Storage power station load tracking method based on GAPSO algorithm

Abstract

The invention provides a storage power station load tracking method based on a GAPSO algorithm, and the method comprises the steps: data collection and preprocessing: collecting and processing related data of a power grid and a power station; constructing a power grid load prediction model based on a GAPSO algorithm, determining variables, and optimizing a neural network; optimizing a charging and discharging strategy of the energy storage power station, and establishing and solving a target function; carrying out real-time scheduling and response, monitoring data, deciding and adjusting the state of the power station; and multi-objective optimization adjustment is carried out, functions are constructed, and solving and continuous improvement are carried out. According to the method, the load prediction accuracy is improved, the operation strategy of the power station is optimized, the real-time scheduling and response capability is enhanced, multi-target comprehensive optimization is realized, and the operation benefit and adaptability of the energy storage power station in a power system are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage in power systems, and particularly to a load tracking method for a storage power station based on the GAPSO algorithm. Background Technique

[0002] With the continuous growth of power demand and the increasing complexity of the power grid structure, the volatility and uncertainty of the power grid load pose severe challenges to the stable operation of the power system. Traditional power generation methods have problems such as slow response speed and poor regulation flexibility when dealing with rapid load changes. Traditional power generation methods, such as thermal power generation and hydropower generation, due to the inherent characteristics of their power generation equipment, have obvious limitations when dealing with rapid load changes. For example, the start-up and regulation processes of thermal power units are relatively slow and it is difficult to make timely responses according to load changes in a short time; although the regulation speed of hydropower units is relatively fast, due to factors such as the distribution of water resources and seasonal changes, their regulation capabilities are also subject to certain constraints. Therefore, relying solely on traditional power generation methods is difficult to meet the requirements of modern power grids for rapid load tracking and precise regulation.

[0003] As an emerging power regulation means, an energy storage power station can charge and store energy during low load periods and discharge to provide power support during high load periods, effectively alleviating the power grid load pressure and improving the stability and reliability of the power grid.

[0004] At the same time, the energy storage power station can also participate in auxiliary services such as frequency modulation and standby of the power grid, providing multi-faceted support for the safe and stable operation of the power grid. However, to achieve the efficient operation of the energy storage power station, the key lies in formulating a reasonable load tracking strategy. Accurately predicting the change trend of the power grid load and optimizing the charge and discharge strategy of the energy storage power station accordingly are the key points and difficulties in the current research in the power field.

[0005] Currently, existing load tracking methods still have many deficiencies in aspects such as prediction accuracy, strategy optimization effect, and multi-objective coordination, and cannot fully utilize the advantages of the energy storage power station in the power system. Summary of the Invention

[0006] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose a load tracking method for a storage power station based on the GAPSO algorithm.

[0007] To achieve the above purpose, the present invention adopts the following technical solutions: A load tracking method for a storage power station based on the GAPSO algorithm, which includes the following steps: S1, data collection and preprocessing; S2, construction of a power grid load prediction model based on the GAPSO algorithm; S3, optimization of the charge and discharge strategy of the energy storage power station; S4, real-time scheduling and response; S5, multi-objective optimization adjustment.

[0008] Preferably, the step S1, data acquisition and preprocessing, includes the following steps:

[0009] S11. Collect data;

[0010] During data collection, collect historical load data of the power grid, real-time load data, operating parameters of the energy storage power station (including battery capacity, charge and discharge efficiency, etc.), electricity price information, and other relevant influencing factor data (such as weather data, time factors, etc.). Obtain load data for different time periods from the power grid monitoring system, obtain battery status information from the energy storage power station management system, and obtain electricity price change data from the power market.

[0011] S12. Data preprocessing;

[0012] Clean the collected data to remove outliers and incorrect data;

[0013] Adopt the 3-sigma criterion to identify and correct data points that deviate significantly from the normal range;

[0014] Supplement missing data, and methods such as mean imputation and linear interpolation can be used;

[0015] If the load data is missing at a certain moment, supplement it according to the average value or linear relationship of the load data at the previous and subsequent moments. Normalize the data and map the data to a specific interval (such as [0,1] or [-1,1]) to improve the subsequent calculation efficiency.

[0016] Let the original data be x, and the normalized data (for mapping to the [0,1] interval), where x min and x max are the minimum and maximum values of the data respectively.

[0017] Preferably, the S2, construction of the power grid load prediction model based on the GAPSO algorithm, includes the following steps:

[0018] S21: Determine the input and output variables;

[0019] Construct the input vector X = [x1, x2, …, x n , which includes the load values at 6 moments, from x1 to x6; the hour number of the current moment is x7; the flag variable indicating whether it is a working day of the current day is x8, where x8 is 0 or 1; the temperature value is x9; the humidity value is x 10 ; the special event flag variable x 11 ; the output variable is y, that is, the load prediction value at the future moment;

[0020] S22: Construct a neural network prediction model optimized by the GAPSO algorithm;

[0021] Let the position vector of particle \(i\) be \(\theta\). i = [w i1 , w i2 , … w ij , b i1 , b i2 , … b ik , where \(w\) ij represents the weight of the \(j\)-th connection of the neural network, and \(b\) ik represents the threshold of the \(k\)-th neuron;

[0022] Calculate the fitness value of each particle. The fitness function uses the mean square error. Let the number of prediction samples be \(N\), the actual load value be \(y\) i , and the predicted load value be Then the fitness function

[0023] The predicted load value is obtained by inputting the input vector \(X\) into the neural network and calculating through forward propagation, that is where \(f\) represents the mapping function of the neural network;

[0024] Update the velocity and position of the particles according to the fitness value. During the update process, for the genetic algorithm part, use dynamic crossover and mutation probabilities;

[0025] Crossover probability:

[0026] P c = MAX(P c × 0.95 t / 4 , P cMIN ),

[0027] Mutation probability:

[0028] P m = MIN(P m × 1.05 t / 4 , P cMAX ),

[0029] where \(t\) is the number of iterations, \(P\) cMIN and \(P\) cMAX are the minimum crossover probability and the maximum mutation probability respectively;

[0030] The initial crossover probability \(P\) c is set between 0.6 - 0.9, the minimum crossover probability is set between 0.3 - 0.5, the initial mutation probability \(P\) cMIN is set between 0.01 - 0.1, and the maximum mutation probability \(P\) m is set between 0.1 - 0.3;

[0031] In the particle swarm optimization algorithm part, the inertia weight \(w = w\) max-k×t, w max is the initial weight, k is a constant, t is the number of iterations, and the learning factors c1 = c2 = 2; the initial weight w max is set between 0.7 and 0.9, and the constant is calculated according to the maximum number of iterations T PSOMAX k = 0.6 / T PSOMAX , where T PSOMAX = 100;

[0032] The velocity update formula is:

[0033] v id (t + 1) = wv id (t) + c1ξ(p id (t) - x id (t)) + c2η(p gd (t) - x id (t)),

[0034] The position update formula is:

[0035] x id (t + 1) = x id (t) + v id (t + 1),

[0036] where v id (t) is the velocity of particle i in the d-th dimension at the t-th iteration, x id (t) is the position of particle i in the d-th dimension at the t-th iteration, p id (t) is the value of the individual extreme value of particle i in the d-th dimension at the t-th iteration, p gd (t) is the value of the global extreme value in the d-th dimension, and ξ and η are random numbers uniformly distributed in the interval [0, 1].

