A power generation and control method based on an integrated photovoltaic-storage-utilization new energy microgrid

Abstract

This invention relates to the field of microgrid power generation technology and proposes a power generation control method for a renewable energy microgrid based on integrated photovoltaic, energy storage, and energy utilization. The method includes the following steps: acquiring initial real-time operating data of the renewable energy microgrid; preprocessing the initial real-time operating data to obtain real-time operating data; predicting the trend of photovoltaic power generation changes in the renewable energy microgrid based on a photovoltaic power generation prediction model, real-time photovoltaic power generation data, and real-time meteorological data; predicting the trend of load demand changes in the renewable energy microgrid based on a load demand prediction model and real-time load data; performing harmonic analysis on the renewable energy microgrid based on voltage and current waveform data to obtain the total harmonic distortion rate; generating a multi-objective scheduling scheme based on a multi-objective optimization algorithm, real-time operating data, and prediction data; and prioritizing the system load of the renewable energy microgrid based on the multi-objective scheduling scheme to regulate the power generation strategy of the renewable energy microgrid. This invention achieves comprehensive optimization and control of photovoltaic power generation, energy storage, load demand, and harmonic distortion in renewable energy microgrids, as well as the integration of photovoltaic, energy storage, and energy utilization in renewable energy microgrids.

Description

Technical Field

[0001] This invention relates to the field of microgrid power generation technology, and in particular to a power generation control method based on an integrated photovoltaic, energy storage and utilization new energy microgrid. Background Technology

[0002] Microgrids can operate in two different modes: islanded mode and grid-connected mode. While the system can operate relatively smoothly and maintain a stable output under traditional control modes, microgrids experience significant system disturbances when disconnected from the grid. Furthermore, renewable energy generation is volatile, and loads are time-varying, resulting in a high total harmonic distortion (THD) rate. Numerous harmonics can damage system components and distort the output voltage and current waveforms, making it impossible to adjust to rated output and severely impacting power quality.

[0003] Existing technologies generally achieve precise control and optimization of energy storage charging and discharging in microgrids by optimizing the charging and discharging of energy storage systems within the microgrid. However, due to the volatility of renewable energy generation, the system operates under high voltage when the load is high, resulting in a high total harmonic distortion (THD). Existing charging and discharging control methods have slow response times and cannot effectively reduce the THD. Summary of the Invention

[0004] In view of this, this invention proposes a power generation and control method for a renewable energy microgrid based on integrated photovoltaic, energy storage, and energy utilization. By acquiring and preprocessing real-time operating data, a photovoltaic power generation prediction model and a load demand prediction model are used to predict photovoltaic power generation and load demand. Harmonic analysis is performed on current and voltage data, and a multi-objective optimization algorithm is used to generate a scheduling scheme. The status of the renewable energy microgrid is monitored in real time, and the power generation and control strategy is adjusted. This achieves comprehensive optimization and control of photovoltaic power generation, energy storage, load demand, and harmonic distortion in the renewable energy microgrid, as well as the integration of photovoltaic, energy storage, and energy utilization in the renewable energy microgrid. This solves the problems of power generation volatility, load time-varying nature, and harmonic distortion in renewable energy microgrids.

[0005] The technical solution of this invention is implemented as follows: This invention provides a power generation and control method based on an integrated photovoltaic, energy storage, and renewable energy microgrid, comprising the following steps:

[0006] S1, acquire the initial real-time operation data of the new energy microgrid, preprocess the initial real-time operation data to obtain real-time operation data, the real-time operation data includes voltage and current waveform data, real-time photovoltaic power generation data, real-time meteorological data and real-time load data;

[0007] S2, based on the photovoltaic power generation prediction model, real-time photovoltaic power generation data and real-time meteorological data, predict the trend of photovoltaic power generation change of the new energy microgrid, based on the load demand prediction model and real-time load data, predict the trend of load demand change of the new energy microgrid, and perform harmonic analysis on the new energy microgrid based on the voltage and current waveform data to obtain the total harmonic distortion rate.

[0008] S3. Based on a multi-objective optimization algorithm, real-time running data, and predicted data, a multi-objective scheduling scheme is generated. The predicted data includes the trend of photovoltaic power generation, the trend of load demand, and the total harmonic distortion rate. The optimization objectives of the multi-objective scheduling scheme include the trend of photovoltaic power generation, the energy storage capacity, the trend of load demand, and the total harmonic distortion rate.

[0009] S4 monitors the current energy storage capacity, photovoltaic power generation, load output, and total harmonic distortion rate of the new energy microgrid in real time, and adjusts the power supply and generation strategy of the new energy microgrid based on the multi-objective scheduling scheme, the current energy storage capacity, the current photovoltaic power generation, the current load output, and the current total harmonic distortion rate.

[0010] Based on the above technical solutions, preferably, step S1 includes:

[0011] The initial real-time running data is preprocessed, including data cleaning, outlier removal, data formatting, and data standardization.

[0012] The data cleaning includes filling in missing values ​​using polynomial interpolation.

[0013] The removal of outliers includes identifying outliers using the quartile range method;

[0014] The data formatting includes converting date and time in different formats into a unified format, converting different units of measurement into a unified unit of measurement, and converting all text data into a unified character encoding format;

[0015] The data standardization includes transforming the data using Z-score standardization.

[0016] Based on the above technical solutions, preferably, the construction process of the photovoltaic power generation prediction model is as follows:

[0017] Historical photovoltaic power generation data and historical meteorological data are obtained, and the historical photovoltaic power generation data and historical meteorological data are cleaned to remove missing values ​​and outliers;

[0018] The seasonal variation coefficient is extracted by time series decomposition method. The difference between historical photovoltaic power generation data and seasonal variation coefficient is tested for unit root. When the difference series is non-stationary, the historical photovoltaic power generation data and seasonal variation coefficient are differentially processed to obtain the photovoltaic power generation difference component. The autocorrelation function graph and partial autocorrelation function graph of the photovoltaic power generation difference component are plotted. Based on the characteristics of the autocorrelation function graph and partial autocorrelation function graph, the initial autoregressive term order set and the initial moving average term order set are determined.

