Distributed optical storage isolated network power supply control method, system, equipment and medium

Abstract

The invention belongs to the technical field of optical storage power supply, and discloses a distributed optical storage isolated network power supply control method, system, device and medium, and the method comprises the steps: processing photovoltaic power generation historical data and current meteorological information through employing a long short-term memory network, and obtaining a photovoltaic power generation prediction error accumulation trend; extracting low-ebb recharging time period data from the energy storage reserved capacity insufficiency risk level, and determining a load peak value and low-ebb recharging overlapping degree quantized value; adjusting an energy storage capacity distribution proportion according to a load peak value and a low ebb recharging overlapping degree quantized value, determining a dynamic energy storage reservation increment, and calculating a capacity allocation parameter; obtaining a real-time power supply reliability index according to the optical storage combined operation preliminary mode, and determining an optical storage system operation cooperation strategy; and generating a light storage system operation adjustment instruction according to the cooperation strategy of the light storage system operation. The problem that the power supply reliability is reduced due to the fact that the photovoltaic power generation prediction error is not matched with the night load peak demand is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optical storage power supply, in particular to a distributed optical storage isolated network power supply control method, system, device and medium. BACKGROUND

[0002] In the field of energy management, distributed optical storage systems, as a new power supply mode, have important significance for ensuring stable power supply in remote areas or independent power supply areas. This combination of photovoltaic power generation and energy storage devices attempts to achieve energy self-sufficiency and reduce dependence on traditional power grids. However, as the application scenarios become more complex, how to maintain power supply reliability in a dynamically changing environment has become a challenge that needs to be addressed.

[0003] Current optical storage systems often face the problem of inconsistency between prediction and actual operating conditions. Many existing methods rely more on static historical data or simple prediction methods when formulating power generation and energy storage plans, ignoring real-time fluctuations in environmental factors and load demand. This oversight results in a lack of flexibility in responding to sudden power generation deviations or load changes, which in turn affects the continuity of power supply.

[0004] The deeper technical difficulty lies in the dynamic balance between photovoltaic power generation prediction error and energy storage capacity allocation. The existence of prediction error may result in a large deviation in estimating photovoltaic power generation, while the allocation of energy storage capacity is directly affected by this prediction result. If the prediction is too high, the energy storage reserved power may not be able to fill the actual power gap; conversely, if the prediction is too low, it may result in waste of energy storage resources. This mismatch, especially when the peak load demand at night overlaps with the low valley power recharge time, further exacerbates the pressure on system deployment.

[0005] Specifically, in a typical independent power supply area, the night-time residential electricity demand often bursts in a certain period, such as the peak period from 7 pm to 9 pm, and at this time if the energy storage device is charging at the low valley power, the system cannot release enough power to meet the demand in time. This time conflict significantly increases the risk of power supply interruption, seriously affecting the user's power consumption experience.

[0006] Therefore, how to reasonably adjust the matching relationship between energy storage capacity and photovoltaic power generation plan in the complex scenario of prediction error accumulation and night-time load peak overlap has become a key problem to improve the reliability of independent power supply. SUMMARY

[0007] The embodiments of the present application provide a distributed optical storage isolated network power supply control method, system, device and medium to solve the problems in the prior art.

[0008] To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or to describe the scope of protection of these embodiments. Its sole purpose is to present some concepts in a simple form as a prelude to the detailed description that follows.

[0009] According to a first aspect of the present invention, a distributed optical-storage islanded grid power supply control method is provided.

[0010] In one embodiment, the distributed optical-storage islanded grid power supply control method includes:

[0011] Historical photovoltaic power generation data and current meteorological information are acquired and combined with energy storage capacity status data. Long short-term memory networks are used for trend analysis to obtain the cumulative trend of photovoltaic power generation prediction error. Based on the cumulative trend of photovoltaic power generation prediction error and energy storage capacity status data, an initial photovoltaic output plan is generated.

[0012] Based on the cumulative trend of photovoltaic power generation prediction errors, nighttime peak load demand distribution data are obtained, and the actual power generation gap level is classified using support vector machine; the risk level of insufficient energy storage reserved capacity is determined based on the classification results of the actual power generation gap level.

[0013] Data on off-peak charging periods are extracted from the risk level of insufficient reserved capacity for energy storage. A time series analysis algorithm is used to calculate the capacity allocation pressure index. Based on the capacity allocation pressure index, a quantitative value of the overlap between peak load and off-peak charging is determined.

[0014] Based on the quantitative value of the overlap between peak and off-peak load recharge and the historical record of allocation mismatch, the energy storage capacity allocation ratio is adjusted to determine the dynamic energy storage reserve increment; based on the capacity allocation pressure index and the dynamic energy storage reserve increment, the capacity allocation parameters are calculated.

[0015] By combining the dynamic energy storage reserve increment and the cumulative trend of photovoltaic power generation prediction error, the initial photovoltaic power output plan is revised; based on the revised initial photovoltaic power output plan, a preliminary photovoltaic-energy storage joint operation mode is established.

[0016] Based on the preliminary model of photovoltaic and energy storage joint operation, real-time power supply reliability indicators are obtained, and capacity allocation parameters are iteratively optimized based on the real-time power supply reliability indicators to determine the collaborative strategy for the operation of the photovoltaic and energy storage system.

[0017] Based on the collaborative strategy of photovoltaic and energy storage system operation, operation adjustment instructions for photovoltaic and energy storage system are generated and deployed to energy storage system to obtain power supply reliability operation status.

[0018] In one embodiment, historical photovoltaic (PV) power generation data and current weather information are acquired, and combined with energy storage capacity status data. A long short-term memory (LSTM) network is used for trend analysis to obtain the cumulative trend of PV power generation prediction errors. Based on the cumulative trend of PV power generation prediction errors and energy storage capacity status data, an initial PV output plan is generated, including the following steps:

[0019] Collect hourly power generation and daily cumulative values ​​from historical photovoltaic power generation data, obtain temperature, humidity, and wind speed parameters from current meteorological information, and extract the remaining power percentage and charging / discharging power limits from energy storage capacity status data;

[0020] The historical data and current meteorological information are extracted using a long short-term memory network. Based on the extracted time-series features, calculations are performed to obtain the cumulative trend of photovoltaic power generation prediction error.

[0021] Based on the cumulative trend of photovoltaic power generation prediction error and the percentage of remaining electricity, the slope of error change is obtained by linear regression fitting. Based on the slope of error change and the percentage of remaining electricity, the correction coefficient of the daytime power generation prediction value is calculated to obtain the adjusted time-sharing power generation value.

[0022] Based on historical load data, a nighttime load demand estimate is generated. Combining the nighttime load demand estimate with the adjusted time-of-use power generation value, energy storage capacity constraints are determined. Based on the constraints and the adjusted time-of-use power generation value, an initial photovoltaic power output plan is generated.

[0023] In one embodiment, based on the cumulative trend of photovoltaic power generation prediction errors, nighttime peak load demand distribution data is obtained, and the actual power generation gap level is classified using a support vector machine. Determining the risk level of insufficient energy storage reserve capacity based on the classification results of the actual power generation gap level includes the following steps:

[0024] Based on the cumulative trend of photovoltaic power generation prediction error, the daily error peak and the time of error occurrence are extracted, and the nighttime load peak demand distribution data is obtained. The distribution data includes hourly load peak, average load and load fluctuation range.