[0037] Preferably, the step S3, the charge and discharge strategy optimization of the energy storage power station, includes the following steps: S31: Establish an optimization objective function:

[0038] The establishment of the optimization objective function is to minimize the operation cost and maximize the revenue of the energy storage power station, while considering the battery life attenuation cost;

[0039] The operation cost mainly includes the charging cost and the discharging cost;

[0040] The charging cost is related to the charging electricity price c charge (t) and the charging power P charge (t), and the discharging cost is related to the discharging electricity price c discharge (t) and the discharging power P discharge (t); the discharging can be regarded as the opposite of the revenue;

[0041] The battery life attenuation cost is calculated based on the charge-discharge cycles and depth of the battery;

[0042] Adopt the model of the Arrhenius formula:

[0043] The battery life attenuation cost is C loss (t),

[0044] Then the objective function is:

[0045]

[0046] Among them, T is the scheduling period; such as one day or one hour, which is set according to specific requirements.

[0047] The calculation of the charging cost and the discharging cost can be determined according to the actual electricity price policy and market rules. If time-of-use electricity price is adopted, c charge (t) and c discharge (t) have different values in different time periods;

[0048] The calculation of the battery life attenuation cost can refer to the formula of the battery life attenuation model based on the Arrhenius formula. According to the relationship between the life loss of the energy storage battery and the discharge power:

[0049] The battery capacity is E bat , the number of cycles is n, the cycle depth is h, and the battery throughput table is:

[0050] A h = E bat ×n×h;

[0051] Then the life loss:

[0052]

[0053] Among them, A is the pre-exponential factor, E a is the experimental activation energy, R is the molar gas constant, T is the absolute temperature, and z is the reaction rate constant;

[0054] Further calculate the battery life attenuation cost (combining factors such as battery replacement cost, etc.). S32: Determine the constraint conditions:

[0055] The said constraint conditions include the charge-discharge power constraint of the energy storage power station, the capacity constraint of the energy storage power station, and the grid power balance constraint;

[0056] The charge-discharge power constraint of the energy storage power station determines the maximum charge-discharge power according to the rated power limit of the energy storage power station equipment;

[0057] Include the maximum charging power P chargeMAX , the maximum discharging power P dischargeMAX ;

[0058] Then 0 ≤ P charge (t) ≤ P chargeMAX , 0 ≤ P discharge (t) ≤ P dischargeMAX ;

[0059] For an energy storage power station with a rated power of 5 MW:

[0060] P chargeMAX = 5 MW, P dischargeMAX = 4 MW (considering a certain power reserve or efficiency factor).

[0061] The capacity constraint of the energy storage power station ensures the safe and stable operation of the energy storage power station, and its state of charge (SOC) needs to be maintained within a certain range;

[0062] Let SOC min be the lower limit of the state of charge, and SOC max be the upper limit of the state of charge,

[0063] Then SOC min < SOC(t) < SOC max ;

[0064] The calculation of SOC is related to the charge and discharge process of the battery:

[0065]

[0066] where σ is the self-discharge rate of the energy storage, η c is the charging efficiency, η d is the discharging efficiency, and Δt is the scheduling time interval;

[0067] The power balance constraint of the power grid means that at any moment, the power exchanged between the power grid and the energy storage power station plus the charge and discharge power of the energy storage power station should be equal to the power grid load, that is

[0068] P grid (t) + P discharge (t) - P charge (t) = P load (t);

[0069] where P grid (t) is the power exchanged between the power grid and the energy storage power station, and P load (t) is the power grid load;

[0070] This constraint ensures the power balance of the power grid operation and avoids the power grid instability problem caused by power imbalance.

[0071] S33: Solve the charge and discharge strategy using the GAPSO algorithm:

[0072] By using the GAPSO algorithm to solve the charge and discharge strategy, the optimization objective function and constraint conditions are transformed into forms that can be processed by the GAPSO algorithm; each particle represents a set of possible charge and discharge strategies, that is, the position vector of the particle represents the charge and discharge power values at different times;

[0073] Let the position vector of the particle be:

[0074]

[0075] where and represent the charging power and discharging power of particle i at time t respectively;

[0076] Then calculate the fitness value of the particle, and calculate the fitness of each particle according to the optimization objective function; the fitness function is

[0077] fitness i =F(x i ), that is, the objective function value corresponding to the charge and discharge strategy represented by particle i. Substitute the position vector of the particle into the objective function to calculate the corresponding fitness value; this value reflects the quality of the charge and discharge strategy.

[0078] Update the velocity and position of the particle according to the fitness value. The formulas and steps used in the particle swarm algorithm part are the same as those used in step S2, including the inertia weight, learning factor, velocity update formula, and position update formula;

[0079] During the update process, ensure that the position of the particle satisfies the constraint conditions. If the updated particle position exceeds the charge and discharge power constraint or capacity constraint range, correct it to the feasible region;

[0080] For example, if the calculated charging power is greater than P chargeMAX , then set it to P chargeMAX ; if the state of charge is lower than SOC min , then adjust the charge and discharge power to make the state of charge return to the feasible range.

[0081] Iteratively optimize until the termination condition is met to obtain the optimal charge and discharge strategy. The charge and discharge power sequence corresponding to the optimal particle position at this time is the "optimal charge and discharge strategy of the energy storage station at different times, which can be used to guide the actual operation of the energy storage station.

[0082] Preferably, step S4, real-time scheduling and response, includes the following steps:

[0083] S41: Real-time monitor the grid load and the state of the energy storage station:

[0084] By means of high-precision sensors installed at key nodes of the power grid and inside energy storage power stations, continuously obtain the real-time load data of the power grid, the real-time state of charge (SOC) of the energy storage power station, and information on charging and discharging power;

[0085] The data acquisition frequency can be set to be acquired once every few minutes (such as 3 - 5 minutes) to ensure that changes in the power grid load and the state of the energy storage power station can be captured in a timely manner.

[0086] The acquired data is processed and transmitted in real time to ensure the accuracy and timeliness of the data. Reliable data communication technologies (such as specific protocols in wired communication or wireless communication) are used to transmit the data to the monitoring center or control system, and the data is verified and error-corrected during the transmission process to avoid the impact of data transmission errors on subsequent decisions.

