Differential standby power control method for hybrid energy storage system based on multi-information fusion

Abstract

The invention discloses a hybrid energy storage system differential standby power control method based on multi-information fusion, and relates to the technical field of hybrid energy storage, and the method comprises the following steps: collecting operation parameter data of a plurality of accessed energy sources in real time, constructing a multi-energy input data set, and transmitting the data to a system central control unit for unified processing. Through a multi-information fusion technology and differentiated standby power control, the system solves the problem of inconsistency of multi-energy data, and improves the power supply stability and efficiency. The power supply gap is predicted and scheduled in real time to guarantee stable operation of the base station; clean energy is intelligently allocated for peak clipping and valley filling, mains supply dependence and energy storage equipment loss are reduced, and the service life is prolonged. Meanwhile, energy distribution is optimized through data preprocessing and load matching, and energy consumption and cost are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of hybrid energy storage, and particularly to a differential power backup control method for a hybrid energy storage system based on multi-information fusion. Background Art

[0002] Differential power backup control of a hybrid energy storage system based on multi-information fusion refers to intelligent regulation in a base station energy storage system by integrating various external information and energy inputs (such as wind turbines, photovoltaics, commercial power, and energy storage), and performing differential power backup management for different requirements and time periods. The usage scenario of the system is base station energy storage, with single-system management, that is, centralized management and control of the operating status of energy storage devices. External access covers wind turbines, photovoltaics, commercial power, and energy storage. The system can flexibly integrate these energy inputs to achieve energy complementarity. Through peak-valley regulation, the system releases energy storage during peak grid loads and charges the energy storage during valley periods, thereby achieving the purpose of peak shaving and valley filling, improving power supply stability and economy. At the same time, the system can provide an energy storage power backup function for critical equipment to ensure power supply continuity and reliability. This differential power backup control can dynamically adjust the energy storage strategy according to different time periods and power supply scenarios to achieve efficient utilization and management of energy.

[0003] The existing technologies have the following deficiencies:

[0004] During the differential power backup control process of a hybrid energy storage system based on multi-information fusion, the problem of system misjudgment caused by inconsistent multi-energy access source information may lead to serious consequences. This system integrates multiple energy inputs such as wind turbines, photovoltaics, commercial power, and energy storage, and realizes energy storage power backup management through real-time monitoring and scheduling. However, in practical applications, if the information of different energy accesses (such as voltage, current, and power data) is inconsistent due to time delay, data loss, or abnormal fluctuations, the system may not be able to correctly judge the current energy supply status and the operating conditions of energy storage devices, resulting in wrong decisions. For example, when the system wrongly judges that external energy (such as photovoltaics or wind turbines) is supplying power but actually fails to provide sufficient power, the energy storage system may over-discharge and fail to replenish the power backup energy storage in time. During peak grid loads, such misjudgment may lead to the exhaustion of power backup energy storage, affecting the power backup function and being unable to provide stable power supply for critical equipment such as base stations, thereby causing serious consequences such as communication interruption. Therefore, ensuring the real-time accuracy and consistency of multi-energy access data and the effective fault tolerance processing of the system for information anomalies are the key technical challenges for ensuring the stable operation of the differential power backup system.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide a differential power backup control method for a hybrid energy storage system based on multi-information fusion. Through multi-information fusion technology and differential power backup control strategies, the system solves the misjudgment problem caused by inconsistent source data of multiple energy accesses, and improves power supply stability and energy utilization efficiency. Under energy fluctuations, the system uses real-time prediction and dynamic scheduling to timely compensate for power supply gaps and ensure the stable operation of base stations; by intelligently allocating wind turbines, photovoltaics, commercial power, and energy storage, clean energy is preferentially used to achieve peak shaving and valley filling, reduce dependence on commercial power and equipment losses, and extend the service life of energy storage devices. At the same time, the system optimizes energy distribution through data preprocessing and load matching, reduces energy consumption and costs, and improves economic benefits to solve the problems in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions: A differential power backup control method for a hybrid energy storage system based on multi-information fusion, including the following steps:

[0008] Collect the operation parameter data of multiple accessed energy sources in real time, construct a multi-energy input data set, and transmit the data to the system central control unit for unified processing;

[0009] Preprocess the collected multi-energy input data to ensure the temporal consistency and numerical accuracy of the data, and form multi-energy power supply state data;

[0010] Based on historical load data and the current power supply state, use a prediction algorithm to conduct real-time dynamic evaluation of the base station load demand and match the power supply capabilities of each energy source;

[0011] Through an energy management algorithm, generate differential power backup strategies according to the power supply demand and energy supply state, including charge and discharge control of energy storage devices and distribution scheduling of each energy source;

[0012] Execute the power backup strategy in real time, and at the same time dynamically adjust and optimize the power backup plan through a feedback algorithm to ensure the stability and reliability of the power backup energy storage power under load changes and energy supply fluctuations.

[0013] Preferably, the specific steps of collecting the operation parameter data of multiple accessed energy sources in real time, constructing a multi-energy input data set, and transmitting the data to the system central control unit for unified processing are as follows:

[0014] Collect multi-energy data through the communication interface of the device accessed by the energy management module;

[0015] After the multi-energy data acquisition is completed, transmit the real-time collected operation parameter data to the system central control unit for unified processing;

[0016] After the data is transmitted to the system central control unit, screen and filter the data;

[0017] After completing data screening and filtering, integrate energy data from different sources to construct a standardized multi-energy input dataset.

[0018] Preferably, the specific steps for preprocessing the collected multi-energy input data to ensure the temporal consistency and numerical accuracy of the data and forming multi-energy power supply status data are as follows:

[0019] First, perform temporal alignment and data synchronization on the collected multi-energy data;

[0020] After temporal alignment is completed, perform data cleaning and outlier removal on the multi-energy input data;

[0021] Perform numerical calibration and consistency adjustment on the processed multi-energy data to unify the measurement standards of different energy data;

[0022] Integrate the multi-energy data that has undergone temporal alignment, data cleaning, and calibration adjustment to construct multi-energy power supply status data.

