Photovoltaic power station energy storage optimal configuration method and system based on big data

Abstract

The invention provides a photovoltaic power station energy storage optimal configuration method and system based on big data, and relates to the technical field of energy storage optimal configuration, and the method comprises the steps: 1, collecting environmental factors, demands and historical data of energy storage configuration; 2, constructing a time sequence model, and predicting environment factors and demand data at future moments; 3, by taking the energy storage configuration data acquisition moment as a reference, aligning the time of the environmental factor data and the demand data; 4, constructing a regression model by taking environmental factors and demand historical data as independent variables and energy storage configuration data as dependent variables; and 5, analyzing the energy storage condition according to the predicted future moment data so as to optimize the energy storage configuration. According to the method, the situation that the energy storage configuration decision is lack of comprehensiveness due to insufficient data utilization in the prior art can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage optimization configuration, and in particular to a photovoltaic power station energy storage optimization configuration method and system based on big data. Background Art

[0002] In the photovoltaic power station system, the energy storage system is like a central link, effectively dealing with the intermittent and volatile problems of photovoltaic power generation, which is crucial to improving the quality of power and ensuring the stable operation of the power grid. Energy storage configuration is a complex task that requires comprehensive consideration of environmental factors such as light intensity and temperature, the time differences in power demand, and the battery characteristics and capacity status of the energy storage system.

[0003] The current photovoltaic power station energy storage configuration technology has obvious shortcomings, insufficient data utilization, and prominent problems. Most solutions rely only on limited basic data, and fail to fully collect and analyze key environmental factors such as cloud thickness and humidity that affect light and thus power generation, as well as detailed user electricity consumption behavior characteristics. At the same time, there is also a lack of sufficient attention and use of data such as the degree of battery aging of the energy storage system itself and performance changes under different working conditions. This makes energy storage configuration decisions lack comprehensiveness. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a photovoltaic power station energy storage optimization configuration method and system based on big data, so as to at least solve the above problems.

[0005] The technical solution adopted by the present invention is as follows: A first aspect of the present invention provides a method for optimizing energy storage configuration of a photovoltaic power station based on big data, the method comprising the following steps: Step 1: Collect historical data on environmental factors, demand, and energy storage configuration; In step 1, the environmental factor data refers to the natural environmental condition data that affects the power generation process of the photovoltaic power station and causes the power generation to fluctuate, expressed as: [E 1 ,E 2 ,…,E n ]; where n is a positive integer, indicating the type of environmental factor, E 1 ~E n They represent the first to n environmental factors respectively; the environmental factor data are collected by deploying sensors or obtained from the meteorological department; the historical data of environmental factors at different times have the same time interval Δt; the data are arranged in chronological order [t 1 ,t 2 ,…,t a ] to form the time series data of environmental factors; where a represents the number of moments, t 1 ~t a Indicates t 1 toa The first to a environmental factor data collection moments in a time period; Demand data refers to the data on the demand for electricity and the characteristics of electricity consumption behavior within a specific period of time, expressed as: 1 ,R 2 ,…,R m ]; where m is a positive integer, indicating the number of demand types, R 1 ~R m They represent the 1st to mth demand types respectively; the demand data is obtained through the electricity meter and the power grid company; the historical demand data at different times have the same time interval ΔT; the data are arranged in chronological order [T 1 ,T 2 ,…,T b ] to form demand time series data; where b represents the number of moments, T 1 ~T b Indicates T 1 To T b The first to b demand data collection moments in the time period; Energy storage configuration data refers to the relevant data such as the power storage status, charging and discharging dynamic process, and system health status of the energy storage system in the photovoltaic power station. The energy storage configuration data is directly related to the formulation of the energy storage optimization configuration strategy, which is expressed as: [D 1 ,D 2 ,…,D a ]; where a is a positive integer, indicating the type of energy storage configuration data, D 1 ~D a They represent the energy storage configuration data of types 1 to a respectively; the energy storage configuration data is obtained from the system monitoring platform and the equipment manufacturer; the energy storage configuration data at different times in history have the same time interval Δτ; the data are arranged in chronological order [τ 1 ,τ 2 ,…,τ c ] to form the energy storage configuration time series data; where c represents the number of moments, τ 1 ~τ c Indicates the 1st to cth energy storage configuration data collection time; Step 2: Build a time series model to predict environmental factors and demand data at future moments; Step 3: Using the energy storage configuration data collection time as a benchmark, align the environmental factor data and demand data time; Step 4: Construct a regression model with environmental factors and historical demand data as independent variables and energy storage configuration data as dependent variables; Step 5: Analyze the energy storage situation based on the predicted future time data to optimize the energy storage configuration.