[0019] An initial photovoltaic power generation prediction model is determined, and the calculation formula for the initial photovoltaic power generation prediction model is as follows:

[0020] ;

[0021] in, For the autoregressive part, a Let the order of the initial autoregressive term be . B For backward shift operators, For difference operators, c This is the initial difference processing order. for t Photovoltaic power generation at time 0 This is the seasonal variation coefficient. d For the difference intercept term, For meteorological data coefficient vectors, for t Meteorological data vector at time 0, For the moving average portion, b The order of the initial moving average term. for t White noise error term at time 0;

[0022] Different combinations of initial autoregressive term order, initial moving average term order, and initial difference processing order are selected. The initial photovoltaic power generation prediction models under different combinations are compared using the information criterion algorithm for preliminary screening. The initial photovoltaic power generation prediction models after preliminary screening are evaluated to obtain the photovoltaic power generation prediction model.

[0023] Based on the above technical solutions, preferably, the construction process of the load demand prediction model includes:

[0024] Obtain initial historical load data and initial historical meteorological data, and normalize the initial historical load data and initial historical meteorological data to obtain historical load data and historical meteorological data;

[0025] The time-series characteristics of historical load data are captured by long short-term memory units, and an initial load demand forecasting model is established by combining it with historical meteorological data.

[0026] The model parameters of the initial load demand prediction model are updated by an adaptive optimization algorithm until the loss function converges, thus obtaining the trained load demand prediction model.

[0027] The formula for the adaptive optimization algorithm is:

[0028] ;

[0029] ;

[0030] in, For the initial load demand forecasting model in t Model parameters at time 1, For the initial load demand forecasting model in t Model parameters at time 1-1 For the initial load demand forecasting model in t Adaptive learning rate at time 1 The gradient of the loss function parameters. , , , The initial load demand forecasting model is respectively in t The weight matrices of the forget gate, input gate, candidate memory unit, and output gate at time 1. , , , The initial load demand forecasting model is respectively in t The biases of the forget gate, input gate, candidate memory unit, and output gate at time 1;

[0031] The formula for calculating the loss function is:

[0032] ;

[0033] in, For the first t The loss function for two iterations N The total number of samples, For the first i Actual value of load demand. For the first i One load demand forecast value, The regularization coefficient is . For regularization terms, For the first t Model parameters of the initial load demand prediction model during the second iteration.

[0034] Based on the above technical solutions, preferably, the calculation formula of the load demand prediction model is:

[0035] ;

[0036] ;

[0037] ;

[0038] ;

[0039] ;

[0040] ;

[0041] ;

[0042] in, for t Forecasted load demand at any given time. This is the weight matrix of the output layer of the Long Short-Term Memory (LSTM) unit. The bias of the output layer of the Long Short Time Memory (LSTM) unit. For Long Short-Term Memory Units in t The hidden state at all times , , , These are the forget gate, input gate, candidate memory unit, and output gate, respectively. t Activation value at time, The current state of the memory cell. for Sigmoid Activation function , , , These are the weight matrices for the forget gate, input gate, candidate memory units, and output gate, respectively. , , , These are the biases for the forget gate, input gate, candidate memory unit, and output gate, respectively. The hyperbolic tangent activation function is used. For Long Short-Term Memory Units in t The hidden state at time -1 For historical load data, Historical meteorological data, For candidate memory units in t The activation value at time -1.

[0043] Based on the above technical solutions, preferably, step S2 includes:

[0044] The photovoltaic power generation forecasting model obtains the photovoltaic power generation at preset time intervals for each moment in the future period, plots the curve of photovoltaic power generation changing with time, and fits the curve to obtain the trend of photovoltaic power generation change. The load demand forecasting model obtains the load demand at preset time intervals for each moment in the future period, plots the curve of load demand changing with time, and fits the curve to obtain the trend of load demand change.

[0045] Based on the above technical solutions, preferably, step S2 further includes:

[0046] Harmonic analysis of the new energy microgrid is performed using the voltage and current waveform data. A Fast Fourier Transform (FFT) is applied to the voltage and current waveform data to convert it to the frequency domain, obtaining the fundamental RMS value and the RMS values ​​of the harmonic components. The total harmonic distortion rate is then calculated based on the fundamental and harmonic RMS values.

[0047] ;

[0048] in, THD Total harmonic distortion (THD) V 1 represents the fundamental effective value. V 2 represents the effective value of the second harmonic component. V 3 represents the effective value of the third harmonic component. V n for n Effective value of the second harmonic component.

[0049] Based on the above technical solutions, preferably, step S3 includes:

[0050] The objective function of the multi-objective optimization algorithm is:

[0051] ;

[0052] ;

[0053] in, Let the objective function of the multi-objective optimization algorithm be... , , , These are the weighting coefficients of the objective sub-functions for the changing trends of photovoltaic power generation, energy storage capacity, load demand, and total harmonic distortion, respectively. Let be the objective sub-function representing the trend of photovoltaic power generation. The objective sub-function is the trend of load demand changes. The objective function is the energy storage capacity value. Let the total harmonic distortion rate be the objective sub-function. X This is the solution to the objective function. , , , These represent the trends in photovoltaic power generation, energy storage capacity, load demand, and total harmonic distortion rate corresponding to the solutions to the objective function.

[0054] The constraints of the multi-objective optimization algorithm are:

[0055] ;

[0056] in, for t Photovoltaic power generation of the new energy microgrid at time 3. for t The power supply capacity of the new energy microgrid at time 3. for t The system load demand of the new energy microgrid at time 3. for t The power generation capacity of the new energy microgrid at time 3. for t The power imbalance at time 3. for t Energy storage capacity at time 3 This represents the minimum power supply value for a new energy microgrid. This represents the maximum power generation value of the new energy microgrid. for t Total harmonic distortion rate of the new energy microgrid at time 3. This is the threshold for total harmonic distortion.

[0057] Based on the above technical solutions, preferably, step S3 further includes:

[0058] Multiple initial solutions are generated by the objective function and constraints of the multi-objective optimization algorithm. The multiple initial solutions satisfy the constraints. The multi-objective optimization algorithm minimizes the objective function based on the constraints in order to select the optimal solution as the multi-objective scheduling scheme.

[0059] The steps of the multi-objective optimization algorithm include:

[0060] Generate initial solution set Each solution in the initial solution set satisfies the constraints.

[0061] For each solution in the initial solution set, calculate the objective function value corresponding to each solution, and select the optimal solution and the worst solution. The objective function value corresponding to the optimal solution is the smallest, and the objective function value corresponding to the worst solution is the largest.

[0062] The update process for each solution in the initial solution set is calculated as follows:

[0063] ;

[0064] in, For the first solution in the initial solution set j One solution. for The updated solution, This is the optimal solution in the initial solution set. This is the worst solution in the initial solution set. and These are the optimal and worst bias coefficients;

[0065] Based on the constraints, all updated solutions are filtered to obtain the first set of updated solutions;

[0066] Repeat the solution update until the preset number of iterations is reached;

[0067] The optimal solution in the final updated solution set is output as a multi-objective scheduling scheme, which includes photovoltaic power generation control value, energy storage power control value, load demand control value and total harmonic distortion rate control value.