[0087] S42: Make a scheduling decision based on the prediction results and optimization strategies:

[0088] Input the real-time monitoring data in step S41 into the load prediction model pre-constructed in step S2 to obtain a short-term load prediction value; then use the neural network prediction model optimized based on the GAPSO algorithm to predict the change trend of the power grid load within the next 15 minutes or half an hour according to the data of the current moment and the previous power grid load, time characteristics, and weather factors;

[0089] S43: Adjust the operating state of the energy storage power station:

[0090] After receiving the charging and discharging power instruction, the energy storage power station realizes the charging and discharging operation by controlling the output of the inverter; the inverter adjusts the output voltage, frequency, and phase according to the instruction to accurately control the charging and discharging power of the energy storage power station to make it conform to the instruction value.

[0091] Preferably, step S5, multi-objective optimization adjustment, includes the following steps:

[0092] S51: Construction of a multi-objective optimization function:

[0093] Comprehensively consider the key objectives of power grid stability, economic benefits of the energy storage power station, and life loss of the energy storage power station, and construct a multi-objective optimization function;

[0094] For power grid stability, a load fluctuation index is used to measure:

[0095] The variance of load fluctuation is as follows:

[0096] where P load,i is the load value at the i-th moment, is the average load value, and N is the number of samples;

[0097] In terms of economic benefits, the difference between the operating cost and the revenue is used as an index:

[0098]

[0099] Where R(t) is the other income of the energy storage power station at time t, including the income obtained from participating in frequency regulation and the reserve market;

[0100] The life loss index is calculated based on the charge-discharge cycle times and depth of the battery. Similar to step S31, the relevant battery life attenuation model, that is, the model of the Arrhenius formula, is adopted. Let the battery life attenuation cost be C loss (t),

[0101] Then

[0102] Construct a multi-objective optimization function:

[0103] O = λ1×f1 + λ2×f2 + λ3×f3, where λ1, λ2, and λ3 are the weight coefficients of the corresponding objectives; they are adjusted according to the actual power grid's emphasis on stability, economy, and energy storage life. In the case of large power grid load fluctuations and high requirements for stability, the value of λ1 can be appropriately increased; if the economic benefit of the energy storage power station is the primary consideration factor, the value of λ2 can be increased; if it is desired to extend the battery life of the energy storage power station, the value of λ3 can be increased.

[0104] S52: Solve the multi-objective optimization problem using the GAPSO algorithm:

[0105] Integrate the multi-objective optimization function and relevant constraint conditions into the GAPSO algorithm framework. Each particle represents a set of parameter combinations that satisfy the constraint conditions, including load prediction model parameters (such as the weights and thresholds of the neural network) and energy storage power station charge-discharge strategy parameters (such as the charge-discharge power at different times);

[0106] Calculate the fitness value of the particle. Since it is a multi-objective optimization problem, methods such as non-dominated sorting and crowding degree calculation are used to evaluate the fitness of the particle;

[0107] Calculate the objective function value vector [f1(x i ), f2(x i ), f3(x i )] of each particle according to the multi-objective optimization function, where x i is the position vector of particle i, and then perform non-dominated sorting on the particles, arranging the non-dominated particles in the front to form different non-dominated levels. For the particles within the same non-dominated level, calculate their crowding degree;

[0108] The crowding degree reflects the density of solutions around a particle. The greater the crowding degree, the sparser the solutions around the particle, and the better its diversity. The fitness value can be determined based on the non-dominated rank and the crowding degree. The particles with a smaller non-dominated rank and a greater crowding degree have a higher fitness value.

[0109] Update the velocity and position of the particles according to the fitness value. The formulas and steps used in the particle swarm optimization part are the same as those used in step S2, including the inertia weight, learning factor, velocity update formula, and position update formula.

[0110] During the update process, ensure that the positions of the particles satisfy the constraint conditions, such as the charge and discharge power constraints, capacity constraints, and grid power balance constraints of the energy storage power station.

[0111] Iteratively optimize until the termination condition is met (such as reaching the maximum iteration or the non-dominated solution set converges. It can be set that the change in the non-dominated solution set in several consecutive iterations is less than a certain threshold, or reaching the preset maximum number of iterations, such as 150 times), and obtain the multi-objective.

[0112] During the iterative optimization process, by continuously adjusting the velocity and position of the particles, the particles search for better parameter combinations in the solution space.

[0113] After each iteration, recalculate the objective function value vector, non-dominated sorting, and crowding degree according to the new particle positions, and update the personal best and global best of the particles.

[0114] The personal best is the parameter combination corresponding to the self-optimal position found by each particle during the iteration process, and the global best is the optimal parameter combination in the entire particle swarm.

[0115] S53: Implement the optimization plan and continuously improve:

[0116] According to the optimal parameter combination obtained by the GAPSO algorithm, implement the new load forecasting model and the charge and discharge strategy of the energy storage power station.

[0117] Apply the optimized load forecasting model parameters to the load forecasting module to update the forecasting model to improve the accuracy of load forecasting.

[0118] Apply the charge and discharge strategy parameters of the energy storage power station to the control system of the energy storage power station to adjust the charge and discharge operations of the energy storage power station.

[0119] Dynamically adjust the weight coefficients λ1, λ2, and λ3 according to the actual operation conditions; if it is found that the power grid stability is poor in actual operation, the value of λ1 can be appropriately increased, and then the GAPSO algorithm is run again for optimization to achieve better comprehensive performance. Through continuous monitoring, evaluation, and adjustment, the operation of the energy storage power station is always adapted to the dynamic changes of the power grid, improving the overall performance and adaptability of the system. At the same time, record and analyze the optimization process and results to provide data support and experience reference for subsequent system optimization and decision-making.

[0120] Compared with the prior art, the beneficial effects of the present invention are as follows: Compared with the prior art, the beneficial effects of the present invention are mainly reflected in the following aspects: First of all, by introducing the GAPSO algorithm, the present invention realizes accurate prediction of the power grid load. The GAPSO algorithm combines the advantages of the genetic algorithm and the particle swarm algorithm, has strong global search ability and fast convergence speed, can significantly improve the accuracy and stability of load prediction, and provides reliable data support for the optimization of the charge and discharge strategy of the energy storage power station.

[0121] Secondly, the present invention has made a breakthrough in the optimization of the charge and discharge strategy of the energy storage power station. By establishing an optimization objective function and imposing conditional constraints on the capacity of the energy storage power station, the optimal charge and discharge strategy is obtained by solving with the GAPSO algorithm. This strategy not only realizes the minimization of the operation cost and the maximization of the revenue of the energy storage power station, but also fully considers the battery life attenuation cost and extends the service life of the energy storage power station.

[0122] In addition, the present invention also realizes real-time scheduling and response. By real-time monitoring the power grid load and the state of the energy storage power station and making scheduling decisions according to the prediction results and optimization strategies, the operation state of the energy storage power station can be adjusted in time to ensure the stable operation of the power grid.