[0006] Furthermore, in step 2, different time series models are constructed based on the different temporal regularities of the environmental factor data and the demand data; For environmental factor data: standardize the data with Z-score; use ADF test, set the stability threshold, perform a stability test on each environmental factor, and determine the number of differences d according to the non-stationary degree of the data; use the difference formula to differentiate the non-stationary series, and perform ADF test again after differentiation until the series is stable; calculate the autocorrelation function ACF and partial autocorrelation function PACF of the stable time series, and preliminarily determine the autoregressive order p and moving average order q based on the graphs of ACF and PACF; determine the optimal combination of p, d, q based on the minimization of the Bayesian Information Criterion BIC; use the maximum likelihood estimation method to evaluate the parameters of the model based on the determined model order (p, d, q); construct a VARIMA model based on the historical data of environmental factors, and use the model to predict data at future moments; For demand data: standardize the data by Z-score; use additive or multiplicative decomposition methods to decompose the demand time series into trend, seasonal and residual parts; build a TBATS model and use the historical demand time series data to train the TBATS model. During the training process, use the maximum likelihood estimation method to optimize the model parameters; use the model to predict future data.

[0007] Further, in step 3, the time of energy storage configuration data [τ 1 ,τ 2 ,…,τ c ] as a benchmark, and use the Lagrange interpolation method to obtain the environmental factor data and demand data at the corresponding time of the energy storage configuration data: For the energy storage configuration time τ k , in the time series of environmental factor data [t 1 ,t 2 ,…,t a ] Select the distance τ k The most recent d time points are expressed as: [t 1 ,t 2 ,…,t d ]; Calculate τ according to the Lagrange interpolation formula k Environmental factors E corresponding to the moment e data:

[0008] Where k represents the energy storage configuration data collection time sequence, k∈{1,2,…,c}, τ k ∈{τ 1 ,τ 2 ,…,τ c}; j, h represent the distance τk The sequence of the most recent time points, j, h∈{1,2,…,d}, t j ,t h ∈{t 1 ,t 2 ,…,t d}; e represents the type sequence of environmental factors, e∈{1,2,…,n}, E e ∈{E 1 ,E 2 ,…,E n}, E e (t j ) indicates t j Environmental factors E corresponding to the moment e data; For the energy storage configuration time τ k , in the time series of demand data [T 1 ,T 2 ,…,T b ] Select the distance τ k The most recent d time points are expressed as: [t 1 ,t 2 ,…,t d ]; Calculate τ according to the Lagrange interpolation formula k The demand data R corresponding to the moment f :

[0009] Where k represents the energy storage configuration data collection time sequence, k∈{1,2,…,c}, τ k ∈{τ 1 ,τ 2 ,…,τ c}; j, h represent the distance τ k The sequence of the most recent time points, j, h∈{1,2,…,d}, t j ,t h ∈{t 1 ,t 2 ,…,t d}; f represents the demand type sequence, f∈{1,2,…,m}, R f ∈{R 1 ,R 2 ,…,R m}, R f (t j ) indicates t j The demand type R corresponding to the moment f data.

[0010] Furthermore, in step 4, a regression model is constructed with the environmental factor data and demand data at the same time as independent variables and the energy storage configuration data as the dependent variable; for a certain type of energy storage configuration data, τ 1 ~τ c The environmental factor data and demand data corresponding to the time are used as independent variables, τ 1 ~τ c The energy storage configuration data corresponding to the time is taken as the dependent variable, and the SVR support vector regression model is constructed; the corresponding SVR support vector regression model is constructed for each energy storage configuration data.

[0011] Furthermore, in step 5, the environmental factor data and demand data at future moments are predicted according to the corresponding time series model; and the future collection time of the energy storage configuration data is calculated according to the time interval Δτ: [τ c +Δτ,τ c +2·Δτ,…,τ c +u·Δτ]; where u is a positive integer, indicating the number of energy storage configuration data collection times; τ is calculated using the Lagrange interpolation method c +Δτ~τ c +u·Δτ corresponding to the environmental factor data and demand data at the moment; calculate the energy storage configuration data corresponding to different moments according to the SVR regression model corresponding to different energy storage configuration data; formulate an energy storage configuration optimization strategy based on the predicted energy storage configuration data.

[0012] A second aspect of the present invention provides a photovoltaic power station energy storage optimization configuration system based on big data, the system comprising an environmental data module, a demand data module, a time alignment module and a prediction optimization module; The environmental data module is used to collect and process environmental factor data; the demand data module is used to collect and process demand data; the time alignment module is used to calculate the environmental factor data and demand data at the corresponding time based on the collection time of the energy storage data; the prediction and optimization module is used to construct a regression model and predict the energy storage configuration data based on the predicted environmental factor data and demand data to formulate an energy storage optimization strategy; The output end of the environment data module is connected to the input end of the time alignment module; the output end of the demand data module is connected to the input end of the time alignment module; the output end of the time alignment module is connected to the input end of the prediction optimization module.

