Distributed energy storage power system optimization scheduling method and system

Abstract

The invention relates to the technical field of energy storage scheduling, and discloses a distributed energy storage power system optimization scheduling method and system. The method comprises the following steps: collecting system operation data, performing normalization processing and anomaly detection, and predicting 24-hour load demand and new energy output in combination with standardized data; performing cost calculation and constraint analysis according to the prediction data, and generating a day-ahead scheduling plan; the operation state is updated according to the 15-minute cycle, and the output data is adjusted; performing multi-time-scale hierarchical processing on the power fluctuation; based on this, the energy storage capacity is optimally configured and the life characteristic is evaluated; and parameter optimization and updating are realized through operation index monitoring and analysis. According to the invention, through multi-dimensional performance evaluation and parameter optimization, efficient utilization and service life prolonging of the energy storage system are realized.

Description

Technical Field

[0001] This application relates to the field of energy storage scheduling, and particularly to an optimized scheduling method and system for a distributed energy storage power system. Background Art

[0002] With the continuous expansion of the scale of new energy power generation and the increasing complexity of the power system structure, energy storage systems, as important regulation means, are widely used in power grid operation. Energy storage systems have the characteristics of fast response speed and wide regulation range, and play an important role in aspects such as new energy consumption, load regulation, and frequency control. Currently, most energy storage systems adopt a centralized deployment method and participate in power grid operation through a unified scheduling control strategy. The capacity configuration and life management of energy storage systems have also received more and more attention, and a large amount of research work on the optimized configuration and life assessment of energy storage systems has been carried out in the industry.

[0003] However, the existing energy storage system scheduling methods have obvious deficiencies: First, the configuration of energy storage capacity lacks in-depth analysis of the characteristics of regional power grids and often uses empirical formulas for simple calculations, resulting in a mismatch between the configuration of the energy storage system and the actual demand; Second, the scheduling control strategy of the energy storage system is too crude and does not fully consider the impact of the charge and discharge process on the battery life, easily causing excessive loss of energy storage equipment; Third, the performance evaluation method of the energy storage system is single and lacks systematic analysis of key parameters such as charge and discharge depth and cycle times, making it difficult to accurately predict the life characteristics of energy storage equipment; Finally, the optimization of the operating parameters of the energy storage system lacks an effective feedback mechanism and cannot dynamically adjust the control strategy according to the actual performance state of the equipment. Summary of the Invention

[0004] This application provides an optimized scheduling method and system for a distributed energy storage power system, which is used to achieve the efficient utilization and life extension of the energy storage system through multi-dimensional performance evaluation and parameter optimization.

[0005] In a first aspect, the present application provides an optimized scheduling method for a distributed energy storage power system. The optimized scheduling method for the distributed energy storage power system includes: collecting real-time operation data of the system through a data acquisition network, performing normalization processing and outlier detection on the collected operation data to obtain standardized data, and performing a negative system prediction with a 24-hour cycle based on the standardized data to generate a system prediction data set; calculating costs and performing data statistics on the fuel consumption of thermal power units, the power generation of new energy, and the charge and discharge amounts of the energy storage system according to the system prediction data set, and outputting day-ahead scheduling plan data through constraint condition screening and power balance analysis; based on the day-ahead scheduling plan data, dynamically updating the system operation state data through sub-period processing with a 15-minute cycle, and adjusting the output data of the power generation unit according to the power deviation calculation result to form real-time scheduling data; according to the real-time scheduling data, performing hierarchical processing of the system power fluctuation at the hourly, minute-level, and second-level, and performing power compensation calculation according to the charge and discharge characteristics of the energy storage system to obtain multi-time scale control data; using the multi-time scale control data to perform regional allocation calculation on the energy storage system capacity, and performing life assessment through statistical analysis of the charge and discharge depth and cycle times to generate energy storage optimization data; based on the energy storage optimization data, performing real-time monitoring and data accumulation on the system operation indicators, and correcting and updating the scheduling parameters through index comparison analysis to obtain system optimization data.

[0006] In a second aspect, the present application provides an optimized scheduling system for a distributed energy storage power system. The optimized scheduling system for the distributed energy storage power system includes: A detection module, configured to collect real-time operation data of the system through a data acquisition network, perform normalization processing and outlier detection on the collected operation data to obtain standardized data, and perform a system prediction with a 24-hour cycle based on the standardized data to generate a system prediction data set; A statistics module, configured to calculate costs and perform data statistics on the fuel consumption of thermal power units, the power generation of new energy, and the charge and discharge amounts of the energy storage system according to the system prediction data set, and output day-ahead scheduling plan data through constraint condition screening and power balance analysis; An update module, configured to dynamically update the system operation state data through sub-period processing with a 15-minute cycle based on the day-ahead scheduling plan data, and adjust the output data of the power generation unit according to the power deviation calculation result to form real-time scheduling data; A hierarchical module, configured to perform hierarchical processing of the system power fluctuation at the hourly, minute-level, and second-level according to the real-time scheduling data, and perform power compensation calculation according to the charge and discharge characteristics of the energy storage system to obtain multi-time scale control data; An allocation module, configured to use the multi-time scale control data to perform regional allocation calculation on the energy storage system capacity, conduct life assessment through statistical analysis of charge-discharge depth and cycle times, and generate energy storage optimization data; A correction module, configured to, based on the energy storage optimization data, perform real-time monitoring and data accumulation on system operation indicators, correct and update scheduling parameters through index comparison analysis, and obtain system optimization data.

[0007] In the technical solution provided by this application, through regional division of multi-time scale control data, accurately calculate the power fluctuation characteristics of each region, and combine with regional power characteristic data to determine the reasonable allocation of energy storage capacity, effectively solving the problem of unreasonable energy storage system capacity configuration; adopt depth statistical data and cycle statistical data to systematically analyze the charge-discharge process, establish a complete performance evaluation system for energy storage units, and realize accurate prediction of the life characteristics of energy storage devices; dynamically optimize the operation parameters of the energy storage system based on performance evaluation data, form a closed-loop regulation and control mechanism, and significantly improve the utilization efficiency of the energy storage system; through a hierarchical and zonal data processing method, achieve precise matching and allocation of energy storage capacity, avoiding waste of energy storage resources; conduct life assessment based on charge-discharge depth and cycle times, establish a health management mechanism for the energy storage system, and extend the service life of the equipment; use real-time monitoring and data accumulation means to continuously track and optimize the adjustment of system operation indicators, ensuring the long-term stable operation of the energy storage system. The technical solution of the present invention, through multi-dimensional data analysis and refined parameter control, not only improves the regulation performance of the energy storage system, but also realizes effective management of the equipment life. Description of the Drawings

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0009] Figure 1 It is a schematic diagram of an embodiment of the optimized scheduling method for a distributed energy storage power system in an embodiment of this application; Figure 2 It is a time series diagram of standardized data obtained by normalizing the collected operation data and detecting outliers; Figure 3 It is a schematic diagram of the process for calculating costs and conducting data statistics in an embodiment of this application; Figure 4 It is a schematic diagram of an embodiment of the optimized scheduling system for a distributed energy storage power system in an embodiment of this application. Detailed Embodiments

