User-side energy storage operation monitoring and analysis method based on source-grid-load collaboration

Abstract

The present invention belongs to the technical field of energy storage operation monitoring, and specifically relates to a user-side energy storage operation monitoring and analysis method based on source-grid-load coordination, the steps of which include obtaining data of the energy storage system, collecting historical data at the same time, and fusing multi-dimensional data; predicting the grid demand and user load within time t, and using a linear algorithm to optimize and adjust the charging and discharging strategy of the energy storage system in real time; comprehensively evaluating the operating benefits of the energy storage system; extracting features from the historical data of the energy storage system, and combining it with the real-time operating data of the energy storage system, comprehensively predicting the potential risks in the operation of the energy storage system, and obtaining risk prediction results; based on the real-time changes in grid demand and user load, as well as the battery status of the energy storage system and the risk prediction results, automatically adjusting the charging and discharging operation strategy to ensure stable operation of the system. The present invention can more accurately predict grid demand and user load, and timely discover and predict potential risks.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy storage operation monitoring, and specifically relates to a user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration. Background Art

[0002] In the power system, source, grid and load are the basic components of the power system. Source refers to the power generation part of the power system, including various types of power stations. Grid refers to the transmission and distribution network of the power system, including high-voltage transmission network, medium-voltage distribution network and low-voltage user access network. The function of the grid is to transmit and distribute the electric energy generated by the power station to various power consumption areas. Load refers to the power consumption part of the power system, including various power loads such as industry, commerce and residents. The power load is the power demand that the power system needs to meet. Its size and characteristics will affect the operation and dispatch of the power system. User-side energy storage refers to the energy storage system installed at the power user end. It is directly connected to the user's power equipment to provide power storage and regulation services for the user's own power demand.

[0003] Existing user-side energy storage systems mostly adopt simple charging and discharging strategies and lack deep coordination with the power grid and user loads, resulting in monitoring and analysis during the operation of the energy storage system, making it difficult to fully exert its regulatory role and economic benefits. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration, which can more accurately predict grid demand and user load, and timely discover and predict potential risks.

[0005] The present invention provides the following technical solution: a user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration, comprising the following steps:

[0006] S1: Obtain the real-time operation data of the energy storage system, real-time grid data, real-time user load data and real-time market electricity price data, collect historical data of the grid, historical data of electricity load and historical data of the energy storage system, and integrate the multi-dimensional data;

[0007] S2: Based on the historical data of the power grid and the historical data of user load, the power grid demand within time t is predicted. and user load , and based on the integrated multi-dimensional data, a linear algorithm is used to optimize and adjust the charging and discharging strategy of the energy storage system in real time;

[0008] S3: Comprehensively evaluate the operating benefits of the energy storage system based on the historical data of the energy storage system;

[0009] S4: By extracting features from the historical data of the energy storage system and combining it with the real-time operation data of the energy storage system, the potential risks in the operation of the energy storage system are comprehensively predicted to obtain risk prediction results;

[0010] S5: Based on the real-time changes in grid demand and user load, as well as the battery status and risk prediction results of the energy storage system, the charging and discharging operation strategy is automatically adjusted to ensure stable operation of the system.

[0011] Preferably, in S1, the multi-dimensional data content is as follows:

[0012] The real-time operating data of the energy storage system includes: battery charging and discharging status, battery health status, temperature data, voltage and current data, and power data;

[0013] Real-time grid data includes: grid load, grid frequency, voltage level, and power flow direction;

[0014] Real-time user load data includes: power demand, power consumption mode, and power equipment status;

[0015] Real-time market electricity price data includes: real-time electricity price, electricity price volatility;

[0016] The historical data of power grid include: historical load, historical fault records;

[0017] The historical data of electricity load include: historical electricity consumption data, historical electricity consumption pattern;

[0018] The historical data of the energy storage system includes: historical charging and discharging data, historical fault records, and historical performance indicators.

[0019] Preferably, in S1, the multi-dimensional data is fused, and the fusion steps are as follows:

[0020] Standardize multi-dimensional data:

[0021] ;

[0022] In the formula, represents the standardized multi-dimensional data, x represents the original data point, and Respectively represent the minimum and maximum values ​​of the original data points in the multi-dimensional data;

[0023] By adding weights and normalization methods, the fusion function F is constructed:

[0024] ;

[0025] In the formula, F represents the comprehensive index after fusion, Represents the weight of each dimension data, ncf represents the total number of weights, They represent the standardized dimensional data respectively, and nmi represents the total number of dimensional data.