[0068] Based on the above technical solutions, preferably, step S4 includes:

[0069] This paper describes a multi-objective scheduling scheme for managing the power generation and supply of a new energy microgrid. The scheme regulates the power generation and supply strategy of the microgrid, which includes power generation and supply status, photovoltaic power generation status, system load status, and harmonic suppression status. The scheme compares the current energy storage capacity of the microgrid with its control value to determine the power generation and supply status; compares the current photovoltaic power generation with its control value to determine the photovoltaic power generation status; compares the current load output with its load demand control value to determine the system load status; and compares the current total harmonic distortion (THD) rate with its control value to determine the harmonic suppression status.

[0070] The power generation and control method of the present invention based on an integrated photovoltaic, energy storage and utilization new energy microgrid has the following advantages over the prior art:

[0071] (1) By acquiring and preprocessing real-time operating data, photovoltaic power generation prediction model and load demand prediction model are used to predict photovoltaic power generation and load demand, and harmonic analysis is performed on current and voltage data. A scheduling scheme is generated by using a multi-objective optimization algorithm, and the status of the new energy microgrid is monitored in real time and the power supply strategy is adjusted. This realizes the comprehensive optimization and control of photovoltaic power generation, energy storage, load demand and harmonic distortion of the new energy microgrid, and realizes the integration of photovoltaic, energy storage and utilization of the new energy microgrid.

[0072] (2) By taking seasonal factors into account in the photovoltaic power generation prediction model, and using historical data and time series analysis methods, a photovoltaic power generation prediction model was constructed to accurately predict the changing trend of photovoltaic power generation and improve the prediction accuracy of the system.

[0073] (3) By using the load demand forecasting model, the changing trend of load demand can be accurately predicted, load management can be optimized, the reliability and efficiency of power supply can be improved, and the power quality can be evaluated by using the fast Fourier transform and harmonic distortion rate calculation, which serves as the basis for harmonic suppression and improves the power quality of new energy microgrids.

[0074] (4) By adopting a multi-objective optimization algorithm, the overall optimization is carried out by comprehensively considering photovoltaic power generation, energy storage, load demand and harmonic distortion, and the power generation strategy is monitored and dynamically adjusted, which enhances the adaptability and flexibility of the new energy microgrid and ensures the efficient operation of the new energy microgrid. Attached Figure Description

[0075] To more clearly illustrate the technical solutions in the embodiments of the present invention or 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 of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0076] Figure 1 This is a flowchart of a power generation and control method based on an integrated photovoltaic, energy storage, and renewable energy microgrid according to the present invention. Detailed Implementation

[0077] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0078] Please see Figure 1 This invention provides a power generation and control method based on an integrated photovoltaic-storage-utilization new energy microgrid, comprising the following steps:

[0079] S1, acquire the initial real-time operation data of the new energy microgrid, preprocess the initial real-time operation data to obtain real-time operation data, the real-time operation data includes voltage and current waveform data, real-time photovoltaic power generation data, real-time meteorological data and real-time load data;

[0080] S2, based on the photovoltaic power generation prediction model, real-time photovoltaic power generation data and real-time meteorological data, predict the trend of photovoltaic power generation change of the new energy microgrid, based on the load demand prediction model and real-time load data, predict the trend of load demand change of the new energy microgrid, and perform harmonic analysis on the new energy microgrid based on the voltage and current waveform data to obtain the total harmonic distortion rate.

[0081] S3. Based on a multi-objective optimization algorithm, real-time running data, and predicted data, a multi-objective scheduling scheme is generated. The predicted data includes the trend of photovoltaic power generation, the trend of load demand, and the total harmonic distortion rate. The optimization objectives of the multi-objective scheduling scheme include the trend of photovoltaic power generation, the energy storage capacity, the trend of load demand, and the total harmonic distortion rate.

[0082] S4 monitors the current energy storage capacity, photovoltaic power generation, load output, and total harmonic distortion rate of the new energy microgrid in real time, and adjusts the power supply and generation strategy of the new energy microgrid based on the multi-objective scheduling scheme, the current energy storage capacity, the current photovoltaic power generation, the current load output, and the current total harmonic distortion rate.

[0083] Specifically, this embodiment proposes a power generation and control method for a renewable energy microgrid based on integrated photovoltaic, energy storage, and energy utilization. By acquiring and preprocessing real-time operating data, photovoltaic power generation and load demand are predicted using photovoltaic power generation prediction models and load demand prediction models. Harmonic analysis is performed on current and voltage data, and a multi-objective optimization algorithm is used to generate a scheduling scheme. The status of the renewable energy microgrid is monitored in real time, and the power generation and control strategy is adjusted. This achieves comprehensive optimization and control of photovoltaic power generation, energy storage, load demand, and harmonic distortion in the renewable energy microgrid, as well as the integration of photovoltaic, energy storage, and energy utilization in the renewable energy microgrid.

[0084] Step S1 includes:

[0085] The initial real-time running data is preprocessed, including data cleaning, outlier removal, data formatting, and data standardization.

[0086] The data cleaning includes filling in missing values ​​using polynomial interpolation.

[0087] The removal of outliers includes identifying outliers using the quartile range method;

[0088] The data formatting includes converting date and time in different formats into a unified format, converting different units of measurement into a unified unit of measurement, and converting all text data into a unified character encoding format;

[0089] The data standardization includes transforming the data using Z-score standardization.

[0090] Specifically, step S1 in this embodiment improves data quality and reliability through preprocessing steps such as data cleaning and outlier removal. Polynomial interpolation fills in missing values ​​to maintain data continuity, while the quartile range method identifies outliers and effectively removes unreasonable data points, thus providing a reliable data foundation.

[0091] Data formatting ensures that data from different sources and in different formats can be processed uniformly. Standardizing date and time, units of measurement, and character encoding can eliminate processing errors caused by inconsistent data formats, improving the accuracy and efficiency of data processing.

[0092] Z-score standardization is used to transform the data, allowing data of different scales and units to be compared and analyzed under the same standard. This eliminates the influence of dimensions, ensuring that each feature has the same weight in the model.

[0093] Step S1 in this embodiment lays a solid data foundation for subsequent prediction and optimization steps through preprocessing operations. It is a crucial step in the entire new energy microgrid power generation and control method, which can improve stability and reliability.