[0010] An embodiment of the present application provides an optimized scheduling method and system for a distributed energy storage power system. Terms such as "first", "second", "third", "fourth", etc. (if any) in the specification, claims, and the above-mentioned drawings of the present application are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the term "including" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0011] For ease of understanding, the specific process of the embodiment of the present application will be described below. Please refer to Figure 1 , an embodiment of the optimized scheduling method for the distributed energy storage power system in the embodiment of the present application includes: Step S101: Collect the real-time operation data of the system through a data acquisition network, perform normalization processing and outlier detection on the collected operation data to obtain standardized data, and perform system prediction for a 24-hour cycle based on the standardized data to generate a system prediction data set; Step S102: Calculate the costs and perform data statistics on the fuel consumption of thermal power units, new energy power generation, and the charge and discharge amount of the energy storage system according to the system prediction data set. After screening through constraint conditions and power balance analysis, output the day-ahead scheduling plan data; Step S103: Based on the day-ahead scheduling plan data, perform dynamic update on the system operation status data through 15-minute period sub-period processing, and adjust the output data of the power generation unit according to the power deviation calculation result to form real-time scheduling data; Step S104: According to the real-time scheduling data, perform hierarchical processing on the system power fluctuation at the hourly, minute-level, and second-level, and perform power compensation calculation according to the charge and discharge characteristics of the energy storage system to obtain multi-time scale control data; Step S105: Use the multi-time scale control data to perform regional allocation calculation on the capacity of the energy storage system, and perform life assessment through statistical analysis of the charge and discharge depth and cycle times to generate energy storage optimization data; Step S106: Based on the energy storage optimization data, perform real-time monitoring and data accumulation on the system operation indicators, and correct and update the scheduling parameters through index comparison analysis to obtain system optimization data.

[0012] It is understandable that the execution entity of this application can be an optimized dispatching system for a distributed energy storage power system, or it can also be a terminal or a server. Specifically, no limitation is imposed here. In the embodiments of this application, the server is taken as an example of the execution entity for illustration.

[0013] Specifically, system operation data is collected through a data acquisition network. This data includes wind farm output data (wind turbine speed, wind wheel torque, output power), photovoltaic power station power generation data (light intensity, component temperature, output power), thermal power unit operation parameter data (boiler pressure, steam turbine speed, generator power), energy storage system state parameter data (state of charge SOC, charge and discharge power, operating temperature), and load demand data (industrial electricity load, commercial electricity load, domestic electricity load). These raw data are normalized to unify data with different dimensions to the [0, 1] interval, facilitating subsequent processing. Outlier detection adopts the 3σ principle, calculates the mean and standard deviation of the data, marks the data outside the range of the mean ± 3 times the standard deviation as outliers and eliminates them, generating standardized data. A sliding window analysis with a 24-hour period is performed on the standardized data. The window length is set to 24 hours and the step size is 1 hour to extract the data features within each window. Load demand prediction is based on the time series characteristics of historical load data, analyzes the load change patterns on weekdays and weekends, and combines environmental factors such as temperature and humidity to predict the load change trend in the next 24 hours. New energy output prediction is based on the wind speed data of the wind farm and the light intensity data of the photovoltaic power station, analyzes the output characteristics, predicts the power generation in the next 24 hours, and generates a system prediction data set.

[0014] After obtaining the system prediction data set, the fuel consumption costs of the thermal power units are calculated, including start-stop costs and operating costs. The new energy power generation is calculated based on the predicted wind power and photovoltaic output data, and the charge and discharge amount of the energy storage system is calculated based on its charge and discharge efficiency and capacity constraints. The data statistics process includes calculating the output distribution, fluctuation characteristics, and correlation analysis of various power generation equipment. Constraint condition screening considers system power balance constraints (total power generation on the power generation side is equal to the total demand on the load side) and equipment operation constraints (such as the ramp rate constraint and output upper and lower limit constraints of thermal power units), and solves the optimal solution through linear programming methods to output the day-ahead dispatching plan data. Based on the day-ahead dispatching plan data, the 24-hour period is divided into 96 15-minute time slots, and the system operation state is dynamically updated before the start of each time slot. By collecting real-time operation data, the power deviation from the planned value is calculated, and according to the magnitude and change trend of the deviation value, the output of the power generation unit is adjusted. The adjustment strategy gives priority to the fast response characteristics of the energy storage system, followed by the output adjustment of new energy power generation, and then the output adjustment of thermal power units, forming real-time dispatching data.

[0015] Perform time-scale partitioning on real-time scheduling data for multi-time-scale data. At the hourly level, mainly handle the economic scheduling of thermal power units. At the minute level, handle the suppression of new energy output fluctuations. At the second level, handle the rapid response of the system frequency. Calculate the power compensation values for different time scales based on these parameters, and then calculate the power compensation values for different time scales. Coordinate the power control at each level to obtain multi-time-scale control data. The regional allocation of the energy storage system capacity is based on the multi-time-scale control data. Analyze the power fluctuation characteristics of each region and calculate the required energy storage capacity. Statistically analyze the charge and discharge depth during the operation of the energy storage system, record the depth value and duration of each charge and discharge cycle, and the cycle count statistics include the accumulation of complete charge and discharge cycles and partial charge and discharge cycles. These data are used to evaluate the life characteristics of the energy storage unit and generate energy storage optimization data.

[0016] Monitor the system operation indicators in real time, including power balance accuracy (the deviation between the planned power and the actual power), system frequency quality (the root mean square value of the frequency deviation), and new energy consumption rate (the ratio of the actual power generation of new energy to the theoretical power generation). By establishing an indicator database, conduct a comparative analysis of historical data, find the optimization space for system operation, dynamically correct and update the scheduling parameters, and obtain system optimization data.

[0017] For example, when performing data processing, the power data of the wind farm is normalized by dividing the original power value by the rated power value. For a 100MW wind farm, if the actual power at a certain moment is 60MW, the normalized value is 0.6. Perform outlier detection through the 3σ principle. Calculate the mean (such as 0.5) and standard deviation (such as 0.1) of the 24-hour data, and mark the data outside the range of [0.2, 0.8] as outliers. For the output data of multiple wind turbines, first calculate the correlation coefficient matrix, identify the wind turbine groups with similar output characteristics, and perform data fusion. During the charge and discharge process of the energy storage system, record the SOC change curve and statistically analyze the charge and discharge depth distribution, such as the proportion of the number of cycles with a depth of 80% and the proportion of the number of cycles with a depth of 50%. These data are directly related to the life assessment of the energy storage unit. During the monitoring of system operation indicators, the power balance accuracy is obtained by calculating the deviation between the planned power and the actual power, the frequency quality is obtained by sampling and statistically analyzing the fluctuations of the system frequency, and the new energy consumption rate needs to calculate the theoretical power generation in combination with meteorological data and compare it with the actual power generation.

