Energy management method for energy storage system based on demand response

Abstract

The application discloses a demand response-based energy management method for an energy storage system, relates to the field of electric energy management, and comprises the following steps: collecting and acquiring a characteristic data set and a power generation data set, performing correlation mapping, and obtaining an energy storage system characteristic-power generation data set; performing demand influence factor extraction on the energy storage system characteristic-power generation data set, determining a demand side influence factor set, performing cluster analysis on the energy storage system characteristic-power generation data set, and obtaining a plurality of energy storage system power generation data clusters; fitting power generation control according to the demand side influence factor set, and building an energy control multi-channel; monitoring demand side response data of a target energy storage system in real time, matching and analyzing by using the energy control multi-channel, determining power generation control parameters, and performing energy feedback control. The application solves the technical problems that the existing energy storage system energy management has poor control strategy adaptability and insufficient response accuracy, and achieves the technical effects of improving the adaptability and response accuracy of the control strategy.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of electric energy management, in particular to an energy management method for energy storage systems based on demand response. BACKGROUND

[0002] Energy management of energy storage systems plays an important role in improving grid stability, reducing electricity costs and promoting renewable energy consumption. At present, the main strategy in this field is to use fixed threshold control or single objective optimization algorithm for power generation scheduling. These methods fail to fully consider the dynamic characteristics of demand side response, resulting in insufficient control accuracy, response lag and limited energy efficiency optimization.

[0003] In the related art, the energy management of the energy storage system has the technical problems of poor adaptability of the control strategy and insufficient response accuracy. SUMMARY

[0004] The present application provides an energy management method for energy storage systems based on demand response, which collects and acquires an energy storage system characteristic data set and an energy storage system power generation data set, and performs correlation mapping on the energy storage system characteristic data set and the energy storage system power generation data set to obtain an energy storage system characteristic-power generation data set. The energy storage system characteristic-power generation data set is subjected to demand influencing factor extraction to determine a demand side influencing factor set, and at the same time, the energy storage system characteristic-power generation data set is subjected to clustering analysis to obtain a plurality of energy storage system power generation data clusters. The plurality of energy storage system power generation data clusters are subjected to power generation control fitting according to the demand side influencing factor set to build an energy storage system energy control multi-channel. The demand side response data of a target energy storage system is monitored in real time, and the demand side response data of the target energy storage system is matched and analyzed using the energy storage system energy control multi-channel to determine energy storage system power generation control parameters, and energy feedback control is performed based on the energy storage system power generation control parameters.

[0005] The present application provides an energy management method for energy storage systems based on demand response, which collects and acquires an energy storage system characteristic data set and an energy storage system power generation data set, and performs correlation mapping on the energy storage system characteristic data set and the energy storage system power generation data set to obtain an energy storage system characteristic-power generation data set. The energy storage system characteristic-power generation data set is subjected to demand influencing factor extraction to determine a demand side influencing factor set, and at the same time, the energy storage system characteristic-power generation data set is subjected to clustering analysis to obtain a plurality of energy storage system power generation data clusters. The plurality of energy storage system power generation data clusters are subjected to power generation control fitting according to the demand side influencing factor set to build an energy storage system energy control multi-channel. The demand side response data of a target energy storage system is monitored in real time, and the demand side response data of the target energy storage system is matched and analyzed using the energy storage system energy control multi-channel to determine energy storage system power generation control parameters, and energy feedback control is performed based on the energy storage system power generation control parameters.

[0006] In a possible implementation, the obtaining of the energy storage system characteristic-power generation dataset comprises: initializing a noise filter according to noise characteristic information of the energy storage system characteristic dataset and the energy storage system power generation dataset; filtering and denoising the energy storage system characteristic dataset and the energy storage system power generation dataset by using the noise filter to obtain an available energy storage system characteristic dataset and an available energy storage system power generation dataset; performing down-sampling on the available energy storage system characteristic dataset and the available energy storage system power generation dataset according to data application requirements to obtain a reduced-dimension energy storage system characteristic dataset and a reduced-dimension energy storage system power generation dataset; and performing data alignment and correlation mapping on the reduced-dimension energy storage system characteristic dataset and the reduced-dimension energy storage system power generation dataset according to time sequences to obtain the energy storage system characteristic-power generation dataset.

[0007] In a possible implementation, the determining of the demand side influence factor set comprises: constructing a demand side influence factor system, wherein the demand side influence factor system comprises a time factor, a meteorological factor, a power price factor, and a load factor; extracting correlation influence factors from the energy storage system characteristic-power generation dataset according to the demand side influence factor system to obtain a demand side correlation influence factor set; performing principal component analysis and importance scoring on the demand side correlation influence factor set based on the energy storage system characteristic-power generation dataset to obtain a correlation influence factor importance coefficient set; and screening the demand side correlation influence factor set based on the correlation influence factor importance coefficient set to determine the demand side influence factor set.

[0008] In a possible implementation, the screening of the demand side correlation influence factor set based on the correlation influence factor importance coefficient set to determine the demand side influence factor set comprises: performing hierarchical analysis and weight distribution on each factor type in the demand side influence factor system to generate an influence factor system weight factor; determining an influence factor type screening proportion according to the influence factor system weight factor; and performing proportional screening on the demand side correlation influence factor set based on the correlation influence factor importance coefficient set according to the influence factor type screening proportion to determine the demand side influence factor set.

[0009] In a possible implementation, the obtaining of the multiple energy storage system power generation data clusters comprises: performing feature extraction and dimension reduction processing on the energy storage system characteristic-power generation dataset to obtain an energy storage power generation key feature set; determining an energy storage power generation feature dimension according to the energy storage power generation key feature set; performing initial K-means clustering on the energy storage system characteristic-power generation dataset according to the energy storage power generation feature dimension to obtain an initial power generation data cluster; and performing silhouette coefficient evaluation and iteration strategy optimization on the initial power generation data cluster to obtain the multiple energy storage system power generation data clusters.

[0010] In a possible implementation, the energy storage system energy control multi-channel is built by performing the following processing: identifying power generation demand of the multi-energy storage system power generation data cluster according to the set of demand side influencing factors, to obtain a multi-energy storage system power generation sample cluster; setting an energy storage system energy control task according to an energy storage system energy management target; and fitting power generation control of the multi-energy storage system power generation sample cluster based on the energy storage system energy control task, to build the energy storage system energy control multi-channel.

[0011] In a possible implementation, the energy storage system energy control multi-channel is built by performing the following processing: fitting power generation control of the multi-energy storage system power generation sample cluster based on the energy storage system energy control task, to build the energy storage system energy control multi-channel.

[0012] In a possible implementation, the energy storage system energy control multi-channel is built by performing the following processing: selecting an energy storage task network structure according to the energy storage system energy control task; fitting power generation control of the energy storage task associated sample cluster data set respectively by using the energy storage task network structure, to generate a multi-energy storage cluster energy control branch channel set; and identifying the multi-energy storage cluster energy control branch channel set according to the multi-energy storage system power generation data cluster, to build the energy storage system energy control multi-channel.