[0025] Based on the error peak and nighttime load peak demand distribution data, the number of overlapping hours between the load peak period and the error exceeding period is calculated. Combined with the percentage of remaining energy storage capacity, a feature vector including the number of overlapping hours and the percentage of remaining energy storage capacity is constructed.

[0026] The feature vectors are classified using a support vector machine, and the actual power generation gap level is determined based on the classification results. The risk level of insufficient reserved energy storage capacity is determined based on the actual power generation gap level and the percentage of remaining energy storage capacity.

[0027] In one embodiment, the energy storage capacity allocation ratio is adjusted based on the quantified value of the overlap between peak and off-peak charging and historical records of allocation mismatch, to determine the dynamic energy storage reserve increment; the calculation of capacity allocation parameters based on the capacity allocation pressure index and the dynamic energy storage reserve increment includes the following steps:

[0028] Based on the quantified value of the overlap between peak and off-peak load recharge, query the historical records of allocation mismatch, and calculate the deviation characteristic value based on the historical records of allocation mismatch.

[0029] The probability of allocation failure is calculated using a Bayesian inference algorithm. Based on the probability of allocation failure, the energy storage capacity allocation ratio is adjusted to determine the dynamic energy storage reserve increment.

[0030] Based on the dynamic energy storage reserved increment and capacity allocation pressure index, the energy storage allocation ratio adjustment value, reserved capacity increment value and recharge time optimization value are calculated and integrated to form capacity allocation parameters.

[0031] In one embodiment, the initial photovoltaic power output plan is revised by combining the dynamic energy storage reserve increment and the cumulative trend of photovoltaic power generation prediction errors; based on the revised initial photovoltaic power output plan, the preliminary photovoltaic-energy storage joint operation mode is established, including the following steps:

[0032] The correction coefficient is determined based on the proportion of the dynamic energy storage reserved increment to the total energy storage capacity. The power generation forecast values ​​for each period are adjusted based on the correction coefficient to obtain the corrected power generation sequence for each period.

[0033] Based on the corrected power generation sequence, the photovoltaic main power supply period and the energy storage power supply period are divided, and the photovoltaic-storage power allocation scheme is determined.

[0034] Based on the photovoltaic power allocation scheme, the parameters of the photovoltaic inverter and the energy storage converter are set, and a preliminary mode for photovoltaic-energy storage joint operation is established.

[0035] In one embodiment, based on the initial joint operation mode of photovoltaic and energy storage, real-time power supply reliability indicators are obtained, and capacity allocation parameters are iteratively optimized based on these indicators to determine the collaborative strategy for the operation of the photovoltaic and energy storage system. This includes the following steps:

[0036] Based on the preliminary model of photovoltaic-storage joint operation, real-time power supply reliability indicators are obtained, and a comprehensive reliability score is calculated.

[0037] The gradient descent method is used to adjust the capacity allocation parameters. Based on the optimized capacity allocation parameters, a photovoltaic output adjustment scheme and an energy storage charging and discharging plan are formulated. A load demand response mechanism is established, and a coordinated strategy for the operation of the photovoltaic and energy storage system is formed.

[0038] In one embodiment, based on the collaborative strategy for the operation of the photovoltaic-storage system, generating operation adjustment instructions for the photovoltaic-storage system and deploying them to the energy storage system to obtain the power supply reliability operating status includes the following steps:

[0039] Based on the collaborative strategy of the operation of the optical storage system, an adjustment instruction set containing execution time and parameter values ​​is generated;

[0040] The adjustment instruction set is deployed to the energy storage system through the data transmission protocol, and the power supply reliability operation status is determined based on the response status data after execution.

[0041] According to a second aspect of the present invention, a distributed optical-storage islanded grid power supply control system is provided.

[0042] In one embodiment, the distributed optical-storage islanded grid power supply control system includes:

[0043] The power generation forecasting and initial scheme generation module is used to acquire historical photovoltaic power generation data and current meteorological information, and combine them with energy storage capacity status data. It uses a long short-term memory network to perform trend analysis to obtain the cumulative trend of photovoltaic power generation forecasting error. Based on the cumulative trend of photovoltaic power generation forecasting error and energy storage capacity status data, it generates an initial scheme for photovoltaic power output plan.

[0044] The energy storage risk level assessment module is used to obtain nighttime peak load demand distribution data based on the cumulative trend of photovoltaic power generation prediction errors, and classify the actual power generation gap level through support vector machine; based on the classification results of the actual power generation gap level, the risk level of insufficient energy storage reserved capacity is determined.

[0045] The load and recharge overlap analysis module is used to extract data on off-peak recharge periods from the risk level of insufficient reserved capacity of energy storage, calculate the capacity allocation pressure index using time series analysis algorithms, and determine the quantitative value of the overlap between peak load and off-peak recharge based on the capacity allocation pressure index.

[0046] The dynamic capacity allocation calculation module is used to adjust the energy storage capacity allocation ratio and determine the dynamic energy storage reserve increment based on the quantitative value of the overlap between peak and off-peak load recharge and the historical record of allocation mismatch; and to calculate the capacity allocation parameters based on the capacity allocation pressure index and the dynamic energy storage reserve increment.

[0047] The photovoltaic-storage joint operation modeling module is used to combine the dynamic energy storage reserve increment and the cumulative trend of photovoltaic power generation prediction error to revise the initial photovoltaic output plan; based on the revised initial photovoltaic output plan, a preliminary photovoltaic-storage joint operation model is established.

[0048] The collaborative operation strategy optimization module is used to obtain real-time power supply reliability indicators based on the initial mode of photovoltaic and energy storage joint operation, iteratively optimize capacity allocation parameters based on the real-time power supply reliability indicators, and determine the collaborative operation strategy of the photovoltaic and energy storage system.

[0049] The instruction deployment and execution module is used to generate operation adjustment instructions for the photovoltaic and energy storage system based on the collaborative strategy of the photovoltaic and energy storage system operation, and deploy them to the energy storage system to obtain the power supply reliability operation status.

[0050] In one embodiment, historical photovoltaic (PV) power generation data and current weather information are acquired, and combined with energy storage capacity status data. A long short-term memory (LSTM) network is used for trend analysis to obtain the cumulative trend of PV power generation prediction errors. Based on the cumulative trend of PV power generation prediction errors and energy storage capacity status data, an initial PV output plan is generated, including the following steps:

[0051] Collect hourly power generation and daily cumulative values ​​from historical photovoltaic power generation data, obtain temperature, humidity, and wind speed parameters from current meteorological information, and extract the remaining power percentage and charging / discharging power limits from energy storage capacity status data;

[0052] The historical data and current meteorological information are extracted using a long short-term memory network. Based on the extracted time-series features, calculations are performed to obtain the cumulative trend of photovoltaic power generation prediction error.

[0053] Based on the cumulative trend of photovoltaic power generation prediction error and the percentage of remaining electricity, the slope of error change is obtained by linear regression fitting. Based on the slope of error change and the percentage of remaining electricity, the correction coefficient of the daytime power generation prediction value is calculated to obtain the adjusted time-sharing power generation value.

[0054] Based on historical load data, a nighttime load demand estimate is generated. Combining the nighttime load demand estimate with the adjusted time-of-use power generation value, energy storage capacity constraints are determined. Based on the constraints and the adjusted time-of-use power generation value, an initial photovoltaic power output plan is generated.