[0013] Further, the environmental data module includes an environmental data unit, a time series model 1 unit and a data prediction 1 unit; The environmental data unit is used to collect and store environmental factor data at fixed time intervals; the time series model 1 unit is used to construct a VARIMA time series model based on the collected environmental factor data; the data prediction 1 unit is used to predict the environmental factor data at future moments based on the constructed time series model; The output end of the environment data unit is connected to the input end of the timing model 1 unit and the time alignment module; the output end of the timing model 1 unit is connected to the input end of the data prediction 1 unit; the output end of the data prediction 1 unit is connected to the input end of the time alignment module; The demand data module includes a demand data unit, a time series model 2 unit and a data prediction 2 unit; The demand data unit is used to collect and store demand data at fixed time intervals; the time series model 2 unit is used to construct a TBATA time series model based on the collected demand data; the data prediction 2 unit is used to predict demand data at future moments based on the constructed time series model; The output end of the demand data unit is connected to the input end of the timing model 2 unit and the time alignment module; the output end of the timing model 2 unit is connected to the input end of the data prediction 2 unit; the output end of the data prediction 2 unit is connected to the input end of the time alignment module.

[0014] Further, the time alignment module includes an energy storage data unit and an interpolation unit; The energy storage data unit is used to collect and store energy storage configuration data at fixed time intervals; the interpolation unit is used to use the Lagrange interpolation method to obtain the environmental factor data and demand data at the corresponding time based on the collection time of the energy storage data; The output end of the energy storage data unit is connected to the input end of the interpolation unit; the output end of the interpolation unit is connected to the input end of the prediction optimization module.

[0015] Furthermore, the prediction optimization module includes a regression model unit and an optimization strategy unit; The regression model unit is used to construct a regression model with environmental factors and historical demand data as independent variables and energy storage configuration data as dependent variables; the optimization strategy unit is used to analyze energy storage conditions according to predicted future time data and formulate a configuration optimization strategy to optimize energy storage configuration; The output end of the regression model unit is connected to the input end of the optimization strategy unit.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a photovoltaic power station energy storage optimization configuration system and method based on big data. A VARIMA model is constructed for environmental factor data, which can effectively capture its complex time series characteristics, such as trends, seasonality and cyclical changes; a TBATS model is constructed for demand data, which can accurately analyze its potential trends, seasonality and residual parts, greatly improving the accuracy of predictions of future environmental factors and demand data; this method uses the Lagrange interpolation method, takes the energy storage configuration data collection time as a benchmark, aligns the other two types of data, and ensures that the data used for model construction and strategy formulation are highly consistent in time. This lays a solid foundation for establishing an accurate regression model, improves the reliability of energy storage configuration decisions, and avoids unreasonable configurations caused by data dislocation; based on accurate prediction and optimization models, the present invention can adjust the energy storage configuration strategy in real time according to changes in the environment and demand. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only preferred embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0018] Figure 1 It is a schematic diagram of the overall flow of the method of an embodiment of the present invention; Figure 2 It is a schematic diagram of the overall structure of the system of an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The enumerated embodiments are only used to explain the present invention and are not used to limit the scope of the present invention.

[0020] Reference Figure 1 , Figure 2 The first aspect of the present invention provides a photovoltaic power station energy storage optimization configuration method based on big data, the method comprising the following steps: Step 1: Collect historical data on environmental factors, demand, and energy storage configuration; In step 1, the environmental factor data refers to the natural environmental condition data that affects the power generation process of the photovoltaic power station and causes the power generation to fluctuate, expressed as: [E 1 ,E 2 ,…,E n ]; where n is a positive integer, indicating the type of environmental factor, E 1 ~E n Represent the 1st to nth environmental factors respectively; Exemplarily, environmental factors include: light intensity: one of the main factors affecting the power generation of photovoltaic power stations; temperature: affecting the working efficiency of photovoltaic modules; wind speed: affecting the heat dissipation efficiency of photovoltaic power stations, thereby affecting the power generation.

[0021] Environmental factor data are collected by deploying sensors or obtained from meteorological departments; the historical data of environmental factors at different times have the same time interval Δt; the data are arranged in chronological order [t 1 ,t 2 ,…,t a ] to form the time series data of environmental factors; where a represents the number of moments, t 1 ~t a Indicates t 1 to a The first to a environmental factor data collection moments in a time period; Demand data refers to the data on the demand for electricity and the characteristics of electricity consumption behavior within a specific period of time, expressed as: 1 ,R 2 ,…,R m ]; where m is a positive integer, indicating the number of demand types, R 1 ~R m Respectively represent the 1st to mth demand types; For example, power load: the main manifestation of power demand; power consumption: the specific power demand; power consumption type: can be subdivided into different types of power demand such as residential, commercial, industrial, etc.