[0013] In a possible implementation, the energy storage system energy control multi-channel is built by performing the following processing: fitting power generation control of the multi-energy storage system power generation sample cluster based on the energy storage system energy control task, to build the energy storage system energy control multi-channel.

[0014] In a possible implementation, the energy feedback control is performed based on the energy storage system power generation control parameter by performing the following processing: performing energy management monitoring on the target energy storage system based on the energy storage system power generation control parameter, to obtain an energy storage control feedback parameter; adjusting and correcting the energy storage system power generation control parameter based on the energy storage control feedback parameter by using a PID controller, and performing energy storage system energy management by using the corrected energy storage system power generation control parameter.

[0015] The energy management method of the energy storage system based on demand response provided in the application can first acquire a set of energy storage system characteristic data and a set of energy storage system power generation data, associate and map the set of energy storage system characteristic data and the set of energy storage system power generation data to obtain a set of energy storage system characteristic-power generation data, then extract demand influencing factors from the set of energy storage system characteristic-power generation data to determine a set of demand side influencing factors, perform clustering analysis on the set of energy storage system characteristic-power generation data to obtain a plurality of energy storage system power generation data clusters, fit power generation control according to the set of demand side influencing factors to the plurality of energy storage system power generation data clusters to build a multi-channel energy control of the energy storage system, and finally monitor demand side response data of a target energy storage system in real time, match and analyze the demand side response data of the target energy storage system by using the multi-channel energy control of the energy storage system, determine energy storage system power generation control parameters, and perform energy feedback control based on the energy storage system power generation control parameters. The technical effect of improving the adaptability and response accuracy of the control strategy is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings of the embodiments of the application will be briefly introduced below. The flowcharts are used in the present application to illustrate the operations performed by the method according to the embodiments of the application. It should be understood that the foregoing or the following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or at the same time according to needs. Meanwhile, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 The flowchart of the energy management method of the energy storage system based on demand response provided in the embodiments of the application is shown.

[0018] Figure 2 The flowchart of obtaining the set of energy storage system characteristic-power generation data in the energy management method of the energy storage system based on demand response provided in the embodiments of the application is shown.

[0019] Figure 3 The flowchart of determining the set of demand side influencing factors in the energy management method of the energy storage system based on demand response provided in the embodiments of the application is shown.

[0020] Figure 4 The flowchart of screening the set of demand side influencing factors in the energy management method of the energy storage system based on demand response provided in the embodiments of the application is shown. DETAILED DESCRIPTION

[0021] The above description is only a summary of the technical solutions of the present application. In order to make the technical means of the present application more clear, the present application can be implemented according to the content of the description, and in order to make the above and other purposes, characteristics and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described.

[0022] In order to make the purposes, technical solutions and advantages of the present application more clear, the following will combine the drawings to make a further detailed description of the present application. The described embodiments should not be regarded as a limitation of the present application. All other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0023] In the following description, "some embodiments" are described, which describe a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subset of all possible embodiments, and can be combined with each other without conflict. The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0024] The embodiments of the present application provide a demand response based energy management method for a storage system, as shown in the method, the method comprises the following steps. Figure 1 The method comprises the following steps.

[0025] In step S100, the storage system characteristic data set and the storage system power generation data set are collected and obtained, and the storage system characteristic data set and the storage system power generation data set are associated and mapped to obtain a storage system characteristic-power generation data set.

[0026] Specifically, the storage system characteristic data set contains various attribute data of the storage device, which can reflect the performance characteristics and operating characteristics of the storage device. The storage system power generation data set includes relevant data of the storage system in the power generation process, which is used to describe the power generation performance of the storage system under different conditions.

[0027] Various sensors are used to collect characteristic data (such as the capacity, self-discharge rate, charge and discharge efficiency, and cycle life of the energy storage system) and power generation data (such as power generation and power generation time) of the energy storage system. Data acquisition cards are used to collect and initially process this data. Then, through data association mapping, the energy storage system characteristic dataset and the energy storage system power generation dataset are associated to generate an energy storage system characteristic-power generation dataset.

[0028] For example, in an energy storage system containing multiple battery packs, current sensors, voltage sensors, temperature sensors, etc., installed on the battery packs can be used to collect characteristic data such as charging and discharging current, voltage, and temperature of the batteries in real time. At the same time, power sensors are used to collect power generation data of the battery packs. After these data are aggregated, correlation mapping is used to associate and bind the characteristic data of different battery packs with the power generation data at the same point in time, forming a complete energy storage system characteristic-power generation dataset.

[0029] like Figure 2 As shown, in one possible implementation, obtaining the energy storage system characteristic-power generation dataset, step S100 further includes step S110, which initializes a noise filter based on the noise characteristic information of the energy storage system characteristic dataset and the energy storage system power generation dataset. Specifically, the noise characteristic information of the energy storage system characteristic dataset and the power generation dataset is analyzed, such as noise type (Gaussian noise, impulse noise, etc.), noise intensity, frequency range, etc. Based on this information, a suitable filter type (such as a Kalman filter, wavelet transform filter, etc.) is selected and the filter parameters are initialized, such as setting the initial state of the filter, the noise covariance matrix, etc.

[0030] For example, if Gaussian noise exists in the power generation data of the energy storage system and the noise intensity is relatively low, a Kalman filter can be selected. Based on prior knowledge of the data, the initial state estimate of the Kalman filter is set to be consistent with the initial value of the actual data, and the noise covariance matrix can be initialized based on empirical values ​​of noise intensity.

[0031] Step S120: The noise filter is used to filter and denoise the energy storage system characteristic dataset and the energy storage system power generation dataset to obtain the available energy storage system characteristic dataset and the available energy storage system power generation dataset. Specifically, the initialized noise filter is applied to the energy storage system characteristic dataset and the power generation dataset. For each data sample, a filter recursion algorithm is used to filter and remove noise components from the data to obtain the available energy storage system characteristic dataset and the available energy storage system power generation dataset.

[0032] For example, Kalman filter is used to filter the power generation data of the energy storage system. At each time point, according to the state estimation of the previous moment and the measurement value at the current moment, the power generation estimation value at the current moment is calculated through the prediction update, measurement update and other steps of Kalman filtering, so as to remove the Gaussian noise in the original data and obtain more accurate power generation data.

[0033] In step S130, the available energy storage system characteristic data set and the available energy storage system power generation data set are down-sampled according to the data application requirements to obtain a reduced dimension energy storage system characteristic data set and a reduced dimension energy storage system power generation data set. Specifically, the available energy storage system characteristic data set and the available energy storage system power generation data set are down-sampled according to the set data application requirements, such as data storage space limitation, calculation complexity requirement, etc. The simple uniform sampling interval method or the non-uniform sampling method based on data characteristics can be used to reduce the data sampling rate to an appropriate level to obtain the reduced dimension energy storage system characteristic data set and the reduced dimension energy storage system power generation data set.