[0094] The process of constructing the photovoltaic power generation prediction model is as follows:

[0095] Historical photovoltaic power generation data and historical meteorological data are obtained, and the historical photovoltaic power generation data and historical meteorological data are cleaned to remove missing values ​​and outliers;

[0096] The seasonal variation coefficient is extracted by time series decomposition method. The difference between historical photovoltaic power generation data and seasonal variation coefficient is tested for unit root. When the difference series is non-stationary, the historical photovoltaic power generation data and seasonal variation coefficient are differentially processed to obtain the photovoltaic power generation difference component. The autocorrelation function graph and partial autocorrelation function graph of the photovoltaic power generation difference component are plotted. Based on the characteristics of the autocorrelation function graph and partial autocorrelation function graph, the initial autoregressive term order set and the initial moving average term order set are determined.

[0097] An initial photovoltaic power generation prediction model is determined, and the calculation formula for the initial photovoltaic power generation prediction model is as follows:

[0098] ;

[0099] in, For the autoregressive part, a Let the order of the initial autoregressive term be . B For backward shift operators, For difference operators, c This is the initial difference processing order. for t Photovoltaic power generation at time 0 This is the seasonal variation coefficient. d For the difference intercept term, For meteorological data coefficient vectors, for t Meteorological data vector at time 0, For the moving average portion, b The order of the initial moving average term. for t White noise error term at time 0;

[0100] Different combinations of initial autoregressive term order, initial moving average term order, and initial difference processing order are selected. The initial photovoltaic power generation prediction models under different combinations are compared using the information criterion algorithm for preliminary screening. The initial photovoltaic power generation prediction models after preliminary screening are evaluated to obtain the photovoltaic power generation prediction model.

[0101] Specifically, the construction process of the photovoltaic power generation prediction model in this embodiment includes extracting seasonal variation coefficients through time series decomposition. The time series decomposition method decomposes the time series into trend components, seasonal components, and stochastic components. The specific process of this step is as follows:

[0102] Calculate a moving average to smooth the data, remove seasonality and random fluctuations, and obtain the trend component:

[0103] ;

[0104] in, for w The trend component of time, M To move the window length, For a moment w - l Historical photovoltaic power generation;

[0105] Calculate the average value for each seasonal cycle, subtract the trend component from the historical photovoltaic power generation data to obtain the seasonal and random components, and average the seasonal and random components to obtain the seasonal variation coefficient.

[0106] ;

[0107] in, This is the seasonal variation coefficient. for w The historical photovoltaic power generation at any given time, where NN is the total number of seasons in the time series of historical photovoltaic power generation data;

[0108] The construction process of the photovoltaic power generation prediction model in this embodiment also includes: performing a unit root test on the difference between historical photovoltaic power generation data and the seasonal variation coefficient. The unit root test is used to detect whether the time series is stationary. In this embodiment, the ADF test is used. The specific process of this step is as follows:

[0109] Assumptions:

[0110] Null hypothesis H 0: The time series has a unit root, indicating that the series is not stationary;

[0111] Alternative Hypothesis H 1. The time series does not have a unit root, indicating that the series is stationary;

[0112] Build ADF Model verification:

[0113] ;

[0114] in, for t The difference between historical photovoltaic power generation data and seasonal variation coefficient at time 4. for t The difference between historical photovoltaic power generation data and seasonal variation coefficient at time 4-1. and They are respectively and The difference, For constant terms, for t The time trend item at time 4. The coefficients of the time series to be tested are... The coefficient of the lagged term, for t Error term at time 4 p The lag order;

[0115] According to the above ADF The test model calculates the test statistic and compares it with the critical value. If the test statistic is less than the critical value, the null hypothesis is rejected, and the series is considered stationary. If the test statistic is not less than the critical value, the null hypothesis is not rejected, and the series is considered non-stationary.

[0116] Plot the autocorrelation function (AFC) and partial autocorrelation function (PIF) of the differential component of photovoltaic power generation. Based on the characteristics of the AFC and PIF, determine the initial set of autoregressive terms and the initial set of moving average terms. The initial set of autoregressive terms and the initial set of moving average terms are determined by the truncation of the AFC and PIF, respectively.

[0117] In the initial photovoltaic power generation prediction model of this embodiment, the autoregressive component Used to represent the relationship between current photovoltaic power generation and its past values; differential processing section. Used to make t Photovoltaic power generation at time 0 and seasonal variation coefficient To stabilize, a seasonal variation coefficient is introduced. Used to capture seasonal fluctuations in photovoltaic power generation; t Meteorological data vector at time 0 This includes various meteorological factors that may affect photovoltaic power generation, such as temperature and light intensity; meteorological data coefficient vectors. Used to indicate the magnitude of the impact on photovoltaic power generation; moving average portion Used to smooth out random fluctuations in historical photovoltaic data over time series; t White noise error term at time 0 Used to represent random errors not captured in the initial photovoltaic power generation prediction model; the difference intercept term d Used to compensate for errors in differential processing.

[0118] This embodiment's initial photovoltaic (PV) power generation forecasting model combines the concepts of autoregression and moving average. The autoregressive component captures the historical dependence of PV power generation, the differencing process eliminates non-stationarity, seasonal variation coefficients and meteorological data capture external influencing factors, and the moving average component smooths out random fluctuations, ultimately yielding a forecast of future PV power generation. This initial PV power generation forecasting model aims to comprehensively consider both the intrinsic characteristics of time series data and external influencing factors to improve forecast accuracy.

[0119] In this embodiment, through AIC The information criterion algorithm compares initial photovoltaic power generation prediction models under different combinations for preliminary screening. AIC The formula for calculating the information criterion algorithm is:

[0120] ;

[0121] in, AIC for AIC Evaluation value of the information criterion algorithm k for AIC The number of parameters in the model evaluated by the information criterion algorithm in this embodiment. AIC The parameters evaluated by the information criterion algorithm include the order of the initial autoregressive term and the order of the initial moving average term. k The value is 2. To obtain the maximum likelihood estimates of the initial photovoltaic power generation prediction model under different combinations of the initial autoregressive term order and the initial moving average term order, we select from all combinations of the initial autoregressive term order set and the initial moving average term order set. AIC The smallest corresponding initial autoregressive term order and initial moving average term order are used as the final autoregressive term order and moving average term order, thus obtaining the final photovoltaic power generation prediction model.

[0122] In one specific embodiment, the final order of the autoregressive term and the order of the moving average term can be 3 and 1, respectively. a The final number is 3. b Ultimately, we choose 1. However, considering that photovoltaic power generation data is usually affected by factors such as weather and seasons, exhibiting short-term non-stationarity, the first-order difference can effectively remove this short-term non-stationarity, making the series more stable. Therefore... c The final value is 1.