[0018] In the embodiments of the present application, by dividing the multi-time scale control data into regions, accurately calculating the power fluctuation characteristics of each region, and determining the reasonable allocation of the energy storage capacity in combination with the regional power characteristic data, the problem of unreasonable capacity configuration of the energy storage system is effectively solved; the charging and discharging process is systematically analyzed using depth statistical data and cyclic statistical data, a complete performance evaluation system for energy storage units is established, and the accurate prediction of the life characteristics of energy storage devices is realized; the operation parameters of the energy storage system are dynamically optimized based on the performance evaluation data, a closed-loop regulation and control mechanism is formed, and the utilization efficiency of the energy storage system is significantly improved; through a hierarchical and zonal data processing method, the accurate matching and allocation of the energy storage capacity are realized, and the waste of energy storage resources is avoided; the life assessment is carried out according to the charge and discharge depth and the number of cycles, a health management mechanism for the energy storage system is established, and the service life of the equipment is extended; by using real-time monitoring and data accumulation means, the system operation indicators are continuously tracked and optimized, ensuring the long-term stable operation of the energy storage system. The technical solution of the present invention not only improves the regulation performance of the energy storage system through multi-dimensional data analysis and refined parameter control, but also realizes the effective management of the equipment life.

[0019] In a specific embodiment, the process of executing step S101 may specifically include the following steps: (1) Collect the wind farm output data, photovoltaic power station generation data, thermal power unit operation parameter data, energy storage system state parameter data, and load demand data respectively according to the data acquisition network, and generate the system real-time operation data; (2) Perform maximum-minimum normalization calculation on the system real-time operation data, detect and eliminate outliers through the 3σ criterion, and generate the standardized data; (3) Divide the standardized data into sliding windows with a 24-hour period, perform time series analysis on the data within each time window, and generate the time series feature data; (4) Extract the load change law and new energy output fluctuation characteristics from the time series feature data, perform data dimensionality reduction processing, and generate the feature data set; (5) Perform correlation analysis on the feature data set, fuse the highly correlated data, and generate the fused data set; (6) Perform prediction analysis on the fused data set according to the 24-hour period, predict and calculate the load demand and new energy output, and generate the system prediction data set.

[0020] Specifically, such as Figure 2As shown in the figure, the time series diagram of the normalized data obtained by normalizing the collected operating data and detecting outliers shows the complete process of data processing and prediction in the distributed energy storage power system. First, the operating data of various power equipment are collected through the data acquisition network, including wind farm output data, photovoltaic power station power generation data, thermal power unit operating parameter data, energy storage system state parameter data and load demand data. The collected raw data is normalized by the maximum and minimum values, and the data of different dimensions and orders of magnitude are uniformly mapped to the [0,1] interval. Then, the 3σ criterion is used for outlier detection, the mean μ and standard deviation σ of the data set are calculated, and outliers beyond the range of [μ-3σ,μ+3σ] are eliminated to generate a standardized data set. The standardized data is divided into sliding windows with a period of 24 hours, the window length is fixed to 24 hours, and the sliding step is 1 hour. The data in each time window is subjected to time series analysis, and statistical features (mean, variance, peak, valley), time domain features (trend, periodicity, randomness) and frequency domain features (main frequency components) are extracted to generate time series feature data. Based on the time series feature data, the load change law and the fluctuation characteristics of new energy output are analyzed, and the principal component analysis method is used for dimensionality reduction to generate a feature data set. The feature data set is subjected to correlation analysis, the Pearson correlation coefficient is calculated, and the high correlation features with an absolute value of the correlation coefficient greater than 0.8 are fused, and the fused data set is generated using the weighted average method. Finally, a 24-hour cycle forecast analysis is performed based on the fused data set, and the system forecast data set is generated through the forecast calculation formula by combining the historical data characteristics, environmental factors characteristics, and operating status characteristics of the same period.

[0021] Among them, a data collection network is established to collect the operating data of various types of power equipment. The wind farm output data includes parameters such as wind turbine speed, wind wheel torque, and real-time power output. The photovoltaic power station power generation data includes parameters such as light intensity, component temperature, and actual power generation. The thermal power unit operation parameter data includes parameters such as boiler steam pressure, turbine speed, and generator real-time power. The energy storage system status parameter data includes parameters such as state of charge SOC value, charge and discharge power, and system temperature. The load demand data includes industrial power load, commercial power load, and residential power load. These data are collected according to a unified data format and sampling period to form an original data set.

[0022] The maximum and minimum values ​​of the collected raw data are normalized and the data of different dimensions and magnitudes are uniformly mapped to the interval [0,1]. The normalization formula is:

[0023] in, represents the normalized data value, X represents the original data value, represents the minimum value in the data set. Represents the maximum value in the data set, where for each type of data (such as wind farm output, photovoltaic power generation, load demand, etc.), the maximum possible value in its historical data is selected as the maximum value. For example, the rated capacity of a wind farm, the installed capacity of a photovoltaic power station, etc. Based on historical data, the minimum possible value of each type of data is selected as the minimum value. For power output data, the minimum value is usually 0. The normalized data is used for outlier detection, using the 3σ criterion, that is, calculating the mean μ and standard deviation σ of the data set, marking the data points outside the range of [μ-3σ,μ+3σ] as outliers and removing them to generate a standardized data set.

[0024] The standardized data is divided into sliding windows according to a 24-hour cycle. The window length is fixed to 24 hours and the sliding step is 1 hour, thus forming a series of continuous data windows. The data in each window is analyzed in time series, and statistical features (mean, variance, peak, valley), time domain features (trend, periodicity, randomness) and frequency domain features (main frequency components) are extracted to generate time series feature data for each window. The load variation law is analyzed according to the time series feature data, including the intraday load curve characteristics, the load difference between working days and rest days, and the seasonal load variation patterns. At the same time, the fluctuation characteristics of new energy output are analyzed, including the wind power fluctuation cycle and the daily variation law of photovoltaic power generation. These feature data are processed by dimensionality reduction, and the main feature vectors are extracted by principal component analysis method to remove redundant information and generate the feature data set after dimensionality reduction. The correlation analysis is performed on the various indicators in the feature data set, and the Pearson correlation coefficient is calculated to identify the features with high correlation. The features with the absolute value of the correlation coefficient greater than 0.8 are fused. The weighted average method is used to merge the highly correlated features. The weight coefficient is determined according to the amount of information of each feature to generate the fused data set.

[0025] Based on the fusion data set, the 24-hour cycle forecast analysis, the load demand and new energy output forecast calculation formula is:

[0026] in, represents the predicted value at time t, Indicates the characteristics of historical data for the same period. Indicates the characteristics of environmental factors. Indicates the operating status characteristics. , , is the comprehensive weight coefficient of various features, , , is the weight coefficient of each feature, n, m, Represents the number of each type of features.

[0027] For example: After the power data of a certain wind farm is normalized, the mean value of 0.45 and the standard deviation of 0.15 are calculated within a certain 24-hour window. According to the 3σ criterion, the power value of 0.95 is determined to be an outlier. Perform time series analysis on the 96 sampled points after standardization, and extract the main frequency components as two periods of 4 hours and 12 hours. After dimensionality reduction, retain the principal component features with an explained variance contribution rate exceeding 85%. Correlation analysis finds that the correlation coefficient between wind speed and output power is 0.92, and weights of 0.6 and 0.4 are respectively assigned during data fusion. According to the prediction formula, combined with historical data, meteorological forecasts, and equipment status, calculate the predicted output values for the next 24 hours and generate a system prediction data set.