[0026] Preferably, in S2, the grid demand within time t is predicted and user load , the specific process is:

[0027] Predict the grid demand at time t , the prediction process is as follows:

[0028] ;

[0029] In the formula, The intercept term represents the grid demand, which indicates the baseline demand level when all eigenvalues ​​are zero. Represents the time feature, indicating the number of hours in a day, Representation and time characteristics The associated coefficient, Represents the week feature, indicating a day of the week, Representatives and characteristics of the week The associated coefficient, Represents the month feature, indicating the month of the year, Representative and monthly characteristics The associated coefficients represent the impact on grid demand in different months;

[0030] Predict user load at time t , the prediction process is as follows:

[0031] ;

[0032] In the formula, represents the intercept term of the electricity load, Representation and time characteristics The associated coefficient, Represents the temperature characteristic, indicating the temperature value at time t, Representative and temperature characteristics The associated coefficient, Represents holiday characteristics, which is a binary variable. It takes the value of 1 on holidays and 0 on non-holidays. Representative and holiday characteristics The associated coefficient represents the impact of holidays on user load.

[0033] Preferably, in S2, based on the fused multi-dimensional data, a linear algorithm is used to optimize and adjust the charging and discharging strategy of the energy storage system in real time, and the optimization and adjustment steps are as follows:

[0034] The charging and discharging strategy is optimized through a linear algorithm. The optimization process is as follows:

[0035] ;

[0036] In the formula, Indicates found The value of makes the charging and discharging cost at all time points t reach the minimum, Represents the amount of electrical energy consumed by the energy storage system during charging in time t, represents the electricity price at time t, Represents the electrical energy provided by the energy storage system during the time t. is the weight coefficient of battery loss, Based on the current charge and discharge status The battery loss function;

[0037] When optimizing the charge and discharge strategy, the impact of charge and discharge depth and frequency on battery loss is fully considered, which can be expressed as:

[0038] ;

[0039] In the formula, It represents the change in charge and discharge between two consecutive time points t and t-1. Indicates the current charge and discharge status and the maximum charge and discharge depth The proportion of For the corresponding The weight of For the corresponding The weight of

[0040] When optimizing the charging and discharging strategy, the charging and discharging state of the battery is constrained, which can be expressed as:

[0041] ;

[0042] In the formula, Represents the minimum charge and discharge depth;

[0043] When optimizing the charging and discharging strategy, balance constraints are imposed on the grid demand and user load:

[0044] .

[0045] Preferably, in S3, a comprehensive evaluation of the operating benefits of the energy storage system is performed, and the evaluation steps are as follows:

[0046] Direct economic benefits to energy storage systems The calculation process is as follows:

[0047] ;

[0048] In the formula, and Respectively represent the battery charging and discharging state and battery conversion efficiency in the i-th time period in the historical data, and Represent the electricity purchase price and electricity sales price in the i-th time period, and ib represents the total number of time periods;

[0049] The operating costs of energy storage systems and benefits The calculation process is as follows:

[0050] ;

[0051] ;

[0052] In the formula, Represents the initial investment cost of the energy storage system, represents the operation and maintenance cost in the i-th time period, and N represents the expected service life of the energy storage system;

[0053] Contribution of energy storage systems to grid stability The calculation process is as follows:

[0054] ;

[0055] In the formula, represents the total demand of the power grid in the i-th time period, that is, the total electrical energy required by the power grid, is the adjustment coefficient;

[0056] Environmental benefits of energy storage systems The calculation process is as follows:

[0057] ;

[0058] In the formula, represents the power generation in the i-th time period, The representative unit is the carbon emission factor of power generation.

[0059] Preferably, in S4, the historical data features of the energy storage system are extracted, and the extraction steps are as follows:

[0060] Establish polynomial features to generate high-order terms and interaction terms of the original features to capture nonlinear relationships:

[0061] ;

[0062] In the formula, represents the original feature variable, Represents a new feature set consisting of the original features and their polynomial combinations;

[0063] Create interaction features and create products of feature pairs to discover interactions between features:

[0064] ;

[0065] In the formula, Represents the interaction items between different features, represents the product of the original feature variables;

[0066] Establish an ARIMA model and use the ARIMA model to decompose the time series to extract trend, seasonality and random components. The establishment process is as follows:

[0067] ;

[0068] In the formula, represents the observed value at time t, e represents a constant term, which is used to adjust the center position of the time series. represents the autoregressive coefficient, Represents the time series arrive The observed value at time, represents the moving average coefficient, represents the white noise error term of the time series at time t, Represents the time series arrive The white noise error term at time t;