[0034] For example, if the sampling frequency of the original energy storage system characteristic data is 100 Hz, and the data application requirement requires a sampling frequency of 10 Hz, the uniform sampling interval method can be used to select a data point every 10 sampling points, so as to reduce the data sampling rate to 10 Hz and realize data dimension reduction.

[0035] In step S140, the reduced dimension energy storage system characteristic data set and the reduced dimension energy storage system power generation data set are aligned and associated according to the time sequence to obtain the energy storage system characteristic-power generation data set. Specifically, the reduced dimension energy storage system characteristic data set and the reduced dimension energy storage system power generation data set are aligned according to the time sequence, that is, it is ensured that each data sample in the two data sets corresponds to the same time point. Then, the correlation mapping algorithm such as principal component analysis (PCA), linear regression, etc. is used to establish the correlation between the two, and finally the energy storage system characteristic-power generation data set is obtained.

[0036] For example, for the reduced dimension energy storage system characteristic data (such as battery capacity, self-discharge rate, etc.) and power generation data (such as power generation power, power generation time, etc.), they are arranged according to the time sequence, so that the data samples at the same time point are aligned in space and time. Then, the linear regression algorithm is used to establish a linear relationship model between the energy storage system characteristic data as the independent variable and the power generation data as the dependent variable, to realize the correlation mapping and obtain the energy storage system characteristic-power generation data set.

[0037] This implementation method effectively removes the noise components in the data by filtering and denoising the original data, reduces the data amount by reducing the sampling rate of the data, reduces the data storage space requirement and calculation complexity, and improves the efficiency of subsequent data processing and analysis.

[0038] Step S200, the demand influencing factor extraction is performed on the energy storage system characteristic-power generation data set to determine the demand side influencing factor set, and clustering analysis is performed on the energy storage system characteristic-power generation data set to obtain multiple energy storage system power generation data clusters.

[0039] Specifically, natural language processing and feature extraction algorithms are used to analyze and process the data in the energy storage system characteristic-power generation data set, and extract the demand side related influencing factors (factors affecting the demand side electricity consumption behavior and demand) therein. These factors will directly or indirectly affect the size and trend of electricity load, such as time (electricity demand characteristics in different time periods), weather (temperature, light, humidity, etc. on electricity demand), user behavior (such as industrial user production period, residential user electricity habit, etc.), etc., forming a demand side influencing factor set. At the same time, clustering algorithms (such as K-means clustering algorithm, DBSCAN clustering algorithm, etc.) are used to perform clustering analysis on the energy storage system characteristic-power generation data set, and according to the similarity and difference of the data, the data is divided into multiple energy storage system power generation data clusters, each data cluster represents a type of situation with similar power generation characteristics and energy storage characteristics.

[0040] For example, for an energy storage system power generation data set containing different weather conditions and different time periods, the timestamp and weather information in the data are analyzed and extracted by natural language processing algorithm to obtain time (such as day, night, peak period, etc.) and weather (sunny, cloudy, rainy, etc.) as demand side influencing factors. Then use K-means clustering algorithm to cluster these data according to the fluctuation range of power generation power, the charging and discharging state change of energy storage equipment, etc. to obtain data clusters of high power generation power and frequent discharging of energy storage equipment in sunny day and peak electricity consumption period, and data clusters of low power generation power and charging of energy storage equipment in rainy night and low electricity consumption period, etc.

[0041] For example, Figure 3In a possible implementation, as shown, the step S200 of determining the set of demand-side influence factors further includes a step S210 of constructing a demand-side influence factor system, which includes time factors, meteorological factors, electricity price factors, and load factors. Specifically, the demand-side influence factor system is defined as follows: the time factors include different time periods in a day (such as peak hours, valley hours, etc.), weekdays and weekends, seasonal changes, and the like. For example, a day is divided into three time periods: peak hours (8:00-12:00, 18:00-22:00), flat hours (6:00-8:00, 12:00-18:00, 22:00-24:00), and valley hours (0:00-6:00), while distinguishing between weekdays and weekends, and between spring, summer, autumn, and winter; the meteorological factors include temperature, humidity, light intensity, wind speed, and other meteorological conditions. For example, the hourly temperature, humidity, solar radiation, and other data are recorded to reflect the influence of weather conditions on the energy storage system and electricity demand; the electricity price factors include time-of-use electricity prices, real-time electricity prices, and tiered electricity prices. For example, the time-of-use electricity price information released by the power grid company is obtained, including the electricity price rates in different time periods, and the real-time electricity price data adjusted according to market supply and demand changes; the load factors include different types of electricity loads (such as industrial loads, commercial loads, residential loads, etc.) and their change characteristics. For example, the load fluctuations caused by the production shifts of industrial users, the load peaks of commercial users during business hours, and the load changes caused by the life habits of residential users are analyzed.

[0042] In step S220, the set of demand-side associated influence factors is obtained by performing influence factor extraction on the set of energy storage system characteristics-power generation data according to the demand-side influence factor system. Specifically, data mining and association rule extraction algorithms, such as the Apriori algorithm or the FP-Growth algorithm in association rule mining, are used to analyze the set of energy storage system characteristics-power generation data according to the constructed demand-side influence factor system, to find out each influence factor related to the demand side, and to form the set of demand-side associated influence factors. For example, it is found through analysis that, in high-temperature weather (meteorological factor) and during peak electricity consumption periods (time factor), the load (load factor) of a specific industrial user will significantly increase, while the electricity price (electricity price factor) in this period is also relatively high, and these factors jointly affect the power generation and operation mode of the energy storage system, and are thus extracted into the set of demand-side associated influence factors.

[0043] Step S230, based on the energy storage system characteristics-power generation dataset, principal component analysis and importance score are performed on the demand side associated influence factor set to obtain an associated influence factor importance coefficient set. Specifically, the principal component analysis (PCA) algorithm is used to reduce the dimension of the demand side associated influence factor set, extract the main components, and calculate the importance score of each influence factor. Specifically, each factor in the demand side associated influence factor set is regarded as a variable, the variance contribution rate and cumulative variance contribution rate of each variable are calculated by PCA, the original influence factors corresponding to the main components are determined, and the importance score of each influence factor is calculated according to the variance contribution rate and other indicators to obtain the associated influence factor importance coefficient set. For example, it is found in the analysis that the time period division in the time factor and the temperature in the meteorological factor have higher variance contribution rates, which are the key factors affecting the power generation of the energy storage system, and their importance coefficients are relatively large.

[0044] Step S240, based on the associated influence factor importance coefficient set, the demand side associated influence factor set is screened to determine the demand side influence factor set. Specifically, according to the associated influence factor importance coefficient set, an importance threshold is set, for example, the influence factors with importance coefficients greater than 0.7 are selected, the demand side associated influence factor set is screened, and finally the demand side influence factor set is determined. In this way, the key factors that have a greater impact on the energy storage system characteristics-power generation data can be retained, and the secondary factors with less influence can be removed, so that the subsequent analysis and control are more targeted and efficient. For example, after screening, the time factor (time period), the meteorological factor (temperature), the electricity price factor (time-of-use electricity price), and the load factor (industrial load) are determined as the final demand side influence factor set.