[0055] In one embodiment, based on the cumulative trend of photovoltaic power generation prediction errors, nighttime peak load demand distribution data is obtained, and the actual power generation gap level is classified using a support vector machine. Determining the risk level of insufficient energy storage reserve capacity based on the classification results of the actual power generation gap level includes the following steps:

[0056] Based on the cumulative trend of photovoltaic power generation prediction error, the daily error peak and the time of error occurrence are extracted, and the nighttime load peak demand distribution data is obtained. The distribution data includes hourly load peak, average load and load fluctuation range.

[0057] Based on the error peak and nighttime load peak demand distribution data, the number of overlapping hours between the load peak period and the error exceeding period is calculated. Combined with the percentage of remaining energy storage capacity, a feature vector including the number of overlapping hours and the percentage of remaining energy storage capacity is constructed.

[0058] The feature vectors are classified using a support vector machine, and the actual power generation gap level is determined based on the classification results. The risk level of insufficient reserved energy storage capacity is determined based on the actual power generation gap level and the percentage of remaining energy storage capacity.

[0059] In one embodiment, the energy storage capacity allocation ratio is adjusted based on the quantified value of the overlap between peak and off-peak charging and historical records of allocation mismatch, to determine the dynamic energy storage reserve increment; the calculation of capacity allocation parameters based on the capacity allocation pressure index and the dynamic energy storage reserve increment includes the following steps:

[0060] Based on the quantified value of the overlap between peak and off-peak load recharge, query the historical records of allocation mismatch, and calculate the deviation characteristic value based on the historical records of allocation mismatch.

[0061] The probability of allocation failure is calculated using a Bayesian inference algorithm. Based on the probability of allocation failure, the energy storage capacity allocation ratio is adjusted to determine the dynamic energy storage reserve increment.

[0062] Based on the dynamic energy storage reserved increment and capacity allocation pressure index, the energy storage allocation ratio adjustment value, reserved capacity increment value and recharge time optimization value are calculated and integrated to form capacity allocation parameters.

[0063] In one embodiment, the initial photovoltaic power output plan is revised by combining the dynamic energy storage reserve increment and the cumulative trend of photovoltaic power generation prediction errors; based on the revised initial photovoltaic power output plan, the preliminary photovoltaic-energy storage joint operation mode is established, including the following steps:

[0064] The correction coefficient is determined based on the proportion of the dynamic energy storage reserved increment to the total energy storage capacity. The power generation forecast values ​​for each period are adjusted based on the correction coefficient to obtain the corrected power generation sequence for each period.

[0065] Based on the corrected power generation sequence, the photovoltaic main power supply period and the energy storage power supply period are divided, and the photovoltaic-storage power allocation scheme is determined.

[0066] Based on the photovoltaic power allocation scheme, the parameters of the photovoltaic inverter and the energy storage converter are set, and a preliminary mode for photovoltaic-energy storage joint operation is established.

[0067] In one embodiment, based on the initial joint operation mode of photovoltaic and energy storage, real-time power supply reliability indicators are obtained, and capacity allocation parameters are iteratively optimized based on these indicators to determine the collaborative strategy for the operation of the photovoltaic and energy storage system. This includes the following steps:

[0068] Based on the preliminary model of photovoltaic-storage joint operation, real-time power supply reliability indicators are obtained, and a comprehensive reliability score is calculated.

[0069] The gradient descent method is used to adjust the capacity allocation parameters. Based on the optimized capacity allocation parameters, a photovoltaic output adjustment scheme and an energy storage charging and discharging plan are formulated. A load demand response mechanism is established, and a coordinated strategy for the operation of the photovoltaic and energy storage system is formed.

[0070] In one embodiment, based on the collaborative strategy for the operation of the photovoltaic-storage system, generating operation adjustment instructions for the photovoltaic-storage system and deploying them to the energy storage system to obtain the power supply reliability operating status includes the following steps:

[0071] Based on the collaborative strategy of the operation of the optical storage system, an adjustment instruction set containing execution time and parameter values ​​is generated;

[0072] The adjustment instruction set is deployed to the energy storage system through the data transmission protocol, and the power supply reliability operation status is determined based on the response status data after execution.

[0073] According to a third aspect of the present invention, a computer device is provided.

[0074] In some embodiments, the computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described above.

[0075] According to a fourth aspect of the present invention, a computer-readable storage medium is provided.

[0076] In one embodiment, a computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the above method.

[0077] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0078] 1. This invention addresses the problem of decreased power supply reliability caused by the mismatch between photovoltaic power generation prediction errors and nighttime peak load demand. This invention processes historical photovoltaic power generation data, meteorological information, and energy storage capacity status using a long short-term memory network to generate an initial photovoltaic output plan. Based on the cumulative trend of prediction errors and support vector machine classification of nighttime peak load demand, it determines the risk level of insufficient reserved energy storage capacity.

[0079] 2. This invention calculates the capacity allocation pressure index through time series analysis, quantifies the overlap between peak and off-peak load recharge, dynamically adjusts the energy storage allocation ratio and reserved increments, optimizes recharge time, and corrects photovoltaic output plans. This invention iteratively optimizes capacity allocation parameters, generates operation adjustment commands for the photovoltaic-energy storage system, and deploys them to the energy storage system, achieving a significant improvement in power supply reliability.

[0080] 3. The most important innovation of this invention is that by dynamically and collaboratively optimizing the prediction error and load demand, it ensures efficient and stable power supply in the scenario of independent power supply of the transformer area, and significantly reduces the risk of power generation gap and peak overlap.

[0081] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0082] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0083] Figure 1 This is a flowchart illustrating a distributed optical-storage islanded grid power supply control method according to an exemplary embodiment;

[0084] Figure 2 This is a schematic diagram of a distributed optical-storage islanded grid power supply control system according to an exemplary embodiment;

[0085] Figure 3 This is a schematic diagram of the structure of a computer device according to an exemplary embodiment. Detailed Implementation

[0086] The following description and accompanying drawings fully illustrate specific embodiments described herein to enable those skilled in the art to practice them. Some portions and features of certain embodiments may be included in or replace portions and features of other embodiments. The scope of the embodiments herein includes the entire scope of the claims and all available equivalents thereof. The various embodiments described herein are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments; similar or identical parts between embodiments can be referred to interchangeably.

[0087] The modules in the apparatus or system of this application can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0088] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0089] Figure 1 The present invention illustrates a distributed optical-storage isolated grid power supply control method.

[0090] In this optional embodiment, the distributed optical-storage islanded grid power supply control method includes:

[0091] Step S101: Obtain historical photovoltaic power generation data and current meteorological information, and combine them with energy storage capacity status data. Use a long short-term memory network to perform trend analysis to obtain the cumulative trend of photovoltaic power generation prediction error. Based on the cumulative trend of photovoltaic power generation prediction error and energy storage capacity status data, generate an initial photovoltaic output plan.