[0069] The periodic pattern in the data is extracted by Fourier transform, and the extraction process is as follows:

[0070] ;

[0071] In the formula, represents the coefficient in the frequency domain, NT represents the length of the time series, k represents the index in the frequency domain, represents the complex exponential function, represents the imaginary unit;

[0072] The local time-frequency characteristics of the signal are captured by wavelet transform. The capture process is as follows:

[0073] ;

[0074] In the formula, represents the coefficients after wavelet transformation, represents the wavelet function, which depends on the scale s, represents the integral symbol, indicating integration over the entire time domain, and dt represents the differential element of the integral;

[0075] The kernel technique is used to deal with nonlinear problems and find the optimal boundary between data points. The search process is as follows:

[0076] ;

[0077] In the formula, represents the optimal boundary of the data point, sign represents the sign function, represents the kernel function, represents the Lagrange multiplier, represents the label of the irlth sample, NY is the total number of sample labels, and r represents The bias term in ;

[0078] The deep neural network is used to process data with sequence characteristics. The processing process is as follows:

[0079] ;

[0080] In the formula, represents the hidden state at time t, RNN represents the activation function, represents the weight matrix input to the hidden layer, Represents the weight matrix of the recursive connection.

[0081] Preferably, in S4, the potential risks in the operation of the energy storage system are comprehensively predicted to obtain risk prediction results, and the prediction steps are as follows:

[0082] Logistic regression was used to calculate the probability of risk occurrence. , the calculation process is as follows:

[0083] ;

[0084] In the formula, Sigmoid represents a function that maps linear output to probability interval, represents a linear combination, b represents The bias term in ;

[0085] Based on the model output and business logic, the risk score Rksre is calculated. The calculation process is as follows:

[0086] ;

[0087] In the formula, Represents the output of different models or features, Respectively represent the corresponding The weight of

[0088] The risk is set to three levels: low, medium and high. and the maximum risk threshold , and based on Rksre, determine which risk level the energy storage system is in:

[0089] if , it is judged as low risk;

[0090] if , it is judged as medium risk;

[0091] if , it is judged as high risk.

[0092] Preferably, in S5, based on the real-time changes in grid demand and user load as well as the battery status and risk prediction results of the energy storage system, the charging and discharging operation strategy is automatically adjusted, and the adjustment steps are as follows:

[0093] The power of basic charge and discharge is calculated as follows:

[0094] ;

[0095] In the formula, Represents the user's electricity demand, Represents the grid load, Represents the power of basic charge and discharge;

[0096] Secondary adjustment is performed based on the power flow direction. The adjustment process is as follows:

[0097] ;

[0098] In the formula, represents the adjusted power, Represents the direction of power flow, and represent adjustment coefficients respectively;

[0099] Three adjustments are made based on the risk level. The adjustment process is as follows:

[0100] ;

[0101] In the formula, represents the power adjusted according to the risk level, , and Risk level adjustment factors representing low risk, medium risk, and high risk, respectively;

[0102] The fourth adjustment is performed based on the battery status. The adjustment process is as follows:

[0103] ;

[0104] In the formula, Represents the final charge and discharge power, SOC represents the current state of charge of the battery, and HSC represents the current health status of the battery. represents the safety factor;

[0105] pass Whether the value is greater than zero, determine whether the charging and discharging strategy needs to be adjusted:

[0106] if If it is greater than 0, it is determined that charging is needed;

[0107] if If it is less than 0, it is determined that discharge is required;

[0108] if If it is equal to 0, it is considered normal.

[0109] Technical effects and advantages of the present invention:

[0110] By acquiring and integrating real-time and historical data, the energy storage system of the present invention can more accurately predict grid demand and user load. This predictive capability enables the system to adjust the charging and discharging strategy in advance to adapt to changes in supply and demand of the grid. The optimized charging and discharging strategy can not only improve the stability of the grid and reduce fluctuations caused by supply and demand imbalance, but also reduce electricity costs and improve economic benefits by charging when electricity prices are low and discharging when electricity prices are high. By conducting in-depth analysis of the historical data of the energy storage system, its operating benefits can be comprehensively evaluated and key factors affecting the benefits can be identified. Combined with real-time operating data and risk prediction models, the energy storage system can promptly discover and predict potential risks, such as battery aging, performance degradation, etc., which helps to take measures in advance to avoid or mitigate risks, thereby extending the service life of the energy storage system, reducing maintenance costs, and improving overall economic benefits and safety.