[0123] Finally, the present invention also considers multi-objective optimization and adjustment. Considering multiple key objectives such as power grid stability, economic benefits of the energy storage power station, and life loss of the energy storage power station, a multi-objective optimization function is constructed, and the optimal parameter combination is obtained by solving with the GAPSO algorithm. This solution not only improves the accuracy of load prediction and optimizes the charge and discharge strategy of the energy storage power station, but also takes into account the requirements of multiple aspects such as power grid stability, economy, and energy storage life. BRIEF DESCRIPTION OF THE DRAWINGS

[0124] Figure 1 It is a schematic diagram of the method framework in the present invention;

[0125] Figure 2 It is a schematic diagram of the method flow in the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0126] To further understand the purpose, structure, features, and functions of the present invention, the following is a detailed description in conjunction with embodiments.

[0127] Please refer to Figure 1 and Figure 2 The present invention provides a load tracking method for a storage power station based on the GAPSO algorithm, which includes the following steps: S1. Data collection and preprocessing; S2. Construction of a power grid load prediction model based on the GAPSO algorithm; S3. Optimization of the charge and discharge strategy of the energy storage power station; S4. Real-time scheduling and response; S5. Multi-objective optimization adjustment.

[0128] Preferably, the step S1, data collection and preprocessing, includes the following steps:

[0129] S11. Data collection;

[0130] In data collection, historical load data of the power grid, real-time load data, operating parameters of the energy storage power station (including battery capacity, charge and discharge efficiency, etc.), electricity price information, and other relevant influencing factor data (such as weather data, time factors, etc.) are collected. Load data for different time periods are obtained from the power grid monitoring system, battery status information is obtained from the energy storage power station management system, and electricity price change data is obtained from the power market.

[0131] S12. Data preprocessing;

[0132] The collected data is cleaned to remove outliers and incorrect data;

[0133] The method of the 3σ criterion is used to identify and correct data points that deviate significantly from the normal range;

[0134] Missing data is supplemented, and methods such as mean imputation and linear interpolation can be used;

[0135] If the load data at a certain moment is missing, it is supplemented according to the average value or linear relationship of the load data at the previous and subsequent moments. The data is normalized to map the data to a specific interval (such as [0,1] or [-1,1]) to improve the subsequent calculation efficiency.

[0136] Let the original data be x, and the normalized data (for mapping to the [0,1] interval), where x min and x max are the minimum and maximum values of the data respectively.

[0137] By comprehensively and deeply analyzing various influencing factors of the power grid load, including historical load data, time characteristics, weather factors, and special events, and quantifying these factors and incorporating them into the scope of input variables, a more scientific and reasonable input vector of the prediction model is constructed.

[0138] Based on the GAPSO algorithm to optimize the neural network prediction model, using the global search ability of the algorithm and the non-linear mapping ability of the neural network, it effectively learns and fits the complex load change law, thus significantly improving the accuracy of power grid load prediction and providing a more reliable basis for formulating the charge and discharge strategies of energy storage power stations.

[0139] Preferably, the S2, construction of the power grid load prediction model based on the GAPSO algorithm, includes the following steps:

[0140] S21: Determine the input and output variables;

[0141] Construct the input vector X = [x1, x2, …, x n , which includes the load values at 6 moments, from x1 to x6; the hour number of the current moment is x7; the flag variable indicating whether it is a working day on the current day is x8, where x8 is 0 or 1; the temperature value is x9; the humidity value is x 10 ; the special event flag variable x 11 ; the output variable is y, that is, the load prediction value at the future moment;

[0142] S22: Construct a neural network prediction model optimized by the GAPSO algorithm;

[0143] Let the position vector of particle i be θ i = [w i1 , w i2 , … w ij , b i1 , b i2 , … b ik , where w ij represents the weight of the jth connection of the neural network, and b ik represents the threshold of the kth neuron;

[0144] Calculate the fitness value of each particle. The fitness function uses the mean square error. Let the number of prediction samples be N, the actual load value be y i , and the predicted load value be Then the fitness function

[0145] The predicted load value is obtained by inputting the input vector X into the neural network and calculating through forward propagation, that is where f represents the mapping function of the neural network;

[0146] Update the velocity and position of the particle according to the fitness value. During the update process, for the genetic algorithm part, adopt dynamic crossover and mutation probabilities;

[0147] Crossover probability:

[0148] Pc = MAX(P c × 0.95 t / 4 , P cMIN ),

[0149] Mutation probability:

[0150] P m = MIN(P m × 1.05 t / 4 , P cMAX ),

[0151] where t is the number of iterations, P cMIN and P cMAX are the minimum crossover probability and the maximum mutation probability respectively;

[0152] The initial crossover probability P c is set between 0.6 - 0.9, the minimum crossover probability is set between 0.3 - 0.5, the initial mutation probability P cMIN is set between 0.01 - 0.1, and the maximum mutation probability P m is set between 0.1 - 0.3;

[0153] In the particle swarm optimization part, the inertia weight w = w max - k × t, w max is the initial weight, k is a constant, t is the number of iterations, and the learning factors c1 = c2 = 2; the initial weight w max is set between 0.7 - 0.9, and the constant is calculated according to the maximum number of iterations T PSOMAX , k = 0.6 / T PSOMAX , where, T PSOMAX = 100;

[0154] The velocity update formula is:

[0155] v id (t + 1) = wv id (t) + c1ξ(p id (t) - x id (t)) + c2η(p gd (t) - x id (t)),

[0156] The position update formula is:

[0157] x id (t + 1) = x id (t) + v id (t + 1),

[0158] where v id (t) is the velocity of particle i in the d-th dimension at the t-th iteration, x id(t) is the position of particle i in the d-th dimension at the t-th iteration, p id (t) is the value of the personal best of particle i in the d-th dimension at the t-th iteration, p gd (t) is the value of the global best in the -th dimension, and ξ and η are random numbers uniformly distributed in the interval [0, 1].

[0159] Preferably, the step S3, optimization of the charge-discharge strategy of the energy storage power station, includes the following steps: S31: Establish an optimization objective function:

[0160] The establishment of the optimization objective function is to minimize the operation cost and maximize the revenue of the energy storage power station, while considering the cost of battery life attenuation;

[0161] The operation cost mainly includes the charging cost and the discharging cost;

[0162] The charging cost is related to the charging electricity price c charge (t) and the charging power P charge (t), and the discharging cost is related to the discharging electricity price c discharge (t) and the discharging power P discharge (t); during discharging, it can be regarded as the opposite of the revenue;

[0163] The cost of battery life attenuation is calculated according to the number of charge-discharge cycles and the depth of the battery;

[0164] Adopt the model of the Arrhenius formula:

[0165] The cost of battery life attenuation is C loss (t),

[0166] Then the objective function is:

[0167]

[0168] Among them, T is the scheduling period; such as one day or one hour, which is set according to specific requirements.