[0092] It should be explained that the process involves real-time collection of hourly power generation records and daily cumulative values ​​from historical photovoltaic power generation data, obtaining three parameters—temperature, humidity, and wind speed—from current meteorological information, extracting the current remaining power percentage and charging / discharging power limits from energy storage capacity status data, and using a Long Short-Term Memory (LSTM) network to extract time-series features from historical data and meteorological information. Based on these extracted time-series features, the cumulative trend of photovoltaic power generation prediction errors is calculated. Based on this cumulative trend, combined with the remaining power percentage from the energy storage capacity status data, a linear regression is performed on the error accumulation trend to obtain the slope of error change. The correction coefficient for the daytime power generation prediction is calculated based on the product of the error change slope and the remaining power percentage, resulting in adjusted time-of-day power generation values. Simultaneously, historical load data is read from the power management database for the distribution area, and the average electricity consumption for the same period over the past seven days is extracted as an estimate of nighttime load demand. Based on the estimated nighttime load demand and the adjusted time-of-day power generation values, energy storage capacity constraints are determined. These constraints include maximum charging power, minimum hold-up capacity, and depth of discharge limits. By comparing the difference between daytime power generation and the estimated nighttime load demand, the energy storage charge / discharge switching time point is calculated. If the difference exceeds a preset threshold for the maximum energy storage capacity, the charge / discharge switching time point is adjusted and advanced by a preset time. Combining the energy storage charge / discharge switching time point and energy storage capacity constraints, an initial photovoltaic power output plan is generated based on the adjusted power generation value for each time period, the estimated nighttime load demand, and the energy storage charge / discharge switching time point. This plan includes the photovoltaic power generation forecast for each time period within 24 hours, the corresponding estimated load demand for that time period, and the energy storage charge / discharge schedule.

[0093] Specifically, the collection of historical photovoltaic power generation data covers the power generation records of all photovoltaic modules within the distribution area, including hourly instantaneous power values, daily cumulative power generation, and monthly summary data. Meteorological information is acquired in real time through meteorological monitoring stations installed at the photovoltaic power station. Temperature sensors record ambient temperature and module backsheet temperature, humidity sensors measure relative humidity, and anemometers monitor wind speed and direction. Energy storage capacity status data is read from the battery management system, including the battery pack's state of charge, available capacity, charge / discharge power limits, and battery health indicators. The temporal feature extraction process of the Long Short-Term Memory (LSTM) network adopts a three-layer LSTM architecture. The input layer receives the normalized historical power generation data sequence and meteorological parameter sequence, and the hidden layer contains 128 LSTM units, each controlling information flow through a forget gate, input gate, and output gate. The network is trained using the backpropagation algorithm to learn the temporal dependencies in historical data, especially the impact patterns of weather changes on power generation. After training, the network output layer generates a power generation prediction sequence for the next 24 hours and calculates the prediction confidence interval for each time point. The error accumulation trend is obtained by comparing the difference between historical forecast values ​​and actual power generation values. The prediction error for each period in the past seven days is calculated using the sliding window method to form an error time series, and then the accumulation characteristics and variation patterns of the error are analyzed.

[0094] It should be noted that the slope of error change is calculated using the least squares method to perform linear regression fitting on the cumulative error trend. Specifically, the error value over the past 168 hours is used as the dependent variable, and the time series is used as the independent variable. The regression coefficient is obtained by solving the system of equations; this coefficient is the slope of error change. The determination of the correction coefficient combines the slope of error change with the current percentage of remaining energy storage capacity. When the error is on an upward trend and the energy storage capacity is sufficient, a larger correction coefficient is used to conservatively estimate power generation; conversely, when the error is on a downward trend or the energy storage capacity is insufficient, a smaller correction coefficient is used.

[0095] In one possible implementation, the district's electricity management database stores all users' electricity consumption records, including time-of-use data for residential, commercial, and industrial electricity consumption. Electricity consumption data for the same time period over the past seven days is extracted from the database, and the hourly average is calculated as a baseline load. Considering the differences in electricity consumption between weekends and weekdays, a date-type weighting factor is introduced, with a weight of 0.7 for weekdays and 0.3 for weekends. A weighted average is then used to obtain a more accurate estimate of nighttime load demand.

[0096] The determination of energy storage capacity constraints is based on the battery's physical characteristics and safe operation requirements. Maximum charging power is limited by the battery's charging rate, typically set between 0.5C and 1C; minimum hold-up capacity, considering battery life protection, is generally maintained at over 20% of the total capacity; depth of discharge is limited according to battery type, reaching up to 90% for lithium iron phosphate batteries and controlled below 80% for ternary lithium batteries. These constraint parameters are dynamically adjusted by real-time monitoring of battery temperature and internal resistance changes.

[0097] For example, the calculation process for the energy storage charging / discharging switching time first analyzes the intersection of the daytime photovoltaic power generation curve and the nighttime load demand curve. In a typical scenario of an independent power supply area, the peak period for photovoltaic power generation is from 9:00 AM to 3:00 PM, during which the power generation exceeds the immediate load demand, and the excess power is used for energy storage charging. After 4:00 PM, photovoltaic power generation gradually decreases, and when the power generation is lower than the load demand, the system switches to energy storage discharging mode. If the forecast shows that the load demand is high the next morning, the switching time is adjusted in advance to ensure that the energy storage equipment is charged before the peak load arrives.

[0098] Understandably, when the difference between daytime power generation and estimated nighttime load demand exceeds a preset threshold for the maximum energy storage capacity, it indicates that the energy storage system may not be able to fully balance the supply and demand imbalance. In this case, adjusting the charging / discharging switching time, extending the energy storage charging period, or starting discharging earlier can optimize the energy allocation strategy. The preset threshold is derived from historical operating data and is typically set between 70% and 85% of the energy storage capacity, with the specific value dynamically adjusted based on season and weather conditions.

[0099] For example, under sunny summer conditions, when photovoltaic power generation is sufficient, the preset threshold can be raised to 85% to allow for more energy fluctuations; while under cloudy and rainy winter conditions, when photovoltaic power generation is unstable, the preset threshold can be lowered to 70% to improve the system's safety margin. The determination of the adjustment duration takes into account the charging and discharging rate of the energy storage equipment and the remaining available time, and is usually selected within the range of 1 to 3 hours.

[0100] The initial photovoltaic power output plan was developed by comprehensively considering energy balance across all time periods. The 24-hour period was divided into 96 15-minute time slots, and each time slot employed a corresponding operational strategy based on adjusted power generation, estimated load demand, and energy storage status. During periods of ample daylight, immediate load demand was prioritized, with surplus electricity used for energy storage charging. In the evening, when sunlight weakened, photovoltaic and energy storage systems worked together to provide power. At night, during periods of no sunlight, power supply relied entirely on energy storage discharge. The plan also included contingency plans; if actual operation deviated significantly from the plan, the system automatically activated a backup scheduling strategy, rapidly adjusting the charging and discharging power of the energy storage to maintain power supply stability.

[0101] Step S102: Based on the cumulative trend of photovoltaic power generation prediction error, obtain the nighttime load peak demand distribution data, and classify the actual power generation gap level using support vector machine; determine the risk level of insufficient energy storage reserved capacity based on the classification results of the actual power generation gap level.