[0111] The present invention utilizes advanced algorithms and models, and the energy storage system can automatically adjust the charging and discharging strategy to cope with the real-time changes in grid demand and user load. This intelligent operation strategy can ensure that the energy storage system can operate efficiently and stably under various working conditions. At the same time, combined with the risk prediction results, the system can achieve self-adjustment to avoid operating under high-risk conditions, thereby ensuring the safety and reliability of the system. This automated operation method reduces the need for manual intervention, improves operational efficiency, and reduces the possibility of human error, providing a solid technical foundation for the future development of the energy storage system; based on the real-time changes in grid demand and user load, as well as the battery status and risk prediction results of the energy storage system, the charging and discharging operation strategy is automatically adjusted to ensure that the energy storage system can operate stably under various conditions, achieving the effect of monitoring and analyzing the operation of the energy storage system based on the three directions of source, grid and load, and improving the reliability and flexibility of the energy storage system. BRIEF DESCRIPTION OF THE DRAWINGS

[0112] Figure 1 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0113] The present invention will be further described below in conjunction with specific embodiments.

[0114] like Figure 1 As shown, the user-side energy storage operation monitoring and analysis method based on source-grid-load coordination includes the following steps:

[0115] S1: Obtain the real-time operation data of the energy storage system, real-time grid data, real-time user load data and real-time market electricity price data, collect historical data of the grid, historical data of electricity load and historical data of the energy storage system, and integrate the multi-dimensional data;

[0116] S2: Based on the historical data of the power grid and the historical data of user load, the power grid demand within time t is predicted. and user load , and based on the integrated multi-dimensional data, a linear algorithm is used to optimize and adjust the charging and discharging strategy of the energy storage system in real time;

[0117] S3: Comprehensively evaluate the operating benefits of the energy storage system based on the historical data of the energy storage system;

[0118] S4: By extracting features from the historical data of the energy storage system and combining it with the real-time operation data of the energy storage system, the potential risks in the operation of the energy storage system are comprehensively predicted to obtain risk prediction results;

[0119] S5: Based on the real-time changes in grid demand and user load, as well as the battery status and risk prediction results of the energy storage system, the charging and discharging operation strategy is automatically adjusted to ensure stable operation of the system.

[0120] In S1, the multi-dimensional data content is as follows:

[0121] The real-time operating data of the energy storage system includes: battery charging and discharging status, battery health status, temperature data, voltage and current data, and power data;

[0122] Real-time grid data includes: grid load, grid frequency, voltage level, and power flow direction;

[0123] Real-time user load data includes: power demand, power consumption mode, and power equipment status;

[0124] Real-time market electricity price data includes: real-time electricity price, electricity price volatility;

[0125] The historical data of power grid include: historical load, historical fault records;

[0126] The historical data of electricity load include: historical electricity consumption data, historical electricity consumption pattern;

[0127] The historical data of the energy storage system includes: historical charging and discharging data, historical fault records, and historical performance indicators.

[0128] In S1, multi-dimensional data is fused, and the fusion steps are as follows:

[0129] Standardize multi-dimensional data:

[0130] ;

[0131] In the formula, represents the standardized multi-dimensional data, x represents the original data point, and Respectively represent the minimum and maximum values ​​of the original data points in the multi-dimensional data;

[0132] By adding weights and normalization methods, the fusion function F is constructed:

[0133] ;

[0134] In the formula, F represents the comprehensive index after fusion, Represents the weight of each dimension data, ncf represents the total number of weights, They represent the standardized dimensional data respectively, and nmi represents the total number of dimensional data.

[0135] In S2, the grid demand is predicted within time t and user load , the specific process is:

[0136] Predict the grid demand at time t , the prediction process is as follows:

[0137] ;

[0138] In the formula, The intercept term represents the grid demand, which indicates the baseline demand level when all eigenvalues ​​are zero. Represents the time feature, indicating the hour of the day, usually from 0 to 23, Representation and time characteristics The associated coefficient, Represents the day of the week, usually from 0 to 6. Representatives and characteristics of the week The associated coefficient, Represents the month feature, indicating the month of the year, usually from 1 to 12, Representative and monthly characteristics The associated coefficients represent the impact on grid demand in different months;

[0139] Predict user load at time t , the prediction process is as follows:

[0140] ;

[0141] In the formula, represents the intercept term of the electricity load, Representation and time characteristics The associated coefficient, Represents the temperature characteristic, indicating the temperature value at a specific time t, Representative and temperature characteristics The associated coefficient, Represents holiday characteristics, which is a binary variable. It takes the value of 1 on holidays and 0 on non-holidays. Representative and holiday characteristics The associated coefficient represents the impact of holidays on user load.