[0045] This implementation can accurately screen out key factors that have a significant impact on the energy storage system characteristics-power generation data from numerous possible influence factors by constructing a comprehensive demand side influence factor system and performing associated influence factor extraction, principal component analysis, and importance scoring, thereby avoiding the waste of computing resources and the increase in complexity caused by blind analysis and processing of all factors.

[0046] As Figure 4As shown, in one possible implementation, the step S240 of screening the set of demand-side related influencing factors based on the set of importance coefficients of the related influencing factors further includes a step S241 of performing analytic hierarchy process and weight distribution on each factor type in the demand-side influencing factor system to generate a weight factor of the influencing factor system. Specifically, the analytic hierarchy process (AHP) is used to hierarchically divide each factor type (time factor, meteorological factor, electricity price factor, and load factor) in the demand-side influencing factor system, construct a hierarchical structure model, and include a target layer (determining the set of demand-side influencing factors), a criterion layer (each factor type), and an index layer (specific factors under each factor type, such as time period, working day / non-working day, and the like under the time factor). The relative importance judgment matrix between each factor is obtained through expert scoring, questionnaire survey, and the like, and then the weight of each factor type is calculated by using the calculation method of AHP, such as normalization processing, consistency check, and the like, to generate the weight factor of the influencing factor system. For example, it is calculated that the weight of the time factor is 0.4, the weight of the meteorological factor is 0.3, the weight of the electricity price factor is 0.2, and the weight of the load factor is 0.1.

[0047] For example, under the target of determining the set of demand-side influencing factors, the time factor, the meteorological factor, the electricity price factor, and the load factor are taken as the criterion layer. The judgment matrix is constructed by expert scoring of two-by-two comparison of each criterion layer factor, such as that the time factor is slightly more important than the meteorological factor (scoring 3), the time factor is equally important as the electricity price factor (scoring 1), the time factor is very important than the load factor (scoring 5), and the like, to construct a complete judgment matrix. Then, the judgment matrix is normalized and checked for consistency, and the weight factor of each factor is calculated, such as that the weight of the time factor is 0.4.

[0048] In step S242, the weight factor of the influencing factor system is used to determine the screening proportion of the factor type. Specifically, the weight factor of the influencing factor system is used to determine the screening proportion of each factor type in the screening process. The screening proportion refers to the share of each factor type in the set of demand-side related influencing factors, and is used to guide the screening operation. For example, according to the weight factor, the weight of the time factor is the largest, and the screening proportion of the time factor can be determined as 40% (corresponding to the weight 0.4), the screening proportion of the meteorological factor is 30% (corresponding to the weight 0.3), the screening proportion of the electricity price factor is 20% (corresponding to the weight 0.2), and the screening proportion of the load factor is 10% (corresponding to the weight 0.1).

[0049] For example, according to the weight factor of the influencing factor system calculated in step S241, the weight factor is directly converted into the screening proportion, that is, the screening proportion of the time factor is 40%, the screening proportion of the meteorological factor is 30%, the screening proportion of the electricity price factor is 20%, and the screening proportion of the load factor is 10%.

[0050] Step S243, the proportion of the demand side associated influence factor set is screened according to the importance coefficient set of the associated influence factor, and the demand side influence factor set is determined. Specifically, the proportion of the demand side associated influence factor set is screened according to the importance coefficient set of the associated influence factor and the screening proportion of each influence factor type. Specifically, for each influence factor type, the number or proportion of influence factors to be retained in this type is determined according to the screening proportion. For example, in the demand side associated influence factor set, there are 10 associated influence factors of time factors, and the top 4 (10*40%) time factors with high importance coefficients are retained according to the screening proportion of 40% and the importance coefficients from high to low; there are 8 associated influence factors of weather factors, and the top 3 (8*30%) weather factors with high importance coefficients are retained; similarly, the influence factors of price and load are screened, and finally the screened factors of each type are collected to determine the final demand side influence factor set.

[0051] For example, in the demand side associated influence factor set, there are 10 specific factors of time factors, which are ranked from high to low according to the importance coefficient of the associated influence factor, and the top 4 factors are retained according to the screening proportion of 40%; there are 8 weather factors, and the top 3 are retained; there are 5 price factors, and the top 1 (5*20%=1) is retained; there are 6 load factors, and the top 0.6 (6*10%=0.6) is retained, and the top 1 is retained after rounding. The final demand side influence factor set contains 4 (time) + 3 (weather) + 1 (price) + 1 (load) = 9 factors.

[0052] This implementation assigns weights to each factor type by the analytic hierarchy process, fully considers the differences in the importance of different factor types to the demand side, avoids the situation that some relatively important but low-weight factor types may be ignored when simply screening according to the importance coefficient, and makes the screened demand side influence factor set more comprehensive, reasonable, and more consistent with the complex influence relationship of the actual demand side.

[0053] In one possible implementation, the plurality of energy storage system power generation data clusters are obtained, and step S200 further includes step S250 of performing feature extraction and dimension reduction processing on the energy storage system characteristic-power generation data set to obtain an energy storage power generation key feature set. Specifically, feature extraction and dimension reduction techniques such as principal component analysis (PCA), linear discriminant analysis (LDA), etc. are used to process the energy storage system characteristic-power generation data set, extract key features that can represent the main characteristics of the data, and form the energy storage power generation key feature set. For example, for time series data and weather data in the energy storage system power generation data, PCA method is used to convert the original multi-dimensional data into a few principal components, which can retain most of the important information of the data, thereby realizing feature extraction and dimension reduction processing.

[0054] For example, assume that the original energy storage system characteristic-power generation dataset contains 10 feature dimensions, such as battery capacity, self-discharge rate, charge-discharge efficiency, cycle life, power generation power, power generation time, ambient temperature, light intensity, humidity, and wind speed. Through PCA analysis, it is found that the first 3 principal components can explain more than 90% of the variance of the data, and then the 3 principal components constitute the energy storage power generation key feature set.

[0055] Step S260, according to the energy storage power generation key feature set, determining the energy storage power generation feature dimension. Specifically, according to the energy storage power generation key feature set, by analyzing the variance contribution rate, cumulative variance contribution rate and other indicators of each principal component, the appropriate energy storage power generation feature dimension is determined, that is, the number of key features used for subsequent clustering analysis is determined. For example, the number of principal components when the cumulative variance contribution rate reaches a certain threshold (such as 85%-95%) is selected as the feature dimension.

[0056] For example, in the above PCA analysis, if the cumulative variance contribution rate of the first 2 principal components has reached 88%, and the variance contribution rate of the 3rd principal component is relatively low, and increasing the feature dimension has a greater impact on the computational complexity of subsequent clustering analysis, then the energy storage power generation feature dimension can be determined as 2.