[0102] It should be explained that, based on the error time series data in the cumulative trend of photovoltaic power generation prediction errors, the daily error peak and its occurrence time are extracted to obtain the nighttime load peak demand distribution data. This distribution data includes the hourly load peak, average load, and load fluctuation range from 6 PM to 6 AM the following morning. If the peak value of the cumulative error trend exceeds a preset peak threshold, the time series exceeding the threshold and the difference between the error value and the threshold within the corresponding time period are extracted. Based on the time series exceeding the threshold and the difference, combined with the load peak of each time period in the nighttime load peak demand distribution data, the number of overlapping hours between the load peak period and the error exceeding the threshold period is calculated. The current remaining capacity percentage is read from the energy storage monitoring data, and a feature vector containing the number of overlapping hours, the difference between the total load demand and the predicted power generation, and the remaining energy storage capacity percentage is constructed. A support vector machine is used to classify the feature vector. By calculating the vertical distance from the feature vector to the classification hyperplane and comparing it with a preset distance threshold, the actual power generation gap level is determined. Based on the actual power generation gap level, combined with the rated capacity of the energy storage device and the current charging status, if the power generation gap level is severe and the remaining energy storage capacity is lower than the first preset ratio of the rated capacity, the risk level of insufficient energy storage reserve capacity is determined to be high risk; if the power generation gap level is moderate and the remaining energy storage capacity is lower than the second preset ratio of the rated capacity, the risk level is determined to be medium risk; otherwise, it is determined to be low risk, thus obtaining the risk level of insufficient energy storage reserve capacity.

[0103] Specifically, the peak error is extracted by scanning the cumulative trend data of photovoltaic power generation prediction errors point by point. Each data point represents the error value for one hour, and local maxima are identified by comparing the error values ​​of adjacent time periods. The nighttime peak load demand distribution data is obtained from the distribution area load monitoring terminal, covering 12 hours of electricity consumption data from 6 pm to 6 am the next day. Each hourly record includes the maximum load value, average load value, and load fluctuation range within that time period.

[0104] Specifically, the calculation of overlapping hours requires first identifying the error-exceeding periods and the peak load periods. Error-exceeding periods are defined as continuous time intervals where the error value exceeds a preset peak threshold, and peak load periods are defined as time intervals where the load value exceeds 1.5 times the daily average load. The number of overlapping hours is then calculated by performing an intersection operation on the two time interval sequences.

[0105] For example, if the error exceeds the limit between 7 PM and 10 PM, and the load peak occurs between 8 PM and 11 PM, the overlap hours are 2 hours. Support Vector Machine (SVM) classification uses radial basis functions as kernel functions to map three-dimensional feature vectors to a high-dimensional space for linear classification. During the training phase, historical data is used to build the classification model, with the overlap hours, load-generation difference, and remaining energy storage capacity percentage as input features, and the severity of the actual power shortage event as the classification label. The classification hyperplane is obtained by solving a quadratic programming problem, and the vertical distance is calculated using the geometric distance formula from a point to the hyperplane. The distance threshold is determined based on the statistical distribution of historical power shortage events. A hierarchical judgment mechanism is used to determine the risk level. The first preset ratio is set at 30%, and the second preset ratio is set at 50%. These ratios are determined based on the safe operating boundaries of the energy storage equipment. When the generation gap level is severe and the remaining energy storage capacity is below 30%, the system faces a high risk of power outage; when the gap level is moderate and the remaining capacity is below 50%, it is in a medium-risk state; other situations are judged as low-risk.

[0106] The risk level also takes into account the impact of seasonal factors. During the summer peak electricity consumption period, the risk level determination threshold is lowered by 5 percentage points, while during the winter heating season, the threshold is raised by 5 percentage points, thus improving the accuracy of risk assessment through dynamic adjustments.

[0107] Step S103: Extract the off-peak charging time period data from the energy storage reserve capacity insufficient risk level, calculate the capacity allocation pressure index using time series analysis algorithm; and determine the quantitative value of the overlap between peak load and off-peak charging based on the capacity allocation pressure index.

[0108] It needs to be explained that the risk level identifier and risk period distribution are extracted from the risk level of insufficient reserved capacity for energy storage. Off-peak charging time period data is obtained, which includes the start and end times of off-peak electricity price periods, the upper limit of charging power for energy storage devices, and historical charging completion time records. Based on the risk level identifier, the time period range requiring key monitoring is determined, resulting in a time series dataset containing energy storage charging and discharging records. An autoregressive moving average method is used to process the time series dataset. The unit time charging demand is obtained by calculating the ratio of energy storage charging demand to available charging time. The change value of remaining energy storage capacity in each period is extracted from the time series dataset. The capacity allocation pressure index is determined based on the absolute value of the difference between the unit time charging demand and the change value of remaining energy storage capacity. Based on the capacity allocation pressure index, the peak load period and the off-peak electricity price charging period are compared. The proportion of their time intersection to the total duration of the peak load period is calculated. If the proportion exceeds a preset overlap threshold, the proportion is added to the pressure index and divided by two to obtain a normalized value, which is then used to determine the quantitative value of the overlap between peak load and off-peak charging.

[0109] Specifically, the risk level is represented numerically, with high risk assigned a value of 3, medium risk assigned a value of 2, and low risk assigned a value of 1. The risk period distribution records the specific time intervals corresponding to each risk level. Off-peak charging time period data is extracted from the grid's time-of-use electricity price table, typically corresponding to the nighttime off-peak electricity price period. The upper limit of the charging power of the energy storage device is dynamically determined by the battery management system based on the current battery temperature and state of charge.

[0110] Specifically, when processing time series data, the autoregressive moving average method first performs a stationarity test on the energy storage charging and discharging records, eliminating the trend term through differencing. The autoregressive part uses observations from the past p time points to predict the current value, while the moving average part uses random error terms from the past q time points to correct the prediction. The model order p and q are determined using the Akaike Information Criterion to ensure that the model accurately fits historical data while avoiding overfitting. The charging demand per unit time is obtained by dividing the total energy storage charging demand by the available charging time, reflecting the urgency of the charging task.

[0111] It should be noted that the calculation of the capacity allocation pressure index takes into account both the urgency of charging demand and the changes in energy storage capacity. When the charging demand per unit time exceeds the change in remaining energy storage capacity, it indicates that the charging speed cannot meet the demand, and the pressure index is the absolute value of the difference between the two. When the charging speed meets the demand, the pressure index decreases accordingly. This index directly reflects the allocation pressure of the energy storage system during a specific period.

[0112] The calculation of time intersection uses set operations. Peak load periods are identified by scanning the 24-hour load curve, pinpointing consecutive periods where the load exceeds a preset multiple of the daily average load. Off-peak charging periods are obtained directly from electricity price information. The intersection of the two time period sequences is the overlapping period, and the proportion of the intersection duration to the total duration of the peak load period reflects the severity of the overlap.

[0113] In one possible implementation, the normalization process maps the proportional value and the pressure index to the same dimensional range. An arithmetic mean of the two values, added together and divided by two, yields a quantified value of overlap between 0 and 1. The closer this value is to 1, the more severe the conflict between peak load and off-peak charging, requiring priority adjustment of energy storage charging and discharging plans to avoid power outage risks.

[0114] Step S104: Based on the quantified value of the overlap between peak and off-peak load recharge and the historical record of allocation mismatch, adjust the energy storage capacity allocation ratio and determine the dynamic energy storage reserve increment; calculate the capacity allocation parameters based on the capacity allocation pressure index and the dynamic energy storage reserve increment.