[0142] In S2, based on the integrated multi-dimensional data, a linear algorithm is used to optimize and adjust the charging and discharging strategy of the energy storage system in real time. The optimization and adjustment steps are as follows:

[0143] The charging and discharging strategy is optimized through a linear algorithm. The optimization process is as follows:

[0144] ;

[0145] In the formula, Indicates found The value of makes the charging and discharging cost at all time points t reach the minimum, Represents the amount of electrical energy consumed by the energy storage system during charging in time t, represents the electricity price at time t, Represents the electrical energy provided by the energy storage system during the time t. is the weight coefficient of battery loss, Based on the current charge and discharge status The battery loss function;

[0146] When optimizing the charge and discharge strategy, the impact of charge and discharge depth and frequency on battery loss is fully considered, which can be expressed as:

[0147] ;

[0148] In the formula, It represents the change in charge and discharge between two consecutive time points t and t-1. Indicates the current charge and discharge status and the maximum charge and discharge depth The proportion of For the corresponding The weight of For the corresponding The weight of

[0149] When optimizing the charging and discharging strategy, the charging and discharging state of the battery is constrained, which can be expressed as:

[0150] ;

[0151] In the formula, Represents the minimum charge and discharge depth;

[0152] When optimizing the charging and discharging strategy, balance constraints are imposed on the grid demand and user load:

[0153] .

[0154] In S3, a comprehensive evaluation of the operating benefits of the energy storage system is conducted. The evaluation steps are as follows:

[0155] Direct economic benefits to energy storage systems The calculation process is as follows:

[0156] ;

[0157] In the formula, and Respectively represent the battery charging and discharging state and battery conversion efficiency in the i-th time period in the historical data, and Represent the electricity purchase price and electricity sales price in the i-th time period, and ib represents the total number of time periods;

[0158] The operating costs of energy storage systems and benefits The calculation process is as follows:

[0159] ;

[0160] ;

[0161] In the formula, Represents the initial investment cost of the energy storage system, represents the operation and maintenance cost in the i-th time period, and N represents the expected service life of the energy storage system;

[0162] Contribution of energy storage systems to grid stability The calculation process is as follows:

[0163] ;

[0164] In the formula, represents the total demand of the power grid in the i-th time period, that is, the total electrical energy required by the power grid, is the adjustment coefficient;

[0165] Environmental benefits of energy storage systems The calculation process is as follows:

[0166] ;

[0167] In the formula, represents the power generation in the i-th time period, The representative unit is the carbon emission factor of power generation.

[0168] In S4, the historical data features of the energy storage system are extracted. The extraction steps are as follows:

[0169] Establish polynomial features to generate high-order terms and interaction terms of the original features to capture nonlinear relationships:

[0170] ;

[0171] In the formula, represents the original feature variable, Represents a new feature set consisting of the original features and their polynomial combinations;

[0172] Create interaction features and create products of feature pairs to discover interactions between features:

[0173] ;

[0174] In the formula, Represents the interaction items between different features, represents the product of the original feature variables;

[0175] Establish an ARIMA model and use the ARIMA model to decompose the time series to extract trend, seasonality and random components. The establishment process is as follows:

[0176] ;

[0177] In the formula, represents the observed value at time t, e represents a constant term, which is used to adjust the center position of the time series. represents the autoregressive coefficient, Represents the time series arrive The observed value at time, represents the moving average coefficient, Represents the white noise error term of the time series at time t (i.e. time t), Represents the time series arrive The white noise error term at time; f and q are constants;

[0178] The periodic pattern in the data is extracted by Fourier transform, and the extraction process is as follows:

[0179] ;

[0180] In the formula, represents the coefficient in the frequency domain, NT represents the length of the time series, k represents the index in the frequency domain, represents the complex exponential function, represents the imaginary unit;

[0181] The local time-frequency characteristics of the signal are captured by wavelet transform. The capture process is as follows:

[0182] ;

[0183] In the formula, represents the coefficients after wavelet transformation, represents the wavelet function, which depends on the scale s, represents the integral symbol, indicating integration over the entire time domain, and dt represents the differential element of the integral;

[0184] The kernel technique is used to deal with nonlinear problems and find the optimal boundary between data points. The search process is as follows:

[0185] ;

[0186] In the formula, represents the optimal boundary of the data point, sign represents the sign function, represents the kernel function, represents the Lagrange multiplier, represents the label of the irlth sample, NY is the total number of sample labels, and r represents The bias term in ;

[0187] The deep neural network is used to process data with sequence characteristics. The processing process is as follows:

[0188] ;

[0189] In the formula, represents the hidden state at time t, RNN represents the activation function, represents the weight matrix input to the hidden layer, Represents the weight matrix of the recursive connection.