[0123] The process of constructing the load demand forecasting model includes:

[0124] Obtain initial historical load data and initial historical meteorological data, and normalize the initial historical load data and initial historical meteorological data to obtain historical load data and historical meteorological data;

[0125] The time-series characteristics of historical load data are captured by long short-term memory units, and an initial load demand forecasting model is established by combining it with historical meteorological data.

[0126] The model parameters of the initial load demand prediction model are updated by an adaptive optimization algorithm until the loss function converges, thus obtaining the trained load demand prediction model.

[0127] The formula for the adaptive optimization algorithm is:

[0128] ;

[0129] ;

[0130] in, For the initial load demand forecasting model in t Model parameters at time 1, For the initial load demand forecasting model in t Model parameters at time 1-1 For the initial load demand forecasting model in t Adaptive learning rate at time 1 The gradient of the loss function parameters. , , , The initial load demand forecasting model is respectively int The weight matrices of the forget gate, input gate, candidate memory unit, and output gate at time 1. , , , The initial load demand forecasting model is respectively in t The biases of the forget gate, input gate, candidate memory unit, and output gate at time 1;

[0131] The formula for calculating the loss function is:

[0132] ;

[0133] in, For the first t The loss function for two iterations N The total number of samples, For the first i Actual value of load demand. For the first i One load demand forecast value, The regularization coefficient is . For regularization terms, For the first t Model parameters of the initial load demand prediction model during the second iteration.

[0134] Specifically, in new energy microgrids, accurate load demand forecasting is crucial for the optimal scheduling of power resources. The load demand forecasting model in this embodiment uses an adaptive optimization algorithm to update the model parameters. Since the historical load data of new energy microgrids contains complex time series data, parameter gradients are introduced. By calculating the gradient of the loss function with respect to the parameters, the model parameters are updated in each iteration in the direction of reducing losses. The adaptive learning rate optimization algorithm can quickly adjust the learning rate, improve the convergence speed of the model, and enable the model to respond to changes in load demand in a timely manner. Through rapid convergence, new energy microgrids can perform energy scheduling and load management more efficiently, maximizing the utilization rate of photovoltaic power generation.

[0135] In renewable energy microgrids, load data can be affected by various factors (such as weather and seasonal changes), leading to data complexity and uncertainty. Regularization terms limit model complexity, prevent overfitting, and thus improve the model's generalization ability under different conditions. The load demand forecasting model in this embodiment is improved from the mean squared error loss function, used to measure the difference between predicted and actual values, by introducing a regularization term. To prevent overfitting of the load demand forecasting model, the generalization ability of the model is improved by penalizing excessively large parameter values.

[0136] The calculation formula for the load demand forecasting model is:

[0137] ;

[0138] ;

[0139] ;

[0140] ;

[0141] ;

[0142] ;

[0143] ;

[0144] in, for t Forecasted load demand at any given time. This is the weight matrix of the output layer of the Long Short-Term Memory (LSTM) unit. The bias of the output layer of the Long Short Time Memory (LSTM) unit. For Long Short-Term Memory Units in t The hidden state at all times , , , These are the forget gate, input gate, candidate memory unit, and output gate, respectively. t Activation value at time, The current state of the memory cell. for Sigmoid Activation function , , , These are the weight matrices for the forget gate, input gate, candidate memory units, and output gate, respectively. , , , These are the biases for the forget gate, input gate, candidate memory unit, and output gate, respectively. The hyperbolic tangent activation function is used. For Long Short-Term Memory Units in t The hidden state at time -1 For historical load data, Historical meteorological data, For candidate memory units in t The activation value at time -1.

[0145] Specifically, the load demand prediction model in this embodiment is based on an improved long short-term memory network model. The load demand prediction model includes an input layer, a long short-term memory network layer, a fully connected layer, and an output layer, which are connected sequentially.

[0146] The input layer is used to receive two input features, namely the initial historical load data and the initial historical meteorological data, and to preprocess the initial historical load data and the initial historical meteorological data.

[0147] The Long Short-Term Memory (LSTM) network layer consists of 64 neurons and is used to capture long-term dependencies in the time-series data of the historical load data. The LSM network layer includes a forget gate, an input gate, an output gate, a memory cell update, and a hidden state. The forget gate is used to determine and forget the unimportant parts of the historical load data. The input gate is used to determine the new information to be written to the historical load data. The memory cell update is used to update the state of the current memory cell. The output gate is used to determine which information to output from the historical load data.

[0148] The fully connected layer consists of 32 neurons, which are used to further process the features extracted by the long short-term memory network layer and generate a non-linear combination of features;

[0149] The output layer consists of one neuron, which outputs the predicted load demand value at future time points.

[0150] Step S2 includes:

[0151] The photovoltaic power generation forecasting model obtains the photovoltaic power generation at preset time intervals for each moment in the future period, plots the curve of photovoltaic power generation changing with time, and fits the curve to obtain the trend of photovoltaic power generation change. The load demand forecasting model obtains the load demand at preset time intervals for each moment in the future period, plots the curve of load demand changing with time, and fits the curve to obtain the trend of load demand change.

[0152] Harmonic analysis of the new energy microgrid is performed using the voltage and current waveform data. A Fast Fourier Transform (FFT) is applied to the voltage and current waveform data to convert it to the frequency domain, obtaining the fundamental RMS value and the RMS values ​​of the harmonic components. The total harmonic distortion rate is then calculated based on the fundamental and harmonic RMS values.

[0153] ;

[0154] in, THD Total harmonic distortion (THD) V 1 represents the fundamental effective value. V 2 represents the effective value of the second harmonic component. V 3 represents the effective value of the third harmonic component. Vn for n Effective value of the second harmonic component.

[0155] Specifically, this embodiment uses a photovoltaic power generation prediction model and a load demand prediction model to accurately predict photovoltaic power generation and load demand in future periods.

[0156] Plot the curves of photovoltaic power generation and load demand over time, fit the trend, and then fit the time-varying functions of photovoltaic power generation and load demand based on the trend.

[0157] By performing harmonic analysis on voltage and current waveform data and calculating the total harmonic distortion rate, the power quality of new energy microgrids can be effectively evaluated, ensuring the stable operation of the power system and the safe use of equipment.

[0158] Using the Fast Fourier Transform to convert time-domain data to the frequency domain allows for a more in-depth analysis of harmonic problems in power systems. Identifying specific harmonic components helps in developing targeted harmonic suppression strategies.