[0023] Preferably, based on historical load data and the current power supply status, the specific steps for using a prediction algorithm to perform real-time dynamic assessment of the base station load demand and match the power supply capabilities of each energy are as follows:

[0024] First, integrate the real-time data with the base station historical load data;

[0025] Conduct in-depth analysis of the base station's historical load data and current power supply status to extract key load characteristics;

[0026] Based on the extracted load characteristics and the current power supply status data, use a prediction algorithm to perform real-time dynamic assessment of the future load demand of the base station;

[0027] Conduct feedback analysis on the matching situation between the predicted load demand result and the current power supply capacity, evaluate the accuracy of the prediction result, and perform dynamic optimization and adjustment.

[0028] Preferably, through an energy management algorithm, generate a differentiated backup power strategy according to the power supply demand and energy supply status, including charge and discharge control of energy storage devices and distribution and scheduling of each energy. The specific steps are as follows:

[0029] First, based on the predicted load demand value and the current supply status data of each energy, construct a mathematical model of power supply demand and supply status, calculate the total power supply, and the calculation expression is as follows:

[0030]

[0031] Where P total (t) is the total power supply of all energies at time t, P i$(t)$ is the real-time output power of the $i$-th type of energy at time $t$, and $n$ is the total number of energies;

[0032] Define the power supply gap, and the calculation expression is as follows:

[0033] $P$ gap $(t)=P$ demand (t) - $P$ total (t)

[0034] In the formula, $P$ gap (t) is the power supply gap at time $t$, and $P$ demand (t) is the predicted value of the load demand at time $t$;

[0035] After the power supply gap $P$ gap (t) is determined, perform charge and discharge control on the energy storage device to optimize the charge and discharge constraint relationship of the energy storage device as follows:

[0036] $0\leq E_s$ tored (t) + $P$ storage (t)·$\Delta t\leq E$ max

[0037] , where $E$ stored (t) is the current remaining power of the energy storage device, $P$ storage (t) is the charge and discharge power of the energy storage device, and $E$ max is the maximum energy storage capacity, and $\Delta t$ is the time interval;

[0038] After the charge and discharge power $P$ storage (t) of the energy storage device is determined, optimize the power distribution of various types of energy for dispatching, and the calculation expression is as follows:

[0039]

[0040] In the formula, $C$ total is the total power supply cost, $\min C$ total is the minimum total power supply cost, and $C$ i is the unit power supply cost of the $i$-th type of energy;

[0041] After completing the optimization of energy distribution and charge and discharge of the energy storage device, generate a differentiated backup power strategy according to the results: the charge and discharge control instructions of the energy storage device and the distribution power instructions of each energy

[0042] Dynamically adjust the strategy in real time through the following formula to ensure power supply stability:

[0043]

[0044] In the formula, $P$ final (t) is the final power supply power;

[0045] Continuously monitor the power supply status and load demand through a feedback mechanism. If the difference between the final power supply power P final (t) and the predicted value of the load demand P demand (t) exceeds the threshold ∈, then re - execute the above steps and dynamically adjust the strategy:

[0046]

[0047] ∈ is the allowable power supply deviation threshold.

[0048] Preferably, the backup power strategy is executed in real - time. At the same time, the backup power plan is dynamically adjusted and optimized through a feedback algorithm to ensure the stability and reliability of the backup power storage capacity under load changes and energy supply fluctuations. The specific steps are as follows:

[0049] After generating the differentiated backup power strategy, execute the backup power strategy in real - time, calculate the output power of each energy source and the discharge power of the energy storage device to ensure meeting the current load demand. The expression is as follows:

[0050] P liad (t) = P pv (t)+P wind (t)+P grid (t)+P bat (t)

[0051] In the formula, P load (t) is the base station load demand at time t, P pv (t) is the photovoltaic power at time t, P wind (t) is the wind turbine output power at time t, P grid (t) is the mains power supply power at time t, P bat (t) is the energy storage discharge power at time t;

[0052] At the same time, in order to prevent over - discharge of the energy storage battery, a minimum energy storage capacity limit is introduced. The calculation expression is as follows:

[0053] P bat (t)·Δt ≤ E bat (t - 1)-E min

[0054] In the formula, Δt is the time interval, E min is the safety threshold of the energy storage capacity, E bat (t - 1) is the remaining power of the energy storage device at the previous moment;

[0055] Continuously monitor the difference between the current load demand and the actual power supply, calculate the power supply deviation, and judge the execution effect of the current backup power strategy. The calculation expression is as follows:

[0056] ΔP error (t) = Psupply (t)-P load (t)

[0057] Wherein, ΔP error (t) is the power supply deviation, and P supply (t) is the current total power supply. The calculation expression is as follows: P supply (t) = P pv (t) + P wind (t) + P grid (t) + P bat (t);

[0058] Based on the real-time power supply deviation ΔP error (t) and the current remaining power of the energy storage, dynamically adjust the discharge power of the energy storage device to ensure energy supply balance and stable energy storage power. The dynamic adjustment expression of the energy storage discharge power is as follows:

[0059]

[0060] Wherein, P bat (t + 1) is the discharge power of the energy storage device at the next moment, K p is the proportional adjustment coefficient, K i is the integral adjustment coefficient, ∑ΔP error (t) is the cumulative value of the real-time power supply deviation, and E bat (t) is the current remaining power of the energy storage;

[0061] After performing the dynamic adjustment, adaptively optimize the backup power plan through the feedback algorithm to improve the system response speed and stability. The calculation expression is as follows:

[0062]

[0063] Wherein, J is the system cost function, P bat (t - 1) is the energy storage discharge power at the next moment t - 1, λ is the adjustment smoothing coefficient, (P bat (t) - P bat (t - 1)) 2 is the cost of the energy storage power fluctuation, and T is the total time interval.

[0064] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0065] Through multi - information fusion technology and differential backup power control strategy, the system effectively solves the misjudgment problem caused by inconsistent source data of multiple energy access, and significantly improves the power supply stability. In the case of energy fluctuations, for example, when the power generation of photovoltaic power is insufficient due to weather changes, the system can perceive the change of load demand in advance through real - time load prediction, and dynamically adjust the discharge power of energy storage devices to compensate for the power supply gap. This rapid response mechanism can effectively prevent the base station from interrupting operation due to power shortage during the peak load period, and ensure the continuity of the communication system. In addition, by optimizing the scheduling of each energy source, the system ensures that the power of the energy storage device always maintains at a reasonable level, so that the base station can also maintain the normal operation of core functions under complex operation scenarios (such as sudden load fluctuations or external power supply failures).