[0190] In S4, the potential risks in the operation of the energy storage system are comprehensively predicted to obtain risk prediction results. The prediction steps are as follows:

[0191] Logistic regression was used to calculate the probability of risk occurrence. , the calculation process is as follows:

[0192] ;

[0193] In the formula, Sigmoid represents a function that maps linear output to probability interval, represents a linear combination, b represents The bias term in ;

[0194] Based on the model output and business logic, the risk score Rksre is calculated. The calculation process is as follows:

[0195] ;

[0196] In the formula, Represents the output of different models or features, Respectively represent the corresponding The weight of

[0197] The risk is set to three levels: low, medium and high. and the maximum risk threshold , and based on Rksre, determine which risk level the energy storage system is in:

[0198] if , it is judged as low risk;

[0199] if , it is judged as medium risk;

[0200] if , it is judged as high risk.

[0201] In S5, the charging and discharging operation strategy is automatically adjusted based on the real-time changes in grid demand and user load, as well as the battery status and risk prediction results of the energy storage system. The adjustment steps are as follows:

[0202] The power of basic charge and discharge is calculated as follows:

[0203] ;

[0204] In the formula, Represents the user's electricity demand, Represents the grid load, Represents the power of basic charge and discharge;

[0205] Secondary adjustment is performed based on the power flow direction. The adjustment process is as follows:

[0206] ;

[0207] In the formula, represents the adjusted power, Represents the direction of power flow, and represent adjustment coefficients respectively;

[0208] Three adjustments are made based on the risk level. The adjustment process is as follows:

[0209] ;

[0210] In the formula, represents the power adjusted according to the risk level, , and Risk level adjustment factors representing low risk, medium risk, and high risk, respectively;

[0211] The fourth adjustment is performed based on the battery status. The adjustment process is as follows:

[0212] ;

[0213] In the formula, Represents the final charge and discharge power, SOC represents the current state of charge of the battery, and HSC represents the current health status of the battery. represents the safety factor;

[0214] pass Whether the value is greater than zero, determine whether the charging and discharging strategy needs to be adjusted:

[0215] if If it is greater than 0, it is determined that charging is needed;

[0216] if If it is less than 0, it is determined that discharge is required;

[0217] if If it is equal to 0, it is considered normal.

[0218] This embodiment realizes comprehensive monitoring and analysis of energy storage operations by integrating real-time and historical data of energy storage systems, power grids, and user loads. Through key technical steps such as multi-dimensional data fusion, prediction models, linear optimization algorithms, benefit evaluation, risk prediction, and adaptive charging and discharging strategy adjustment, the operating efficiency and economic benefits of the energy storage system can be effectively improved, while ensuring the stability of the power grid and the safety of the energy storage system itself. Through standardized processing and weighted fusion of multi-dimensional data, a comprehensive indicator is constructed, which provides a basis for subsequent prediction and optimization. Historical data is used to predict power grid demand and user load, and combined with real-time data, a linear algorithm is used to optimize the charging and discharging strategy to minimize costs. At the same time, through in-depth analysis of the historical data of the energy storage system, the operating benefits of the energy storage system are comprehensively evaluated, including economic benefits, operating costs, contribution to power grid stability, and environmental benefits.

[0219] This embodiment also predicts the potential risks in the operation of the energy storage system through feature extraction and risk prediction models, providing a guarantee for the safe operation of the system. Finally, according to the grid demand, user load, battery status and risk prediction results, the charging and discharging operation strategy is automatically adjusted to ensure the stable operation of the system.

