Power supply management method and system based on new energy charging pile

Abstract

The invention discloses a power supply management method and system based on a new energy charging pile, and relates to the technical field of power supply management, and the method comprises the steps: collecting power supply source data, calculating a charging pile power supply risk, calculating a risk weight through an entropy weight method, and constructing a two-stage LSTM model to predict photovoltaic power generation and EV demand data. And determining the power supply priority of the charging pile based on the risk weight, constructing a dynamic energy distribution model, determining the power distribution of a power supply source by using a Pareto optimization algorithm, and carrying out charging power distribution by using a stable matching algorithm according to a utility function. According to the method, influence degrees of photovoltaic power generation fluctuation risks, energy storage performance risks and power grid price fluctuation risks are quantified through entropy calculation, importance of risk indexes is accurately reflected as weights, data optimization of an input layer of a double-stage LSTM model is guided, the risk indexes are brought into priority calculation, the power supply weight of high-risk energy is reduced, and the power supply efficiency is improved. And power supply fluctuation or cost increase caused by high risk is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of power supply management, and in particular to a power supply management method and system based on a new energy charging pile. Background Art

[0002] Renewable energy such as photovoltaic power generation has gradually been widely used due to its advantages of being green, environmentally friendly and sustainable. However, due to the intermittent and volatile nature of new energy, its large-scale access to the power supply network has brought many challenges to power supply management, especially in the field of electric vehicle charging piles. The deployment of new energy charging piles has significantly improved the green level of charging infrastructure, but it has also caused problems such as uneven power distribution, imbalance in supply and demand, and low charging efficiency.

[0003] In the existing technology, traditional power supply management methods usually rely on static allocation strategies or single optimization algorithms, which are difficult to fully cope with the dual uncertainties of fluctuations in renewable energy power generation and changes in EV charging demand. They lack comprehensive consideration of power supply priorities and dynamic weight adjustments. At the power supply distribution level, most methods are limited to fixed optimization models and fail to fully combine real-time risk assessment with multi-objective optimization strategies. Summary of the invention

[0004] In view of the problems existing in the above-mentioned existing power supply management method and system based on new energy charging piles, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is that traditional power supply management methods usually rely on static allocation strategies or a single optimization algorithm, which makes it difficult to fully cope with the dual uncertainties of fluctuations in renewable energy power generation and changes in EV charging demand, and lacks comprehensive consideration of power supply priority and dynamic weight adjustment. At the power supply distribution level, most methods are limited to fixed optimization models and fail to fully combine real-time risk assessment with multi-objective optimization strategies.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a power supply management method based on a new energy charging pile, which comprises:

[0007] Collect power supply source data, calculate the power supply risk of charging piles, calculate risk weights through entropy weight method, and build a two-stage LSTM model to predict photovoltaic power generation and EV demand data;

[0008] Determine the power supply priority of charging piles based on risk weights, build a dynamic energy allocation model, and use the Pareto optimization algorithm to determine the power allocation of power supply sources;

[0009] A stable matching algorithm is used according to the utility function to allocate charging power and record the charging pile status data for cloud upload.

[0010] As a preferred solution of the power supply management method based on new energy charging piles described in the present invention, wherein: the collected power supply source data includes:

[0011] The power supply sources of new energy charging piles include photovoltaic power generation systems, energy storage systems and power grid systems. The power output data of the photovoltaic power generation system is collected through photovoltaic inverters, and meteorological data, including direct normal radiation DNI, is obtained based on the meteorological service platform.

[0012] Collect the discharge efficiency of the energy storage system, collect the current cycle number to determine the remaining cycle number data, collect the actual capacity information to determine the attenuation according to the factory capacity;

[0013] Collect time-of-use electricity price data.

[0014] As a preferred solution of the power supply management method based on new energy charging piles described in the present invention, wherein: the power supply risk of the charging pile is calculated, the risk weight is calculated by the entropy weight method, and a two-stage LSTM model is constructed to predict photovoltaic power generation and EV demand data, including:

[0015] The power generation risk index is calculated based on the ratio of the maximum and minimum direct normal radiation values ​​of the collected data as the weather fluctuation coefficient and the normalized photovoltaic power output data.

[0016] Calculate the energy storage risk index based on the collected discharge efficiency, remaining cycle number data and attenuation of the energy storage system;

[0017] Based on the standard deviation and mean of the electricity purchase price of the collected time-of-use electricity price data, determine the grid load price fluctuation risk index;

[0018] A data matrix is ​​constructed based on the data of photovoltaic power generation volatility risk, energy storage equipment performance risk, and grid load price volatility risk, where each row represents the risk value data at different times, each column represents a different risk indicator, and each column of data is standardized;

[0019] The proportion of standardized data in each column is calculated based on the entropy weight method, and the entropy value is calculated, which is expressed as:

[0020]

[0021] Where pij represents the normalized ratio of the jth column and the ith row of the data matrix, zij represents the value of the jth column and the ith row of the data matrix, n represents the total number of samples collected in the data matrix, k represents the range constant, and ej represents the entropy value of the jth column in the data matrix;

[0022] Calculate the entropy value of each risk indicator, form an entropy value vector, and calculate the risk weight;

[0023] Constructing a two-stage LSTM model including a photovoltaic power generation prediction model and an EV demand prediction model;

[0024] The photovoltaic power generation prediction model includes an input layer, a hidden layer, and an output layer. The input layer inputs meteorological data weighted by the photovoltaic power generation volatility risk weight. The hidden layer retains important information and filters irrelevant noise through the LSTM memory unit to generate time series features related to photovoltaic power generation. The output layer outputs the time series features of the LSTM hidden layer.