[0066] Through intelligent prediction algorithms and energy management technologies, the present invention optimizes the power supply scheduling among wind turbines, photovoltaics, commercial power, and energy storage devices, greatly improving the utilization rate of clean energy, while reducing the dependence on commercial power and unnecessary charging and discharging of energy storage devices. This energy distribution optimization strategy realizes peak shaving and valley filling. It preferentially releases the stored energy of the energy storage device during the peak electricity consumption period and uses clean energy for charging during the valley period, which not only reduces the operating energy consumption, but also reduces the excessive loss of the energy storage device and prolongs the device life. In addition, through the dynamic pre - processing and unified analysis of multi - energy data, the system matches the load demand and power supply capacity in advance, avoiding energy waste and ineffective allocation, and further reducing the operating cost and improving the economic benefits while ensuring reliable power supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required to be used in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0068] Figure 1 It is a method flow chart of the differential backup power control method for a hybrid energy storage system based on multi - information fusion according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0069] Now, the exemplary embodiments will be described more fully with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided so that this disclosure will be more complete and comprehensive, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0070] The present invention provides a differential backup power control method for a hybrid energy storage system based on multi - information fusion as shown in Figure 1 and includes the following steps:

[0071] Real-time collect the operation parameter data of multiple accessed energy sources, construct a multi-energy input data set, and transmit the data to the central control unit of the system for unified processing;

[0072] The specific steps for real-time collecting the operation parameter data of multiple accessed energy sources, constructing a multi-energy input data set, and transmitting the data to the central control unit of the system for unified processing are as follows:

[0073] Collect multi-energy data through the communication interface of the device accessed by the energy management module;

[0074] In this process, multiple energy sources including wind turbines, photovoltaics, mains electricity, and energy storage systems are accessed in real time, and the key operation parameter data of each energy source are obtained respectively, including information such as voltage value, current value, power value, and energy output status. The data acquisition frequency is dynamically adjusted according to system requirements. For example, for the data of photovoltaics and wind turbines with large power fluctuations, the system can set a high-frequency sampling mechanism to ensure that changes in energy supply can be captured in a timely manner; for data sources with stable operation such as mains electricity or energy storage devices, a lower sampling frequency is adopted to reduce data redundancy. At the same time, through the multi-source protocol adaptation function, the system formats the energy data uniformly to ensure the standardized output of the data for subsequent processing.

[0075] After the multi-energy data acquisition is completed, transmit the real-time collected operation parameter data to the central control unit of the system for unified processing;

[0076] During the data transmission process, to ensure the continuity and timeliness of the data, an efficient data transmission protocol (such as Modbus, CAN bus, or wireless communication protocol) is adopted to quickly upload the data to the central control unit by wired or wireless means. In addition, to prevent data loss or delay during transmission, the system is equipped with a caching mechanism at the local data acquisition terminal to temporarily store the data that has not been completed for transmission, and a data retransmission strategy is set to make up for the data loss caused by sudden failures. At the same time, during the transmission process, the system identifies and classifies the data stream to ensure that data from different energy sources can be accurately identified and distinguished by the central control unit, thus avoiding data chaos.

[0077] After the data is transmitted to the central control unit of the system, screen and filter the data;

[0078] Perform timestamp verification on the transmitted multi-energy data, filter out data with poor real-time performance or incomplete data, and prevent abnormal data from affecting the overall analysis. In addition, the system uses data filtering algorithms (such as mean filtering, Kalman filtering, etc.) to smooth the random noise and mutation values in the data, and eliminate data anomalies caused by external interference or transmission errors. For continuously lost data, the system repairs it through interpolation algorithms to ensure data continuity. After screening and filtering, the data already has high accuracy and reliability, laying a solid foundation for the subsequent construction of the multi-energy data set.

[0079] After completing data screening and filtering, integrate energy data from different sources to construct a standardized multi-energy input data set;

[0080] In this process, align the data in time sequence according to the timestamp to ensure that parameters such as voltage, current, and power from different energies are synchronized within the same time period; subsequently, the system stores the data in layers according to energy types (wind turbines, photovoltaic, mains power, energy storage), and uniformly archives it in a structured data format (such as two-dimensional arrays, database tables). In addition, the system also labels the source information and data characteristics for each data set, such as wind turbine output data, photovoltaic power fluctuation data, mains power stability data, etc., so as to provide a clear and complete data basis for subsequent data analysis and decision-making scheduling. Finally, the constructed data set will be stored in the central control unit of the system and has the functions of calling, analyzing, and real-time monitoring to ensure that the system can quickly respond to changes in different energy supplies and achieve efficient management and scheduling of multiple energies.

[0081] Preprocess the collected multi-energy input data to ensure the time sequence consistency and numerical accuracy of the data, and form multi-energy power supply status data;

[0082] The specific steps for preprocessing the collected multi-energy input data to ensure the time sequence consistency and numerical accuracy of the data and form multi-energy power supply status data are as follows:

[0083] First, perform time sequence alignment and data synchronization on the collected multi-energy data;

[0084] Specifically, the system uniformly standardizes the timestamps of each data source to eliminate time differences caused by transmission delays. Subsequently, through linear interpolation algorithms or time window overlapping mechanisms, it matches data with a higher sampling frequency to that with a lower frequency. For example, for data with a higher frequency such as photovoltaic and wind turbines, the system takes the average value to reduce it to the standard sampling frequency, while for low-frequency data such as mains power and energy storage, interpolation algorithms are used to supplement missing values, ultimately ensuring that all data sources are consistent on the same time axis. Through time series alignment and data synchronization, it is ensured that energy parameter data can be accurately compared and analyzed at the same time point, forming multi-energy power supply data with consistent time series.

[0085] After the time series alignment is completed, data cleaning and outlier removal are performed on the multi-energy input data;

[0086] Parameter data such as voltage, current, and power that exceed the normal operating range are detected through threshold filtering methods and removed. For example, when the instantaneous voltage fluctuation of the wind turbine output far exceeds the rated value, the system marks it as abnormal data. Secondly, smoothing filtering algorithms (such as Kalman filtering or median filtering) are used to denoise the data and eliminate data instability caused by random fluctuations. In addition, the system also identifies mutation points and duplicate data through trend detection algorithms (such as Z-score or mean square deviation detection) to ensure data continuity and stability. Through data cleaning and outlier removal, the system effectively improves the reliability of the power supply status data, ensuring the authenticity and accuracy of the data relied on for subsequent analysis.