[0159] The system is monitored in real time by acquiring and analyzing data at preset time intervals.

[0160] The trends in photovoltaic power generation and load demand, along with the calculated total harmonic distortion rate, provide the necessary input characteristics for multi-objective optimal scheduling, thereby enabling precise scheduling and optimized control.

[0161] Through accurate forecasting and real-time monitoring, potential supply-demand imbalances or power quality problems can be identified in advance, allowing for preventative measures to be taken and improving system stability.

[0162] Accurate forecasting and analysis can help the system better balance supply and demand, improve the efficiency of renewable energy utilization, and reduce energy waste.

[0163] Harmonic analysis can not only assess power quality but also serve as a tool for fault diagnosis. Abnormal harmonic components can be used to indicate the presence of certain faults or anomalies in the system.

[0164] Step S3 includes:

[0165] The objective function of the multi-objective optimization algorithm is:

[0166] ;

[0167] ;

[0168] in, Let the objective function of the multi-objective optimization algorithm be... , , , These are the weighting coefficients of the objective sub-functions for the changing trends of photovoltaic power generation, energy storage capacity, load demand, and total harmonic distortion, respectively. Let be the objective sub-function representing the trend of photovoltaic power generation. The objective sub-function is the trend of load demand changes. The objective function is the energy storage capacity value. Let the total harmonic distortion rate be the objective sub-function. X This is the solution to the objective function. , , , These represent the trends in photovoltaic power generation, energy storage capacity, load demand, and total harmonic distortion rate corresponding to the solutions to the objective function.

[0169] The constraints of the multi-objective optimization algorithm are:

[0170] ;

[0171] in, for t Photovoltaic power generation of the new energy microgrid at time 3. for t The power supply capacity of the new energy microgrid at time 3. for t The system load demand of the new energy microgrid at time 3. for t The power generation capacity of the new energy microgrid at time 3. for t The power imbalance at time 3. for t Energy storage capacity at time 3 This represents the minimum power supply value for a new energy microgrid. This represents the maximum power generation value of the new energy microgrid. for t Total harmonic distortion rate of the new energy microgrid at time 3. This is the threshold for total harmonic distortion.

[0172] Specifically, this embodiment achieves comprehensive optimization of multiple key indicators of new energy microgrids by setting a multi-objective function that includes the trend of photovoltaic power generation, energy storage capacity, load demand, and total harmonic distortion.

[0173] Use weighting coefficients , , , Adjust the importance of photovoltaic power generation change trend, energy storage capacity value, load demand change trend and total harmonic distortion rate according to actual needs, so that the optimization process is more flexible and can adapt to different operating scenarios and priorities.

[0174] Through multi-objective optimization, the best balance can be found between improving the utilization rate of photovoltaic power generation, optimizing the use of energy storage, meeting load demand and controlling harmonic distortion, so as to achieve the overall optimal operation of the system.

[0175] The established constraints ensure that the system operates within a safe and reliable range. Specifically, power balance constraints guarantee supply and demand balance, energy storage capacity constraints ensure the rational use of the energy storage system, and total harmonic distortion (THD) constraints guarantee power quality. THD is used as one of the optimization objectives to improve the system's power quality and reduce the adverse effects of harmonics on equipment.

[0176] Step S3 also includes:

[0177] Multiple initial solutions are generated by the objective function and constraints of the multi-objective optimization algorithm. The multiple initial solutions satisfy the constraints. The multi-objective optimization algorithm minimizes the objective function based on the constraints in order to select the optimal solution as the multi-objective scheduling scheme.

[0178] The steps of the multi-objective optimization algorithm include:

[0179] Generate initial solution set Each solution in the initial solution set satisfies the constraints.

[0180] For each solution in the initial solution set, calculate the objective function value corresponding to each solution, and select the optimal solution and the worst solution. The objective function value corresponding to the optimal solution is the smallest, and the objective function value corresponding to the worst solution is the largest.

[0181] The update process for each solution in the initial solution set is calculated as follows:

[0182] ;

[0183] in, For the first solution in the initial solution set j One solution. for The updated solution, This is the optimal solution in the initial solution set. This is the worst solution in the initial solution set. and These are the optimal and worst bias coefficients;

[0184] Based on the constraints, all updated solutions are filtered to obtain the first set of updated solutions;

[0185] Repeat the solution update until the preset number of iterations is reached;

[0186] The optimal solution in the final updated solution set is output as a multi-objective scheduling scheme, which includes photovoltaic power generation control value, energy storage power control value, load demand control value and total harmonic distortion rate control value.

[0187] Specifically, in this embodiment, based on the global search of the Jaya algorithm, the Jaya algorithm is applied from single-objective optimization to the field of multi-objective optimization by adding a local search strategy and introducing an adaptive weight adjustment mechanism to dynamically adjust the weights according to the current performance of each objective.

[0188] By generating multiple initial solutions and optimizing them under constraints, the algorithm can explore multiple regions of the solution space, increasing the likelihood of finding the global optimum and avoiding getting trapped in local optima.

[0189] During the generation of the initial solution and the updating of the solution, it is always ensured that the solution satisfies the constraints, and that every step of the optimization process is carried out within the feasible solution space, thus ensuring the feasibility and practicality of the final solution.

[0190] This embodiment improves convergence speed and accuracy by iteratively updating the solution set and dynamically adjusting the direction and magnitude of feasible solutions to gradually approach the optimal solution. By introducing the optimal bias coefficient and the worst bias coefficient, the solution update strategy is improved, making the algorithm more directional and efficient in the search process.

[0191] The improved multi-objective optimization algorithm retains the simplicity and efficiency of the Jaya algorithm while enhancing its applicability and performance in complex multi-objective problems, thus realizing the optimized control of new energy microgrids.

[0192] Step S4 includes:

[0193] This paper describes a multi-objective scheduling scheme for managing the power generation and supply of a new energy microgrid. The scheme regulates the power generation and supply strategy of the microgrid, which includes power generation and supply status, photovoltaic power generation status, system load status, and harmonic suppression status. The scheme compares the current energy storage capacity of the microgrid with its control value to determine the power generation and supply status; compares the current photovoltaic power generation with its control value to determine the photovoltaic power generation status; compares the current load output with its load demand control value to determine the system load status; and compares the current total harmonic distortion (THD) rate with its control value to determine the harmonic suppression status.

[0194] Specifically, step S4 in this embodiment achieves refined management of the new energy microgrid by comparing the current state with the control value. Each state (power generation, photovoltaic power generation, system load, harmonic suppression) can be monitored and adjusted individually to ensure that the system operates in the optimal state.