[0025] The EV demand forecasting model includes an input layer, a hidden layer, and an output layer. The input layer inputs the EV demand data of new energy charging piles weighted by energy storage performance risk and grid price fluctuation risk. The hidden layer uses the memory unit of the LSTM network to extract important features in the time series. The output layer outputs specific forecast values.

[0026] The two-stage LSTM model is trained using the training set. The cross entropy loss function is selected to calculate the difference between the output and actual value of the photovoltaic power generation prediction model and the EV demand prediction model. The Adam optimizer is used for gradient descent optimization to update the weights of the photovoltaic power generation prediction model and the EV demand prediction model. If the model loss no longer decreases significantly during the continuous iteration process, the iteration output model parameters is stopped.

[0027] A two-stage LSTM model is used to output forecast data for PV power generation and EV demand.

[0028] As a preferred solution of the power supply management method based on new energy charging piles of the present invention, wherein: the power supply priority of the charging piles is determined based on the risk weight, including:

[0029] The available power of the energy storage system is calculated based on the product of the current discharge efficiency, the remaining available capacity and the maximum discharge rate of the energy storage system;

[0030] Based on the capacity limit of the current charging pile connected to the grid as the power available to the grid;

[0031] Based on the forecast data of photovoltaic power generation as the available power of photovoltaic power generation;

[0032] The EV charging demand of the new energy charging pile is taken as the total power demand of the power supply, and the sequence priority is calculated for photovoltaic power generation energy, energy storage energy and grid energy, which is expressed as:

[0033]

[0034] Where Si represents the priority score of the i-th energy source, wi represents the risk index of the i-th energy source, pi(t) represents the available power of the i-th energy source, and pz represents the total required power.

[0035] As a preferred solution of the power supply management method based on new energy charging piles described in the present invention, wherein: the dynamic energy allocation model is constructed, and the power allocation of the power supply source is determined using the Pareto optimization algorithm, including:

[0036] A dynamic energy allocation model is constructed, and the objective function is constructed based on minimizing the total power supply cost, maximizing the photovoltaic utilization rate, and minimizing the energy storage load fluctuation, which is expressed as:

[0037]

[0038] Where Cto represents the total power supply cost, Cgr represents the unit grid power purchase cost, Cess represents the discharge loss cost of the energy storage device, pgr(t) represents the grid power available at time t, Cgr represents the grid power purchase cost, ηpv represents the photovoltaic utilization rate, ppv(t) represents the photovoltaic power available at time t, ppv,max represents the maximum photovoltaic power available, Less represents the energy storage load smoothing value, pess(t) and pess(t-1) represent the energy storage system available at time t and time t-1 respectively;

[0039] Define constraints, including determining that the total power supply of the charging pile is the sum of the power available from the photovoltaic power plant, the power available from the grid, and the power available from the energy storage system. The state of charge of the energy storage system must be kept within a calibrated safety range.

[0040] Prioritize the available power of photovoltaic power, the available power of the grid and the available power of the energy storage system based on the priority scores;

[0041] Based on the mixed integer linear programming MILP optimizer, the available power of energy is gradually allocated in order of priority. If the available power of photovoltaic power generation is not fully utilized, the allocated photovoltaic available power is gradually increased, and the available power allocation of the power grid is reduced until the photovoltaic power generation is fully loaded;

[0042] If the state of charge of the energy storage system exceeds the calibrated safety range, the available power allocated by the energy storage system is reduced;

[0043] Based on each adjustment allocation, the target value is calculated based on the objective function, expressed as:

[0044] Z = Cto-ηpv+Less;

[0045] Where Z represents the allocation target value;

[0046] Use the Pareto optimization method to perform Pareto sorting on all calculated allocation target values. If at least one of the corresponding total power supply cost, photovoltaic utilization rate, and energy storage load smoothing value of the two compared allocation target values ​​is better, it is marked as a non-inferior solution, and the non-inferior solution is retained to form a non-inferior solution set;

[0047] Based on the priority score sorting, the solutions with the minimum total power supply cost, the highest photovoltaic utilization rate and the minimum energy storage load smoothing value are searched in the non-inferior solution set as the final solution, and the available power of grid generation, photovoltaic generation and energy storage system are determined.

[0048] As a preferred solution of the power supply management method based on the new energy charging pile of the present invention, wherein: the charging power allocation is performed using a stable matching algorithm according to the utility function, including:

[0049] Determine different charging pile alliances based on the charging pile clusters, and determine the total available power of the charging piles based on the available power of the grid power generation, photovoltaic power generation and energy storage system of different charging piles;

[0050] The EV user utility function and the charging pile alliance utility function are defined respectively and expressed as:

[0051]

[0052] Where Uev represents the user's satisfaction when charging at the current charging pile, Es represents the actual charging amount obtained by the user at the current moment, Ex represents the total charging amount required by the user, Tj represents the user's expected charging completion time, Ts represents the user's actual charging completion time, Upcp represents the current charging pile load efficiency relative to its maximum load, Pev represents the total load power provided by the current charging pile for EV, Pmax represents the maximum power supply capacity of the charging pile, and Cy represents the operating cost of the current charging pile;

[0053] Create a utility matrix based on the EV user utility function and the charging pile alliance utility function. The elements of the matrix are the sum of the utility function of the i-th EV user and the utility function of the j-th charging pile alliance, which is the total utility of user i and charging pile alliance j.

[0054] Using the stable matching algorithm Gale-Shapley, EV users are prioritized based on the size of total utility, and matching requests are made to the charging pile alliance with the largest utility for unmatched EV users according to priority.