[0087] Perform numerical calibration and consistency adjustment on the processed multi-energy data to unify the measurement standards of different energy data;

[0088] First, perform baseline calibration based on the rated parameters of each energy device. For example, calibrate the rated voltage value of the wind turbine, the power output standard of the photovoltaic system, etc., and remove error data deviating from the baseline. Secondly, standardize the numerical ranges of each energy through proportional scaling algorithms, normalizing data in different dimensions such as voltage, current, and power to the same dimension to ensure data consistency. For example, adjust the power data output by wind turbines, photovoltaics, and mains power to the same unit (such as kilowatts) for subsequent unified analysis. In addition, the system dynamically adjusts the calibration parameters based on historical operation data and real-time feedback to compensate for equipment output deviations caused by environmental factors (such as temperature and light changes), further improving data accuracy and consistency.

[0089] Integrate the multi-energy data that has undergone time series alignment, data cleaning, and calibration adjustment to construct multi-energy power supply status data;

[0090] In this process, the system stores the data in segments according to the time series and classifies and summarizes it based on each energy type. For example, key parameter data such as voltage, current, and power of wind turbines, photovoltaics, mains electricity, and energy storage are stored in layers, and tags (such as timestamps, energy types, data status identifiers, etc.) are added. At the same time, the system uses data mapping technology to associate the power supply data of each energy source with the current load demand to form a "multi-energy - load demand" data mapping table, providing a basis for subsequent power supply capacity analysis and the generation of differentiated backup power strategies. In addition, the system continuously and dynamically updates the processed power supply status data to ensure that the system can reflect the power supply status and change trends of each energy source in real time, providing accurate and reliable data support for the stable operation of the differential backup power control.

[0091] Based on historical load data and the current power supply status, use prediction algorithms to dynamically evaluate the load demand of the base station in real time and match the power supply capabilities of each energy source;

[0092] Based on historical load data and the current power supply status, the specific steps of using prediction algorithms to dynamically evaluate the load demand of the base station in real time and match the power supply capabilities of each energy source are as follows:

[0093] First, integrate the real-time data with the historical load data of the base station;

[0094] The historical load data includes the load change curves of the base station in different time periods (such as daily, weekly, monthly), covering parameters such as voltage, current, power, and load rate; the current power supply status data is the data obtained by real-time collection and preprocessing of the input of multiple energy sources such as wind turbines, photovoltaics, mains electricity, and energy storage, including the power supply capabilities, operating states, and load contribution situations of each energy source. The system matches the two types of data in time series through timestamps to construct a "historical load - current power supply" data set, providing a data basis for subsequent prediction modeling. This step not only continues the results of the multi-energy data preprocessing stage but also provides multi-dimensional and high-accuracy data support for the prediction algorithm.

[0095] Deeply analyze the historical load data and the current power supply status of the base station to extract key load characteristics;

[0096] And identify the changing patterns of the base station load demand. First, using time series analysis methods, conduct trend analysis, periodic analysis, and load peak-valley identification on historical data to extract the electricity consumption patterns of the base station at different time periods (for example, the load is relatively large during the morning and evening peak hours, and relatively small at noon and at night). Second, through clustering analysis or data dimensionality reduction algorithms, classify the historical load data to identify the load characteristics under different climates, seasons, and base station operation modes. In addition, combined with the current power supply status data, the system analyzes the volatility of the power supply capabilities of various energy sources in real time. For example, the power supply fluctuation patterns of photovoltaic power generation affected by light intensity and wind turbines affected by wind speed, providing high-quality input data for the application of the next-step prediction algorithm. This process realizes the intelligent processing and feature extraction of data, enabling the system to more accurately identify the internal patterns of the base station load demand changes.

[0097] Based on the extracted load characteristics and the current power supply status data, conduct real-time dynamic assessment of the future load demand of the base station through a prediction algorithm;

[0098] Specifically, the system adopts a long short-term memory network (LSTM) or an ARIMA time series prediction model, inputs the historical load data and the current power supply status data into the prediction model for multi-dimensional analysis, and real-time predicts the load demand in the future short time periods (such as 15 minutes, 30 minutes, 1 hour). At the same time, the prediction results will be dynamically updated and corrected according to the actual operation situation of the base station to improve the prediction accuracy.

[0099] After obtaining the future load demand, the system further matches the power supply capabilities of various energy sources, analyzes the current power supply status of each energy source, the remaining power of the energy storage device, and its charge and discharge capabilities, and calculates the load contribution ratio that each energy source can provide. The system identifies potential power supply gaps based on the difference between the power supply capabilities and the predicted load, and provides a decision-making basis for the generation of differential backup power strategies. This process realizes the precise matching of the load demand and the power supply capabilities, laying a foundation for the optimal scheduling and stable operation of the system.

[0100] Conduct feedback analysis on the matching situation between the predicted load demand result and the current power supply capabilities, evaluate the accuracy of the prediction results, and conduct dynamic optimization and adjustment;

[0101] By real-time monitoring the deviation between the actual load and the predicted load of the base station, adaptively train the prediction model in combination with the feedback data, continuously optimize the algorithm parameters, and improve the accuracy and real-time performance of the load prediction. In addition, the system dynamically optimizes the power supply capability matching scheme using the feedback results. For example, when the actual load is higher than the predicted value, the system will immediately allocate the energy storage device to increase the discharge power to supplement the power supply gap; when the load is lower than the predicted value, the system will reduce the output of the energy storage device to avoid energy waste.

[0102] Through feedback and dynamic optimization, the system forms a closed-loop control mechanism of "prediction - evaluation - feedback - optimization", continuously improving the reliability of load prediction and the accuracy of power supply capacity matching, and ensuring the stability of base station power supply and the adequacy of backup power energy storage.