[0028] In a specific embodiment, the process of executing step S102 may specifically include the following steps: (1) Extract the operating condition data of the thermal power units according to the system prediction data set, calculate the fuel consumption per hour and count the daily cumulative value, divide and accumulate the new energy power generation data according to the time series, and segment and count the charge and discharge data of the energy storage system; (2) Use the fuel consumption of the thermal power units, the new energy power generation, and the charge and discharge volume of the energy storage system to calculate the unit power generation cost of each type of unit and summarize the power generation cost data; (3) Optimally screen the power generation cost data of each type of unit according to the system power balance constraint, and eliminate the data combinations that do not meet the power balance requirements; (4) Perform secondary screening on the screened data in combination with the unit ramp rate constraint, the upper and lower limits of the output constraint, and the power factor constraint, and eliminate the data combinations that do not meet the equipment operation constraints; (5) Sort the data that meets the dual constraint conditions according to the principle of minimum cost, and extract the output data of the optimal power generation combination plan; (6) Split and re-organize the data of the optimal power generation combination plan in a 24-hour cycle to generate the day-ahead scheduling plan data.

[0029] Specifically, such as Figure 3As shown, it is a schematic flow chart for cost calculation and data statistics in the embodiment of the present application, showing the generation process of the day-ahead scheduling plan. First, the operating condition data of thermal power units, the power generation data of new energy sources, and the charge and discharge data of energy storage systems are extracted from the system prediction dataset for processing. For thermal power unit data, fuel consumption is calculated considering boiler efficiency, steam turbine efficiency, and generator efficiency; for new energy data, the cumulative power generation of wind farms and photovoltaic power stations is statistically calculated; for energy storage systems, charge and discharge cycle data is recorded. Then, the power generation cost is calculated. For thermal power units, fuel cost, start-stop cost, and maintenance cost are considered; for new energy sources, equipment depreciation and operation and maintenance costs are considered; for energy storage systems, charge and discharge losses and equipment usage costs are considered. The cost data of all units are summarized. Next, double constraint checks are performed: first, the power balance constraint is to ensure that the total output on the power generation side is equal to the load demand; second, the unit operation constraints include ramp rate, upper and lower output limits, and power factor constraints. Data combinations that do not meet the constraint conditions are excluded. Finally, the solutions that meet the constraints are sorted according to the principle of minimum cost, the optimal power generation combination plan is extracted, and it is split and reorganized according to a 24-hour cycle to generate the final day-ahead scheduling plan data. The arrows in the figure represent the data flow direction, and the subgraphs represent the specific steps of each processing link.

[0030] Extract the operating condition data of thermal power units from the system prediction dataset. The operating condition data includes key parameters such as boiler steam pressure, steam turbine speed, and generator power output. The fuel consumption per hour is calculated through these parameters. The calculation of fuel consumption needs to consider the thermal characteristic curve and efficiency curve of the unit, mainly involving the losses in three links: boiler efficiency, steam turbine efficiency, and generator efficiency. The fuel consumption per hour is recorded in the time series data table, and the cumulative consumption value for 24 hours is calculated. The new energy power generation data is processed according to the time series, and the output data of wind farms and photovoltaic power stations are statistically accumulated hourly. The output data of wind farms needs to consider the conversion characteristics of wind speed and power, and the output data of photovoltaic power stations needs to combine the change law of light intensity. The charge and discharge data of the energy storage system are segmented according to the operation cycle, and information such as the start and end time, charge and discharge power, and state of charge change of each charge and discharge cycle is recorded. Calculate the unit power generation cost of various units based on fuel consumption, new energy power generation, and energy storage charge and discharge. The power generation cost of thermal power units mainly includes fuel cost, start-stop cost, and equipment maintenance cost. The new energy power generation cost mainly considers equipment depreciation and operation and maintenance costs. The energy storage system cost needs to calculate the charge and discharge loss cost and equipment usage cost. Summarize these cost data according to the power generation type and time series to form a cost data table.

[0031] The system power balance constraint requires that the total output on the power generation side must be equal to the total demand on the load side. Therefore, it is necessary to optimize and screen the power generation cost data of various types of units. The total power generation and total load demand in each time period are statistically calculated, the power balance deviation is calculated, and the data combinations with deviations exceeding the allowable range are excluded. The calculation of power balance needs to consider line losses and substation equipment losses. The unit operation constraint is an important condition to ensure the safe operation of equipment. The ramp rate constraint limits the output change rate of thermal power units, the upper and lower output limits constraint specifies the operation range of the units, and the power factor constraint is related to the power quality. The data that meet the power balance constraint are screened a second time, and the operation plans that exceed the physical constraints of the units are excluded.

[0032] The data screened by the double constraint conditions are sorted according to the principle of minimum cost. The sorting process needs to comprehensively consider the power generation costs of various types of units, the operation costs of the system, and the degree of satisfaction of the constraint conditions. The power generation combination plan with the lowest cost and meeting all constraint conditions is extracted from the sorting results to obtain the optimal output data of each unit. The optimal power generation combination plan is processed according to a 24-hour cycle. The overall plan is split into 24 hourly segments, and each hourly segment includes the output data of various types of units, the charge and discharge plans of the energy storage system, and the system operation parameters. Then these data are reorganized according to the time sequence logic to form the day-ahead scheduling plan data.

[0033] For example: When the boiler steam pressure of a thermal power unit is 13 MPa, the steam turbine speed is 3000 rpm, and the generator power output is 300 MW, the hourly fuel consumption is calculated to be 110 tons of standard coal according to the thermal characteristics curve, and the daily consumption is obtained by accumulating for 24 hours. At the same time, the output data of 6 wind turbines in the wind farm at different wind speeds, the power generation of 500 photovoltaic modules in the photovoltaic power station at different light intensities, and the power change and energy change of the energy storage system during a complete charge and discharge cycle are recorded. According to the fuel unit price, equipment depreciation, and maintenance costs, the power generation cost of the thermal power unit is calculated to be 0.4 yuan / kWh, the new energy power generation cost is 0.2 yuan / kWh, and the energy storage system cost is 0.1 yuan / kWh. Through power balance calculation, it is found that the deviation between the total power generation and the load demand in a certain time period exceeds 5%, and the data combination in this time period is excluded. Considering that the ramp rate limit of the thermal power unit is 2 MW per minute, some schemes with excessive output changes are excluded. The power generation combination plan with the lowest total cost is selected and split into a 24-hour segment scheduling plan.

[0034] In a specific embodiment, the process of executing step S103 may specifically include the following steps: (1) Divide the day-ahead scheduling plan data at 15-minute time intervals to generate the reference output data for 96 time segments; (2) Sample the system operating status for each time period, collect the real-time output of the wind farm, the power generation power of the photovoltaic power station, the output of the thermal power unit, and the charge and discharge power of the energy storage system, and generate real-time operation data; (3) Compare and calculate the real-time operation data with the reference output data to obtain the power deviation value of each power generation unit, and generate power deviation data; (4) Judge the positive and negative values of the power deviation data, determine the output adjustment direction of the power generation unit, and calculate the adjustment power value of each unit; (5) Generate the output command of the thermal power unit and the charge and discharge command of the energy storage system according to the adjustment power value, and correct the output data of the power generation unit in real time; (6) Combine the corrected output data of the power generation unit in time series, update the system operating status according to a 15-minute cycle, and generate real-time dispatching data.