[0055] The Charging Pile Alliance sorts all received requests from highest to lowest priority;

[0056] Based on the total available power of the charging pile alliance and the required power of the EV users, the EV user request is rejected, and the unmatched EV users are asked to send a request to the charging pile alliance of the next priority, and the above process is repeated;

[0057] After completing the matching of EV users, stop matching and record the matching results of total utility;

[0058] According to the matching results, the power supply is allocated according to the utility ratio of EV users, expressed as:

[0059]

[0060] Where Pi(t) represents the power allocated to EV user i at time t, P'i(t) represents the required power of EV user i at time t, Pz(t) represents the total available power at time t, Uev,i represents the utility of EV user i, Uev,j represents the utility of EV user j, and N represents the total number of EV users;

[0061] The charging power is dynamically allocated based on the power allocated to EV user i at time t.

[0062] As a preferred solution of the power supply management method based on new energy charging piles of the present invention, wherein: the recording of charging pile status data for cloud uploading includes:

[0063] Calculate the remaining available power of a single charging pile of the charging pile alliance in real time, monitor the charging status of each EV user at the current moment, and record the current load utilization of the charging pile;

[0064] Based on the report generation tool, the recorded data is filled in to generate data reports, and uploaded to the cloud server via the wireless network for information recording.

[0065] Another object of the present invention is to provide a system for power supply management method based on new energy charging piles, which comprises:

[0066] Data acquisition module, collects power supply source data,

[0067] Power supply risk assessment module, which calculates risk indicators and risk weights through entropy weight method;

[0068] The power generation prediction module builds a two-stage LSTM model to predict the photovoltaic power generation forecast value and the EV user demand forecast value;

[0069] The dynamic energy allocation module calculates the priority scores of photovoltaic, energy storage and power grid based on risk weights and available power, uses the Pareto optimization algorithm to determine the power allocation of power supply sources, and dynamically adjusts the power allocation of photovoltaic, energy storage and power grid;

[0070] The charging allocation module constructs a utility matrix and uses a stable matching algorithm to dynamically allocate charging power;

[0071] The data recording module records the status data of each charging pile, uploads it to the cloud server, and generates data reports.

[0072] A computer device comprises: a memory and a processor; the memory stores a computer program, and the processor implements the steps of the above-mentioned power supply management method based on new energy charging piles when executing the computer program.

[0073] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned power supply management method based on new energy charging piles.

[0074] The beneficial effects of the present invention are as follows: the impact of photovoltaic power generation volatility risk, energy storage performance risk and power grid price volatility risk is quantified through entropy calculation, the importance of risk indicators is accurately reflected as weights, the data optimization of the input layer of the two-stage LSTM model is guided, the risk indicators are incorporated into the priority calculation, the power supply weight of high-risk energy is reduced, and power supply fluctuations or cost increases due to high risks are avoided. Based on the combination of dynamic energy allocation model, MILP optimizer and Pareto optimization method, multi-objective optimization can be achieved in complex energy distribution scenarios, which not only effectively improves the economy, environmental protection and equipment reliability, but also enhances the flexibility of power supply strategy and the stability of system operation, provides a dynamic power allocation strategy combined with a stable matching algorithm, ensures supply and demand balance and efficient resource utilization, and enhances the robustness of system operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0076] Figure 1 The figure is a flow chart of a power supply management method based on new energy charging piles.

[0077] Figure 2 This is a structural diagram of the power supply management system based on new energy charging piles. DETAILED DESCRIPTION

[0078] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0079] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0080] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0081] Example 1, reference Figure 1 , which is the first embodiment of the present invention, and this embodiment provides a power supply management method based on a new energy charging pile. The power supply management method based on a new energy charging pile includes:

[0082] S1, collect power supply source data, calculate the power supply risk of charging piles, calculate the risk weight through entropy weight method, and build a two-stage LSTM model to predict photovoltaic power generation and EV demand data;

[0083] Preferably, collecting power supply source data includes:

[0084] The power supply sources of new energy charging piles include photovoltaic power generation systems, energy storage systems and power grid systems. The power output data of the photovoltaic power generation system is collected through photovoltaic inverters, and meteorological data, including direct normal radiation DNI, is obtained based on the meteorological service platform.

[0085] Collect the discharge efficiency of the energy storage system, collect the current cycle number to determine the remaining cycle number data, collect the actual capacity information to determine the attenuation according to the factory capacity;

[0086] Collect time-of-use electricity price data.

[0087] By collecting photovoltaic output power in real time and combining it with DNI data to predict future short-term power generation, power supply scheduling can be made more accurate, avoiding waste or insufficient load caused by excess or insufficient photovoltaic power generation. According to photovoltaic power and meteorological forecast results, the proportion of photovoltaics in the total power supply capacity of new energy charging piles can be adjusted, giving priority to the use of renewable energy, reducing dependence on energy storage and power grids, and estimating the future power supply capacity of the energy storage system through remaining capacity and health status, providing accurate input for the power supply distribution model.

[0088] Furthermore, the charging pile power supply risk is calculated, the risk weight is calculated by the entropy weight method, and a two-stage LSTM model is constructed to predict photovoltaic power generation and EV demand data, including:

[0089] Based on the ratio of the maximum and minimum direct normal radiation values ​​of the collected data as the weather fluctuation coefficient, combined with the normalized photovoltaic power output data, the power generation risk index is calculated and expressed as:

[0090]

[0091] Among them, Rpv represents the volatility risk of photovoltaic power generation, RSDpv represents the normalized standard deviation of photovoltaic power generation, and CFLpv represents the coefficient of fluctuation of weather conditions;

[0092] Based on the collected discharge efficiency, remaining cycle number data and attenuation of the energy storage system, the energy storage risk index is calculated and expressed as:

[0093]

[0094] Where Ress represents the performance risk of the energy storage device, ηess represents the discharge efficiency of the energy storage device, Less represents the remaining cycle life ratio, and ΔCess represents the capacity attenuation of the device. Indicates the maximum number of cycles of the device;