[0103] Through the energy management algorithm, according to the power supply demand and energy supply status, generate differentiated backup power strategies, including charge and discharge control of energy storage devices and distribution and scheduling of various energy sources;

[0104] Through the energy management algorithm, according to the power supply demand and energy supply status, the specific steps for generating differentiated backup power strategies, including charge and discharge control of energy storage devices and distribution and scheduling of various energy sources, are as follows:

[0105] First, based on the load demand prediction value and the current supply status data of each energy source, construct a mathematical model of power supply demand and supply status, calculate the total power supply, and the calculation expression is as follows:

[0106]

[0107] In the formula, P total (t) is the total power supply of all energy sources at time t, P i (t) is the real-time output power of the i-th type of energy source at time t, and n is the total number of energy sources;

[0108] Define the power supply gap, which is the difference between the demand power and the current total power supply. The calculation expression is as follows:

[0109] P gap (t) = P demand (t) - P total (t)

[0110] In the formula, P gap (t) is the power supply gap at time t. If P gap (t) > 0, it means the power supply is insufficient; if P gap (t) ≤ 0, it means the power supply meets the demand, and P demand (t) is the load demand prediction value at time t;

[0111] The output parameter P gap (t) of this step will be used as an important input basis for subsequent energy distribution and energy storage regulation.

[0112] After the power supply gap P gap (t) is determined, optimize the charge and discharge control of the energy storage device to compensate for insufficient power supply or store excess power. The charge and discharge constraint relationship of the energy storage device is as follows:

[0113] 0 ≤ E stored (t) + P storage(t)·Δt ≤ E max

[0114] where E stored (t) is the current remaining power of the energy storage device, P storage (t) is the charging and discharging power of the energy storage device, E max is the maximum energy storage capacity, and Δt is the time interval;

[0115] Among them, the charging and discharging power P storage (t) of the energy storage device is dynamically calculated according to the power supply gap P gap (t): when P gap (t) > 0 (when the power supply is insufficient): When P gap (t) ≤ 0 (when the power supply is excessive):

[0116] The output parameter P storage (t) of this step will be used as the charging and discharging control instruction of the energy storage device and participate in the calculation in the next energy distribution.

[0117] After the charging and discharging power P storage (t) of the energy storage device is determined, the power distribution of various energy sources is optimized and scheduled to meet the power supply demand to the greatest extent and reduce energy waste. The energy distribution problem can be solved through the minimum cost optimization model, and the calculation expression is as follows:

[0118]

[0119] where C total is the total power supply cost, min C total is the minimum total power supply cost, and C i is the unit power supply cost of the i-th type of energy source;

[0120] Among them, the power supply power constraint condition is: and it satisfies the power constraint of each energy source output: 0 ≤ P i (t) ≤ P max，i , where P max，i is the maximum output power of the i-th type of energy source;

[0121] By solving the above model, the system can preferentially allocate energy sources with lower costs (such as photovoltaic and wind turbines) on the premise of meeting the power supply demand, and output the power distribution P i (t) of each energy source in the result for the generation of the next backup power strategy.

[0122] After completing the optimization of energy distribution and charge-discharge of energy storage devices, differential backup power strategies are generated based on the results: charge-discharge control instructions for energy storage devices and distribution power instructions for each energy source

[0123] The strategy is dynamically adjusted in real time through the following formula to ensure power supply stability:

[0124]

[0125] In the formula, P final (t) is the final power supply;

[0126] Through the feedback mechanism, the power supply status and load demand are continuously monitored. If the difference between the final power supply P final (t) and the predicted value of the load demand P demand (t) exceeds the threshold ∈, then the above steps are re-executed to dynamically adjust the strategy:

[0127]

[0128] ∈ is the allowable power supply deviation threshold.

[0129] Through the above steps, the optimal management and scheduling of energy are realized, accurate differential backup power strategies are generated, the coordination and efficiency of charge-discharge control of energy storage devices and the distribution of each energy source are ensured, the load demand of the base station is met, and the power supply stability is improved.

[0130] The backup power strategy is executed in real time, and at the same time, the backup power plan is dynamically adjusted and optimized through the feedback algorithm to ensure the stability and reliability of the backup energy storage power under the conditions of load changes and energy supply fluctuations;

[0131] The specific steps to execute the backup power strategy in real time and dynamically adjust and optimize the backup power plan through the feedback algorithm to ensure the stability and reliability of the backup energy storage power under the conditions of load changes and energy supply fluctuations are as follows:

[0132] After generating the differential backup power strategy, the backup power strategy is executed in real time, and the output power of each energy source and the discharge power of the energy storage device are calculated to ensure that the current load demand is met. The expression is as follows:

[0133] P liad (t) = P pv (t) + P wind (t) + P grid (t) + P bat (t)

[0134] In the formula, P load (t) is the base station load demand at time t, P pv (t) is the photovoltaic power at time t, P windP(t) is the output power of the wind turbine at time t grid P(t) is the power supplied by the mains at time t bat P(t) is the power of energy storage discharging at time t;

[0135] Meanwhile, in order to prevent over-discharge of the energy storage battery, a minimum energy storage power limit is introduced, and the calculation formula is as follows:

[0136] P bat (t)·Δt ≤ E bat (t - 1) - E min

[0137] In the formula, Δt is the time interval, and E min is the safety threshold of the energy storage power, and E bat (t - 1) is the remaining power of the energy storage device at the previous moment;

[0138] When implementing the backup power strategy, the output power of clean energy such as photovoltaic and wind turbines is preferentially allocated, and the remaining demand is supplemented by the mains and the energy storage system to ensure power supply stability and efficient use of energy.

[0139] The difference between the current load demand and the actual power supply is monitored in real time, and the power supply deviation is calculated to judge the execution effect of the current backup power strategy. The calculation formula is as follows:

[0140] ΔP error (t) = P supply (t) - P load (t)

[0141] In the formula, ΔP error (t) is the power supply deviation. A positive value indicates an excess of power supply, and a negative value indicates a shortage of power supply. P supply (t) is the current total power supply. The calculation formula is as follows: P supply (t) = P pv (t) + P wind (t) + P grid (t) + P bat (t);

[0142] The ΔP error (t) monitored in real time is fed back to the central control unit, and the cumulative deviation value ∑ΔP error (t) is recorded to provide input data for dynamic adjustment. This parameter can be used to detect the degree of imbalance between system supply and demand and provide a decision-making basis for optimizing scheduling.