[0115] It should be explained that if the overlap quantification value is higher than the overlap threshold, the historical data of allocation mismatch is queried based on the overlap quantification value of peak load and off-peak charging. This historical data includes the timestamp, allocation ratio deviation value, and actual power shortage value of each energy storage capacity allocation failure event within a preset time period. The historical allocation deviation characteristic value is obtained by calculating the average and standard deviation of the deviation values. Based on this historical allocation deviation characteristic value, a Bayesian inference method is used to calculate the posterior probability of energy storage capacity allocation failure in the current period. The adjustment range of the energy storage capacity allocation ratio is determined based on the comparison between the posterior probability and the preset probability threshold. The adjustment range is multiplied by the overlap quantification value to obtain the initial adjustment value. The initial adjustment value is mapped to the zero-to-one interval to determine the dynamic energy storage reserve increment. Combining the dynamic energy storage reserve increment and the capacity allocation pressure index, a weighted summation method is used to calculate the energy storage allocation ratio adjustment value. The weight coefficients of the dynamic energy storage reserve increment and the capacity allocation pressure index are determined based on historical operating data. The reserve capacity increment value is calculated based on the energy storage allocation ratio adjustment value. Based on the energy storage allocation ratio adjustment value and the reserved capacity increment value, the off-peak recharge period is segmented, and the charging start time that minimizes the overlap between the load peak period and the charging period in each segment is calculated to obtain the recharge time optimization value. The energy storage allocation ratio, the reserved capacity increment and the recharge time optimization value are combined to form the capacity allocation parameter.

[0116] Specifically, in one implementation, the overlap threshold is set based on statistical analysis of historical operating data from independent power supply areas. By collecting the quantified overlap values ​​of all power outage events occurring within the past year, calculating their distribution characteristics, and setting the quantified value corresponding to a cumulative probability of 85% as the threshold, the system is considered to face a high risk of power outage when the actual overlap value exceeds this threshold, requiring the activation of the capacity allocation parameter adjustment mechanism. The process of constructing the allocation mismatch history involves the integration of multiple data sources. The energy storage management system records detailed information for each charge / discharge switching event in real time, including the switching time, remaining energy storage capacity, load demand, and whether a power outage occurred. When the actual load demand exceeds the capacity available from the energy storage, leading to a power outage, the system records this event as an allocation failure event. The allocation ratio deviation value is obtained by calculating the difference between the actual demand capacity and the pre-allocated capacity, while the actual power shortage is recorded as the cumulative power gap during the power outage. The historical allocation deviation characteristic value includes two dimensions: the mean deviation and the standard deviation. The mean reflects systematic deviation, and the standard deviation reflects the degree of fluctuation in deviation.

[0117] It should be noted that Bayesian inference methods have unique advantages in this scenario. The prior probability is determined based on the frequency of historical allocation failure events, and the likelihood function is constructed by analyzing the correlation between historical allocation deviation characteristic values ​​and allocation failure events. Once the allocation deviation characteristic value for the current period is obtained, the posterior probability is calculated according to Bayes' theorem, which is the probability of allocation failure given the current characteristic value. The comparison between the posterior probability and a preset probability threshold determines the adjustment magnitude; a higher posterior probability results in a larger adjustment magnitude. The adjustment magnitude is multiplied by the quantified overlap value to obtain the initial adjustment value; this product operation considers the amplification effect of overlap on adjustment demand. Normalization uses a linear mapping method to map the initial adjustment value to the 0-1 interval, ensuring that the dynamic energy storage reserve increment is within a reasonable range.

[0118] The weighting coefficients in the weighted summation method are determined through regression analysis of historical operating data. Records of all capacity allocation events over the past six months are collected, including dynamic energy storage reserve increments, capacity allocation pressure indices, and power supply reliability indicators. A multiple linear regression method is used to fit the relationship between these three factors, and the regression coefficients are the weighting coefficients. The weighting coefficients reflect the degree of contribution of each factor to the final allocation effect.

[0119] In one possible implementation, the calculation of the energy storage allocation ratio adjustment value takes into account the priority of different types of loads within the distribution area. Residential electricity consumption has the highest priority, followed by commercial electricity consumption, and interruptible industrial loads have the lowest priority. Based on the comprehensive value obtained by weighted summation, combined with the load priority allocation matrix, the energy storage capacity allocation ratio corresponding to each type of load is determined.

[0120] For example, the segmentation of the off-peak charging period uses an equal-length segmentation method, dividing the entire off-peak electricity price period into multiple 15-minute sub-periods. For each sub-period, the expected charging end time if charging starts at that time, as well as the overlap duration between that charging period and the peak load period, are calculated. By iterating through all possible charging start times, the time that minimizes the overlap duration is found as the preferred charging start time. The optimized charging time is the difference between the preferred charging start time and the originally planned charging time.

[0121] Understandably, the combination of capacity allocation parameters needs to consider the mutual influence between parameters. The energy storage allocation ratio determines the energy storage capacity allocation for each time period, the reserved capacity increment adds a safety margin to the original allocation, and the optimized recharge time reduces conflicts with load peaks by adjusting the charging timing. These three parameters together constitute a complete capacity allocation scheme. Furthermore, the verification of capacity allocation parameters is achieved through simulation testing. The determined parameters are input into the power supply simulation model of the distribution area to simulate the operating scenario for the next 24 hours and evaluate the power supply reliability under different weather conditions and load fluctuations. If the simulation results show that the probability of power outage is still higher than acceptable, the parameters are fine-tuned until the power supply reliability requirements are met.

[0122] For example, in a practical application of an independent power supply area, when a potential conflict is detected between the peak load period from 7 PM to 9 PM and the off-peak charging period from 2 AM to 5 AM, the capacity allocation parameter adjustment process is automatically initiated. Analysis of historical data revealed that the average power shortage in similar scenarios is 50 kWh, with a standard deviation of 15 kWh. Bayesian inference calculated the current allocation failure probability to be 0.72, exceeding the threshold of 0.6, thus triggering the adjustment mechanism. The energy storage allocation ratio was increased from 60% to 75%, the reserved capacity increment was set at 20 kWh, and the recharge start time was advanced by 1.5 hours. This dynamic adjustment effectively reduced the risk of power outages and improved the power supply reliability of the independent power supply area.

[0123] Step S105: Combine the dynamic energy storage reserved increment and the cumulative trend of photovoltaic power generation prediction error to revise the initial photovoltaic output plan; based on the revised initial photovoltaic output plan, establish a preliminary photovoltaic-energy storage joint operation mode;

[0124] It should be explained that, by combining the dynamic energy storage reserved increment and the prediction error accumulation trend, the power generation forecast values ​​for each time period in the initial photovoltaic power output plan are obtained. A correction coefficient is determined based on the proportion of the dynamic energy storage reserved increment to the total energy storage capacity. This correction coefficient is multiplied by the power generation forecast values ​​for each time period, and then the error accumulation trend value is subtracted to obtain the corrected power generation sequence for each time period. Based on this corrected power generation sequence, periods when power generation exceeds load demand are designated as photovoltaic main supply periods, and periods when power generation is less than load demand are designated as energy storage supply periods. During the photovoltaic main supply period, the difference between power generation and load demand is calculated as the energy storage charging power. During the energy storage supply period, the discharge power is determined based on the ratio of load demand to the remaining energy storage capacity, thus forming a photovoltaic-energy storage power allocation scheme. According to the photovoltaic-storage power allocation scheme, the output power limit of the photovoltaic inverter is set as a preset percentage of the rated power, and the charging and discharging switching threshold of the energy storage converter is set as a preset ratio of the rated capacity of the energy storage. When the real-time power generation is greater than the load demand and the energy storage is not fully charged, the photovoltaic priority power supply state is entered. When the power generation is insufficient and the energy storage has power, the energy storage supplementary power supply state is entered. When there is no photovoltaic power generation, the energy storage independent power supply state is entered, thus establishing a preliminary mode of photovoltaic-storage joint operation.

