traditional algorithm search strategies, such as insufficient exploration or excessively rapid convergence, and efficiently obtains the optimal power generation and energy storage scheduling plan. Furthermore, it improves upon the SSA algorithm by introducing Cauchy mutation perturbations when optimizing LSTM hyperparameters, preventing premature convergence and enhancing the accuracy of cost prediction by the surrogate model, thus aiding in optimization. Through a carbon trading and green certificate interaction mechanism, the carbon emission reductions from wind and solar renewable energy are converted into virtual carbon credits to offset carbon emission quota shortfalls. Simultaneously, it clarifies that renewable energy output cannot repeatedly apply for green certificates, thus incentivizing renewable energy consumption to reduce wind and solar curtailment and promoting carbon emission reduction in the power system, achieving a synergistic improvement in both economic efficiency and environmental protection.

[0024] System Implementation Examples According to embodiments of the present invention, a system for constructing a multi-source collaborative power balance optimization model is provided. Figure 2 This is a schematic diagram of a multi-source collaborative power balance optimization model construction system according to an embodiment of the present invention. Figure 2 As shown, the multi-source collaborative power balance optimization model construction system of this embodiment of the invention specifically includes: The data acquisition module is used to acquire historical and real-time operating data of thermal power generation, wind power generation, photovoltaic power generation and energy storage systems in the power system; The model building module, based on the data acquired by the data acquisition module, constructs an optimization model with the objective function of minimizing the total system operating cost. The total system operating cost specifically includes: the generation and start-up / shutdown costs of thermal power, the subsidy and curtailment penalty costs of new energy sources, and the operating loss costs of energy storage. The constraint set of the optimization model includes: power balance constraints, operating constraints of each power source, and grid security constraints. The model solving module uses an improved intelligent optimization algorithm to solve the model in order to obtain a power generation and energy storage scheduling plan that satisfies the set of constraints. In each iteration, the improved intelligent optimization algorithm dynamically divides the population into an exploratory subgroup that performs global exploration and a development subgroup that performs local development, based on the evolutionary state of the objective function value, and adopts different position update strategies for different subgroups.

[0025] Device Example 1 According to an embodiment of the present invention, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the program to implement the steps of the multi-source collaborative power balance optimization model construction method described above.

[0026] Device Example 2 According to an embodiment of the present invention, a computer-readable storage medium is provided, on which a computer program is stored, characterized in that, when the program is executed by a processor, it implements the steps of the multi-source collaborative power balance optimization model construction method described above.

[0027] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for constructing a multi-source coordinated power balance optimization model, characterized in that The application comprises the following steps: S1, obtaining historical and real-time operation data of thermal power generation, wind power generation, photovoltaic power generation and energy storage system in a power system; S2, based on the data obtained in S1, an optimization model is constructed with the minimization of total system operation cost as the objective function, wherein the total system operation cost specifically includes generation and start-stop cost of thermal power, subsidy and penalty cost of new energy, and operation loss cost of energy storage; the constraint condition set of the optimization model includes power balance constraint, operation constraint of each power source and grid safety constraint; S3, an improved intelligent optimization algorithm is used to solve the model to obtain a generation and energy storage scheduling plan that meets the constraint condition set; In each iteration of the improved intelligent optimization algorithm, the population is dynamically divided into an exploration subgroup for global exploration and a development subgroup for local development according to the evolution state of the objective function value, and different position updating strategies are used for different subgroups.

2. The method of claim 1, wherein, The dynamic division of the population into the exploration subgroup for global exploration and the development subgroup for local development specifically includes: calculating the sum and average of the objective function values of all individuals in the current population; individuals with objective function values better than the average are assigned to the development subgroup, and the remaining individuals are assigned to the exploration subgroup; wherein, for the exploration subgroup, a updating strategy based on Levy flight is used to enhance the global search ability; for the development subgroup, a deterministic search strategy based on the gradient information of the current Pareto front solution is used to accelerate convergence.

3. The method of claim 1, wherein, The operation loss cost of the energy storage in the objective function is quantified by a nonlinear function related to the cycle depth and charging and discharging rate, and is specifically represented as: ； wherein, wherein, L is the life loss cost, DOD is the depth of discharge, CR is the charge and discharge rate, k is the cost coefficient, a and β are the accelerated aging factors obtained by fitting experimental data, and a > 1, β > 0.