[0143] Based on the real-time power supply deviation ΔP error (t) and the current remaining power of the energy storage, the discharging power of the energy storage device is dynamically adjusted to ensure energy supply balance and stable energy storage power. The dynamic adjustment expression of the energy storage discharging power is as follows:

[0144]

[0145] Wherein, P bat (t + 1) is the discharge power of the energy storage device at the next moment, K p is the proportional adjustment coefficient, reflecting the influence degree of the load deviation, K i is the integral adjustment coefficient, used to smooth the influence of long-term deviation, ∑ΔP error (t) is the cumulative value of the real-time power supply deviation, E bat (t) is the remaining power of the current energy storage;

[0146] This formula is a dynamic adjustment model based on proportional-integral control (PI control). By adjusting the energy storage discharge power in real time, it compensates for the problems of insufficient or excessive power supply, ensuring the balance of energy supply and demand in the system. In addition, the system dynamically verifies the adjusted energy storage power to prevent the energy storage device power from being lower than the safety threshold E min .

[0147] After performing the dynamic adjustment, the backup power plan is adaptively optimized through the feedback algorithm to improve the system response speed and stability. The calculation expression is as follows:

[0148]

[0149] Wherein, J is the system cost function, measuring the smoothness of the load deviation and the energy storage power adjustment, P bat (t - 1) is the energy storage discharge power at the next moment t - 1, λ is the adjustment smoothness coefficient, balancing the load deviation and the volatility of the energy storage adjustment, (P bat (t) - P bat (t - 1)) 2 is the cost of the energy storage power fluctuation, and T is the total time interval.

[0150] Through the cost function J, the feedback algorithm adaptively adjusts the proportional-integral coefficients K p and K i , optimizes the energy storage adjustment strategy, and ensures the power supply stability and the efficient management of the energy storage power. The finally optimized parameter values will be fed back to the power distribution link of the next cycle to form a closed-loop control and continuously improve the system performance.

[0151] Embodiment 1: Load demand prediction and energy scheduling optimization based on long short-term memory network (LSTM)

[0152] This embodiment realizes the high-precision prediction of the base station load demand through the long short-term memory network (LSTM) algorithm, and combines the multi-energy power supply status data to dynamically generate the optimal energy scheduling plan to ensure the accurate matching of the system power supply and the load demand.

[0153] First, in order to make load predictions, the system uses historical load data and current power supply status data as input sources. Historical load data includes long-term operation information of base stations, such as daily, weekly, and monthly load changes, including voltage, current, power, peak load, valley load and other data; while current power supply status data is obtained by real-time monitoring of wind turbines, photovoltaics, city power and energy storage systems, such as voltage, current, output power, and energy storage capacity. The system uses the aforementioned data preprocessing methods to clean and filter the input data, remove outliers and noise, and standardize and normalize the data to ensure data quality and consistency, providing high-quality input data for the LSTM model.

[0154] The long short-term memory network (LSTM) is a recurrent neural network designed for time series data, which is suitable for the time series problem of base station load demand prediction. The system first inputs the processed historical load data and the current power supply status data into the LSTM network for training. The LSTM model can capture the long-term dependencies in the data through its unique "memory unit" mechanism, while avoiding the gradient vanishing problem in traditional neural networks. Specifically, the LSTM network is divided into four parts: input gate, forget gate, output gate and memory unit. The input gate determines the weight of the current input information, the forget gate determines which information needs to be discarded, and the output gate outputs the predicted value of the current time step. The system divides the training data, uses 80% of the data for model training, and 20% of the data for verification to ensure the generalization ability of the model. During the training process, the system adjusts the network parameters through the back-propagation algorithm and optimizer (such as Adam optimizer), gradually reduces the error function value, and improves the prediction accuracy.

[0155] After the model training is completed, the system inputs the current power supply status data and some historical load data into the trained LSTM network to perform short-term load demand forecasting. The forecast time period can be flexibly set according to the needs of the base station, such as 15 minutes, 30 minutes, 1 hour, etc. The system outputs the base station load demand value in the future short period of time. In addition, to improve the real-time performance of the forecast, the system will regularly update the LSTM model and use the latest load data for incremental training to ensure that the model always adapts to the changing trend of the base station load, especially when there are seasonal changes, special weather conditions or sudden load fluctuations, it can still maintain a high forecast accuracy.

[0156] Based on the load demand prediction results output by the LSTM model, the system combines the power supply capabilities of current wind turbines, photovoltaics, mains power, and energy storage devices, and uses energy management algorithms (such as linear programming algorithms and minimum cost algorithms) to generate an optimal energy scheduling plan. The system reasonably arranges the output power of each energy source according to the power supply capabilities of each energy source, the remaining power of the energy storage device, and the charge and discharge rates. For example, when photovoltaic power generation is sufficient, it preferentially uses photovoltaic power supply, and at the same time, the energy storage device is charged; when the load demand increases and the photovoltaic or wind turbine power supply is insufficient, the system automatically deploys the energy storage device to discharge to supplement the power supply gap. In addition, the system dynamically optimizes the energy scheduling plan by using feedback control algorithms through real-time monitoring of the load and power supply status, ensuring the power supply stability of the base station and the efficient management of the energy storage device.

[0157] During the execution of the energy scheduling plan, the system adjusts and optimizes in a timely manner by real-time monitoring the deviation between the actual load demand of the base station and the prediction results. When the prediction error is large, the system feeds the actual load data back to the LSTM model to retrain and update the network parameters to improve the prediction accuracy. In addition, the system dynamically adjusts the energy scheduling plan according to the operating status of each energy source and the change of load demand, ensuring stable and reliable power supply, and avoiding the depletion of energy storage power or energy waste.

[0158] Through this embodiment, the system can achieve high-precision load demand prediction and energy scheduling optimization, ensure the dynamic balance between the power supply demand of the base station and the multi-energy supply, and improve the power supply stability and the utilization efficiency of the energy storage device.

[0159] Embodiment 2: Dynamic data acquisition and consistency calibration based on multi-sensor fusion

[0160] This embodiment realizes the dynamic acquisition of multi-energy data through multi-sensor fusion technology, and uses data preprocessing and consistency calibration methods to form high-quality power supply status data, providing reliable data support for system load prediction and energy scheduling.