[0022] Demand data is obtained through electricity meters and power grid companies; the historical demand data at different times have the same time interval ΔT; the data is in chronological order [T 1 ,T 2 ,…,T b ] to form demand time series data; where b represents the number of moments, T 1 ~T b Indicates T 1 To T b The first to b demand data collection moments in the time period; Energy storage configuration data refers to the relevant data such as the power storage status, charging and discharging dynamic process, and system health status of the energy storage system in the photovoltaic power station. The energy storage configuration data is directly related to the formulation of the energy storage optimization configuration strategy, which is expressed as: [D 1 ,D 2 ,…,D a ]; where a is a positive integer, indicating the type of energy storage configuration data, D 1 ~D a Respectively represent the energy storage configuration data of types 1 to a; Exemplary, for example: battery energy storage status: the current charge status of the battery in the energy storage system; charging and discharging power: the power of charging or discharging the energy storage system; energy storage efficiency: the efficiency of the energy storage process.

[0023] Energy storage configuration data is obtained from the system monitoring platform and equipment manufacturers; the energy storage configuration data at different times in history have the same time interval Δτ; the data is in chronological order [τ 1 ,τ 2 ,…,τ c ] to form the energy storage configuration time series data; where c represents the number of moments, τ 1 ~τ c Indicates the time of collecting the 1st to cth energy storage configuration data.

[0024] Step 2: Build a time series model to predict environmental factors and demand data at future moments; In step 2, different time series models are constructed based on the different time regularities of environmental factor data and demand data; For environmental factor data: standardize the data with Z-score; use ADF test, set the stability threshold, perform a stability test on each environmental factor, and determine the number of differences d according to the non-stationary degree of the data; use the difference formula to differentiate the non-stationary series, and perform ADF test again after differentiation until the series is stable; calculate the autocorrelation function ACF and partial autocorrelation function PACF of the stable time series, and preliminarily determine the autoregressive order p and moving average order q based on the graphs of ACF and PACF; determine the optimal combination of p, d, q based on the minimization of the Bayesian Information Criterion BIC; use the maximum likelihood estimation method to evaluate the parameters of the model based on the determined model order (p, d, q); construct a VARIMA model based on the historical data of environmental factors, and use the model to predict data at future moments; For demand data: standardize the data by Z-score; use additive or multiplicative decomposition methods to decompose the demand time series into trend, seasonal and residual parts; build a TBATS model and use the historical demand time series data to train the TBATS model. During the training process, use the maximum likelihood estimation method to optimize the model parameters; use the model to predict future data.

[0025] Step 3: Using the energy storage configuration data collection time as a benchmark, align the environmental factor data and demand data time; In step 3, the time of energy storage configuration data [τ 1 ,τ 2 ,…,τ c ] as a benchmark, and use the Lagrange interpolation method to obtain the environmental factor data and demand data at the corresponding time of the energy storage configuration data: For the energy storage configuration time τ k , in the time series of environmental factor data [t 1 ,t 2 ,…,t a ] Select the distance τ k The most recent d time points are expressed as: [t 1 ,t 2 ,…,t d ]; Calculate τ according to the Lagrange interpolation formula k Environmental factors E corresponding to the moment e data:

[0026] Where k represents the energy storage configuration data collection time sequence, k∈{1,2,…,c}, τ k ∈{τ 1 ,τ 2 ,…,τ c}; j, h represent the distance τ k The sequence of the most recent time points, j, h∈{1,2,…,d}, t j ,t h ∈{t 1 ,t 2 ,…,t d}; e represents the type sequence of environmental factors, e∈{1,2,…,n}, E e ∈{E 1 ,E 2 ,…,E n}, E e (t j ) indicates t j Environmental factors E corresponding to the moment e data; For the energy storage configuration time τ k , in the time series of demand data [T 1 ,T 2 ,…,T b ] Select the distance τ k The most recent d time points are expressed as: [t 1 ,t 2 ,…,t d ]; Calculate τ according to the Lagrange interpolation formula k The demand data R corresponding to the moment f :

[0027] Where k represents the energy storage configuration data collection time sequence, k∈{1,2,…,c}, τ k ∈{τ 1 ,τ 2,…,τ c}; j, h represent the distance τ k The sequence of the most recent time points, j, h∈{1,2,…,d}, t j ,t h ∈{t 1 ,t 2 ,…,t d}; f represents the demand type sequence, f∈{1,2,…,m}, R f ∈{R 1 ,R 2 ,…,R m}, R f (t j ) indicates t j The demand type R corresponding to the moment f data.