Claims

1. A user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration, characterized in that: The following steps are involved: S1: Obtain the real-time operation data of the energy storage system, real-time grid data, real-time user load data and real-time market electricity price data, collect historical data of the grid, historical data of electricity load and historical data of the energy storage system, and integrate the multi-dimensional data; S2: Based on the historical data of the power grid and the historical data of user load, the power grid demand within time t is predicted. and user load , and based on the integrated multi-dimensional data, a linear algorithm is used to optimize and adjust the charging and discharging strategy of the energy storage system in real time; The charging and discharging strategy is optimized through a linear algorithm. The optimization process is as follows: ； In the formula, Indicates found The value of makes the charging and discharging cost at all time points t reach the minimum, Represents the amount of electrical energy consumed by the energy storage system during charging in time t, represents the electricity price at time t, Represents the electrical energy provided by the energy storage system during the time t. is the weight coefficient of battery loss, Based on the current charge and discharge status The battery loss function; S3: Comprehensively evaluate the operating benefits of the energy storage system based on the historical data of the energy storage system; S4: By extracting features from the historical data of the energy storage system and combining it with the real-time operation data of the energy storage system, the potential risks in the operation of the energy storage system are comprehensively predicted to obtain risk prediction results; S5: Based on the real-time changes in grid demand and user load, as well as the battery status and risk prediction results of the energy storage system, the charging and discharging operation strategy is automatically adjusted to ensure stable operation of the system. The adjustment steps include calculating the basic charging and discharging power, making a secondary adjustment based on the power flow direction, making a tertiary adjustment based on the risk level, and making a fourth adjustment based on the battery status.

2. The user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration according to claim 1 is characterized in that: In the above-mentioned S1, the multi-dimensional data content is as follows: The real-time operating data of the energy storage system includes: battery charging and discharging status, battery health status, temperature data, voltage and current data, and power data; Real-time grid data includes: grid load, grid frequency, voltage level, and power flow direction; Real-time user load data includes: power demand, power consumption mode, and power equipment status; Real-time market electricity price data includes: real-time electricity price, electricity price volatility; The historical data of power grid include: historical load, historical fault records; The historical data of electricity load include: historical electricity consumption data, historical electricity consumption pattern; The historical data of the energy storage system includes: historical charging and discharging data, historical fault records, and historical performance indicators.

3. The user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration according to claim 1 is characterized in that: In the above-mentioned S1, multi-dimensional data is fused, and the fusion steps are as follows: Standardize multi-dimensional data: ； In the formula, represents the standardized multi-dimensional data, x represents the original data point, and Respectively represent the minimum and maximum values ​​of the original data points in the multi-dimensional data; By adding weights and normalization methods, the fusion function F is constructed: ； In the formula, F represents the comprehensive index after fusion, Represents the weight of each dimension data, ncf represents the total number of weights, They represent the standardized dimensional data respectively, and nmi represents the total number of dimensional data.

4. The user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration according to claim 1 is characterized in that: In S2, the grid demand is predicted within time t and user load , the specific process is: Predict the grid demand at time t , the prediction process is as follows: ； In the formula, The intercept term represents the grid demand, which indicates the baseline demand level when all eigenvalues ​​are zero. Represents the time feature, indicating the number of hours in a day, Representation and time characteristics The associated coefficient, Represents the week feature, indicating a day of the week, Representatives and characteristics of the week The associated coefficient, Represents the month feature, indicating the month of the year, Representative and monthly characteristics The associated coefficients represent the impact on grid demand in different months; Predict user load at time t , the prediction process is as follows: ； In the formula, represents the intercept term of the electricity load, Representation and time characteristics The associated coefficient, Represents the temperature characteristic, indicating the temperature value at time t, Representative and temperature characteristics The associated coefficient, Represents holiday characteristics, which is a binary variable. It takes the value of 1 on holidays and 0 on non-holidays. Representative and holiday characteristics The associated coefficient represents the impact of holidays on user load.

5. The user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration according to claim 4 is characterized in that: In S2, when optimizing the charge and discharge strategy, the influence of the charge and discharge depth and frequency on the battery loss during battery loss is fully considered, which is expressed as: ； In the formula, It represents the change in charge and discharge between two consecutive time points t and t-1. Indicates the current charge and discharge status and the maximum charge and discharge depth The proportion of For the corresponding The weight of For the corresponding The weight of When optimizing the charging and discharging strategy, the charging and discharging state of the battery is constrained, which can be expressed as: ； In the formula, Represents the minimum charge and discharge depth; When optimizing the charging and discharging strategy, balance constraints are imposed on the grid demand and user load: 。 6. The user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration according to claim 1 is characterized in that: In S3, a comprehensive evaluation of the operating benefits of the energy storage system is conducted, and the evaluation steps are as follows: Direct economic benefits to energy storage systems The calculation process is as follows: ； In the formula, and Respectively represent the battery charging and discharging state and battery conversion efficiency in the i-th time period in the historical data, and Represent the electricity purchase price and electricity sales price in the i-th time period, and ib represents the total number of time periods; The operating costs of energy storage systems and benefits The calculation process is as follows: ； ； In the formula, Represents the initial investment cost of the energy storage system, represents the operation and maintenance cost in the i-th time period, and N represents the expected service life of the energy storage system; Contribution of energy storage systems to grid stability The calculation process is as follows: ； In the formula, represents the total demand of the power grid in the i-th time period, that is, the total electrical energy required by the power grid, is the adjustment coefficient; Environmental benefits of energy storage systems The calculation process is as follows: ； In the formula, represents the power generation in the i-th time period, The representative unit is the carbon emission factor of power generation.