[0169] The calculation of the charging cost and the discharging cost can be determined according to the actual electricity price policy and market rules. If time-of-use electricity price is adopted, c charge (t) and c discharge (t) have different values at different time periods;

[0170] The calculation of the cost of battery life attenuation can refer to the formula of the battery life attenuation model based on the Arrhenius formula, according to the relationship between the life loss of the energy storage battery and the discharging power:

[0171] The battery capacity is E bat , the number of cycles is n, the cycle depth is h, and the battery throughput table is:

[0172] Ah = E bat × n × h;

[0173] Then the life loss:

[0174]

[0175] Where A is the pre-exponential factor, E a is the experimental activation energy, R is the molar gas constant, T is the absolute temperature, and z is the reaction rate constant;

[0176] Further calculate the battery life attenuation cost (combining factors such as battery replacement cost). S32: Determine the constraint conditions:

[0177] The constraint conditions include the charge and discharge power constraint of the energy storage power station, the capacity constraint of the energy storage power station, and the grid power balance constraint;

[0178] The charge and discharge power constraint of the energy storage power station determines the maximum charge and discharge power according to the rated power limit of the energy storage power station equipment;

[0179] Including the maximum charge power P chargeMAX , the maximum discharge power P dischargeMAX ;

[0180] Then 0 ≤ P charge (t) ≤ P chargeMAX , 0 ≤ P discharge (t) ≤ P dischargeMAX ;

[0181] For an energy storage power station with a rated power of 5 MW:

[0182] P chargeMAX = 5 MW, P dischargeMAX = 4 MW (considering a certain power reserve or efficiency factor).

[0183] The capacity constraint of the energy storage power station ensures the safe and stable operation of the energy storage power station, and its state of charge (SOC) needs to be maintained within a certain range;

[0184] Let SOC min be the lower limit of the state of charge, and be SOC max the upper limit of the state of charge,

[0185] Then SOC min < SOC(t) < SOC max ;

[0186] The calculation of SOC is related to the charge and discharge process of the battery:

[0187]

[0188] where σ is the self-discharge rate of the energy storage, and η c is the charging efficiency, and η d is the discharging efficiency, and Δt is the scheduling time interval;

[0189] The power balance constraint of the power grid means that at any moment, the power exchanged between the power grid and the energy storage power station plus the charging and discharging power of the energy storage power station should be equal to the power grid load, that is

[0190] P grid (t)+P discharge (t)-P charge (t)=P load (t);

[0191] where P grid (t) is the power exchanged between the power grid and the energy storage power station, and P load (t) is the power grid load;

[0192] This constraint ensures the power balance of the power grid operation and avoids the power grid instability problem caused by power imbalance.

[0193] S33: Solve the charging and discharging strategy using the GAPSO algorithm:

[0194] By using the GAPSO algorithm to solve the charging and discharging strategy, the optimization objective function and constraint conditions are transformed into forms that can be processed by the GAPSO algorithm; each particle represents a set of possible charging and discharging strategies, that is, the position vector of the particle represents the charging and discharging power values at different times;

[0195] Let the position vector of the particle be:

[0196]

[0197] where and respectively represent the charging power and discharging power of particle i at time t;

[0198] Then calculate the fitness value of the particle, and calculate the fitness of each particle according to the optimization objective function; the fitness function is

[0199] fitness i =F(x i ), that is, the objective function value corresponding to the charging and discharging strategy represented by particle i. Substitute the position vector of the particle into the objective function to calculate the corresponding fitness value; this value reflects the quality of the charging and discharging strategy.

[0200] Update the velocity and position of the particle according to the fitness value. The formulas and steps used in the particle swarm algorithm part are the same as those used in step S2, including the inertia weight, learning factor, velocity update formula, and position update formula;

[0201] During the update process, ensure that the positions of the particles satisfy the constraint conditions. If the updated particle positions exceed the charging and discharging power constraints or the capacity constraint range, correct them to the feasible region.

[0202] For example, if the calculated charging power is greater than P chargeMAX , set it to P chargeMAX ; if the state of charge is lower than SOC min , adjust the charging and discharging power to bring the state of charge back within the feasible range.

[0203] Iteratively optimize until the termination condition is met to obtain the optimal charging and discharging strategy. The charging and discharging power sequence corresponding to the optimal particle position at this time is the "optimal charging and discharging strategy of the energy storage power station at different times, which can be used to guide the actual operation of the energy storage power station.

[0204] An optimization objective function that comprehensively considers the operating cost (charging cost and discharging cost), revenue (discharging revenue), and battery life attenuation cost is constructed, enabling the operating strategy of the energy storage power station to achieve a balance between economic benefits and equipment life.

[0205] Combined with the actual operating conditions such as the charging and discharging power constraints, capacity constraints, and grid power balance constraints of the energy storage power station, the GAPSO algorithm is used to solve for the optimal charging and discharging strategy, ensuring that the energy storage power station minimizes the operating cost and maximizes the revenue while operating safely and stably, and at the same time extending the battery service life and improving the overall operating efficiency of the energy storage power station.

[0206] Preferably, the step S4, real-time scheduling and response, includes the following steps:

[0207] S41: Real-time monitor the grid load and the status of the energy storage power station:

[0208] Through high-precision sensors installed at key grid nodes and inside the energy storage power station, continuously obtain the real-time load data of the grid, the real-time state of charge (SOC) of the energy storage power station, and the information of the charging and discharging power;

[0209] The data acquisition frequency can be set to be acquired once every few minutes (such as 3 - 5 minutes) to ensure that the changes in the grid load and the status of the energy storage power station can be captured in a timely manner.

[0210] The acquired data is processed and transmitted in real time to ensure the accuracy and timeliness of the data. Use reliable data communication technologies (such as specific protocols in wired communication or wireless communication) to transmit the data to the monitoring center or control system, and check and correct the data during the transmission process to avoid the impact of data transmission errors on subsequent decisions.

[0211] S42: Make scheduling decisions based on the prediction results and optimization strategies:

[0212] Input the real-time monitoring data in step S41 into the load prediction model pre-constructed in step S2 to obtain the short-term load prediction value; then use the neural network prediction model optimized by the GAPSO algorithm to predict the change trend of the grid load within the next 15 minutes or half an hour based on the data of the current moment and the previous grid load, time characteristics, and weather factors.

[0213] S43: Adjust the operating state of the energy storage power station:

[0214] After receiving the charge and discharge power command, the energy storage power station realizes the charge and discharge operation by controlling the output of the inverter; the inverter adjusts the output voltage, frequency, and phase according to the command to accurately control the charge and discharge power of the energy storage power station to make it consistent with the command value.