[0057] Step S270, performing initial K-means clustering on the energy storage system characteristic-power generation dataset according to the energy storage power generation feature dimension to obtain initial power generation data clusters. Specifically, according to the determined energy storage power generation feature dimension, the initial K-means clustering is performed on the energy storage system characteristic-power generation dataset. First, randomly initialize K cluster centers, then calculate the distance of each data point to each cluster center, assign each data point to the cluster to which the nearest cluster center belongs, then recalculate the cluster center of each cluster, repeat the above process until the cluster center no longer changes or reaches the set iteration number, and obtain the initial power generation data clusters.

[0058] For example, assume that the energy storage power generation feature dimension is determined to be 2, and the energy storage system characteristic-power generation dataset is mapped to a two-dimensional space, and 3 cluster centers are randomly initialized. Calculate the Euclidean distance of each data point to the 3 cluster centers, and assign the data points to the cluster to which the nearest cluster center belongs. Recalculate the cluster center coordinates of each cluster, and perform data point assignment again. After several iterations, the cluster center no longer changes, and 3 initial power generation data clusters are obtained.

[0059] Step S280, profile coefficient evaluation and iterative strategy optimization are performed on the initial power generation data cluster to obtain the multi-energy storage system power generation data cluster. Specifically, the profile coefficient is used to evaluate the quality of the initial power generation data cluster, and the value of the profile coefficient ranges from -1 to 1. The larger the value, the better the clustering effect. According to the evaluation result of the profile coefficient, iterative strategy is used to optimize the initial power generation data cluster, such as adjusting the cluster center, changing the parameters of the clustering algorithm, etc., until the optimal multi-energy storage system power generation data cluster is obtained. For example, if the profile coefficient of the initial clustering is low, optimization can be performed by reselecting the initial clustering center, increasing the number of clustering iterations, etc., to improve the profile coefficient and improve the clustering effect.

[0060] For example, the profile coefficient of the initial power generation data cluster is calculated to be 0.5. In order to optimize the clustering effect, different initial clustering centers are used for multiple K-means clustering, and the profile coefficient is calculated each time. The clustering result with the highest profile coefficient is selected as the final multi-energy storage system power generation data cluster. It is assumed that the optimized profile coefficient is improved to 0.7.

[0061] This implementation determines the energy storage power generation feature dimension based on the set of key energy storage power generation features, ensuring that the main features and information of the data are retained while avoiding excessive feature dimensions that introduce noise and complexity. This allows the clustering analysis to more accurately reflect the true structure and regularity of the data, improving the reliability and interpretability of the clustering results.

[0062] Step S300, fitting power generation control according to the set of demand side influencing factors on the multi-energy storage system power generation data cluster, building energy storage system energy control multi-channel.

[0063] Specifically, power generation control fitting refers to establishing a power generation control model based on existing data and demand side influencing factors, and training the model through algorithms, so that the model can better adapt to the power generation control requirements of the energy storage system under different demand side influencing factors. Energy storage system energy control multi-channel is a multiple independent energy storage system power generation control channel established for different combinations of demand side influencing factors. Each channel can achieve energy control of the energy storage system under specific conditions.

[0064] Based on the set of demand side influencing factors, different power generation control models are established, and machine learning algorithms (such as neural networks, support vector machines, etc.) are used to train the models, so that they can fit the energy storage system power generation data cluster according to different demand side influencing factors, thereby building energy storage system energy control multi-channel. Each channel corresponds to a specific demand side influencing factor combination of the energy storage system power generation control mode, and through these channels, the energy storage system can be controlled for power generation.

[0065] For example, multiple power generation control models are established for different weather and time combinations of demand side influencing factors. For the combination of sunny day, daytime and peak electricity demand period, the model is trained by machine learning algorithm to enable it to fit the optimal power generation control mode of the energy storage system under this condition, such as letting the energy storage device preferentially release the stored electric energy to meet the peak demand when the power generation is low, and appropriately controlling the state of charge of the energy storage device when the power generation is high and can meet the electricity demand, to avoid overcharging. The channel corresponding to such a model is the energy storage system power generation control channel for sunny day, daytime and peak electricity demand period, and other combinations of channels can be established in the same way to form the energy storage system energy control multi-channel.

[0066] In a possible implementation, the step S300 of establishing the energy storage system energy control multi-channel further includes a step S310 of performing power generation demand identification on the multiple energy storage system power generation data clusters according to the set of demand side influencing factors, to obtain a multiple energy storage system power generation sample cluster. Specifically, according to the determined set of demand side influencing factors, each data cluster in the multiple energy storage system power generation data cluster is analyzed to determine its corresponding power generation demand characteristics, and the corresponding power generation demand identification is performed. For example, according to the time factor, weather factor, electricity price factor and load factor in the set of demand side influencing factors, a certain power generation data cluster is identified as a power generation data cluster under the "peak period, high temperature weather, high electricity price, high industrial load demand" scenario.

[0067] For example, for a cluster containing multiple energy storage system power generation data, it is found through analysis that the data in the cluster is mostly concentrated in the case of summer afternoon 14:00-16:00 (time factor), temperature higher than 35°C (weather factor), peak period high price (electricity price factor) and industrial user load at peak (load factor), and therefore it is identified as a "summer high temperature industrial peak load high electricity price power generation demand" data cluster.

[0068] The step S320 sets the energy storage system energy control task according to the energy management target of the energy storage system. Specifically, according to the energy management target of the energy storage system, such as improving the stability of the power grid, reducing the electricity cost, prolonging the service life of the energy storage device, maximizing the utilization rate of renewable energy, etc., the corresponding energy storage system energy control task is set. For example, for the "summer high temperature industrial peak load high electricity price power generation demand" scenario, the energy control task is set as: on the premise of meeting the industrial load demand, preferentially use the low-price electric energy stored in the energy storage system to discharge at the high-price period, and reasonably control the charging and discharging depth of the energy storage device, to reduce the electricity cost and prolong the service life of the device.

[0069] For example, for the identified "summer high-temperature industrial peak load high-price power generation demand" data cluster, setting the energy control task includes: fully charging the energy storage device during the low-price period (such as 23:00-5:00 the next day at night); during the 14:00-16:00 peak period, control the energy storage device to discharge at the maximum discharge power to meet the industrial load demand; ensure that the charge and discharge depth of the energy storage device does not exceed its safety threshold (such as charging not exceeding 90% and discharging not lower than 10%) to protect the device.

[0070] Step S330, based on the energy control task of the energy storage system, the multi-energy storage system power generation sample cluster is fitted for power generation control, and the energy storage system energy control multi-channel is built. Specifically, based on the set energy control task of the energy storage system, for each data cluster in the multi-energy storage system power generation sample cluster, a machine learning algorithm (such as neural network, support vector machine, etc.) or a control strategy optimization algorithm (such as dynamic programming, genetic algorithm, etc.) is used for power generation control fitting. Through training the model or optimizing the control strategy, the energy storage system can accurately manage the energy according to the preset energy control task under different power generation demand scenarios, so as to build the energy storage system energy control multi-channel. Each channel corresponds to a specific power generation demand scenario and the corresponding control task.