[0095] Based on the standard deviation and mean of the electricity purchase price of the collected time-of-use electricity price data, the grid load price fluctuation risk index is determined, which is expressed as:

[0096]

[0097] Where Rg represents the risk of grid load price fluctuation, σc represents the standard deviation of the electricity purchase price, and μc represents the mean of the electricity purchase price, which is 0;

[0098] A data matrix is ​​constructed based on the data of photovoltaic power generation volatility risk, energy storage equipment performance risk, and grid load price volatility risk, where each row represents the risk value data at different times, each column represents a different risk indicator, and each column of data is standardized;

[0099] Based on the entropy weight method, the proportion of each column of standardized data is calculated, and the entropy value is calculated, which is expressed as:

[0100]

[0101] Where pij represents the normalized ratio of the jth column and the ith row of the data matrix, zij represents the value of the jth column and the ith row of the data matrix, n represents the total number of samples collected in the data matrix, k represents the range constant, and ej represents the entropy value of the jth column in the data matrix;

[0102] Calculate the entropy value of each risk indicator, form an entropy vector, and calculate the risk weight, expressed as:

[0103]

[0104] Where wj represents the weight of the j-th column risk indicator;

[0105] Constructing a two-stage LSTM model including a photovoltaic power generation prediction model and an EV demand prediction model;

[0106] The photovoltaic power generation prediction model includes an input layer, a hidden layer, and an output layer. The input layer inputs meteorological data weighted by the photovoltaic power generation volatility risk weight. The hidden layer retains important information and filters irrelevant noise through the LSTM memory unit to generate time series features related to photovoltaic power generation. The output layer outputs the time series features of the LSTM hidden layer.

[0107] The EV demand forecasting model includes an input layer, a hidden layer, and an output layer. The input layer inputs the EV demand data of new energy charging piles weighted by energy storage performance risk and grid price fluctuation risk. The hidden layer uses the memory unit of the LSTM network to extract important features in the time series. The output layer outputs specific forecast values.

[0108] The two-stage LSTM model is trained using the training set. The cross entropy loss function is selected to calculate the difference between the output and actual value of the photovoltaic power generation prediction model and the EV demand prediction model. The Adam optimizer is used for gradient descent optimization to update the weights of the photovoltaic power generation prediction model and the EV demand prediction model. If the model loss no longer decreases significantly during the continuous iteration process, the iteration output model parameters is stopped.

[0109] A two-stage LSTM model is used to output forecast data for PV power generation and EV demand.

[0110] By quantifying the risk of photovoltaic power generation volatility, energy storage equipment performance risk and grid load price volatility risk and establishing a data matrix, the potential uncertainty of various power supply sources can be systematically analyzed. The weight of each type of risk is calculated by the entropy weight method, which avoids the limitations of subjective and artificial weight setting and realizes the objectivity of risk importance evaluation. By introducing meteorological data weighted by photovoltaic power generation volatility risk, the model can focus on weather factors (such as solar radiation, temperature, weather volatility, etc.) that are highly related to photovoltaic power generation, significantly improving the accuracy of power generation forecasting. The memory unit of the LSTM network is used to extract time series features, which can capture short-term fluctuations and long-term trends in photovoltaic power generation data. By introducing EV demand data weighted by energy storage performance risk and grid price volatility risk, the model can more accurately capture the key factors related to demand changes. Through the design of the two-stage model, it can simultaneously focus on the dynamic volatility of photovoltaic power generation and the load changes on the demand side, realizing the coordinated forecasting of supply and demand. The impact of photovoltaic power generation volatility risk, energy storage performance risk and grid price volatility risk is quantified by entropy value calculation, and the importance of risk indicators is accurately reflected as weights to guide the data optimization of the input layer of the two-stage LSTM model.

[0111] S2, determine the power supply priority of charging piles based on risk weights, build a dynamic energy allocation model, and use the Pareto optimization algorithm to determine the power allocation of power supply sources;

[0112] Preferably, determining the power supply priority of the charging pile based on the risk weight includes:

[0113] The available power of the energy storage system is calculated based on the product of the current discharge efficiency, the remaining available capacity and the maximum discharge rate of the energy storage system;

[0114] Based on the capacity limit of the current charging pile connected to the grid as the power available to the grid;

[0115] Based on the forecast data of photovoltaic power generation as the available power of photovoltaic power generation;

[0116] The EV charging demand of the new energy charging pile is taken as the total power demand of the power supply, and the sequence priority is calculated for photovoltaic power generation energy, energy storage energy and grid energy, which is expressed as:

[0117]

[0118] Where Si represents the priority score of the i-th energy source, wi represents the risk index of the i-th energy source (obtained based on the risk index of different columns of data in the data matrix), pi(t) represents the available power of the i-th energy source, and pz represents the total required power.

[0119] By calculating the priority score of each energy source based on photovoltaic power generation forecast data, the available power of the energy storage system and the capacity limit of the power grid, it is possible to achieve dynamic matching of resources on the power supply side, avoid excessive use or idle resources, give priority to the use of low-cost, renewable energy such as photovoltaic power generation, and reduce dependence on high-cost grid energy. At the same time, the proportion of power supply of each energy source can be adjusted in real time according to demand to improve the overall power supply efficiency of the system. In the energy priority sequence calculation, risk indicators are included in the priority calculation to reduce the power supply weight of high-risk energy (such as photovoltaic power generation with large fluctuations or energy storage systems with degraded performance) to avoid power supply fluctuations or cost increases due to high risks.