[0215] An efficient data acquisition system is established to obtain the grid load and the status information of the energy storage power station in real time through high-precision sensors, and reliable data communication technology is used for real-time processing and transmission to ensure the accuracy and timeliness of the data.

[0216] According to the real-time monitoring data and the optimized prediction model, accurate scheduling decisions can be made in a timely manner, and the charge and discharge power of the energy storage power station can be adjusted to enable it to quickly respond to the change of the grid load, realize accurate load tracking and power regulation, effectively ensure the stable operation of the grid, and reduce the risk of grid faults caused by load fluctuations.

[0217] Preferably, step S5, multi-objective optimization adjustment, includes the following steps:

[0218] S51: Construction of multi-objective optimization function:

[0219] Comprehensively consider the key objectives of grid stability, economic benefits of the energy storage power station, and life loss of the energy storage power station, and construct a multi-objective optimization function;

[0220] For grid stability, the load fluctuation index is used to measure:

[0221] The variance of load fluctuation is as follows:

[0222] where P load,i is the load value at the i-th moment, is the average load value, and N is the number of samples;

[0223] In terms of economic benefits, the difference between the operating cost and the revenue is used as the index:

[0224]

[0225] Where R(t) is the other income of the energy storage power station at time t, including the income obtained from participating in frequency regulation and the reserve market;

[0226] The life loss index is calculated based on the charge and discharge cycle times and depth of the battery. Consistent with step S31, the relevant battery life attenuation model, that is, the model of the Arrhenius formula, is adopted. Let the battery life attenuation cost be C loss (t),

[0227] Then

[0228] Construct a multi-objective optimization function:

[0229] O = λ1×f1 + λ2×f2 + λ3×f3, where λ1, λ2, and λ3 are the weight coefficients of the corresponding objectives; they are adjusted according to the actual power grid's emphasis on stability, economy, and energy storage life. In the case of large power grid load fluctuations and high requirements for stability, the value of λ1 can be appropriately increased; if the economic benefits of the energy storage power station are the primary consideration factor, the value of λ2 can be increased; if it is desired to extend the battery life of the energy storage power station, the value of λ3 can be increased.

[0230] S52: Solve the multi-objective optimization problem with the GAPSO algorithm:

[0231] Integrate the multi-objective optimization function and relevant constraint conditions into the GAPSO algorithm framework. Each particle represents a set of parameter combinations that meet the constraint conditions, including load prediction model parameters (such as the weights and thresholds of the neural network) and energy storage power station charge and discharge strategy parameters (such as the charge and discharge power at different times);

[0232] Calculate the fitness value of the particle. Since it is a multi-objective optimization problem, methods such as non-dominated sorting and crowding degree calculation are used to evaluate the fitness of the particle;

[0233] Calculate the objective function value vector [f1(x i ), f2(x i ), f3(x i )] of each particle according to the multi-objective optimization function, where x i is the position vector of particle i, and then perform non-dominated sorting on the particles, arranging the non-dominated particles in the front to form different non-dominated levels. For the particles within the same non-dominated level, calculate their crowding degree;

[0234] The crowding degree reflects the density of the solutions around the particle. The larger the crowding degree, the sparser the solutions around the particle, and the better its diversity. The fitness value can be determined according to the non-dominated level and the crowding degree. The particles with smaller non-dominated levels and larger crowding degrees have higher fitness values

[0235] Update the velocity and position of the particles according to the fitness value. The formulas and steps used in the particle swarm optimization part are the same as those used in step S2, including the inertia weight, learning factor, velocity update formula, and position update formula;

[0236] During the update process, ensure that the positions of the particles satisfy the constraint conditions, such as the charge and discharge power constraints, capacity constraints, and grid power balance constraints of the energy storage power station.

[0237] Iteratively optimize until the termination condition is met (such as reaching the maximum number of iterations or the non-dominated solution set converges. It can be set that the change in the non-dominated solution set in several consecutive iterations is less than a certain threshold, or reaching the preset maximum number of iterations, such as 150 times), to obtain the multi-objective.

[0238] During the iterative optimization process, by continuously adjusting the velocity and position of the particles, the particles search for better parameter combinations in the solution space.

[0239] After each iteration, recalculate the objective function value vector, non-dominated sorting, and crowding degree according to the new particle positions, and update the personal best and global best of the particles.

[0240] The personal best is the parameter combination corresponding to the optimal position found by each particle during the iteration, and the global best is the optimal parameter combination in the entire particle swarm.

[0241] S53: Implement the optimization plan and continuously improve:

[0242] According to the optimal parameter combination obtained by the GAPSO algorithm, implement the new load forecasting model and the charge and discharge strategy of the energy storage power station;

[0243] Apply the optimized load forecasting model parameters to the load forecasting module to update the forecasting model to improve the accuracy of load forecasting;

[0244] Apply the charge and discharge strategy parameters of the energy storage power station to the control system of the energy storage power station to adjust the charge and discharge operations of the energy storage power station.

[0245] Dynamically adjust the weight coefficients λ1, λ2, λ3 according to the actual operation situation; if it is found that the grid stability is poor in the actual operation, the value of λ1 can be appropriately increased, and then run the GAPSO algorithm again for optimization to achieve better comprehensive performance. Through continuous monitoring, evaluation, and adjustment, make the operation of the energy storage power station always adapt to the dynamic changes of the power grid, improve the overall performance and adaptability of the system. At the same time, record and analyze the optimization process and results to provide data support and experience reference for subsequent system optimization and decision-making.

[0246] Compared with the prior art, the beneficial effects of the present invention are as follows: Compared with the prior art, the beneficial effects of the present invention are mainly reflected in the following aspects: First, by introducing the GAPSO algorithm, the present invention realizes the accurate prediction of the power grid load. The GAPSO algorithm combines the advantages of the genetic algorithm and the particle swarm algorithm, has strong global search ability and fast convergence speed, can significantly improve the accuracy and stability of load prediction, and provides reliable data support for the optimization of the charge and discharge strategies of energy storage power stations.

[0247] Secondly, the present invention has made a breakthrough in the optimization of the charge and discharge strategies of energy storage power stations. By establishing an optimization objective function and imposing conditional constraints on the capacity of the energy storage power station, the optimal charge and discharge strategy is obtained by using the GAPSO algorithm. This strategy not only realizes the minimization of the operating cost and the maximization of the revenue of the energy storage power station, but also fully considers the cost of battery life attenuation and extends the service life of the energy storage power station.

[0248] In addition, the present invention also realizes real-time scheduling and response. By real-time monitoring the power grid load and the state of the energy storage power station and making scheduling decisions according to the prediction results and optimization strategies, the operating state of the energy storage power station can be adjusted in time to ensure the stable operation of the power grid.