[0125] Specifically, in one implementation, the dynamic energy storage reserve increment reflects the additional capacity reserves required by the energy storage system to cope with uncertainties. The prediction error accumulation trend records the evolution of historical prediction deviations. The correction coefficient is calculated as the ratio of the dynamic energy storage reserve increment to the total energy storage capacity; a higher ratio indicates greater uncertainty faced by the system, requiring a more conservative correction strategy.

[0126] Specifically, the correction process employs a two-step calculation method. The first step multiplies the correction coefficient by the original power generation forecast for each time period to obtain the preliminary adjustment value. The second step subtracts the cumulative error trend value for the corresponding time period from the preliminary adjustment value to obtain the final corrected power generation. This correction method considers both the overall adjustment needs arising from energy storage capacity constraints and the individualized characteristics of errors in each time period. The corrected power generation sequence more closely reflects actual operating conditions, providing a reliable basis for subsequent power allocation.

[0127] It should be noted that the criteria for dividing time periods are based on a real-time comparison of power generation and load demand. When the corrected power generation for a given period exceeds the predicted load demand for that period, that period is designated as the photovoltaic (PV) main supply period; otherwise, it is designated as the energy storage supply period. During the PV main supply period, excess power generation is used for energy storage charging. The charging power equals the difference between power generation and load demand, but does not exceed the maximum charging power limit of the energy storage device. During the energy storage supply period, the discharge power is determined by comprehensively considering the load demand and the remaining energy storage capacity. The actual discharge power value is obtained by calculating the ratio of the two and multiplying it by the rated discharge power of the energy storage.

[0128] Once the photovoltaic-storage power allocation scheme is finalized, the system parameters are set according to the principle of safety margin. The output power limit of the photovoltaic inverter is typically set to 90% to 95% of its rated power to avoid long-term full-load operation of the equipment. The charge / discharge switching thresholds of the energy storage converter are set to 20% and 80% of the rated energy storage capacity, respectively, corresponding to the judgment conditions for discharge termination and charge termination.

[0129] In one possible implementation, the switching between the three operating states is achieved through real-time monitoring and logical judgment. The photovoltaic priority power supply state is suitable for the daytime when there is sufficient sunlight, at which time the photovoltaic directly supplies the load, and the excess power charges the energy storage; the energy storage supplementary power supply state is activated when the sunlight weakens in the evening, and the photovoltaic and energy storage jointly supply power to meet the load demand; the energy storage independent power supply state operates at night under the condition of no sunlight, and relies entirely on energy storage to maintain power supply.

[0130] Step S106: Based on the preliminary mode of photovoltaic-storage joint operation, obtain real-time power supply reliability indicators, iteratively optimize capacity allocation parameters based on real-time power supply reliability indicators, and determine the collaborative strategy for the operation of the photovoltaic-storage system.

[0131] It should be explained that, according to the initial model of photovoltaic-storage joint operation, real-time power supply reliability indicators are obtained from the power supply monitoring terminal. These indicators include values ​​for three dimensions: power supply continuity, voltage qualification rate, and frequency stability. The weights of the three dimensions are set to preset weight values. A comprehensive reliability score is calculated by weighted summation. If the score is lower than the target threshold, an iterative optimization process is initiated to obtain the optimized values ​​for energy storage allocation ratio, reserved capacity increment, and recharge time in the current capacity allocation parameters. The gradient descent method is used to adjust each current capacity allocation parameter. The adjustment amount of the energy storage allocation ratio is equal to the difference between the reliability score and the target threshold multiplied by a preset adjustment coefficient. The reserved capacity increment is corrected according to the ratio of the historical power shortage event frequency to a preset frequency threshold. The optimized recharge time value is adjusted by identifying the load peak and valley time intervals. The iterative process continues until the reliability score reaches the target threshold or the number of iterations reaches a preset upper limit, resulting in an optimized set of capacity allocation parameters. Based on the optimized capacity allocation parameter set, including the energy storage allocation ratio and reserved capacity increment, a photovoltaic (PV) output adjustment scheme is formulated. Output regulation is achieved by setting the upper limit of PV inverter output power for each time period. Simultaneously, an energy storage charge / discharge plan is generated based on the optimized recharge time, determining the charge / discharge power and duration for each time period. This charge / discharge plan corresponds sequentially with the PV output adjustment scheme. Based on the PV output adjustment scheme and the energy storage charge / discharge plan, a load demand response mechanism is established. Time-adjustable loads and interruptible loads are identified from the load management database. During periods of power shortage, time-adjustable loads are shifted, and power supply to interruptible loads is suspended. The PV output adjustment scheme, energy storage charge / discharge plan, and load demand response mechanism are integrated sequentially to determine a coordinated strategy for the operation of the PV-energy storage system.

[0132] Specifically, in one implementation, a power supply monitoring terminal is deployed on the low-voltage side of the distribution transformer in the transformer substation, and a high-precision power quality analyzer collects voltage, current, and frequency data in real time. Power supply continuity is calculated by statistically analyzing the number of power outages and their duration per unit time. Voltage qualification rate is determined based on the proportion of time the voltage deviation exceeds the rated value by ±7%. Frequency stability is assessed based on the root mean square error of the frequency deviation from the 50Hz standard value. The preset weight values ​​for the three dimensions are determined according to the power consumption characteristics of the transformer substation: the weight for power supply continuity in residential areas is set to 0.5, the weight for voltage qualification rate is set to 0.3, and the weight for frequency stability is set to 0.2.

[0133] Specifically, the application of the gradient descent method requires constructing an objective function and calculating the gradient. The objective function is defined as the squared difference between the comprehensive reliability score and the target threshold. The gradient is obtained by calculating the partial derivatives of each parameter with respect to the objective function using numerical differentiation methods. The adjustment coefficient for the energy storage allocation ratio is set based on historical optimization experience, typically between 0.01 and 0.05, to ensure the stability of parameter updates. The correction of the reserved capacity increment is based on statistical analysis. When the frequency of historical power shortage events exceeds a preset frequency threshold, the increment is amplified by the square root of the excess ratio; conversely, it is reduced proportionally. The adjustment of the recharge time optimization value is achieved by analyzing the peak and valley characteristics of the load curve, identifying the time interval between the load trough and peak periods, and adjusting the charging period to the center of the trough period to maximize the use of low-load periods for energy storage replenishment. During the iteration process, the reliability score is recalculated after each parameter update. Optimization is terminated when the score improvement is less than 0.001 after three consecutive iterations or when the number of iterations reaches 100. The formulation of the photovoltaic output adjustment scheme needs to balance power generation efficiency and system stability. The upper limit of output power for each time period is dynamically determined based on the load demand and energy storage status for that period. When the energy storage is close to full charge, the photovoltaic output is reduced to avoid curtailment. When the energy storage capacity is insufficient, the photovoltaic output is increased to increase the proportion of direct supply. The upper limit of power is set using a piecewise linear function. In the energy storage state of charge range of 0-30%, the photovoltaic output is at full power. In the range of 30-70%, it is linearly reduced to 80% of the rated power. In the range of 70-100%, it is further reduced to 60%.