[0028] Exemplarily, τ is calculated according to the Lagrange interpolation formula 1 ~τ c Environmental factor data and demand data corresponding to each moment.

[0029] Step 4: Construct a regression model with environmental factors and historical demand data as independent variables and energy storage configuration data as dependent variables; In step 4, a regression model is constructed with the environmental factor data and demand data at the same time as independent variables and the energy storage configuration data as the dependent variable; for a certain type of energy storage configuration data, τ 1 ~τ c The environmental factor data and demand data corresponding to the time are used as independent variables, τ 1 ~τ c The energy storage configuration data corresponding to the time is taken as the dependent variable, and the SVR support vector regression model is constructed; the corresponding SVR support vector regression model is constructed for each energy storage configuration data.

[0030] Step 5: Analyze the energy storage situation based on the predicted future time data to optimize the energy storage configuration.

[0031] In step 5, the environmental factor data and demand data at future moments are predicted according to the corresponding time series model; the future collection time of the energy storage configuration data is calculated according to the time interval Δτ: [τ c +Δτ,τ c +2·Δτ,…,τ c +u·Δτ]; where u is a positive integer, indicating the number of energy storage configuration data collection times; τ is calculated using the Lagrange interpolation method c +Δτ~τ c+u·Δτ corresponding to the environmental factor data and demand data at the moment; calculate the energy storage configuration data corresponding to different moments according to the SVR regression model corresponding to different energy storage configuration data; formulate an energy storage configuration optimization strategy based on the predicted energy storage configuration data.

[0032] Exemplarily, a reasonable energy storage configuration strategy is formulated through an optimization algorithm based on the predicted energy storage configuration data. First, the load fluctuation of the power grid and the fluctuation of photovoltaic power generation are evaluated in combination with environmental factors and demand forecast results to determine the charging and discharging requirements of the energy storage system. Then, the energy storage system can be discharged in time during high-demand periods and charged when photovoltaic power generation is in excess. In addition, considering limitations such as battery temperature and efficiency, the service life and performance of the energy storage equipment are optimized. Finally, the optimal strategy is found through mathematical models (such as linear programming, genetic algorithms, etc.) to reduce power waste, improve grid stability, and maximize the power generation utilization of photovoltaic power stations.

[0033] A second aspect of the present application provides a photovoltaic power station energy storage optimization configuration system based on big data, the system comprising an environmental data module, a demand data module, a time alignment module and a prediction optimization module; The environmental data module is used to collect and process environmental factor data; the demand data module is used to collect and process demand data; the time alignment module is used to calculate the environmental factor data and demand data at the corresponding time based on the collection time of the energy storage data; the prediction and optimization module is used to construct a regression model and predict the energy storage configuration data based on the predicted environmental factor data and demand data to formulate an energy storage optimization strategy; The output end of the environment data module is connected to the input end of the time alignment module; the output end of the demand data module is connected to the input end of the time alignment module; the output end of the time alignment module is connected to the input end of the prediction optimization module.

[0034] The environmental data module includes an environmental data unit, a time series model 1 unit and a data prediction 1 unit; The environmental data unit is used to collect and store environmental factor data at fixed time intervals; the time series model 1 unit is used to construct a VARIMA time series model based on the collected environmental factor data; the data prediction 1 unit is used to predict the environmental factor data at future moments based on the constructed time series model; The output end of the environment data unit is connected to the input end of the timing model 1 unit and the time alignment module; the output end of the timing model 1 unit is connected to the input end of the data prediction 1 unit; the output end of the data prediction 1 unit is connected to the input end of the time alignment module; The demand data module includes a demand data unit, a time series model 2 unit and a data prediction 2 unit; The demand data unit is used to collect and store demand data at fixed time intervals; the time series model 2 unit is used to construct a TBATA time series model based on the collected demand data; the data prediction 2 unit is used to predict demand data at future moments based on the constructed time series model; The output end of the demand data unit is connected to the input end of the timing model 2 unit and the time alignment module; the output end of the timing model 2 unit is connected to the input end of the data prediction 2 unit; the output end of the data prediction 2 unit is connected to the input end of the time alignment module.

[0035] The time alignment module includes an energy storage data unit and an interpolation unit; The energy storage data unit is used to collect and store energy storage configuration data at fixed time intervals; the interpolation unit is used to use the Lagrange interpolation method to obtain the environmental factor data and demand data at the corresponding time based on the collection time of the energy storage data; The output end of the energy storage data unit is connected to the input end of the interpolation unit; the output end of the interpolation unit is connected to the input end of the prediction optimization module.