[0120] Furthermore, a dynamic energy allocation model is constructed, and the power allocation of power supply sources is determined using the Pareto optimization algorithm, including:

[0121] A dynamic energy allocation model is constructed, and the objective function is constructed based on minimizing the total power supply cost, maximizing the photovoltaic utilization rate, and minimizing the energy storage load fluctuation, which is expressed as:

[0122] minCto=Cess·pess(t)+Cgr·pgr(t);

[0123]

[0124] minLess=|pess(t)-pess(t-1)|;

[0125] Where Cto represents the total power supply cost, Cgr represents the unit grid power purchase cost, Cess represents the discharge loss cost of the energy storage device, pgr(t) represents the grid power available at time t, Cgr represents the grid power purchase cost, ηpv represents the photovoltaic utilization rate, ppv(t) represents the photovoltaic power available at time t, ppv,max represents the maximum photovoltaic power available, Less represents the energy storage load smoothing value, pess(t) and pess(t-1) represent the energy storage system available at time t and time t-1 respectively;

[0126] Define constraints, including determining that the total power supply of the charging pile is the sum of the power available from the photovoltaic power plant, the power available from the grid, and the power available from the energy storage system. The state of charge of the energy storage system must be kept within a calibrated safety range.

[0127] Prioritize the available power of photovoltaic power, the available power of the grid and the available power of the energy storage system based on the priority scores;

[0128] Based on the mixed integer linear programming MILP optimizer, the available power of energy is gradually allocated in order of priority. If the available power of photovoltaic power generation is not fully utilized, the allocated photovoltaic available power is gradually increased, and the available power allocation of the power grid is reduced until the photovoltaic power generation is fully loaded;

[0129] If the state of charge of the energy storage system exceeds the calibrated safety range, the available power allocated by the energy storage system is reduced;

[0130] Based on each adjustment allocation, the target value is calculated based on the objective function, expressed as:

[0131] Z = Cto-ηpv+Less;

[0132] Where Z represents the allocation target value;

[0133] Use the Pareto optimization method to perform Pareto sorting on all calculated allocation target values. If at least one of the corresponding total power supply cost, photovoltaic utilization rate, and energy storage load smoothing value of the two compared allocation target values ​​is better, it is marked as a non-inferior solution, and the non-inferior solution is retained to form a non-inferior solution set;

[0134] Based on the priority score sorting, the solutions with the minimum total power supply cost, the highest photovoltaic utilization rate and the minimum energy storage load smoothing value are searched in the non-inferior solution set as the final solution, and the available power of grid generation, photovoltaic generation and energy storage system are determined.

[0135] By dynamically adjusting the available power of photovoltaic power generation, grid power supply and energy storage system, it is ensured that the new energy charging piles meet the total power demand of EV users at different times. By combining the dynamic energy allocation model with priority sorting, the power supply strategy is made more flexible and effectively adapts to the volatility of photovoltaic power generation, the attenuation of energy storage performance and the fluctuation of grid prices. The low cost and zero carbon emission characteristics of photovoltaic power generation are given priority to reduce the demand for grid power purchase and reduce the system operation cost. When the available power of photovoltaic power is not fully utilized, the photovoltaic allocation power is gradually increased and the proportion of grid power supply is reduced to improve the utilization rate of photovoltaic power, which is conducive to maximizing the use of clean energy. By incorporating the unit grid power purchase cost and the discharge loss cost of energy storage equipment into the total power supply cost calculation, the energy allocation ratio is dynamically adjusted to avoid unnecessary high-cost grid power purchase. By smoothing the discharge power of the energy storage system, the load fluctuation of the energy storage equipment is reduced. Photovoltaic, grid and energy storage are sorted by priority score to provide a clear optimization order for energy allocation and reduce the calculation complexity. The priority score is combined with the weight of risk indicators to restrict high-risk energy (such as energy storage system in low charge state) during allocation to avoid system instability caused by high-risk use.

[0136] Through the MILP optimizer, the optimal combination that meets the constraints can be found in the distribution of photovoltaic, power grid and energy storage, ensuring the global optimality of the distribution plan. Through Pareto sorting, a set of non-inferior solutions with better performance can be quickly screened out, significantly reducing the calculation time in the optimization process. Based on the combination of dynamic energy distribution model, MILP optimizer and Pareto optimization method, multi-objective optimization can be achieved in complex energy distribution scenarios, which not only effectively improves the economy, environmental protection and equipment reliability, but also enhances the flexibility of power supply strategy and the stability of system operation.

[0137] S3, using a stable matching algorithm based on the utility function to allocate charging power and record charging pile status data for cloud upload;

[0138] Preferably, a stable matching algorithm is used according to the utility function to allocate charging power, including:

[0139] Determine different charging pile alliances based on the charging pile clusters, and determine the total available power of the charging piles based on the available power of the grid power generation, photovoltaic power generation and energy storage system of different charging piles;

[0140] The EV user utility function and the charging pile alliance utility function are defined respectively and expressed as:

[0141]

[0142] Where Uev represents the user's satisfaction when charging at the current charging pile, Es represents the actual charging amount obtained by the user at the current moment, Ex represents the total charging amount required by the user, Tj represents the user's expected charging completion time, Ts represents the user's actual charging completion time, Upcp represents the current charging pile load efficiency relative to its maximum load, Pev represents the total load power provided by the current charging pile for EV, Pmax represents the maximum power supply capacity of the charging pile, and Cy represents the operating cost of the current charging pile;

[0143] Create a utility matrix based on the EV user utility function and the charging pile alliance utility function. The elements of the matrix are the sum of the utility function of the i-th EV user and the utility function of the j-th charging pile alliance, which is the total utility of user i and charging pile alliance j.

[0144] Using the stable matching algorithm Gale-Shapley, EV users are prioritized based on the size of total utility, and matching requests are made to the charging pile alliance with the largest utility for unmatched EV users according to priority.