[0195] By monitoring and adjusting various parameters of the microgrid in real time, it can quickly respond to changes in the external environment and internal demand, improve the dynamic adaptability of the system, regulate power generation strategies, optimize the utilization of photovoltaic power generation and energy storage resources, reduce energy waste, and improve energy efficiency.

[0196] By managing harmonic suppression, the harmonic distortion rate in the power grid is reduced, power quality is improved, and the stability and reliability of the system are enhanced. By comparing the load output with the demand control value, load management is optimized to ensure that load demand is effectively met, while avoiding overload or resource waste.

[0197] Specifically, the power generation strategy in this embodiment includes calculations based on the power generation status, photovoltaic power generation status, system load status, and harmonic suppression status. The calculation formulas for the power generation status, photovoltaic power generation status, system load status, and harmonic suppression status are as follows:

[0198] ;

[0199] ;

[0200] ;

[0201] ;

[0202] in, , , , These are respectively the power generation status, photovoltaic power generation status, system load status, and harmonic suppression status. , , , These are the control values ​​for energy storage capacity, photovoltaic power generation, load demand, and total harmonic distortion rate. , , , These are the current energy storage capacity, current photovoltaic power generation, current load output, and current total harmonic distortion rate, respectively.

[0203] This refers to the power generation status of a new energy microgrid. It is the power generation state that drives the new energy microgrid to supply power to the outside;

[0204] It refers to the photovoltaic power generation status that improves the photovoltaic power generation efficiency of new energy microgrids. This refers to the photovoltaic power generation status that reduces the photovoltaic power generation efficiency of new energy microgrids;

[0205] It is the system load status that improves the load output of new energy microgrids. This refers to the system load state that reduces the load output of the new energy microgrid;

[0206] It is a harmonic suppression state that maintains the total harmonic distortion rate of the new energy microgrid. It is a harmonic suppression state that reduces the total harmonic distortion rate of new energy microgrids.

[0207] By considering the power supply and generation status, photovoltaic power generation status, system load status, and harmonic suppression status, this embodiment realizes the integration of photovoltaic, energy storage, and utilization in the new energy microgrid, and rationally controls the power supply and generation of the new energy microgrid.

[0208] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A power generation and control method based on an integrated photovoltaic-storage-utilization microgrid, characterized in that, Includes the following steps: S1, acquire the initial real-time operation data of the new energy microgrid, preprocess the initial real-time operation data to obtain real-time operation data, the real-time operation data includes voltage and current waveform data, real-time photovoltaic power generation data, real-time meteorological data and real-time load data; S2, based on the photovoltaic power generation prediction model, real-time photovoltaic power generation data and real-time meteorological data, predict the trend of photovoltaic power generation change of the new energy microgrid, based on the load demand prediction model and real-time load data, predict the trend of load demand change of the new energy microgrid, and perform harmonic analysis on the new energy microgrid based on the voltage and current waveform data to obtain the total harmonic distortion rate. The process of constructing the photovoltaic power generation prediction model includes: An initial photovoltaic power generation prediction model is determined, and the calculation formula for the initial photovoltaic power generation prediction model is as follows: ； in, For the autoregressive part, a Let the order of the initial autoregressive term be . B For backward shift operators, For difference operators, c This is the initial difference processing order. for t Photovoltaic power generation at time 0 This is the seasonal variation coefficient. d For the difference intercept term, For meteorological data coefficient vectors, for t Meteorological data vector at time 0, For the moving average portion, b The order of the initial moving average term. for t White noise error term at time 0; Different combinations of initial autoregressive term order, initial moving average term order, and initial difference processing order are selected. The initial photovoltaic power generation prediction models under different combinations are compared using the information criterion algorithm for preliminary screening. The initial photovoltaic power generation prediction models after preliminary screening are evaluated to obtain the photovoltaic power generation prediction model. S3. Based on a multi-objective optimization algorithm, real-time running data, and predicted data, a multi-objective scheduling scheme is generated. The predicted data includes the trend of photovoltaic power generation, the trend of load demand, and the total harmonic distortion rate. The optimization objectives of the multi-objective scheduling scheme include the trend of photovoltaic power generation, the energy storage capacity, the trend of load demand, and the total harmonic distortion rate. S4 monitors the current energy storage capacity, photovoltaic power generation, load output, and total harmonic distortion rate of the new energy microgrid in real time, and adjusts the power supply and generation strategy of the new energy microgrid based on the multi-objective scheduling scheme, the current energy storage capacity, the current photovoltaic power generation, the current load output, and the current total harmonic distortion rate.

2. The power generation and control method based on an integrated photovoltaic-storage-utilization microgrid as described in claim 1, characterized in that, Step S1 includes: The initial real-time running data is preprocessed, including data cleaning, outlier removal, data formatting, and data standardization. The data cleaning includes filling in missing values ​​using polynomial interpolation. The removal of outliers includes identifying outliers using the quartile range method; The data formatting includes converting date and time in different formats into a unified format, converting different units of measurement into a unified unit of measurement, and converting all text data into a unified character encoding format; The data standardization includes transforming the data using Z-score standardization.

3. The power generation and control method based on an integrated photovoltaic-storage-utilization new energy microgrid as described in claim 1, characterized in that, The process of constructing the photovoltaic power generation prediction model is as follows: Historical photovoltaic power generation data and historical meteorological data are obtained, and the historical photovoltaic power generation data and historical meteorological data are cleaned to remove missing values ​​and outliers; Seasonal variation coefficients are extracted using the time-series decomposition method. Unit root tests are performed on the differences between historical photovoltaic power generation data and seasonal variation coefficients. When the series of differences is non-stationary, the historical photovoltaic power generation data and seasonal variation coefficients are differentially processed to obtain photovoltaic power generation difference components. The autocorrelation function graph and partial autocorrelation function graph of the photovoltaic power generation difference components are plotted. Based on the characteristics of the autocorrelation function graph and partial autocorrelation function graph, the initial set of autoregressive term orders and the initial set of moving average term orders are determined.