[0036] The prediction optimization module includes a regression model unit and an optimization strategy unit; The regression model unit is used to construct a regression model with environmental factors and historical demand data as independent variables and energy storage configuration data as dependent variables; the optimization strategy unit is used to analyze energy storage conditions according to predicted future time data and formulate a configuration optimization strategy to optimize energy storage configuration; The output end of the regression model unit is connected to the input end of the optimization strategy unit.

[0037] In this embodiment, we implement the energy storage optimization configuration scheme for a medium-sized photovoltaic power station; the photovoltaic power station has an impact capacity of 5MW and is equipped with an energy storage system with a total capacity of 2MWh.

[0038] Environmental factor data: Sensors deployed around the PV power station collected environmental data such as light intensity, temperature, and humidity over the past year. These data were collected at 15-minute intervals to form time series data of environmental factors. For example, on a summer day, from 6 a.m. to 6 p.m., the light intensity data was [200, 300, 450, …, 100], and the temperature data was [25, 26, 28, …, 30]. At the same time, long-term historical meteorological data was obtained from the local meteorological department as a supplement. The hourly electricity demand data and user electricity consumption behavior characteristics of the area in the past year were obtained through the local power grid company. The commercial electricity demand is higher during the daytime on weekdays, and the proportion of residential electricity demand increases at night. On a typical weekday, from 8 am to 8 pm, the electricity demand data is [800, 900, 1200, …, 1000] in sequence; Data on the energy storage system’s power storage status, charging and discharging dynamic process, and system health status over the past year were collected from the energy storage system’s monitoring platform and equipment manufacturers. The data collection interval was 30 minutes. In a certain period of time, the energy storage system’s power storage status data was [0.8, 0.75, 0.7, …], and the charging and discharging power data was [100, -80, 120, …]. After Z-score standardization of the collected environmental factor data, the ADF test found that some of the light intensity data had first-order non-stationarity; after a difference process, the data reached a stable state; the autocorrelation function ACF and partial autocorrelation function PACF were calculated, and the autoregressive order p=2 and the moving average order q=1 were preliminarily determined; according to the minimization of the Bayesian information criterion BIC, the final (p, d, q) combination was determined to be (2, 1, 1), and after the model parameters were evaluated using the maximum likelihood estimation method, a VARIMA (2, 1, 1) model was constructed; the model was used to predict the environmental factor data for the next week, for example, it was predicted that the light intensity would reach 1000W / m² at around 12 noon three days later; After Z-score standardization of demand data, the additive decomposition method is used to decompose the demand time series into trend, seasonal and residual parts; the TBATS model is built and trained using historical data, and the maximum likelihood estimation method is used to optimize the model parameters. The model is used to predict the power demand for the next week. For example, it is predicted that there will be a peak in power demand from 7 to 9 p.m. on weekends, and the power demand will reach 1500kW; Taking the collection time of energy storage configuration data as the benchmark, it is assumed that energy storage configuration data is collected at every hour of the day. For a certain energy storage configuration time, such as 10 a.m., select the three time points closest to 10 a.m. in the time series of environmental factor data, namely 9:45, 10:00, and 10:15, and calculate the environmental factor data corresponding to that time according to the Lagrange interpolation formula. The same method is used for demand data to ensure that all data are aligned in time, providing an accurate data basis for subsequent model construction. The environmental factor data and demand data aligned at the same time are used as independent variables, and the energy storage configuration data is used as the dependent variable to build the SVR support vector regression model. For example, for the energy storage system's charging power, which is the energy storage configuration data, the environmental factors (such as light intensity, temperature) and demand data (power demand) corresponding to 10 a.m. every day in the past year are used as independent variables, and the charging power at that time is used as the dependent variable to build the corresponding SVR model. The optimal parameters of the model are determined through multiple training and optimization. Based on the predicted environmental factors and demand data for the next week, combined with the SVR regression model, the energy storage configuration data for different times of the day in the next week is calculated. For example, it is predicted that the light intensity will be high and the power demand will be relatively low at 12 noon three days later. The regression model calculates that the energy storage system should be charged at a power of 150kW at this time. Based on these calculation results, a detailed energy storage configuration optimization strategy is formulated, including charging and discharging power settings at different times of the day, energy storage capacity reservation plans, etc. After running according to this strategy for a week, compared with the previous unoptimized operating data, the charging and discharging efficiency of the energy storage system has improved, and the abandonment rate of the photovoltaic power station has decreased, effectively improving the overall operating efficiency and economic benefits of the photovoltaic power station.