[0035] Specifically, divide the day-ahead dispatching plan data with 15 minutes as the basic time unit. There are 96 time periods in 24 hours. Each time period contains the planned output values of each power generation unit, including the rated output value of the thermal power unit, the predicted output value of the wind farm, the predicted power generation power value of the photovoltaic power station, and the planned charge and discharge power value of the energy storage system. These data serve as the reference output data in the real-time dispatching process and are used for subsequent deviation calculation and adjustment control. Perform real-time status sampling on 96 time periods respectively, with a sampling frequency of once per minute. The collected data includes the current actual power output of the wind farm (wind turbine speed, wind wheel torque, actual power generation), the actual power generation data of the photovoltaic power station (light intensity, module temperature, real-time power), the operating parameters of the thermal power unit (boiler pressure, steam turbine speed, generator power), and the operating status of the energy storage system (state of charge SOC, real-time charge and discharge power, system temperature). After data preprocessing and format unification, these sampling data form the real-time operation data of the current time period.

[0036] Compare and analyze the real-time operation data with the reference output data, and calculate the power deviation of each power generation unit. The calculation formula for power deviation is:

[0037] Among them, represents the power deviation value of the power generation unit, and respectively represent the actual output and reference output of the i-th thermal power unit, and respectively represent the actual output and reference output of the j-th new energy unit, and respectively represent the actual power and reference power of the k-th energy storage unit, , , is the weight coefficient of various devices, and n, m, and l respectively represent the quantities of various devices.

[0038] Judge the positive and negative values of the calculated power deviation data. A positive value indicates that the actual output is greater than the planned value and the output needs to be reduced, and a negative value indicates that the actual output is less than the planned value and the output needs to be increased. Calculate the adjusted power value according to the direction and magnitude of the deviation. The calculation formula is:

[0039] Among them, represents the adjusted power value, is the overall adjustment coefficient, is the adjustment weight of each unit, is the power deviation of each unit, is the upper limit of the adjustment ability of each unit, is the time decay coefficient, is the adjustment duration. According to the calculated adjusted power value, generate the output adjustment instruction of the thermal power unit and the charge and discharge instruction of the energy storage system respectively. The instruction signal includes three elements: adjustment direction, adjustment amplitude, and execution time. Real-time correct the output data of each power generation unit, and the correction value needs to meet the physical constraints of the device (such as ramp rate limit, output upper and lower limits, etc.). Recombine the corrected output data of each unit in chronological order, and perform a rolling update of the system operation state in a 15-minute cycle to generate real-time scheduling data.

[0040] For example: The reference output of a thermal power unit is 300MW, and the actual sampled output is 280MW. The calculated power deviation is -20MW, indicating that the output needs to be increased. Considering that the ramp rate limit of this unit is 2MW / minute, the maximum output that can be increased within 15 minutes is 30MW. At the same time, due to the decrease in wind speed, the actual output of the wind farm is 15MW lower than the planned value, and the energy storage system needs to switch to the discharge mode to supplement the power deficit. After adjustment calculation, the thermal power unit increases the output by 20MW, and the energy storage system discharges 15MW to balance the power deviation. These corrected data are recorded in a new 15-minute scheduling cycle and continue to be optimized by rolling.

[0041] In a specific embodiment, the process of executing step S104 may specifically include the following steps: (1) Divide the real-time scheduling data into hourly data, minute-level data, and second-level data according to the time scale to generate hierarchical processing data; (2) Conduct a trend analysis on the power fluctuations in the hourly data, calculate the output adjustment amount of the thermal power unit, and generate hourly power control data; (3) Perform statistical analysis on the fluctuation amplitude of minute-level data, calculate the output change of the wind farm and photovoltaic power station, and generate minute-level power control data; (4) Conduct frequency fluctuation analysis on second-level data, extract the instantaneous power fluctuation value of the system, and generate second-level power control data; (5) Calculate the power compensation value corresponding to each time scale according to the charging efficiency and discharging efficiency of the energy storage system, and generate energy storage compensation data; (6) Perform time-series fusion on the hourly power control data, minute-level power control data, second-level power control data, and energy storage compensation data to generate multi-time-scale control data.

[0042] Specifically, the real-time scheduling data is divided into three levels according to the time resolution: the sampling period of the hourly data is 60 minutes, which is mainly used for economic scheduling and load tracking; the sampling period of the minute-level data is 1 minute, which is mainly used for power balance and fluctuation suppression; the sampling period of the second-level data is 1 second, which is mainly used for frequency regulation and transient response. When performing trend analysis on the hourly data, the power curve is fitted by the linear regression method, and the slope value is calculated to characterize the change trend. A positive slope indicates an increasing power trend, and a negative slope indicates a decreasing power trend. According to the trend analysis results, calculate the output adjustment amount of the thermal power unit. The adjustment amount calculation needs to consider the ramp rate constraint and operating cost characteristics of the unit. Record the adjusted thermal power unit output data in the hourly power control data table.

[0043] The statistical analysis of the fluctuation amplitude of the minute-level data is mainly for the new energy generation units. The calculation formula for the output change of the wind farm and photovoltaic power station is:

[0044] Among them, represents the total output change of the new energy, and respectively represent the power outputs of the wind farm and photovoltaic power station, and are the weight coefficients of each unit, is the time decay factor, is the response time, and q and r represent the numbers of the wind farm and photovoltaic power station respectively.

[0045] Conduct frequency fluctuation analysis on the second-level data, use the fast Fourier transform to extract the frequency characteristics of the power fluctuation, and focus on the fluctuation components in the range of 0.1 - 1 Hz. Conduct segmented statistics on the power fluctuation value, and record the fluctuation amplitude, frequency, and duration of each statistical interval. These data are used for subsequent energy storage compensation control.

[0046] The calculation formula for the power compensation value of the energy storage system at different time scales is:

[0047] Among them, represents the total compensation power value, , , respectively represent the power deviations at the hourly, minute - level, and second - level, , , are the compensation coefficients for each time scale, , , are the charge - discharge efficiencies of the energy storage system at each time scale. a, b, and c respectively represent the number of compensation points at each time scale.

[0048] The control data at each time scale is fused in time series according to a unified data format, and a control instruction is synthesized by using the weighted superposition method. The setting of the weight coefficient needs to consider the response characteristics and priorities of the control at each time scale. High - frequency fluctuations are preferentially responded to by the energy storage system, while low - frequency fluctuations mainly rely on the regulation of conventional units.

[0049] For example: In a certain distributed energy storage power system within a dispatching period, the hourly - level data records that the output of the thermal power unit shows an upward trend. Through linear regression calculation, the slope is positive, indicating an increase in load demand. The minute - level data shows the power fluctuations caused by gusts in the wind farm, and the change in output per minute needs to be calculated. The second - level data finds an obvious power fluctuation component near 0.5 Hz through spectrum analysis. The energy storage system calculates the compensation values according to the fluctuation characteristics of these three time scales respectively: providing capacity support for the hourly - level trend change, suppressing the new - energy fluctuations at the minute - level, and providing a fast response to the second - level frequency disturbances. These control data are combined in time series to generate a hierarchical and coordinated control instruction.