[0145] The Charging Pile Alliance sorts all received requests from highest to lowest priority;

[0146] Based on the total available power of the charging pile alliance and the required power of the EV users, the EV user request is rejected, and the unmatched EV users are asked to send a request to the charging pile alliance of the next priority, and the above process is repeated;

[0147] After completing the matching of EV users, stop matching and record the matching results of total utility;

[0148] According to the matching results, the power supply is allocated according to the utility ratio of EV users, expressed as:

[0149]

[0150] Where Pi(t) represents the power allocated to EV user i at time t, P'i(t) represents the required power of EV user i at time t, Pz(t) represents the total available power at time t, Uev,i represents the utility of EV user i, Uev,j represents the utility of EV user j, and N represents the total number of EV users;

[0151] The charging power is dynamically allocated based on the power allocated to EV user i at time t.

[0152] By calculating the total available power of the charging pile alliance according to the power supply capacity of photovoltaic, power grid and energy storage system, the full utilization of power supply resources can be ensured. The total available power of charging piles in the alliance can disperse the risk of insufficient resources of a single charging pile, improve the collaborative power supply capacity of the charging pile cluster, and avoid the decline of user experience due to single point failure or resource fluctuation. By introducing the utility matrix, the total utility of EV users and charging pile alliances is quantified, and the decision-making process of the power supply side and the demand side is unified. Through the use of Gale-Shapley algorithm, through the dynamic request mechanism to the next priority alliance, resource allocation has higher flexibility and adaptability, avoiding long-term idleness or repeated allocation of resources, and improving the stability of the system supply and demand balance. The dynamic allocation strategy effectively balances the loads of different charging piles, avoids overload or inefficient operation of a single charging pile, prolongs the service life of the equipment, ensures balanced load distribution of the charging pile cluster at different times, and improves the stability of the system. The dynamic power allocation strategy is combined with the Gale-Shapley stable matching algorithm to ensure supply and demand balance and efficient utilization of resources, and enhance the robustness of system operation.

[0153] Furthermore, the charging pile status data is recorded and uploaded to the cloud, including:

[0154] Calculate the remaining available power of a single charging pile of the charging pile alliance in real time, monitor the charging status of each EV user at the current moment, and record the current load utilization of the charging pile;

[0155] Based on the report generation tool, the recorded data is filled in to generate data reports, and uploaded to the cloud server via the wireless network for information recording.

[0156] By calculating the remaining available power of a single charging pile in real time, the charging pile alliance can instantly understand the resource usage of each charging pile and monitor the charging status of EV users (such as current charging amount and remaining demand) in real time, ensuring that users can understand the charging progress at any time and reduce waiting anxiety. By using report generation tools, the real-time calculated remaining available power, EV user charging status and load utilization are displayed in the form of tables or charts, allowing operators to quickly obtain the system operation status.

[0157] Example 2, reference Figure 2 , which is the second embodiment of the present invention, and which is different from the previous embodiment, provides a system for power supply management method based on new energy charging piles, including:

[0158] Data acquisition module, collects power supply source data,

[0159] Power supply risk assessment module, which calculates risk indicators and risk weights through entropy weight method;

[0160] The power generation prediction module builds a two-stage LSTM model to predict the photovoltaic power generation forecast value and the EV user demand forecast value;

[0161] The dynamic energy allocation module calculates the priority scores of photovoltaic, energy storage and power grid based on risk weights and available power, uses the Pareto optimization algorithm to determine the power allocation of power supply sources, and dynamically adjusts the power allocation of photovoltaic, energy storage and power grid;

[0162] The charging allocation module constructs a utility matrix and uses a stable matching algorithm to dynamically allocate charging power;

[0163] The data recording module records the status data of each charging pile, uploads it to the cloud server, and generates data reports.

[0164] Example 3

[0165] This is the third embodiment of the present invention. This embodiment is different from the first two embodiments. In order to verify the beneficial effects of the method of the present invention, the specific operation process of the present invention will be demonstrated through experiments.

[0166] Based on the collected data on the power supply sources of charging piles, various risk indicators of the power supply sources of charging piles are calculated. The data are as follows:

[0167] Time t Photovoltaic Fluctuation Risk Energy storage performance risks Grid fluctuation risk t1 0.15 0.10 0.12 t2 0.10 0.15 0.10 t3 0.20 0.20 0.15

[0168] The entropy value is calculated based on the entropy weight method, and the entropy value of each risk indicator is calculated;

[0169] Based on the weight data of photovoltaic risk, the two-stage LSTM model is used as input and outputs the forecast data of photovoltaic power generation and EV demand, and the sequence priority of photovoltaic power generation energy, energy storage energy and grid energy is calculated. The table is as follows:

[0170]

[0171] Based on the priority ranking and the actual available power, the power supply of photovoltaic, energy storage and grid energy is allocated on demand by optimizing the objective function, and the discharge loss cost of the energy storage equipment is determined to be 0.5, and the unit grid power purchase cost is 1;

[0172] Based on the objective function and constraints, the initial allocation is performed, and the table is as follows:

[0173]

[0174]

[0175] The target value at time point t1 is calculated as Zt1 = [29, -1.0, 6];

[0176] The target value at time point t2 is calculated as Zt2 = [39.25, -1.0, 10.5];

[0177] The target value at time point t3 is calculated as Zt3 = [67, -1.0, 5.5];

[0178] According to the definition of non-inferior solution, the non-inferior solution set is determined to be {Zt1, Zt3}, and based on the non-inferior solution set, the solution with the minimum total power supply cost, the highest photovoltaic utilization rate and the minimum energy storage load smoothing value is determined as the final solution. The table is as follows:

[0179] time Total power supply cost Photovoltaic utilization rate Energy storage load fluctuations Assignment Target t1 29 1 6.0 Minimum cost t2 37 1 5.5 Minimum volatility