[0039] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A photovoltaic power station energy storage optimization configuration method based on big data, characterized in that: The method comprises the following steps: Step 1: Collect historical data on environmental factors, demand, and energy storage configuration; In step 1, the environmental factor data refers to the natural environmental condition data that affects the power generation process of the photovoltaic power station and causes the power generation to fluctuate, which is expressed as: [E1, E2, …, E n ]; where n is a positive integer, representing the type of environmental factor, E1~E n They represent the first to n environmental factors respectively; the environmental factor data are collected by deploying sensors or obtained from the meteorological department; the historical data of environmental factors at different times have the same time interval Δt; the data are arranged in the time sequence [t1, t2, …, t a ] to form the time series data of environmental factors; where a represents the number of moments, t1~t a Indicates t1 to t a The first to a environmental factor data collection moments in a time period; Demand data refers to the demand for electricity and the characteristics of electricity consumption behavior within a specific time, expressed as: [R1, R2, …, R m ]; where m is a positive integer, indicating the number of demand types, R1~R m They represent the 1st to mth demand types respectively; the demand data is obtained through the electricity meter and the power grid company; the historical demand data at different times have the same time interval ΔT; the data are arranged in the time sequence [T1, T2, …, T b ] to form demand time series data; where b represents the number of moments, T1~T b Indicates T1 to T b The first to b demand data collection moments in the time period; Energy storage configuration data refers to the relevant data such as the power storage status, charging and discharging dynamic process, and system health status of the energy storage system in the photovoltaic power station. The energy storage configuration data is directly related to the formulation of the energy storage optimization configuration strategy, which is expressed as: [D1, D2, …, D a ]; where a is a positive integer, indicating the type of energy storage configuration data, D1~D a They represent the energy storage configuration data of types 1 to a respectively; the energy storage configuration data is obtained from the system monitoring platform and the equipment manufacturer; the energy storage configuration data at different times in history have the same time interval Δτ; the data are arranged in the time sequence [τ1, τ2, …, τ c ] to form the energy storage configuration time series data; where c represents the number of moments, τ1~τ c Indicates the 1st to cth energy storage configuration data collection time; Step 2: Build a time series model to predict environmental factors and demand data at future moments; Step 3: Using the energy storage configuration data collection time as a benchmark, align the environmental factor data and demand data time; Step 4: Construct a regression model with environmental factors and historical demand data as independent variables and energy storage configuration data as dependent variables; Step 5: Analyze the energy storage situation based on the predicted future time data to optimize the energy storage configuration.

2. A photovoltaic power station energy storage optimization configuration method based on big data according to claim 1, characterized in that: In step 2, different time series models are constructed based on the different time regularities of environmental factor data and demand data; For environmental factor data: perform Z-score standardization on the data; use the ADF test, set the stability threshold, perform a stability test on each environmental factor, and determine the number of differences d based on the degree of non-stationarity of the data; Use the difference formula to differentiate the non-stationary series, and perform the ADF test again after differentiation until the series is stationary; calculate the autocorrelation function ACF and partial autocorrelation function PACF of the stationary time series, and preliminarily determine the autoregressive order p and the moving average order q based on the graphs of ACF and PACF; determine the optimal combination of p, d, q based on the minimization of the Bayesian Information Criterion BIC; use the maximum likelihood estimation method to evaluate the parameters of the model based on the determined model order (p, d, q); construct a VARIMA model based on the historical data of environmental factors, and use the model to predict the data at future moments; For demand data: standardize the data by Z-score; use additive or multiplicative decomposition methods to decompose the demand time series into trend, seasonal and residual parts; build a TBATS model and use the historical demand time series data to train the TBATS model. During the training process, use the maximum likelihood estimation method to optimize the model parameters; use the model to predict future data.

3. The photovoltaic power station energy storage optimization configuration method based on big data according to claim 2 is characterized in that: In step 3, the time [τ1, τ2, …, τ c ] as a benchmark, and use the Lagrange interpolation method to obtain the environmental factor data and demand data at the corresponding time of the energy storage configuration data: For the energy storage configuration time τ k , in the time series of environmental factor data [t1, t2, …, t a ] Select the distance τ k The most recent d time points are expressed as: [t 1 ,t 2 ,…,t d ]; Calculate τ according to the Lagrange interpolation formula k Environmental factors E corresponding to the moment e data: Where k represents the energy storage configuration data collection time sequence, k∈{1,2,…,c}, τ k ∈{τ1,τ2,…,τ c }; j, h represent the distance τ k The sequence of the most recent time points, j, h∈{1,2,…,d}, t j ,t h ∈{t 1 ,t 2 ,…,t d }; e represents the type sequence of environmental factors, e∈{1,2,…,n}, E e ∈{E1,E2,…,E n }, E e (t j ) indicates t j Environmental factors E corresponding to the moment e data; For the energy storage configuration time τ k , in the time series of demand data [T1,T2,…,T b ] Select the distance τ k The most recent d time points are expressed as: [t 1 ,t 2 ,…,t d ]; Calculate τ according to the Lagrange interpolation formula k The demand data R corresponding to the moment f : Where k represents the energy storage configuration data collection time sequence, k∈{1,2,…,c}, τ k ∈{τ1,τ2,…,τ c }; j, h represent the distance τ k The sequence of the most recent time points, j, h∈{1,2,…,d}, t j ,t h ∈{t 1 ,t 2 ,…,t d }; f represents the demand type sequence, f∈{1,2,…,m}, R f ∈{R1,R2,…,R m }, R f (t j ) indicates t j The demand type R corresponding to the moment f data.