[0050] In a specific embodiment, the process of executing step S105 may specifically include the following steps: (1) Divide the multi - time - scale control data according to the regional power grid structure, calculate the power fluctuation characteristics of each region, and generate regional power characteristic data; (2) Conduct statistical analysis on the regional power characteristic data, calculate the capacity of the energy storage system required for each region, and generate capacity allocation data; (3) Record the operating state of the energy storage system according to the capacity allocation data, and statistically analyze the charge depth and discharge depth to generate depth statistical data; (4) Conduct time - series accumulation on the depth statistical data, calculate the charge - discharge cycle times of the energy storage system, and generate cycle statistical data; (5) Calculate the performance degradation value of the energy storage unit according to the depth statistical data and the cycle statistical data, and generate life evaluation data; (6) Correlate and analyze the life assessment data with the capacity allocation data, perform optimization calculations on the operating parameters of the energy storage system, and generate energy storage optimization data.

[0051] Specifically, according to the topological structure of the regional power grid, allocate the control data to different regional units. Each regional unit includes wind farm power data, photovoltaic power station power data, thermal power unit operation data, and energy storage system status data. Analyze the power fluctuation characteristics of each region, mainly including the amplitude, frequency, and duration of power changes. By calculating the power fluctuation characteristics of various power generation units within the region, form regional power characteristic data, which records the power fluctuation patterns, regulation requirements, and energy storage application scenarios of each region. Based on the regional power characteristic data, optimize the configuration of the energy storage capacity, statistically analyze the new energy generation fluctuation laws, load change characteristics, and grid regulation requirements of each region. By integrating the hourly trend changes, minute-level power fluctuations, and second-level frequency disturbances, determine the energy storage capacity required for each region. Capacity calculation needs to consider the power response demand and energy support requirements, and the generated capacity allocation data includes the rated power and rated capacity of the energy storage system in each region.

[0052] After deploying the energy storage system according to the capacity allocation data, it is necessary to record the operating state parameters of the energy storage unit. Mainly focus on two key indicators: the state of charge (SOC) (the ratio of the charge state to the rated capacity) and the depth of discharge (DOD) (the ratio of the discharge amount to the rated capacity). Statistically analyze these depth data, record the occurrence frequency, duration, and conversion rules in different depth intervals, and form depth statistical data, which reflects the usage intensity and working mode of the energy storage unit. Cumulatively process the depth statistical data in chronological order to calculate the charge-discharge cycle times of the energy storage system. The cycle times are divided into full cycles (both the charge and discharge depths reach the set threshold) and partial cycles (the charge and discharge depths do not reach the threshold). Weighted cumulative of different types of cycles, record the depth distribution, time interval, and temperature conditions during the cycle, and generate cycle statistical data.

[0053] Evaluate the performance degradation of the energy storage unit based on the depth statistical data and cycle statistical data. The performance degradation is mainly manifested in two aspects: capacity degradation and internal resistance increase, and it needs to be comprehensively analyzed in combination with factors such as charge-discharge depth, cycle times, and operating temperature. By establishing a degradation model, calculate the remaining capacity and equivalent life of the energy storage unit to form life assessment data. Finally, correlate and analyze the life assessment data with the capacity allocation data to optimize the operation strategy of the energy storage system. According to the performance state of the energy storage unit, adjust the operating parameters such as charge-discharge power limits, depth thresholds, and response times. In the optimization process, it is necessary to balance the relationship between energy storage life and regulation performance to generate the final energy storage optimization data.

[0054] For example, a regional power grid includes a 300 MW wind farm, a 200 MW photovoltaic power station, and a 600 MW thermal power unit. By analyzing the power fluctuation characteristics of this region, it is found that the hourly load change is about 50 MW, the new energy fluctuation at the minute level is about 30 MW, and the frequency disturbance at the second level requires a rapid response capacity of 10 MW. Based on these characteristics, an energy storage system is configured with a rated power of 50 MW and a rated capacity of 100 MWh. During actual operation, it is recorded that the charge depth of the energy storage system mostly concentrates in the range of 20% - 80%, and the discharge depth is mainly distributed in the range of 30% - 70%. After one month of operation, the cumulative number of complete cycles is about 90 times, and the number of partial cycles is about 270 times. Through performance evaluation, it is found that the capacity retention rate of the energy storage unit is 99%. Based on this, the charge and discharge depth thresholds are optimized, and the working range is appropriately reduced to delay performance degradation. These optimized parameters are updated to the controller of the energy storage system as the control basis for the new round of operation.

[0055] In a specific embodiment, the process of executing step S106 may specifically include the following steps: (1) Extract the system power balance accuracy, frequency quality, and new energy consumption rate according to the energy storage optimization data, sample and record them at a 15 - minute cycle, and generate operation index data; (2) Perform time - series accumulation on the operation index data, calculate the statistical characteristic values and change trends of each index, and generate index statistical data; (3) Compare and analyze the index statistical data with the historical operation data, calculate the deviation values of each index, and generate index deviation data; (4) Correct the output parameters of the thermal power unit and the charge - discharge parameters of the energy storage system according to the index deviation data, and generate parameter correction data; (5) Verify the rationality of the parameter correction data, screen the parameter combinations that meet the system constraint conditions, and generate parameter update data; (6) Import the parameter update data into the system dispatching control unit, update the operation parameters of the generator set and the energy storage system, and generate system optimization data.

[0056] Specifically, three key operating indicators are extracted from the energy storage optimization data: system power balance accuracy (the absolute value of the difference between the planned power value and the actual power value), frequency quality (the root mean square value of the system frequency deviation), and new energy consumption rate (the ratio of the actual power generation of new energy to the theoretical power generation). These indicators are recorded by sampling once every 15 minutes. The sampled data includes timestamps, measured values, and data quality identifiers. The power balance accuracy is obtained by comparing the dispatching plan value with the measured value. The frequency quality is calculated through frequency sampling. The new energy consumption rate is obtained by calculating the theoretical power generation in combination with meteorological data and then comparing it with the actual power generation. The collected operating indicator data is processed by time series accumulation, and statistical characteristic values are calculated, including mean, standard deviation, skewness, and kurtosis. The mean reflects the overall level of the indicator, the standard deviation characterizes the degree of fluctuation, the skewness indicates the asymmetry of the data distribution, and the kurtosis describes the sharpness of the data distribution. The change trends are extracted through time series analysis methods, including long-term trends, periodic changes, and random fluctuation components. These statistical characteristics and trend information are organized into indicator statistical data.

[0057] The generated indicator statistical data is compared and analyzed with historical operating data, which includes concurrent data (the same period last year, the same period last month) and typical operating data (optimal operating conditions, standard operating conditions). By calculating the deviation between the current indicator and the historical indicator, the deviation degree of each indicator is obtained. Deviation analysis needs to consider influencing factors such as seasonal factors, load characteristics, and equipment status, and a dataset containing the deviation values of each indicator is generated. The operating parameters are corrected according to the indicator deviation data. The correction objects include parameters such as the given power of thermal power units, ramp rate limits, upper and lower output limits, as well as the charge and discharge power limits, state of charge thresholds, response times, etc. of the energy storage system. The parameter correction follows the principle of "large deviation, large adjustment; small deviation, small adjustment", and the adjustment amount is proportional to the deviation value. The corrected parameter values are recorded as parameter correction data.