[0161] In a multi-energy access system, the operating state parameters of devices such as wind turbines, photovoltaics, mains power, and energy storage have time-varying and diverse characteristics. The system collects parameters such as voltage, current, and power in real time by deploying a variety of high-precision sensors. Since there are differences in the data sampling frequencies and transmission paths of different energy sources, the system uses a time synchronization mechanism (such as NTP synchronization technology) to ensure that the timestamps of all data sources are consistent, eliminating the problem of inconsistent data time sequences. In addition, to improve the stability and efficiency of data acquisition, the system designs a multi-level caching mechanism. When data transmission is delayed or lost, the cache module can temporarily store the data, and supplement the missing data through the data retransmission mechanism, ensuring the integrity and continuity of the data.

[0162] The original data collected may be affected by external environmental noise, sensor errors, or communication failures, resulting in data anomalies and noise problems. The system adopts a variety of data denoising methods, including mean filtering, median filtering, and Kalman filtering, etc., to smooth the random noise in the data and eliminate outliers and mutation points. In addition, the system uses data cleaning algorithms to detect and eliminate invalid or duplicate data to ensure the accuracy and consistency of the data.

[0163] To eliminate the amplitude deviation between different data sources, the system performs numerical calibration based on the rated parameters of the equipment. For example, the voltage and power data output by wind turbines, photovoltaic, and energy storage devices are benchmarked and calibrated to ensure that all data can be compared and analyzed under the same measurement standard. At the same time, the system maps multi-energy data to the same numerical range through normalization and standardization methods to form a unified high-quality data set.

[0164] The data after dynamic collection, denoising, and calibration processing will be uniformly stored in the central control unit of the system to form a structured multi-energy power supply status data set. The data set contains voltage, current, power, and power supply status information of each energy source, and is stored in segments according to the time series to ensure that the data can be quickly called and analyzed.

[0165] Through this implementation method, the system realizes the efficient collection and standardized processing of multi-energy data, ensures the time series consistency and numerical accuracy of the data, and provides reliable data support for subsequent load demand prediction and backup power scheduling.

[0166] Embodiment 3: Execution and Optimization of Differentiated Backup Power Control Strategy Based on Energy Management Algorithm

[0167] This embodiment realizes the execution and dynamic optimization of the differentiated backup power control strategy through the energy management algorithm, ensuring stable and reliable power supply for the base station under complex load and energy fluctuation conditions.

[0168] Based on the load demand prediction results and the current power supply status data, the system uses energy management algorithms (such as linear programming algorithms or minimum cost algorithms) to generate differentiated backup power strategies. The strategies include power output control of wind turbines, photovoltaic, mains power, and energy storage devices, as well as charge and discharge plans for energy storage devices, ensuring that the load demand is accurately met.

[0169] The system executes the differentiated backup power strategy to each energy management module in real time, and dynamically allocates the output power of each energy source through the scheduling controller. For example, when the photovoltaic power supply is insufficient, the system automatically increases the discharge power of the energy storage device to compensate for the power supply gap; when the load demand decreases, the system preferentially uses clean energy for charging and energy storage to reduce the use of mains power.

[0170] The system dynamically adjusts the backup power strategy by monitoring the power supply status and load demand changes in real time. When there is a deviation between the load demand and the actual power supply, the system quickly adjusts the output power through a feedback algorithm to ensure the reasonable distribution and stable supply of the backup power storage capacity.

[0171] Through this embodiment, the system realizes the precise execution and dynamic optimization of backup power control, ensuring the power supply stability and energy utilization efficiency of the base station.

[0172] Through the multi-information fusion technology and the differentiated backup power control strategy of the present invention, the system effectively solves the misjudgment problem caused by inconsistent source data of multiple energy accesses, and significantly improves the power supply stability. In the case of energy fluctuations, for example, when the power generation of photovoltaic power is insufficient due to weather changes, the system can perceive the changes in load demand in advance through real-time load prediction, and dynamically adjust the discharge power of energy storage devices to compensate for the power supply gap. This rapid response mechanism can effectively prevent the base station from interrupting operation due to power shortage during the peak load period, and ensure the continuity of the communication system. In addition, the system optimizes the scheduling of each energy source to ensure that the power of the energy storage device always maintains at a reasonable level, so that the base station can also maintain the normal operation of its core functions under complex operating scenarios (such as sudden load fluctuations or external power supply failures).

[0173] Through the intelligent prediction algorithm and energy management technology of the present invention, the power supply scheduling among the fan, photovoltaic power, commercial power and energy storage devices is optimized, greatly improving the utilization rate of clean energy, while reducing the dependence on commercial power and unnecessary charging and discharging of energy storage devices. This energy distribution optimization strategy realizes peak shaving and valley filling, preferentially releasing the energy storage power during the peak electricity consumption period and using clean energy for charging during the valley period, which not only reduces the operating energy consumption, but also reduces the excessive loss of energy storage devices and prolongs the device life. In addition, the system dynamically preprocesses and uniformly analyzes the multi-energy data, matches the load demand and power supply capacity in advance, avoids energy waste and ineffective distribution, and further reduces the operating cost and improves the economic benefits while ensuring reliable power supply.

[0174] Only some exemplary embodiments of the present invention are described by way of illustration above. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. A hybrid energy storage system differential backup power control method based on multi-information fusion, characterized in that: The following steps are involved: Collect operating parameter data of multiple connected energy sources in real time, build a multi-energy input data set, and transmit the data to the system central control unit for unified processing; Pre-process the collected multi-energy input data to ensure the data's time consistency and numerical accuracy, and form multi-energy power supply status data; Based on historical load data and current power supply status, a prediction algorithm is used to dynamically evaluate base station load demand in real time and match the power supply capacity of each energy source; Through the energy management algorithm, differentiated backup power strategies are generated according to power supply demand and energy supply status, including charge and discharge control of energy storage equipment and distribution and scheduling of various energy sources; The backup power strategy is executed in real time, and the backup power plan is dynamically adjusted and optimized through feedback algorithms to ensure the stability and reliability of the backup power storage capacity under load changes and energy supply fluctuations.