4. The photovoltaic power station energy storage optimization configuration method based on big data according to claim 3 is characterized in that: In step 4, a regression model is constructed with the environmental factor data and demand data at the same time as independent variables and the energy storage configuration data as the dependent variable; for a certain type of energy storage configuration data, τ1~τ c The environmental factor data and demand data corresponding to the time are used as independent variables, τ1~τ c The energy storage configuration data corresponding to the time is taken as the dependent variable, and the SVR support vector regression model is constructed; the corresponding SVR support vector regression model is constructed for each energy storage configuration data.

5. A photovoltaic power station energy storage optimization configuration method based on big data according to claim 4, characterized in that: In step 5, the environmental factor data and demand data at future moments are predicted according to the corresponding time series model; the future collection time of the energy storage configuration data is calculated according to the time interval Δτ: [τ c +Δτ,τ c +2·Δτ,…,τ c +u·Δτ]; where u is a positive integer, indicating the number of energy storage configuration data collection times; τ is calculated using the Lagrange interpolation method c +Δτ~τ c +u·Δτ corresponding to the environmental factor data and demand data at the moment; calculate the energy storage configuration data corresponding to different moments according to the SVR regression model corresponding to different energy storage configuration data; formulate an energy storage configuration optimization strategy based on the predicted energy storage configuration data.

6. A photovoltaic power station energy storage optimization configuration system based on big data, characterized in that: The system includes an environment data module, a demand data module, a time alignment module and a prediction optimization module; The environmental data module is used to collect and process environmental factor data; the demand data module is used to collect and process demand data; the time alignment module is used to calculate the environmental factor data and demand data at the corresponding time based on the collection time of the energy storage data; the prediction and optimization module is used to construct a regression model and predict the energy storage configuration data based on the predicted environmental factor data and demand data to formulate an energy storage optimization strategy; The output end of the environmental data module is connected to the input end of the time alignment module; The output end of the demand data module is connected to the input end of the time alignment module; the output end of the time alignment module is connected to the input end of the prediction optimization module.

7. A photovoltaic power station energy storage optimization configuration system based on big data according to claim 6, characterized in that: The environmental data module includes an environmental data unit, a time series model 1 unit and a data prediction 1 unit; The environmental data unit is used to collect and store environmental factor data at fixed time intervals; the time series model 1 unit is used to construct a VARIMA time series model based on the collected environmental factor data; the data prediction 1 unit is used to predict the environmental factor data at future moments based on the constructed time series model; The output end of the environment data unit is connected to the input end of the timing model 1 unit and the time alignment module; the output end of the timing model 1 unit is connected to the input end of the data prediction 1 unit; the output end of the data prediction 1 unit is connected to the input end of the time alignment module; The demand data module includes a demand data unit, a time series model 2 unit and a data prediction 2 unit; The demand data unit is used to collect and store demand data at fixed time intervals; the time series model 2 unit is used to construct a TBATA time series model based on the collected demand data; the data prediction 2 unit is used to predict demand data at future moments based on the constructed time series model; The output end of the demand data unit is connected to the input end of the timing model 2 unit and the time alignment module; the output end of the timing model 2 unit is connected to the input end of the data prediction 2 unit; the output end of the data prediction 2 unit is connected to the input end of the time alignment module.

8. A photovoltaic power station energy storage optimization configuration system based on big data according to claim 7, characterized in that: The time alignment module includes an energy storage data unit and an interpolation unit; The energy storage data unit is used to collect and store energy storage configuration data at fixed time intervals; the interpolation unit is used to use the Lagrange interpolation method to obtain the environmental factor data and demand data at the corresponding time based on the collection time of the energy storage data; The output end of the energy storage data unit is connected to the input end of the interpolation unit; the output end of the interpolation unit is connected to the input end of the prediction optimization module.

9. A photovoltaic power station energy storage optimization configuration system based on big data according to claim 8, characterized in that: The prediction optimization module includes a regression model unit and an optimization strategy unit; The regression model unit is used to construct a regression model with environmental factors and historical demand data as independent variables and energy storage configuration data as dependent variables; the optimization strategy unit is used to analyze energy storage conditions according to predicted future time data and formulate a configuration optimization strategy to optimize energy storage configuration; The output end of the regression model unit is connected to the input end of the optimization strategy unit.

Images