4. The power generation and control method based on an integrated photovoltaic-storage-utilization new energy microgrid as described in claim 3, characterized in that, The process of constructing the load demand forecasting model includes: Obtain initial historical load data and initial historical meteorological data, and normalize the initial historical load data and initial historical meteorological data to obtain historical load data and historical meteorological data; The time-series characteristics of historical load data are captured by long short-term memory units, and an initial load demand forecasting model is established by combining it with historical meteorological data. The model parameters of the initial load demand prediction model are updated by an adaptive optimization algorithm until the loss function converges, thus obtaining the trained load demand prediction model. The formula for the adaptive optimization algorithm is: ； ； in, For the initial load demand forecasting model in t Model parameters at time 1, For the initial load demand forecasting model in t Model parameters at time 1-1 For the initial load demand forecasting model in t Adaptive learning rate at time 1 The gradient of the loss function parameters. , , , The initial load demand forecasting model is respectively in t The weight matrices of the forget gate, input gate, candidate memory unit, and output gate at time 1. , , , The initial load demand forecasting model is respectively in t The biases of the forget gate, input gate, candidate memory unit, and output gate at time 1; The formula for calculating the loss function is: ； in, For the first t The loss function for two iterations N The total number of samples, For the first i Actual value of load demand. For the first i One load demand forecast value, The regularization coefficient is . For regularization terms, For the first t Model parameters of the initial load demand prediction model during the second iteration.

5. The power generation and control method based on an integrated photovoltaic-storage-utilization microgrid as described in claim 4, characterized in that, The calculation formula for the load demand forecasting model is as follows: ； ； ； ； ； ； ； in, for t Forecasted load demand at any given time. This is the weight matrix of the output layer of the Long Short-Term Memory (LSTM) unit. The bias of the output layer of the Long Short Time Memory (LSTM) unit. For Long Short-Term Memory Units in t The hidden state at all times , , , These are the forget gate, input gate, candidate memory unit, and output gate, respectively. t Activation value at time, The current state of the memory cell. for Sigmoid Activation function , , , These are the weight matrices for the forget gate, input gate, candidate memory units, and output gate, respectively. , , , These are the biases for the forget gate, input gate, candidate memory unit, and output gate, respectively. The hyperbolic tangent activation function is used. For Long Short-Term Memory Units in t The hidden state at time -1 For historical load data, Historical meteorological data, For candidate memory units in t The activation value at time -1.

6. The power generation and control method based on an integrated photovoltaic-storage-utilization microgrid as described in claim 5, characterized in that, Step S2 includes: The photovoltaic power generation forecasting model obtains the photovoltaic power generation at preset time intervals for each moment in the future period, plots the curve of photovoltaic power generation changing with time, and fits the curve to obtain the trend of photovoltaic power generation change. The load demand forecasting model obtains the load demand at preset time intervals for each moment in the future period, plots the curve of load demand changing with time, and fits the curve to obtain the trend of load demand change.

7. The power generation and control method based on an integrated photovoltaic-storage-utilization microgrid as described in claim 6, characterized in that, Step S2 also includes: Harmonic analysis of the new energy microgrid is performed using the voltage and current waveform data. A Fast Fourier Transform (FFT) is applied to the voltage and current waveform data to convert it to the frequency domain, obtaining the fundamental RMS value and the RMS values ​​of the harmonic components. The total harmonic distortion rate is then calculated based on the fundamental and harmonic RMS values. ； in, THD Total harmonic distortion (THD) V 1 represents the fundamental effective value. V 2 represents the effective value of the second harmonic component. V 3 represents the effective value of the third harmonic component. V n for n Effective value of the second harmonic component.

8. The power generation and control method based on an integrated photovoltaic-storage-utilization microgrid as described in claim 7, characterized in that, Step S3 includes: The objective function of the multi-objective optimization algorithm is: ； ； in, Let the objective function of the multi-objective optimization algorithm be... , , , These are the weighting coefficients of the objective sub-functions for the changing trends of photovoltaic power generation, energy storage capacity, load demand, and total harmonic distortion, respectively. Let be the objective sub-function representing the trend of photovoltaic power generation. The objective sub-function is the trend of load demand changes. The objective function is the energy storage capacity value. Let the total harmonic distortion rate be the objective sub-function. X This is the solution to the objective function. , , , These represent the trends in photovoltaic power generation, energy storage capacity, load demand, and total harmonic distortion rate corresponding to the solutions to the objective function. The constraints of the multi-objective optimization algorithm are: ； in, for t Photovoltaic power generation of the new energy microgrid at time 3. for t The power supply of the new energy microgrid at time 3. for t The system load demand of the new energy microgrid at time 3. for t The power generation capacity of the new energy microgrid at time 3. for t The power imbalance at time 3. for t Energy storage capacity at time 3 This represents the minimum power supply value for a new energy microgrid. This represents the maximum power generation value of the new energy microgrid. for t Total harmonic distortion rate of the new energy microgrid at time 3. This is the threshold for total harmonic distortion.

9. The power generation and control method based on an integrated photovoltaic-storage-utilization microgrid as described in claim 8, characterized in that, Step S3 also includes: Multiple initial solutions are generated by the objective function and constraints of the multi-objective optimization algorithm. The multiple initial solutions satisfy the constraints. The multi-objective optimization algorithm minimizes the objective function based on the constraints in order to select the optimal solution as the multi-objective scheduling scheme. The steps of the multi-objective optimization algorithm include: Generate initial solution set Each solution in the initial solution set satisfies the constraints. For each solution in the initial solution set, calculate the objective function value corresponding to each solution, and select the optimal solution and the worst solution. The objective function value corresponding to the optimal solution is the smallest, and the objective function value corresponding to the worst solution is the largest. The update process for each solution in the initial solution set is calculated as follows: ； in, For the first solution in the initial solution set j One solution. for The updated solution, This is the optimal solution in the initial solution set. This is the worst solution in the initial solution set. and These are the optimal and worst bias coefficients; Based on the constraints, all updated solutions are filtered to obtain the first set of updated solutions; Repeat the solution update until the preset number of iterations is reached; The optimal solution in the final updated solution set is output as a multi-objective scheduling scheme, which includes photovoltaic power generation control value, energy storage power control value, load demand control value and total harmonic distortion rate control value.

10. The power generation and control method based on an integrated photovoltaic-storage-utilization microgrid as described in claim 9, characterized in that, Step S4 includes: This paper describes a multi-objective scheduling scheme for managing the power generation and supply of a new energy microgrid. The scheme regulates the power generation and supply strategy of the microgrid, which includes power generation and supply status, photovoltaic power generation status, system load status, and harmonic suppression status. The scheme compares the current energy storage capacity of the microgrid with its control value to determine the power generation and supply status; compares the current photovoltaic power generation with its control value to determine the photovoltaic power generation status; compares the current load output with its load demand control value to determine the system load status; and compares the current total harmonic distortion (THD) rate with its control value to determine the harmonic suppression status.

Images