[0058] The rationality of the parameter correction data is verified. The verification content includes: whether the parameter values are within the physical constraints of the equipment, whether the parameter combination meets the system power balance requirements, whether the parameter changes conform to the dynamic characteristics of the equipment, etc. By setting multiple verification conditions, the parameter combinations that meet all the constraint conditions are screened out to form parameter update data. The screening of parameter combinations needs to consider the priorities and mutual influences of various constraint conditions. The verified parameter update data is imported into the dispatching control unit to update the operating parameters of thermal power units and the control parameters of the energy storage system. The parameter update adopts a smooth transition method to avoid system fluctuations caused by parameter mutations. The set of updated operating parameters constitutes the system optimization data, which serves as the basic data for the next round of optimal dispatching.

[0059] For example, during the operation of a certain distributed energy storage system, it is found through 15-minute cycle sampling records that there is a deviation in the power balance accuracy index. Statistical analysis is performed on the operation data for the past 4 hours to calculate the mean and standard deviation. It is found that the standard deviation is relatively large, indicating obvious power fluctuations. Comparing these statistical values with the historical data for the same period, it is found that the current power fluctuations exceed the normal range. After analyzing the reasons, it is found that the ramp rate setting of the thermal power unit is unreasonable, resulting in insufficiently rapid response to power changes. To address this issue, the ramp rate parameter of the thermal power unit is adjusted, and at the same time, the participation of the energy storage system is increased. After the new parameter combination is verified through the constraint conditions, it is confirmed that it meets the operation requirements of the equipment and the system. The adjusted parameters are updated to the control system, improving the power balance control effect.

[0060] The above describes the optimal dispatching method for the distributed energy storage power system in the embodiments of the present application. Next, the optimal dispatching system for the distributed energy storage power system in the embodiments of the present application will be described. Please refer to Figure 4 One embodiment of the optimal dispatching system for the distributed energy storage power system in the embodiments of the present application includes: A detection module 201, configured to collect real-time operation data of the system through a data acquisition network, perform normalization processing and outlier detection on the collected operation data to obtain standardized data, and perform system prediction for a 24-hour cycle based on the standardized data to generate a system prediction data set; A statistics module 202, configured to perform cost calculation and data statistics on the fuel consumption of the thermal power unit, the new energy power generation, and the charge and discharge amount of the energy storage system according to the system prediction data set, and output the day-ahead dispatching plan data through constraint condition screening and power balance analysis; An update module 203, configured to dynamically update the system operation status data through sub-period processing with a 15-minute cycle based on the day-ahead dispatching plan data, and adjust the output data of the power generation unit according to the power deviation calculation result to form real-time dispatching data; A layering module 204, configured to perform hierarchical processing of the system power fluctuations at the hourly, minute-level, and second-level according to the real-time dispatching data, and perform power compensation calculation according to the charge and discharge characteristics of the energy storage system to obtain multi-time scale control data; An allocation module 205, configured to use the multi-time scale control data to perform regional allocation calculation of the energy storage system capacity, and perform life assessment through statistical analysis of the charge and discharge depth and cycle times to generate energy storage optimization data; A correction module 206, configured to perform real-time monitoring and data accumulation on the system operation indicators based on the energy storage optimization data, and correct and update the dispatching parameters through index comparison analysis to obtain system optimization data.

[0061] Through the collaborative cooperation of the above-mentioned various components, by dividing the multi-time-scale control data into regions, accurately calculating the power fluctuation characteristics of each region, and determining the reasonable allocation of energy storage capacity in combination with the regional power characteristic data, the problem of unreasonable energy storage system capacity configuration is effectively solved; by using deep statistical data and cyclic statistical data to systematically analyze the charging and discharging process, a complete energy storage unit performance evaluation system is established, realizing the accurate prediction of the life characteristics of energy storage devices; based on the performance evaluation data, the operation parameters of the energy storage system are dynamically optimized, forming a closed-loop regulation and control mechanism, significantly improving the utilization efficiency of the energy storage system; through the hierarchical and zonal data processing method, the accurate matching and allocation of energy storage capacity are realized, avoiding the waste of energy storage resources; based on the charge and discharge depth and the number of cycles for life assessment, a health management mechanism of the energy storage system is established, extending the service life of the equipment; by using real-time monitoring and data accumulation means, continuously tracking and optimizing the system operation indicators, ensuring the long-term stable operation of the energy storage system. The technical solution of the present invention, through multi-dimensional data analysis and refined parameter control, not only improves the regulation performance of the energy storage system, but also realizes the effective management of the equipment life.

[0062] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for optimizing the scheduling of a distributed energy storage power system, characterized in that The optimized scheduling method for the distributed energy storage power system includes: Collecting the real-time operation data of the system through the data acquisition network, performing normalization processing and outlier detection on the collected operation data to obtain standardized data, and generating a system prediction data set through system prediction with a 24-hour cycle based on the standardized data; Calculating the costs and conducting data statistics for the fuel consumption of thermal power units, new energy power generation, and the charge and discharge amounts of the energy storage system according to the system prediction data set, and outputting the day-ahead scheduling plan data through constraint condition screening and power balance analysis; Based on the day-ahead scheduling plan data, dynamically updating the system operation state data through sub-period processing with a 15-minute cycle, and adjusting the output data of the power generation units according to the power deviation calculation results to form real-time scheduling data; According to the real-time scheduling data, performing hierarchical processing of the system power fluctuations at the hourly, minute-level, and second-level, and calculating power compensation according to the charge and discharge characteristics of the energy storage system to obtain multi-time scale control data; Using the multi-time scale control data to calculate the regional allocation of the energy storage system capacity, and conducting life assessment through statistical analysis of the charge and discharge depth and cycle times to generate energy storage optimization data; Based on the energy storage optimization data, performing real-time monitoring and data accumulation of the system operation indicators, and correcting and updating the scheduling parameters through index comparison analysis to obtain system optimization data.

2. The optimized scheduling method for a distributed energy storage power system according to claim 1, wherein The step of collecting the real-time operation data of the system through the data acquisition network, performing normalization processing and outlier detection on the collected operation data to obtain standardized data, and generating a system prediction data set through system prediction with a 24-hour cycle based on the standardized data includes: Respectively collecting the output data of the wind farm, the power generation data of the photovoltaic power station, the operation parameter data of the thermal power unit, the state parameter data of the energy storage system, and the load demand data according to the data acquisition network to generate the system real-time operation data; Performing maximum-minimum normalization calculation on the system real-time operation data, detecting and removing outliers through the 3σ criterion to generate standardized data; Dividing the standardized data into sliding windows with a 24-hour cycle, and performing time series analysis on the data within each time window to generate time series feature data; Extracting the load change law and new energy output fluctuation characteristics according to the time series feature data, and performing data dimensionality reduction processing to generate a feature data set; Performing correlation analysis on the feature data set, and fusing the highly correlated data to generate a fused data set; Performing prediction analysis on the fused data set with a 24-hour cycle to generate the system prediction data set.