[0180] Based on the final solution, the photovoltaic power, energy storage system power and grid power provided by t1 to the charging pile are determined respectively, and the power distribution of different charging is carried out based on the total supply power. The table is as follows:

[0181] time user Required power Distributing power t1 1 40 32.96 t1 2 50 31.96 t1 3 30 30

[0182] By substituting the actual power of the power supply source and the needs of EV users, the utility allocation formula is used to dynamically allocate power for each time period to ensure that the needs of each user are maximized within the total supply power range, while achieving a fair and efficient energy allocation strategy;

[0183] Power is allocated through the utility function and limited resources are allocated according to the proportion of user utility. This can ensure that different users receive fair charging power. Even if the total power supply is insufficient, high-utility users are given priority, while ensuring that each user obtains reasonable charging resources. In order to avoid exceeding the load limit of the energy storage device or the instability caused by fluctuations in photovoltaic power generation, the allocated power is limited by the power supply equipment capacity and optimization constraints. This mechanism contributes to the long-term stable operation of the system.

[0184] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0185] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0186] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0187] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0188] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A power supply management method based on a new energy charging pile, characterized in that: include, Collect power supply source data, calculate the power supply risk of charging piles, calculate risk weights through entropy weight method, and build a two-stage LSTM model to predict photovoltaic power generation and EV demand data; Determine the power supply priority of charging piles based on risk weights, build a dynamic energy allocation model, and use the Pareto optimization algorithm to determine the power allocation of power supply sources; A stable matching algorithm is used according to the utility function to allocate charging power and record the charging pile status data for cloud upload.

2. The power supply management method based on the new energy charging pile according to claim 1, characterized in that: The collecting of power supply source data includes: The power supply sources of new energy charging piles include photovoltaic power generation systems, energy storage systems and power grid systems. The power output data of the photovoltaic power generation system is collected through photovoltaic inverters, and meteorological data, including direct normal radiation DNI, is obtained based on the meteorological service platform. Collect the discharge efficiency of the energy storage system, collect the current cycle number to determine the remaining cycle number data, collect the actual capacity information to determine the attenuation according to the factory capacity; Collect time-of-use electricity price data.

3. The power supply management method based on the new energy charging pile according to claim 2, characterized in that: The charging pile power supply risk is calculated, the risk weight is calculated by the entropy weight method, and a two-stage LSTM model is constructed to predict photovoltaic power generation and EV demand data. include, The power generation risk index is calculated based on the ratio of the maximum and minimum direct normal radiation values ​​of the collected data as the weather fluctuation coefficient and the normalized photovoltaic power output data. Calculate the energy storage risk index based on the collected discharge efficiency, remaining cycle number data and attenuation of the energy storage system; Based on the standard deviation and mean of the electricity purchase price of the collected time-of-use electricity price data, determine the grid load price fluctuation risk index; A data matrix is ​​constructed based on the data of photovoltaic power generation volatility risk, energy storage equipment performance risk, and grid load price volatility risk, where each row represents the risk value data at different times, each column represents a different risk indicator, and each column of data is standardized; Based on the entropy weight method, the proportion of each column of standardized data is calculated, and the entropy value is calculated, which is expressed as: Where pij represents the normalized ratio of the jth column and the ith row of the data matrix, zij represents the value of the jth column and the ith row of the data matrix, n represents the total number of samples collected in the data matrix, k represents the range constant, and ej represents the entropy value of the jth column in the data matrix; Calculate the entropy value of each risk indicator, form an entropy value vector, and calculate the risk weight; Constructing a two-stage LSTM model including a photovoltaic power generation prediction model and an EV demand prediction model; The photovoltaic power generation prediction model includes an input layer, a hidden layer, and an output layer. The input layer inputs meteorological data weighted by the photovoltaic power generation volatility risk weight. The hidden layer retains important information and filters irrelevant noise through the LSTM memory unit to generate time series features related to photovoltaic power generation. The output layer outputs the time series features of the LSTM hidden layer. The EV demand forecasting model includes an input layer, a hidden layer, and an output layer. The input layer inputs the EV demand data of new energy charging piles weighted by energy storage performance risk and grid price fluctuation risk. The hidden layer uses the memory unit of the LSTM network to extract important features in the time series. The output layer outputs specific forecast values. The two-stage LSTM model is trained using the training set. The cross entropy loss function is selected to calculate the difference between the output and actual value of the photovoltaic power generation prediction model and the EV demand prediction model. The Adam optimizer is used for gradient descent optimization to update the weights of the photovoltaic power generation prediction model and the EV demand prediction model. If the model loss no longer decreases significantly during the continuous iteration process, the iteration output model parameters is stopped. A two-stage LSTM model is used to output forecast data for PV power generation and EV demand.

4. The power supply management method based on the new energy charging pile according to claim 3 is characterized in that: The determining of the power supply priority of the charging pile based on the risk weight includes: The available power of the energy storage system is calculated based on the product of the current discharge efficiency, the remaining available capacity and the maximum discharge rate of the energy storage system; Based on the capacity limit of the current charging pile connected to the grid as the grid available power; Based on the forecast data of photovoltaic power generation as the available power of photovoltaic power generation; The EV charging demand of new energy charging piles is taken as the total power demand of power supply, and the sequence priority is calculated for photovoltaic power generation energy, energy storage energy and grid energy, which is expressed as: Where Si represents the priority score of the i-th energy source, wi represents the risk index of the i-th energy source, pi(t) represents the available power of the i-th energy source, and pz represents the total required power.