[0071] For example, for the "summer high-temperature industrial peak load high-price power generation demand" data cluster, a neural network algorithm is used for power generation control fitting. The charge and discharge power, charge and discharge time, etc. of the energy storage device are used as input variables, and the output targets are to meet the industrial load demand, reduce the electricity cost, and prolong the service life of the device, etc. The neural network model is trained. After a large amount of data training, the model can output the optimal charge and discharge control strategy according to the real-time power generation demand and energy storage state, thereby building the energy storage system energy control channel for this scenario.

[0072] This implementation can clearly identify the characteristics of power generation demand under different scenarios by identifying the power generation demand of the multi-energy storage system power generation data cluster, so that the setting of the energy control task of the energy storage system is more targeted. The close association of the control task with the specific power generation demand scenario can realize the accurate response of the energy storage system to different demand scenarios and improve the fine degree of energy management.

[0073] In a possible implementation, the step S330 of fitting the power generation control of the multiple energy storage system power generation sample clusters based on the energy storage system energy control task further includes a step S331 of sequentially associating the multiple energy storage system power generation sample clusters with the energy storage task based on the energy storage system energy control task, to obtain an energy storage task associated sample cluster data set. Specifically, based on the set energy storage system energy control task, the multiple energy storage system power generation sample clusters are sequentially analyzed, the power generation data in each power generation sample cluster is associated and bound with the corresponding energy control task, and an energy storage task associated sample cluster data set is formed. The association can be achieved by adding labels, establishing mapping relationships, and the like, to ensure that each power generation data sample is explicitly associated with a specific energy control task.

[0074] For example, for the "summer high-temperature industrial peak load high-price power generation demand" data cluster, the corresponding energy control task is "fully charge the energy storage device during the low-price period, control the energy storage device to discharge at the maximum discharge power during the 14:00-16:00 peak period, and ensure that the charge and discharge depth is within the safety threshold". Each power generation data sample in the data cluster is labeled with this energy control task label, thereby establishing an energy storage task associated sample cluster data set.

[0075] The step S332 of fitting the power generation control of the energy storage task associated sample cluster data set respectively, to build the energy storage system energy control multi-channel. Specifically, for each energy storage task associated sample cluster data set, a power generation control fitting algorithm (such as a linear regression, a decision tree, or the like, which is used to fit the control strategy in a machine learning algorithm) is used for power generation control fitting. Through the training model, the energy storage system can output appropriate charge and discharge control parameters (such as charge and discharge power, time, and the like) according to the feature information in the power generation data sample and according to the corresponding energy control task, thereby building multiple independent energy storage system energy control channels, each channel corresponding to a specific energy control task and a power generation scene.

[0076] For example, for the energy storage task associated sample cluster data set with the "summer high-temperature industrial peak load high-price power generation demand" energy control task label described above, a linear regression algorithm is used for power generation control fitting. The environmental temperature, industrial load size, current price, and the like in the power generation data are used as input variables, and the charge and discharge power, charge and discharge start time, and the like of the energy storage device are used as output variables, to train a linear regression model. After a large number of sample training, the model can output the charge and discharge control parameters that meet the requirements of the energy control task according to the real-time input power generation data features, and thereby build a corresponding energy storage system energy control channel.

[0077] This implementation mode combines the energy storage system energy control task with the power generation data by associating the power generation data of the multiple energy storage system power generation sample clusters, so that the control strategy of the energy storage system can be closely related to the specific energy control task. This can achieve precise control of the energy storage system under different power generation scenarios, ensuring that the energy storage system operates according to the established energy management objectives under various operating conditions.

[0078] In one possible implementation, the respective power generation control fitting of the energy storage task association sample cluster data set is performed to build a multi-channel energy storage system energy control, and step S332 further includes step S3321 of selecting an energy storage task network structure according to the energy storage system energy control task. Specifically, according to the characteristics and requirements of the energy storage system energy control task, a suitable energy storage task network structure is selected, including neural networks, deep learning networks, support vector machines, etc. For example, for complex nonlinear control tasks, a multilayer perceptron (MLP) neural network can be selected; for power generation control tasks with time series characteristics, a recurrent neural network (RNN) or a long short-term memory network (LSTM) can be selected.

[0079] For example, if the energy storage system energy control task is to control charging and discharging according to real-time electricity prices and load demand, and the power generation data has obvious time series characteristics, an LSTM network is selected as the energy storage task network structure because LSTM can effectively handle and predict long-term dependencies in time series data.

[0080] Step S3322 uses the energy storage task network structure to respectively perform power generation control fitting on the energy storage task association sample cluster data set to generate a multi-energy storage cluster energy control branch channel set. Specifically, the selected energy storage task network structure is used to train and fit the energy storage task association sample cluster data set. Through iterative training, the network learns the mapping relationship between the power generation data and the energy control task, thereby generating a multi-energy storage cluster energy control branch channel set. Each branch channel corresponds to a specific energy storage task association sample cluster data set and can independently perform power generation control.

[0081] For example, for the energy storage task association sample cluster data set with the "summer high-temperature industrial peak load high electricity price power generation demand" energy control task label, the selected LSTM network is used for training. During training, the time series characteristics in the power generation data (such as electricity prices, loads, and environmental temperatures in the past few hours) are used as inputs, and the charging and discharging power and time of the energy storage device are used as outputs. The weights and parameters of the network are continuously adjusted to make the network's prediction output as close as possible to the actual control task requirements. After sufficient training, the corresponding energy control branch channel is generated, which can automatically output the charging and discharging control parameters that meet the energy control task requirements according to the real-time input power generation data characteristics.

[0082] Step S3323, the multi-energy storage cluster energy control branch channel set is identified according to the multi-energy storage system power generation data cluster, and a multi-channel energy storage system energy control is built. Specifically, the generated multi-energy storage cluster energy control branch channel set is identified according to the multi-energy storage system power generation data cluster, that is, each branch channel is explicitly associated with the corresponding power generation data cluster, so that the corresponding control channel can be quickly called according to the power generation data in actual operation. Finally, a multi-channel energy storage system energy control is built to realize precise energy management of the energy storage system under different power generation scenarios.

[0083] For example, the energy control branch channel corresponding to the "summer high temperature industrial peak load high price power generation demand" data cluster is identified as channel A. In actual operation, when it is monitored that the power generation data of the energy storage system meets the characteristics of the data cluster, the system can quickly identify and call channel A, and control the charging and discharging of the energy storage system according to the control strategy in channel A to meet the energy management target in this scenario.

[0084] This implementation can customize highly matched control strategies for each power generation scenario by selecting appropriate network structures according to different energy storage system energy control tasks and fitting the sample cluster data set for each energy storage task, which can more accurately meet the energy management needs in different scenarios and improve the operation efficiency and performance of the energy storage system. The generation and identification of the multi-energy storage cluster energy control branch channel set enable the energy storage system to quickly switch and call the corresponding control channel for different power generation data clusters, realizing efficient energy management. In the face of complex and variable power generation environment and demand side conditions, the system can respond in time, optimize the charging and discharging process of the energy storage system, reduce energy loss, and improve energy utilization efficiency.