4. The method of claim 1, wherein, The constraint condition set specifically includes: The power balance constraint specifically includes that the sum of the generation power of all power sources is equal to the sum of the load demand and the network loss at any scheduling period t, which is obtained by the following formula: ； wherein, Pti represents the actual power generation of the ith power source at time t, Pdis(t) represents the discharging power of the energy storage system at time t, Pch(t) represents the charging power of the energy storage system at time t, Pload(t) represents the total load demand power at time t, Ploss(t) represents the total line loss power of the power grid at time t. The operation constraints of the thermal power unit include: an output range constraint, a ramp rate constraint and a minimum start-stop time constraint, the output range constraint is that the power generation of the thermal power unit g at time t needs to be between the minimum technical output and the maximum rated output, the ramp rate constraint is to limit the power variation rate of the thermal power unit g in adjacent time periods, and the minimum start-stop time constraint is that the thermal power unit g needs to be continuously operated for at least period after starting, and needs to be continuously stopped for at least period after stopping. The output range constraint expression is: ； wherein, and MinTechPowergand MaxRatedPowergare the minimum technical power and the maximum rated power of the thermal power unit g, respectively, is the start-stop state of the thermal power unit g at time t. The climbing rate constraint expression is: ； wherein, and are the upward and downward ramping rate limits for the thermal power unit g, respectively. The minimum start-stop time constraint expression is: ; The new energy generation constraint includes the wind power output constraint and the photovoltaic output constraint, the wind power output constraint determines the output range of the wind farm w at time t according to the wind speed-power characteristic curve and the prediction uncertainty, and the photovoltaic output constraint determines the output range of the photovoltaic power station v at time t according to the illumination intensity-power characteristic and the temperature correction coefficient; The wind power output constraint is obtained by the following formula: ； wherein is the power prediction value of the wind farm w at time t, is the power prediction error tolerance of the wind farm w; The photovoltaic output constraint is obtained by the following formula: ； wherein, is the light-to-electricity conversion efficiency of the photovoltaic power plant v, is the actual light intensity at time t, is the total installed area of the photovoltaic power plant v, is the temperature coefficient of the photovoltaic cell, is the actual ambient temperature at time t, is the standard test temperature; The energy storage system constraint includes the charging and discharging power constraint and the state of charge constraint, the charging and discharging power constraint ensures that the charging and discharging power of the energy storage system at time t is within the rated range, and the state of charge constraint ensures that the SOC of the energy storage system is within the safe range to avoid overcharging and overdischarging; The charging and discharging power constraint is obtained by the following formula: ； ； wherein, and Pmax,chargeand Pmax,dischargeare the maximum charge and discharge power of the energy storage system, respectively. The state of charge constraint is obtained by the following formula: ； ； wherein, and are minimum and maximum safety limits of SOC, respectively, and are charging and discharging efficiencies of the energy storage system, respectively, is the length of the dispatch period, is the state of charge of the energy storage system at time t.

5. The method of claim 1, wherein, Before solving by using the improved intelligent optimization algorithm, the step of constructing a surrogate model is further included: An LSTM is used to construct a surrogate model, the input of which is the planned output of each power source and the load sequence, and the output of which is the predicted value of the total system operation cost; The hyperparameters of the LSTM are optimized by using an improved SSA, wherein a Cauchy mutation disturbance is introduced to the position update of the leader to jump out of local optimization; In the optimization solving process, the predicted value of the objective function of the agent model and the calculated value of the original objective function are weighted and fused as the fitness evaluation basis for algorithm iteration, and the weighting coefficient tilts towards the original objective function as the iteration number increases.

6. The method of claim 5, wherein, The Cauchy mutation disturbance introduced to the position update of the leader is specifically: After the position update of the leader in the SSA, a Cauchy mutation operation is performed on the generated candidate leader position with an adaptive probability; The disturbance step length of the Cauchy mutation operation is controlled by an adaptive mutation intensity coefficient, which is set to a larger value in the early stage of algorithm iteration to promote global exploration, and is nonlinearly decreased with the iteration process to focus on local development in the later stage of iteration; The value of the adaptive probability is dynamically adjusted according to the diversity index of the current population: when the population diversity is detected to decrease below a preset threshold, the adaptive probability is increased to enhance the disturbance and avoid premature convergence of the algorithm; when the population diversity is maintained at a high level, the adaptive probability is reduced to maintain the stability of the convergence process.

7. The method of claim 1, wherein, The multi-source collaborative power balance optimization model construction method further comprises a modeling step of a carbon trading and green certificate interaction mechanism: Carbon trading cost items and green certificate trading cost items are introduced into the objective function; An interaction mechanism of the carbon trading cost items and the green certificate trading cost items is established, specifically: the actual carbon emission reduction generated by the wind and light new energy output is converted into virtual carbon credits that can be used to offset the system carbon emission quota shortage through a tradable conversion factor; The virtual carbon credits can be directly deducted in the carbon trading cost items, and at the same time, the part of new energy output is no longer repeated for applying for green certificates; The set of constraint conditions further comprises carbon emission quota constraints and renewable energy quota constraints.

8. A multi-source collaborative power balance optimization system for implementing the multi-source collaborative power balance optimization model construction method of any one of claims 1-7, characterized in that, It comprises: A data acquisition module for acquiring historical and real-time operation data of thermal power generation, wind power generation, photovoltaic power generation and energy storage systems in a power system; A model construction module constructs an optimization model with the minimum system total operation cost as the objective function based on the data acquired by the data acquisition module, wherein the system total operation cost specifically includes the generation and start-stop cost of thermal power, the subsidy and penalty cost of new energy, and the operation loss cost of energy storage; the constraint condition set of the optimization model includes power balance constraints, power supply operation constraints and power grid safety constraints; A model solving module solves the model by using an improved intelligent optimization algorithm to obtain a generation and energy storage scheduling plan that satisfies the set of constraint conditions. In each iteration of the improved intelligent optimization algorithm, the population is dynamically divided into an exploration subpopulation for global exploration and a development subpopulation for local development according to the evolution state of the objective function value, and different position update strategies are used for different subpopulations.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor implements the steps of the multi-source collaborative power balance optimization model construction method of any one of claims 1-7 when executing the program. The processor implements the steps of the multi-source collaborative power balance optimization model construction method of any one of claims 1-7 when executing the program.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the steps of the multi-source coordinated power balance optimization model construction method according to any one of claims 1-7.

Images