2. The hybrid energy storage system differential backup power control method based on multi-information fusion according to claim 1 is characterized in that: The specific steps for real-time collection of operating parameter data of multiple connected energy sources, building a multi-energy input data set, and transmitting the data to the system central control unit for unified processing are as follows: Access the communication interface of the equipment through the energy management module to collect multi-energy data; After the multi-energy data acquisition is completed, the real-time collected operating parameter data is transmitted to the system central control unit for unified processing; After the data is transmitted to the system's central control unit, the data is screened and filtered; After completing data screening and filtering, energy data from different sources are integrated to construct a standardized multi-energy input data set.

3. The hybrid energy storage system differential backup power control method based on multi-information fusion according to claim 1 is characterized in that: The specific steps for preprocessing the collected multi-energy input data to ensure the time consistency and numerical accuracy of the data and form the multi-energy power supply status data are as follows: First, the collected multi-energy data is time-series aligned and data synchronized; After the time series alignment is completed, the multi-energy input data is cleaned and outliers are eliminated; Perform numerical calibration and consistency adjustment on the processed multi-energy data to unify the measurement standards of different energy data; Integrate the multi-energy data after time alignment, data cleaning and calibration adjustment to construct multi-energy power supply status data.

4. The hybrid energy storage system differential backup power control method based on multi-information fusion according to claim 1 is characterized in that: Based on historical load data and current power supply status, the specific steps for using the prediction algorithm to dynamically evaluate the base station load demand in real time and match the power supply capacity of each energy source are as follows: First, the real-time data is integrated with the historical load data of the base station; Conduct in-depth analysis of the base station's historical load data and current power supply status to extract key load characteristics; Based on the extracted load characteristics and current power supply status data, the future load demand of the base station is dynamically evaluated in real time through the prediction algorithm; Feedback analysis is conducted on the matching of the predicted load demand results with the current power supply capacity to evaluate the accuracy of the predicted results and make dynamic optimization adjustments.

5. The hybrid energy storage system differential backup power control method based on multi-information fusion according to claim 1 is characterized in that: Through the energy management algorithm, differentiated backup power strategies are generated according to power supply demand and energy supply status, including the specific steps of charging and discharging control of energy storage equipment and energy allocation and scheduling as follows: First, based on the load demand forecast value and the current supply status data of each energy source, a mathematical model of power supply demand and supply status is constructed to calculate the total power supply power. The calculation expression is as follows: Where P total (t) is the total power supply of all energy sources at time t, P i (t) is the real-time output power of the i-th energy source at time t, and n is the total number of energy sources; Define the power supply gap, and the calculation expression is as follows: P gap (t)=P demand (t)-P total (t) Where P gap (t) is the power supply gap at time t, P demand (t) is the load demand forecast value at time t; In the power supply gap P gap (t) After determination, the charge and discharge control optimization of the energy storage device is performed, and the charge and discharge constraint relationship of the energy storage device is as follows: 0≤E stored (t)+P storage (t)·Δt≤E max In the formula, E stored (t) is the current remaining power of the energy storage device, P storage (t) is the charging and discharging power of the energy storage device, E max is the maximum energy storage capacity, Δt is the time interval; The charging and discharging power P of the energy storage device storage After (t) is determined, the power allocation of various energy sources is optimized and scheduled. The calculation expression is as follows: In the formula, C total is the total power supply cost, min C total is the minimum total power supply cost, C i is the unit power supply cost of the i-th energy source; After completing the optimization of energy distribution and energy storage device charging and discharging, generate differentiated backup power strategies based on the results: charging and discharging control instructions for energy storage devices and power distribution instructions for each energy source The strategy is adjusted dynamically in real time through the following formula to ensure power supply stability: Where P final (t) is the final power supply; The power supply status and load demand are continuously monitored through the feedback mechanism. If the final power supply power P final (t) and load demand forecast value P demand (t) If the difference exceeds the threshold ∈, re-execute the above steps and dynamically adjust the strategy: ∈ is the allowable power supply deviation threshold.

6. The hybrid energy storage system differential backup power control method based on multi-information fusion according to claim 1 is characterized in that: The specific steps for executing the backup power strategy in real time and dynamically adjusting and optimizing the backup power plan through feedback algorithms to ensure the stability and reliability of backup power storage under load changes and energy supply fluctuations are as follows: After generating the differentiated backup power strategy, the backup power strategy is executed in real time to calculate the output power of each energy source and the discharge power of the energy storage device to ensure that the current load demand is met. The expression is as follows: P load (t)=P pv (t)+P wind (t)+P grid (t)+P bat (t) Where P load (t) is the base station load demand at time t, P pv (t) is the photovoltaic power at time t, P wind (t) is the wind turbine output power at time t, P grid (t) is the mains power supply power at time t, P bat (t) is the energy storage discharge power at time t; At the same time, in order to prevent the energy storage battery from over-discharging, the minimum energy storage capacity limit is introduced, and the calculation expression is as follows: P bat (t)·Δt≤E bat (t-1)-E min Where Δt is the time interval, E min is the safety threshold of energy storage capacity, E bat (t-1) is the remaining power of the energy storage device at the previous moment; Monitor the difference between the current load demand and the actual power supply in real time, calculate the power supply deviation, and judge the execution effect of the current backup power strategy. The calculation expression is as follows: ΔP error (t)=P supply (t)-P load (t) Where ΔP error (t) is the power supply deviation, P supply (t) is the current total power supply, and the calculation expression is as follows: P supply (t) = P pv (t)+P wind (t)+P grid (t)+P bat (t); Based on real-time power supply deviation ΔP error (t) and the current remaining energy storage capacity, dynamically adjust the discharge power of the energy storage device to ensure balanced energy supply and stable energy storage capacity. The expression for dynamic adjustment of energy storage discharge power is as follows: Where P bat (t+1) is the discharge power of the energy storage device at the next moment, K p is the proportional adjustment coefficient, K i is the integral adjustment coefficient, ∑ΔP error (t) is the real-time power supply deviation cumulative value, E bat (t) is the remaining amount of energy stored at present; After dynamic adjustment, the backup power solution is adaptively optimized through the feedback algorithm to improve the system response speed and stability. The calculation expression is as follows: Where J is the system cost function, P bat (t-1) is the energy storage discharge power at the next moment t-1, λ is the adjustment smoothing coefficient, (P bat (t)-P bat (t-1)) 2 is the cost of energy storage power fluctuation, and T is the total time interval.

Images