[0085] In one possible implementation, the step S3323 of identifying the multi-energy storage cluster energy control branch channel set according to the multi-energy storage system power generation data cluster and building a multi-channel energy storage system energy control further includes a step S33231 of classifying and identifying the multi-energy storage cluster energy control branch channel set according to the multi-energy storage system power generation data cluster to obtain a multi-energy storage cluster associated branch channel set. Specifically, the multi-energy storage cluster energy control branch channel set is classified and identified according to the multi-energy storage system power generation data cluster, that is, each branch channel is explicitly labeled with its corresponding power generation data cluster category to form a multi-energy storage cluster associated branch channel set. For example, the energy control branch channel corresponding to the "summer high temperature industrial peak load high price power generation demand" data cluster is identified as category A, and the energy control branch channel corresponding to the "night low valley load low price" data cluster is identified as category B. Through classification and identification, the system can clearly identify the specific power generation scenario to which each branch channel is directed.

[0086] For example, after training multiple energy control branch channels, the source of the energy storage task associated sample cluster data set corresponding to each branch channel is assigned a corresponding class label. For example, if a branch channel is trained based on the "spring weekday commercial load mid-tariff power generation demand" data cluster, it is identified as class C.

[0087] Step S33232, respectively, parallel integration of the multiple energy storage cluster associated branch channel set, build energy storage system energy control multi-channel, wherein the number of channels of the energy storage system energy control multi-channel corresponds one-to-one to the multiple energy storage system power generation data clusters. Specifically, the multiple energy storage cluster associated branch channel set is integrated in parallel, that is, multiple branch channels are connected in parallel to form an energy storage system energy control multi-channel. Each channel operates independently and does not interfere with each other, and the input end of each channel receives the corresponding power generation data and the output end outputs the corresponding energy control instruction. The number of channels of the final formed energy storage system energy control multi-channel corresponds one-to-one to the multiple energy storage system power generation data clusters, ensuring that the system can quickly and accurately control energy through the corresponding channel when facing different power generation scenarios.

[0088] This implementation integrates multiple channels in parallel to enable each energy control channel to operate independently, avoiding the impact of a single channel failure on the entire system. Even if an abnormality occurs in a channel, other channels can still work normally, ensuring the reliability and stability of the energy storage system and reducing the risk of energy management failure caused by system failure.

[0089] Step S400, real-time monitoring of the demand side response data of the target energy storage system, using the energy storage system energy control multi-channel to match and analyze the demand side response data of the target energy storage system, determine the energy storage system power generation control parameters, and perform energy feedback control based on the energy storage system power generation control parameters.

[0090] Specifically, the demand side response data includes real-time electricity load data, price signals, user electricity demand changes and other information, which reflects the demand and dynamic changes of the demand side for power resources. Real-time monitoring of the demand side response data of the target energy storage system is achieved using Internet of Things technology, such as real-time electricity load changes and price fluctuation signals, etc. These data are sent to the energy storage system energy control multi-channel through the data transmission module. In the multi-channel, a matching algorithm (such as a pattern matching algorithm) is used to analyze the real-time demand side response data, and the energy storage system power generation control parameters are determined according to the analysis results, such as the charging and discharging power of the energy storage device, the charging and discharging time, etc. Then, through the actuator (such as power regulation device, switch control device, etc.), the energy storage system is controlled based on these parameters to achieve dynamic response of the energy storage system to the demand side.

[0091] For example, in a distributed energy system comprising multiple energy storage devices, real-time demand-side response data such as changes in power consumption load and electricity price signals are obtained by smart meters and monitoring devices installed at user terminals, and these data are transmitted in real time to the energy control multi-channel of the energy storage system. When the power consumption load suddenly increases and the electricity price is at a peak period, the matching algorithm in the multi-channel matches the corresponding energy storage system power generation control parameters according to the set rules, such as increasing the discharge power of the energy storage device, so that the energy storage device quickly releases electric energy to meet the increase in power consumption demand, and at the same time, the energy storage device is controlled to charge at an appropriate charging power during a period of lower electricity price, so as to realize real-time energy feedback control of the energy storage system to the demand side.

[0092] In a possible implementation, the step S400 of performing energy feedback control based on the energy storage system power generation control parameters further includes a step S410 of performing energy management monitoring on the target energy storage system based on the energy storage system power generation control parameters to obtain energy storage control feedback parameters. Specifically, based on the energy storage system power generation control parameters, real-time energy management monitoring is performed on the target energy storage system by various sensors and monitoring devices. The monitoring content includes key indicators such as the state of charge, charging and discharging power, charging and discharging efficiency, system frequency, voltage, and other parameters related to power generation control of the energy storage device. After the monitoring data are collected and arranged, the energy storage control feedback parameters are formed, which are used for subsequent control adjustment.

[0093] For example, in actual operation, for a lithium ion battery energy storage system, the voltage, current, temperature and other parameters of the battery are monitored in real time by devices such as voltage sensors, current sensors and temperature sensors installed on the battery pack. At the same time, the real-time state of charge (SOC), charging and discharging efficiency and other key indicators of the battery are calculated based on the charging and discharging control parameters (such as the set charging and discharging power, time, etc.) of the energy storage system, as energy storage control feedback parameters.

[0094] In step S420, a PID controller is used to adjust and correct the energy storage system power generation control parameters based on the energy storage control feedback parameters, and the energy storage system is managed by using the corrected energy storage system power generation control parameters. Specifically, a PID controller (proportional-integral-derivative controller) is used to adjust and correct the energy storage system power generation control parameters. The energy storage control feedback parameters are used as the input of the PID controller, and the deviation between the set target value (such as the expected state of charge, charging and discharging power, etc.) and the actual feedback value is used as the error input of the PID controller. By adjusting the proportional coefficient (Kp), integral coefficient (Ki) and differential coefficient (Kd) of the PID controller, the corrected energy storage system power generation control parameters are calculated, and the energy storage system is managed according to these corrected parameters to realize precise control of the energy storage system power generation.

[0095] For example, assume the goal is to maintain the state of charge (SOC) of an energy storage battery at around 50%. At a certain moment, it is monitored that the actual SOC value is 45%, deviating from the target value by -5%. This deviation is input into the PID controller, which calculates the corrected charging and discharging power according to the pre-tuned Kp, Ki, and Kd parameters. For example, the PID controller can output a small charging power, causing the battery to charge at an appropriate rate and gradually restore the SOC to the vicinity of the target value. In this way, the charging and discharging process of the energy storage system is adjusted in real time, achieving precise energy management.

[0096] This implementation can effectively reduce fluctuations and oscillations in the charging and discharging process of the energy storage system by introducing a PID controller, enhancing the stability of the system. When the energy storage system faces interference factors such as load changes and fluctuations in power generation, it can maintain a relatively stable operating state, avoiding performance degradation or failure due to excessive parameter fluctuations.