7. The user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration according to claim 4 is characterized in that: In the above-mentioned S4, the historical data features of the energy storage system are extracted, and the extraction steps are as follows: Establish polynomial features to generate high-order terms and interaction terms of the original features to capture nonlinear relationships: ； In the formula, represents the original feature variable, Represents a new feature set consisting of the original features and their polynomial combinations; Create interaction features and create products of feature pairs to discover interactions between features: ； In the formula, Represents the interaction items between different features, represents the product of the original feature variables; Establish an ARIMA model and use the ARIMA model to decompose the time series to extract trend, seasonality and random components. The establishment process is as follows: ； In the formula, represents the observed value at time t, e represents a constant term, which is used to adjust the center position of the time series. represents the autoregressive coefficient, Represents the time series arrive The observed value at time, represents the moving average coefficient, represents the white noise error term of the time series at time t, Represents the time series arrive The white noise error term at time t; The periodic pattern in the data is extracted by Fourier transform, and the extraction process is as follows: ； In the formula, represents the coefficient in the frequency domain, NT represents the length of the time series, k represents the index in the frequency domain, represents the complex exponential function, represents the imaginary unit; The local time-frequency characteristics of the signal are captured by wavelet transform. The capture process is as follows: ； In the formula, represents the coefficients after wavelet transformation, represents the wavelet function, which depends on the scale s, represents the integral symbol, indicating integration over the entire time domain, and dt represents the differential element of the integral; The kernel technique is used to deal with nonlinear problems and find the optimal boundary between data points. The search process is as follows: ； In the formula, represents the optimal boundary of the data point, sign represents the sign function, represents the kernel function, represents the Lagrange multiplier, represents the label of the irlth sample, NY is the total number of sample labels, and r represents The bias term in ; The deep neural network is used to process data with sequence characteristics. The processing process is as follows: ； In the formula, represents the hidden state at time t, RNN represents the activation function, represents the weight matrix input to the hidden layer, Represents the weight matrix of the recursive connection.

8. The user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration according to claim 7 is characterized in that: In S4, the potential risks in the operation of the energy storage system are comprehensively predicted to obtain risk prediction results. The prediction steps are as follows: Logistic regression was used to calculate the probability of risk occurrence. , the calculation process is as follows: ； In the formula, Sigmoid represents a function that maps linear output to probability interval, represents a linear combination, b represents The bias term in ; Based on the model output and business logic, the risk score Rksre is calculated. The calculation process is as follows: ； In the formula, Represents the output of different models or features, Respectively represent the corresponding The weight of The risk is set to three levels: low, medium and high. and the maximum risk threshold , and based on Rksre, determine which risk level the energy storage system is in: if , it is judged as low risk; if , it is judged as medium risk; if , it is judged as high risk.

9. The user-side energy storage operation monitoring and analysis method based on source-grid-load collaboration according to claim 1 is characterized in that: In S5, the charging and discharging operation strategy is automatically adjusted based on the real-time changes in grid demand and user load as well as the battery status and risk prediction results of the energy storage system. The adjustment steps are as follows: The power of basic charge and discharge is calculated as follows: ； In the formula, Represents the user's electricity demand, Represents the grid load, Represents the power of basic charge and discharge; Secondary adjustment is performed based on the power flow direction. The adjustment process is as follows: ； In the formula, represents the adjusted power, Represents the direction of power flow, and represent adjustment coefficients respectively; Three adjustments are made based on the risk level. The adjustment process is as follows: ； In the formula, represents the power adjusted according to the risk level, , and Risk level adjustment factors representing low risk, medium risk, and high risk, respectively; The fourth adjustment is performed based on the battery status. The adjustment process is as follows: ； In the formula, Represents the final charge and discharge power, SOC represents the current state of charge of the battery, and HSC represents the current health status of the battery. represents the safety factor; pass Whether the value is greater than zero, determine whether the charging and discharging strategy needs to be adjusted: if If it is greater than 0, it is determined that charging is needed; if If it is less than 0, it is determined that discharge is required; if If it is equal to 0, it is considered normal.

Images