3. The optimized scheduling method for a distributed energy storage power system according to claim 1, wherein The step of calculating the costs and conducting data statistics for the fuel consumption of thermal power units, new energy power generation, and the charge and discharge amounts of the energy storage system according to the system prediction data set, and outputting the day-ahead scheduling plan data through constraint condition screening and power balance analysis includes: Extracting the operation condition data of the thermal power unit according to the system prediction data set, calculating the hourly fuel consumption and statistically accumulating the daily total value, dividing the new energy power generation data according to the time series and accumulating it, and performing segmented statistics on the charge and discharge data of the energy storage system; Calculate the unit power generation cost of various types of units by using the fuel consumption of thermal power units, the power generation of new energy, and the charge and discharge of energy storage systems, and summarize the power generation cost data; Optimize and screen the power generation cost data of various types of units according to the system power balance constraint, and eliminate the data combinations that do not meet the power balance requirements; Perform a secondary screening on the screened data in combination with the unit ramp rate constraint, the upper and lower limits of output constraint, and the power factor constraint, and eliminate the data combinations that do not meet the equipment operation constraints; Sort the data that meet the dual constraint conditions according to the principle of minimum cost, and extract the output data of the optimal power generation combination plan; Perform data splitting and time series recombination on the optimal power generation combination plan in a 24-hour cycle to generate the day-ahead scheduling plan data.

4. The optimized scheduling method for a distributed energy storage power system according to claim 1, wherein Based on the day-ahead scheduling plan data, dynamically update the system operation state data through sub-period processing in a 15-minute cycle, and adjust the output data of the power generation unit according to the power deviation calculation result to form real-time scheduling data, including: Divide the day-ahead scheduling plan data into cycles at 15-minute time intervals to generate the reference output data for 96 time periods; Sample the system operation state for each time period, collect the real-time output of the wind farm, the power generation power of the photovoltaic power station, the output of the thermal power unit, and the charge and discharge power of the energy storage system to generate real-time operation data; Compare and calculate the real-time operation data with the reference output data to obtain the power deviation value of each power generation unit and generate power deviation data; Judge the positive and negative values of the power deviation data to determine the output adjustment direction of the power generation unit, and calculate the adjustment power value of each unit; Generate the output command of the thermal power unit and the charge and discharge command of the energy storage system according to the adjustment power value, and perform real-time correction on the output data of the power generation unit; Perform time series combination on the corrected output data of the power generation unit, update the system operation state according to the 15-minute cycle, and generate the real-time scheduling data.

5. The optimized scheduling method for a distributed energy storage power system according to claim 1, characterized in that According to the real-time scheduling data, perform hierarchical processing on the system power fluctuations at the hourly, minute-level, and second-level, and perform power compensation calculations based on the charge and discharge characteristics of the energy storage system to obtain multi-time scale control data, including: Divide the real-time scheduling data into hourly data, minute-level data, and second-level data according to the time scale to generate hierarchical processing data; Conduct trend analysis on the power fluctuations in the hourly data, calculate the output adjustment amount of the thermal power unit, and generate hourly power control data; Perform fluctuation amplitude statistics on the minute-level data, calculate the output change amount of the wind farm and the photovoltaic power station, and generate minute-level power control data; Conduct frequency fluctuation analysis on the second-level data, extract the instantaneous power fluctuation value of the system, and generate second-level power control data; Calculate the power compensation value corresponding to each time scale according to the charge efficiency and discharge efficiency of the energy storage system to generate energy storage compensation data; Perform time series fusion on the hourly power control data, minute-level power control data, second-level power control data, and energy storage compensation data to generate the multi-time scale control data.

6. The optimal scheduling method for a distributed energy storage power system according to claim 1, characterized in that Using the multi-time scale control data to perform regional allocation calculation on the energy storage system capacity, and conducting life assessment through statistical analysis of charge and discharge depth and cycle times to generate energy storage optimization data, including: Dividing the multi-time scale control data according to the regional power grid structure, calculating the power fluctuation characteristics of each region, and generating regional power characteristic data; Conducting statistical analysis on the regional power characteristic data, calculating the energy storage system capacity required for each region, and generating capacity allocation data; Recording the operating status of the energy storage system according to the capacity allocation data, statistically analyzing the charge depth and discharge depth, and generating depth statistical data; Performing time series accumulation on the depth statistical data, calculating the charge and discharge cycle times of the energy storage system, and generating cycle statistical data; Calculating the performance attenuation value of the energy storage unit according to the depth statistical data and cycle statistical data, and generating life assessment data; Performing correlation analysis on the life assessment data and the capacity allocation data, and performing optimization calculation on the operating parameters of the energy storage system to generate the energy storage optimization data.

7. The optimal scheduling method for a distributed energy storage power system according to claim 1, wherein Based on the energy storage optimization data, performing real-time monitoring and data accumulation on the system operation indicators, and correcting and updating the dispatching parameters through index comparison analysis to obtain system optimization data, including: Extracting the system power balance accuracy, frequency quality, and new energy consumption rate according to the energy storage optimization data, sampling and recording them at a 15-minute interval to generate operation index data; Performing time series accumulation on the operation index data, calculating the statistical characteristic values and change trends of each index, and generating index statistical data; Performing comparison analysis on the index statistical data and historical operation data, calculating the deviation values of each index, and generating index deviation data; Correcting the output parameters of thermal power units and the charge and discharge parameters of the energy storage system according to the index deviation data to generate parameter correction data; Performing rationality verification on the parameter correction data, screening the parameter combinations that meet the system constraint conditions, and generating parameter update data; Importing the parameter update data into the system dispatching control unit to update the operating parameters of the generating units and the energy storage system, and generating the system optimization data.

8. A distributed energy storage power system optimal scheduling system for implementing the distributed energy storage power system optimal scheduling method as described in any one of claims 1-7, characterized in that, The distributed energy storage power system optimal dispatching system includes: A detection module for collecting real-time operation data of the system through a data acquisition network, performing normalization processing and outlier detection on the collected operation data to obtain standardized data, and performing load demand prediction and new energy output prediction with a 24-hour cycle according to the standardized data to generate a system prediction data set; A statistics module for calculating the costs and performing data statistics on the fuel consumption of thermal power units, new energy power generation, and charge and discharge amounts of the energy storage system according to the system prediction data set, and outputting day-ahead dispatching plan data through constraint condition screening and power balance analysis; An update module for dynamically updating the system operation status data through sub-period processing with a 15-minute cycle based on the day-ahead dispatching plan data, and adjusting the output data of the generating units according to the power deviation calculation results to form real-time dispatching data; A hierarchical module for performing hierarchical processing of system power fluctuations at hourly, minute-level, and second-level based on the real-time scheduling data, calculating power compensation according to the charge and discharge characteristics of the energy storage system, and obtaining multi-time scale control data; An allocation module for calculating the regional allocation of the energy storage system capacity by using the multi-time scale control data, performing life assessment through statistical analysis of the charge and discharge depth and cycle times, and generating energy storage optimization data; A correction module for performing real-time monitoring and data accumulation of system operation indicators based on the energy storage optimization data, correcting and updating scheduling parameters through index comparison analysis, and obtaining system optimization data.

Images