[0097] The embodiments of the present application adopt parallel collection of energy storage system characteristic data sets and power generation data sets, generate a fused data set through association mapping, synchronously perform demand influencing factor extraction (obtain the influencing factor set) and clustering analysis (generate power generation data clusters), input both into the power generation control fitting module, construct a multi-channel energy control model, match and analyze the optimal power generation control parameters through real-time monitoring of demand side response data, and perform energy feedback control and other technical means, solving the technical problems of poor adaptability and insufficient response accuracy of the existing energy storage system energy management control strategy, and achieving the technical effects of improving the adaptability and response accuracy of the control strategy.

[0098] The above specific embodiments do not constitute a limitation on the scope of protection of the present application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present application should be included within the scope of protection of the present application. In some cases, the actions or steps described in the present application can be performed in a different order from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

Claims

1. A method for energy management of an energy storage system based on demand response, characterized in that, The method comprises: Collecting and obtaining energy storage system characteristic data set and energy storage system power generation data set, and performing correlation mapping on the energy storage system characteristic data set and the energy storage system power generation data set to obtain an energy storage system characteristic-power generation data set; Performing demand influencing factor extraction on the energy storage system characteristic-power generation data set to determine a demand side influencing factor set, and performing cluster analysis on the energy storage system characteristic-power generation data set to obtain a plurality of energy storage system power generation data clusters; Performing power generation control fitting on the plurality of energy storage system power generation data clusters according to the demand side influencing factor set to build an energy storage system energy control multi-channel; Real-time monitoring of the demand side response data of a target energy storage system, matching and analyzing the demand side response data of the target energy storage system by using the energy storage system energy control multi-channel, determining energy storage system power generation control parameters, and performing energy feedback control based on the energy storage system power generation control parameters; The building of the energy storage system energy control multi-channel comprises: Performing power generation demand identification on the plurality of energy storage system power generation data clusters according to the demand side influencing factor set to obtain a plurality of energy storage system power generation sample clusters; Setting an energy storage system energy control task according to an energy storage system energy management target; Performing power generation control fitting on the plurality of energy storage system power generation sample clusters based on the energy storage system energy control task to build an energy storage system energy control multi-channel; Wherein, the identification of the plurality of energy storage cluster energy control branch channel set according to the plurality of energy storage system power generation data clusters to build an energy storage system energy control multi-channel comprises: Classifying and identifying the plurality of energy storage cluster energy control branch channel set according to the plurality of energy storage system power generation data clusters to obtain a plurality of energy storage cluster associated branch channel sets; Parallelly integrating the plurality of energy storage cluster associated branch channel sets respectively to build an energy storage system energy control multi-channel, wherein the number of channels of the energy storage system energy control multi-channel corresponds to the plurality of energy storage system power generation data clusters one by one.

2. The demand response based energy storage system energy management method of claim 1, wherein, The obtaining of the energy storage system characteristic-power generation data set comprises: Initializing a noise filter according to noise characteristic information of the energy storage system characteristic data set and the energy storage system power generation data set; Filtering and denoising the energy storage system characteristic data set and the energy storage system power generation data set by using the noise filter to obtain an available energy storage system characteristic data set and an available energy storage system power generation data set; According to the data application requirement, the available energy storage system characteristic data set and the available energy storage system power generation data set are down-sampled to obtain a reduced dimension energy storage system characteristic data set and a reduced dimension energy storage system power generation data set; The reduced dimension energy storage system characteristic data set and the reduced dimension energy storage system power generation data set are data-aligned and correlated according to time sequence to obtain the energy storage system characteristic-power generation data set.

3. The demand response based energy storage system energy management method of claim 1, wherein, The determination of the demand side influencing factor set comprises: Building a demand side influencing factor system, the demand side influencing factor system comprising time factor, meteorological factor, electricity price factor and load factor; Performing correlation influencing factor extraction on the energy storage system characteristic-power generation data set according to the demand side influencing factor system to obtain a demand side correlation influencing factor set; Perform principal component analysis and importance scoring on the set of demand-side associated impact factors based on the set of energy storage system characteristic-power generation data, to obtain a set of associated impact factor importance coefficients; Screen the set of demand-side associated impact factors based on the set of associated impact factor importance coefficients, to determine a set of demand-side impact factors.

4. The demand response based energy storage system energy management method of claim 3, wherein, The screening of the set of demand-side associated impact factors based on the set of associated impact factor importance coefficients to determine a set of demand-side impact factors comprises: Perform hierarchical analysis and weight distribution on each factor type in the demand-side impact factor system to generate an impact factor system weight factor; Determine an impact factor type screening proportion based on the impact factor system weight factor; Perform proportional screening of the set of demand-side associated impact factors based on the set of associated impact factor importance coefficients according to the impact factor type screening proportion, to determine the set of demand-side impact factors.

5. The demand response based energy storage system energy management method of claim 1, wherein, The obtaining of the multiple energy storage system power generation data clusters comprises: Perform feature extraction and dimension reduction processing on the set of energy storage system characteristic-power generation data to obtain a set of energy storage power generation key features; Determine an energy storage power generation feature dimension based on the set of energy storage power generation key features; Perform initial K-means clustering on the set of energy storage system characteristic-power generation data according to the energy storage power generation feature dimension, to obtain initial power generation data clusters; Perform contour coefficient evaluation and iteration strategy optimization on the initial power generation data clusters to obtain the multiple energy storage system power generation data clusters.

6. The demand response based energy storage system energy management method of claim 1, wherein, The power generation control fitting of the multiple energy storage system power generation sample clusters based on the energy storage system energy control task to build an energy storage system energy control multi-channel comprises: Perform power generation data association on the multiple energy storage system power generation sample clusters based on the energy storage system energy control task in sequence, to obtain an energy storage task associated sample cluster data set; Perform power generation control fitting on the energy storage task associated sample cluster data set respectively to build an energy storage system energy control multi-channel.

7. The demand response based energy storage system energy management method of claim 6, wherein, The power generation control fitting of the energy storage task associated sample cluster data set respectively to build an energy storage system energy control multi-channel comprises: Select an energy storage task network structure according to the energy storage system energy control task; Perform power generation control fitting on the energy storage task associated sample cluster data set respectively using the energy storage task network structure to generate a set of multiple energy storage cluster energy control branch channels; Identify the set of multiple energy storage cluster energy control branch channels according to the multiple energy storage system power generation data clusters to build an energy storage system energy control multi-channel.

8. The demand response based energy storage system energy management method of claim 1, wherein, The energy feedback control based on the energy storage system power generation control parameters comprises: Perform energy management monitoring on the target energy storage system based on the energy storage system power generation control parameters to obtain energy storage control feedback parameters; Adjust and correct the energy storage system power generation control parameters based on the energy storage control feedback parameters using a PID controller, and perform energy storage system energy management through the corrected energy storage system power generation control parameters.

Images