[0249] An optimization function is constructed that comprehensively considers multiple key objectives such as the stability of the power grid (measured by the load fluctuation index), the economic benefits of the energy storage power station (measured by the difference between the operating cost and the revenue), and the life loss of the energy storage power station (calculated based on the number of charge and discharge cycles and depth of the battery), and the flexible emphasis on different objectives is achieved by adjusting the weight coefficients.

[0250] The GAPSO algorithm is used to solve the multi-objective optimization problem, and a set of non-dominated solutions is obtained, clearly presenting the trade-off relationship between different objectives, which is convenient for selecting a suitable optimization scheme according to actual needs. By continuously monitoring and dynamically adjusting the weight coefficients, the operation of the energy storage power station can adapt to the dynamic changes of the power grid, realize the comprehensive optimization of multiple objectives, and improve the overall performance and adaptability of the power system.

[0251] The present invention has been described by the above related embodiments. However, the above embodiments are only examples for implementing the present invention. It must be pointed out that the disclosed embodiments do not limit the scope of the present invention. On the contrary, modifications and refinements made without departing from the spirit and scope of the present invention fall within the scope of the patent protection of the present invention.

Claims

1. A load tracking method for a pumped storage power station based on the GAPSO algorithm, characterized in that: The steps of the method are as follows: S1: Data collection and preprocessing; In data collection, historical load data of the power grid, real-time load data, operating parameters of the energy storage power station, electricity price information, and other relevant influencing factor data are collected, including weather data and time factors; Load data for different time periods are obtained from the power grid monitoring system, battery state information is obtained from the energy storage power station management system, and electricity price change data are obtained from the power market; Then, the collected and obtained data are cleaned to remove outliers and incorrect data; S2: Construction of a power grid load prediction model based on the GAPSO algorithm; S21: Determine input and output variables; Construct the input vector X = [x1, x2, …, x n , which includes the load values at 6 moments, from x1 to x6; the hour number at the current moment is x7; the flag variable indicating whether it is a working day on that day is x8, where x8 is 0 or 1; the temperature value is x9; the humidity value is x 10 ; the special event flag variable x 11 ; the output variable is y, that is, the load prediction value at the future moment; S22: Construct a neural network prediction model optimized by the GAPSO algorithm; Let the position vector of particle i be θ i = [w i1 , w i2 , … w ij , b i1 , b i2 , … b ik , where w ij represents the weight of the j-th connection of the neural network, and b ik represents the threshold of the k-th neuron; Calculate the fitness value of each particle. The mean square error is used as the fitness function. Let the number of prediction samples be N, and the actual load value be y i , and the predicted load value is Then the fitness function Predicted load value It is obtained by inputting the input vector X into the neural network and performing forward propagation calculation, that is where f represents the mapping function of the neural network; Update the velocity and position of the particles according to the fitness value. During the update process, for the genetic algorithm part, dynamic crossover and mutation probabilities are adopted; Crossover probability: P c = MAX(P c × 0.95 t / 4 , P cMIN ), Mutation probability: P m = MIN(P m × 1.05 t / 4 , P cMAX ), where t is the number of iterations, P cMIN and P cMAX are the minimum crossover probability and the maximum mutation probability respectively; Initial crossover probability P c is set between 0.6 - 0.9, the minimum crossover probability is set between 0.3 - 0.5, and the initial mutation probability P cMIN is set between 0.01 - 0.1, and the maximum mutation probability P m is set between 0.1 - 0.3; In the particle swarm optimization algorithm part, the inertia weight w = w max - k × t, where w max is the initial weight, k is a constant, t is the number of iterations, and the learning factors c1 = c2 = 2; the initial weight w max is set between 0.7 and 0.9, and the constant is calculated according to the maximum number of iterations T PSOMAX as k = 0.6 / T PSOMAX , where T PSOMAX = 100; The velocity update formula is: v id (t + 1)= wv id (t)+ c1ξ(p id (t)- x id (t))+ c2η(p gd (t)- x id (t)) The position update formula is: x id (t + 1) = x id (t) + v id (t + 1), where v id (t) is the velocity of particle i in the d-th dimension at the t-th iteration, x id (t) is the position of particle i in the d-th dimension at the t-th iteration, p id (t) is the value of the personal best of particle i in the d-th dimension at the t-th iteration, p gd (t) is the value of the global best in the d-th dimension, ξ and η are random numbers uniformly distributed in the interval [0, 1]; When the termination condition is met, output the optimal weights and values to construct the final load prediction model; the combination of weights and thresholds corresponding to the optimal particle position at this time is what is sought, and then it is applied to the neural network for subsequent load prediction; S3: Optimization of the charging and discharging strategy of the energy storage power station; By establishing an optimization objective function, imposing conditional constraints on the capacity of the energy storage power station, and using the GAPSO algorithm in step S2 to solve the charging and discharging strategy, transform the optimization objective function and constraint conditions into a form that can be processed by the GAPSO algorithm, recalculate the fitness value of each particle according to the optimization objective function, and update the velocity and position of the particles according to the recalculated fitness value, and perform iterative optimization until the termination condition is met to obtain the optimal charging and discharging strategy; S4: Real-time scheduling and response; By real-time monitoring the power grid load and the status of the energy storage power station, make scheduling decisions according to the prediction results of step S2 and the optimization strategy of step S3. After the energy storage power station receives the charging and discharging power command, realize the charging and discharging operation by controlling the output of the inverter equipment; S5: Multi-objective optimization adjustment; Comprehensively consider multiple key objectives such as power grid stability, economic benefits of the energy storage power station, and life loss of the energy storage power station, construct a multi-objective optimization function, integrate the multi-objective optimization function and relevant constraint conditions into the GAPSO algorithm framework, and implement a new load prediction model and the charging and discharging strategy of the energy storage power station according to the optimal parameter combination obtained by the GAPSO algorithm.

2. The load tracking method of a pumped storage power station based on the GAPSO algorithm according to claim 1, characterized in that: In step S1, the data collection and preprocessing are specifically as follows: S11: Data collection: In data collection, historical load data of the power grid, real-time load data, operating parameters of the energy storage power station, electricity price information, and other relevant influencing factor data are collected, including weather data and time factors; Load data for different time periods are obtained from the power grid monitoring system, battery state information is obtained from the energy storage power station management system, and electricity price change data are obtained from the power market; S12: Data preprocessing: Clean the collected data to remove outliers and incorrect data; Adopt the 3σ criterion to identify and correct data points that deviate significantly from the normal range; supplement the missing data by using the methods of mean imputation and linear interpolation; Normalize the data and map it to a specific interval, [0, 1] or [-1, 1].