5. The power supply management method based on the new energy charging pile according to claim 4, characterized in that: The dynamic energy allocation model is constructed and the power allocation of the power supply source is determined using the Pareto optimization algorithm, including: A dynamic energy allocation model is constructed, and the objective function is constructed based on minimizing the total power supply cost, maximizing the photovoltaic utilization rate, and minimizing the energy storage load fluctuation, which is expressed as: minCto=Cess·pess(t)+Cgr·pgr(t); minLess=|pess(t)-pess(t-1)|; Where Cto represents the total power supply cost, Cgr represents the unit grid power purchase cost, Cess represents the discharge loss cost of the energy storage device, pgr(t) represents the grid power available at time t, Cgr represents the grid power purchase cost, ηpv represents the photovoltaic utilization rate, ppv(t) represents the photovoltaic power available at time t, ppv,max represents the maximum photovoltaic power available, Less represents the energy storage load smoothing value, pess(t) and pess(t-1) represent the energy storage system available at time t and time t-1 respectively; Define constraints, including determining that the total power supply of the charging pile is the sum of the power available from the photovoltaic power plant, the power available from the grid, and the power available from the energy storage system. The state of charge of the energy storage system must be kept within a calibrated safety range. Prioritize the available power of photovoltaic power, the available power of the grid and the available power of the energy storage system based on the priority scores; Based on the mixed integer linear programming MILP optimizer, the available power of energy is gradually allocated in order of priority. If the available power of photovoltaic power generation is not fully utilized, the allocated photovoltaic available power is gradually increased, and the available power allocation of the power grid is reduced until the photovoltaic power generation is fully loaded; If the state of charge of the energy storage system exceeds the calibrated safety range, the available power allocated by the energy storage system is reduced; Based on each adjustment allocation, the target value is calculated based on the objective function, expressed as: Z = Cto-ηpv+Less; Where Z represents the allocation target value; Use the Pareto optimization method to perform Pareto sorting on all calculated allocation target values. If at least one of the corresponding total power supply cost, photovoltaic utilization rate, and energy storage load smoothing value of the two compared allocation target values ​​is better, it is marked as a non-inferior solution, and the non-inferior solution is retained to form a non-inferior solution set; Based on the priority score sorting, the solutions with the minimum total power supply cost, the highest photovoltaic utilization rate and the minimum energy storage load smoothing value are searched in the non-inferior solution set as the final solution, and the available power of grid generation, photovoltaic generation and energy storage system are determined.

6. The power supply management method based on the new energy charging pile according to claim 5, characterized in that: The method of using a stable matching algorithm according to the utility function to distribute charging power includes: Determine different charging pile alliances based on the charging pile clusters, and determine the total available power of the charging piles based on the available power of the grid power generation, photovoltaic power generation and energy storage system of different charging piles; The EV user utility function and the charging pile alliance utility function are defined respectively and expressed as: Where Uev represents the user's satisfaction when charging at the current charging pile, Es represents the actual charging amount obtained by the user at the current moment, Ex represents the total charging amount required by the user, Tj represents the user's expected charging completion time, Ts represents the user's actual charging completion time, Upcp represents the current charging pile load efficiency relative to its maximum load, Pev represents the total load power provided by the current charging pile for EV, Pmax represents the maximum power supply capacity of the charging pile, and Cy represents the operating cost of the current charging pile; Create a utility matrix based on the EV user utility function and the charging pile alliance utility function. The elements of the matrix are the sum of the utility function of the i-th EV user and the utility function of the j-th charging pile alliance, which is the total utility of user i and charging pile alliance j. Using the stable matching algorithm Gale-Shapley, EV users are prioritized based on the size of total utility, and matching requests are made to the charging pile alliance with the largest utility for unmatched EV users according to priority. The Charging Pile Alliance sorts all received requests from highest to lowest priority; Based on the total available power of the charging pile alliance and the required power of the EV users, the EV user request is rejected, and the unmatched EV users are asked to send a request to the charging pile alliance of the next priority, and the above process is repeated; After completing the matching of EV users, stop matching and record the matching results of total utility; According to the matching results, the power supply is allocated according to the utility ratio of EV users, expressed as: Where Pi(t) represents the power allocated to EV user i at time t, P'i(t) represents the required power of EV user i at time t, Pz(t) represents the total available power at time t, Uev,i represents the utility of EV user i, Uev,j represents the utility of EV user j, and N represents the total number of EV users; The charging power is dynamically allocated based on the power allocated to EV user i at time t.

7. The power supply management method based on the new energy charging pile according to claim 6, characterized in that: The recorded charging pile status data is uploaded to the cloud, include, Calculate the remaining available power of a single charging pile of the charging pile alliance in real time, monitor the charging status of each EV user at the current moment, and record the current load utilization of the charging pile; Based on the report generation tool, the recorded data is filled in to generate data reports, and uploaded to the cloud server via the wireless network for information recording.

8. A system based on the power supply management method based on the new energy charging pile according to any one of claims 1 to 7, characterized in that: include, Data acquisition module, collects power supply source data, Power supply risk assessment module, which calculates risk indicators and risk weights using the entropy weight method; The power generation prediction module builds a two-stage LSTM model to predict the photovoltaic power generation forecast value and the EV user demand forecast value; The dynamic energy allocation module calculates the priority scores of photovoltaic, energy storage and power grid based on risk weights and available power, uses the Pareto optimization algorithm to determine the power allocation of power supply sources, and dynamically adjusts the power allocation of photovoltaic, energy storage and power grid; The charging allocation module constructs a utility matrix and uses a stable matching algorithm to dynamically allocate charging power; The data recording module records the status data of each charging pile, uploads it to the cloud server, and generates data reports.

9. A computer device comprising: Memory and processor; The memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of the power supply management method based on the new energy charging pile described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the power supply management method based on a new energy charging pile described in any one of claims 1 to 7 are implemented.

Images