3. The load tracking method of a storage power station based on the GAPSO algorithm according to claim 1, characterized in that: In step S3, the charging and discharging strategy of the energy storage power station is optimized, and the specific steps are as follows: S31: Establish an optimization objective function: The establishment of the optimization objective function aims to minimize the operating cost and maximize the revenue of the energy storage power station, while considering the cost of battery life attenuation; The operating cost mainly includes the charging cost and the discharging cost; Charging cost is related to the charging electricity price c charge (t) and the charging power P charge Discharging cost is related to the discharging electricity price c discharge (t) and the discharging power P discharge (t); The battery life attenuation cost is calculated according to the charging and discharging cycle times and depths of the battery; Adopt the model of the Arrhenius formula: The battery life attenuation cost is C loss (t), Then the objective function is: Among them, T is the scheduling period; S32: Determine the constraint conditions: The constraint conditions include the charging and discharging power constraints of the energy storage power station, the capacity constraints of the energy storage power station, and the grid power balance constraints; The charging and discharging power constraints of the energy storage power station determine the maximum charging and discharging power according to the rated power limit of the energy storage power station equipment; including a maximum charging power P chargeMAX , a maximum discharging power P dischargeMAX ; Then 0 ≤ P charge (t) ≤ P chargeMAX , 0 ≤ P discharge (t) ≤ P dischargeMAX ; The capacity constraints of the energy storage power station ensure the safe and stable operation of the energy storage power station, and its state of charge (SOC) needs to be maintained within a certain range; Let SOC min be the lower limit of the state of charge, and SOC max be the upper limit of the state of charge, Then the SOC min <SOC(t)<SOC max ; The calculation of SOC is related to the charging and discharging process of the battery: where σ is the self-discharge rate of energy storage, η c is the charging efficiency, η d is the discharging efficiency, and Δt is the scheduling time interval; The grid power balance constraint means that at any moment, the power exchanged between the grid and the energy storage power station plus the charging and discharging power of the energy storage power station should be equal to the grid load, that is P grid (t) + P discharge (t) - P charge (t) = P load (t); Among which P grid (t) is the power exchanged between the power grid and the energy storage power station, and P load (t) is the power grid load S33: Solve the charge and discharge strategy using the GAPSO algorithm: By using the GAPSO algorithm to solve the charging and discharging strategy, the optimization objective function and the constraint conditions are transformed into a form that can be processed by the GAPSO algorithm; each particle represents a group of possible charging and discharging strategies, that is, the position vector of the particle represents the charging and discharging power values at different times; Let the position vector of the particle be: wherein and respectively represent the charging power and discharging power of particle i at time t; Then calculate the fitness value of the particles, and calculate the fitness of each particle according to the optimization objective function; the fitness function is fitness i = F(x i ), that is, the objective function value corresponding to the charge and discharge strategy represented by particle i. Substitute the position vector of the particle into the objective function to calculate the corresponding fitness value. Update the velocity and position of the particle according to the fitness value. The formulas and steps used in the particle swarm algorithm part are the same as those used in step S2, including the inertia weight, learning factor, velocity update formula, and position update formula; During the update process, ensure that the position of the particle satisfies the constraint conditions. If the updated particle position exceeds the charging and discharging power constraint or the capacity constraint range, correct it to the feasible region.

4. A load tracking method for a pumped storage power station based on the GAPSO algorithm according to claim 1, characterized in that: In step S4, real-time scheduling and response include the following steps: S41: Real-time monitor the grid load and the state of the energy storage power station: Through high-precision sensors installed at key grid nodes and inside the energy storage power station, continuously obtain the real-time load data of the grid, the real-time state of charge (SOC) of the energy storage power station, and the information of the charging and discharging power; S42: Make scheduling decisions according to the prediction results and the optimization strategy: Input the real-time monitoring data in step S41 into the load prediction model pre-constructed in step S2 to obtain the short-term load prediction value; then use the neural network prediction model optimized based on the GAPSO algorithm to predict the future grid load change trend according to the data of the current moment and the previous grid load, time characteristics, and weather factors; S43: Adjust the operating state of the energy storage power station: After receiving the charging and discharging power instruction, the energy storage power station realizes the charging and discharging operation by controlling the output of the inverter; the inverter adjusts the output voltage, frequency, and phase according to the instruction to accurately control the charging and discharging power of the energy storage power station to make it consistent with the instruction value.

5. The load tracking method for a storage power station based on the GAPSO algorithm according to claim 3, characterized in that: The specific steps of the multi-objective optimization adjustment in step S5 are as follows: S51: Construct a multi-objective optimization function: Comprehensively consider the key objectives of grid stability, the economic benefits of the energy storage power station, and the life loss of the energy storage power station, and construct a multi-objective optimization function; For grid stability, use the load fluctuation index to measure: The variance of the load fluctuation is as follows: where P load,i is the load value at the i-th moment, is the average load value, and N is the number of samples; In terms of economic benefits, the difference between operating costs and revenues is used as an indicator: Among them, R(t) is the other revenue of the energy storage power station at time t, including the revenue obtained from participating in frequency regulation and the reserve market; The life loss index is calculated based on the charge and discharge cycle times and depth of the battery. Similar to that in step S31, a relevant battery life attenuation model, namely the Arrhenius formula model, is adopted. Let the battery life attenuation cost be C loss (t), Then Construct a multi-objective optimization function: O = λ1×f1 + λ2×f2 + λ3×f3, where λ1, λ2, and λ3 are the weight coefficients of the corresponding objectives; they are adjusted according to the actual power grid's emphasis on stability, economy, and energy storage life; S52: Solve the multi-objective optimization problem using the GAPSO algorithm: Integrate the multi-objective optimization function and relevant constraint conditions into the GAPSO algorithm framework. Each particle represents a set of parameter combinations that satisfy the constraint conditions, including the parameters of the load forecasting model and the charge and discharge strategies of the energy storage power station; Calculate the fitness value of the particle. Since it is a multi-objective optimization problem, methods such as non-dominated sorting and crowding degree calculation are used to evaluate the fitness of the particle; Calculate the objective function value vector [f1(x i ), f2(x i ), f3(x i )] for each particle according to the multi-objective optimization function, where x i is the position vector of particle i, and then perform non-dominated sorting on the particles, arranging the non-dominated particles in the front to form different non-dominated ranks. For the particles within the same non-dominated rank, calculate their crowding degree; Update the velocity and position of the particle according to the fitness value. The formulas used in the particle swarm algorithm part are the same as those used in step S2, including the inertia weight, learning factor, velocity update formula, and position update formula; During the update process, ensure that the position of the particle satisfies the constraint conditions, such as the charge and discharge power constraints, capacity constraints, and power balance constraints of the power grid for the energy storage power station; S53: Implement the optimization plan and continuously improve: According to the optimal parameter combination obtained by the GAPSO algorithm, implement a new load forecasting model and the charge and discharge strategies of the energy storage power station; Apply the optimized load forecasting model parameters to the load forecasting module and update the forecasting model to improve the accuracy of load forecasting; Apply the charge and discharge strategy parameters of the energy storage power station to the control system of the energy storage power station and adjust the charge and discharge operations of the energy storage power station.

Images