[0134] The energy storage charge / discharge plan follows the principle of battery life protection. The charging power employs a constant current / constant voltage charging strategy based on the current state of charge (SOC). When the SOC is below 80%, constant current charging is used with a charging current of 0.5C. Once the SOC exceeds 80%, constant voltage charging is switched, gradually reducing the charging current. The discharge power is determined by considering both the urgency of the load demand and the sufficiency of the remaining energy storage capacity. The base discharge power is calculated by the ratio of these two factors, and then multiplied by a temperature correction factor to obtain the actual discharge power.

[0135] In one possible implementation, the load management database records the type, rated power, usage period, and dispatchability of all electrical equipment within the distribution area. Time-adjustable loads mainly include devices with flexible usage times, such as electric water heaters, washing machines, and charging stations; the operating hours of these devices can be flexibly adjusted within a 4-8 hour range. Interruptible loads include non-critical electrical equipment such as landscape lighting, billboards, and some air conditioning loads; these can be suspended for 15-30 minutes during periods of power shortage without affecting basic power needs.

[0136] The load demand response mechanism employs a tiered response strategy. A Level 1 response is initiated when the power supply margin falls below 20%, delaying the operation of time-adjustable loads by 1-2 hours. A Level 2 response is initiated when the margin falls below 10%, suspending power supply to some interruptible loads. A Level 3 response is initiated when the margin falls below 5%, fully implementing the load reduction plan. Each response level has minimum duration and maximum execution frequency limits to avoid frequent adjustments impacting users.

[0137] Understandably, the time-series integration of the three schemes is a key aspect of the collaborative strategy. The integration process employs a time-slicing method, dividing 24 hours into 96 15-minute time slices. Within each time slice, the photovoltaic output power, energy storage charging / discharging power, and load adjustment are clearly defined. Transitions between time slices are smoothly handled to avoid sudden power surges impacting the system. Furthermore, the effectiveness of the collaborative strategy is evaluated through simulation verification and comparison with actual operating data. The simulation model considers various uncertainties such as weather changes, load fluctuations, and equipment failures, simulating a one-month operating scenario. In actual operation, power supply reliability indicators are automatically calculated daily and compared with simulation predictions; when the deviation exceeds 5%, adaptive adjustments to the strategy parameters are triggered.

[0138] For example, during the high temperatures of summer, a surge in air conditioning load at night in a certain independent power supply area caused the power supply reliability to drop to 92%, below the target threshold of 95%. Iterative optimization was initiated, and after 12 iterations, the energy storage allocation ratio increased from 60% to 72%, the reserved capacity increment increased from 15kWh to 25kWh, and the recharge time was moved forward by 2 hours to 1:00 AM. Simultaneously, a coordinated strategy was implemented: limiting the output power of photovoltaic systems to 85% of their rated power in the afternoon to prioritize charging energy storage; adjusting the heating time of some electric water heaters to the early morning hours; and suspending landscape lighting for 30 minutes during peak evening electricity consumption. Through this comprehensive regulation, the power supply reliability improved to 96.5%, meeting the power supply requirements of the independent power supply area.

[0139] Step S107: Based on the collaborative strategy of the photovoltaic-storage system operation, generate photovoltaic-storage system operation adjustment instructions and deploy them to the energy storage system to obtain the power supply reliability operation status.

[0140] It should be explained that, based on the collaborative strategy of the photovoltaic-storage system operation, the photovoltaic output adjustment scheme, energy storage charging and discharging plan, and load demand response mechanism are converted into control command format. These commands include execution time, action type, target device address, and parameter values. For the actual power shortage and peak overlap periods in independent power supply scenarios for distribution areas, priority identifiers and execution condition judgment logic are added to these commands to obtain the photovoltaic-storage system operation adjustment command set. This command set is deployed to the energy storage system through a preset data transmission protocol. During deployment, commands are sent to the energy storage controller in a time sequence. After execution, the response status data of the energy storage system is obtained, including actual charging and discharging power, state of charge changes, and power supply continuity records. By comparing the number and duration of power outages before and after execution, the degree of improvement in power supply reliability is determined, resulting in the improved power supply reliability operating status.

[0141] Specifically, the control command format adopts a standardized data structure design. Execution time is represented by a timestamp, accurate to the second; action types include three basic states: charging, discharging, and standby; the target device address uses a unique device identifier; parameter values ​​include power settings and duration. Priority is divided into three levels: urgent, normal, and delayed. The urgent level is used to handle actual power shortages, the normal level for routine scheduling, and the delayed level for non-critical operations. The data transmission protocol adopts an industry-standard communication protocol to ensure the reliability of command transmission. During deployment, commands are sequentially sent to the energy storage controller according to their execution time. Each command waits for an acknowledgment signal after being sent; if no acknowledgment is received, it is resent, with a maximum of three retries. Response status data is obtained through real-time monitoring. The number of power outages is calculated by counting the number of events where the voltage drops below 80% of the rated value per unit time, and the duration is recorded as the difference between the start and end times of each outage. The degree of improvement in power supply reliability is calculated through comparative indicators. Before execution, record the baseline number of power outages and the total duration. After execution, record the corresponding data within the same duration. The degree of improvement is equal to the average of the percentage reduction in the number of outages and the percentage reduction in the duration. When the degree of improvement exceeds the preset improvement threshold, it is determined to be an effective improvement in power supply reliability.

[0142] Figure 2 The distributed optical-storage isolated grid power supply control system of the present invention is shown.

[0143] In this optional embodiment, the distributed optical-storage islanded grid power supply control system includes:

[0144] The power generation forecasting and initial scheme generation module 201 is used to acquire historical data of photovoltaic power generation and current meteorological information, and combine it with energy storage capacity status data. It uses a long short-term memory network to perform trend analysis to obtain the cumulative trend of photovoltaic power generation forecasting error. Based on the cumulative trend of photovoltaic power generation forecasting error and energy storage capacity status data, it generates an initial scheme for photovoltaic power output plan.

[0145] The energy storage risk level assessment module 202 is used to obtain nighttime load peak demand distribution data based on the cumulative trend of photovoltaic power generation prediction error, and classify the actual power generation gap level through support vector machine; and determine the risk level of insufficient energy storage reserved capacity based on the classification results of the actual power generation gap level.

[0146] The load and recharge overlap analysis module 203 is used to extract data on the off-peak recharge period from the risk level of insufficient reserved capacity of energy storage, calculate the capacity allocation pressure index using time series analysis algorithm, and determine the quantitative value of the overlap between peak load and off-peak recharge based on the capacity allocation pressure index.

[0147] The dynamic capacity allocation calculation module 204 is used to adjust the energy storage capacity allocation ratio and determine the dynamic energy storage reserve increment based on the quantitative value of the overlap between peak load and valley charging and the historical record of allocation mismatch; and to calculate the capacity allocation parameters based on the capacity allocation pressure index and the dynamic energy storage reserve increment.

[0148] The photovoltaic-storage joint operation modeling module 205 is used to combine the dynamic energy storage reserved increment and the cumulative trend of photovoltaic power generation prediction error to revise the initial photovoltaic output plan; and to establish a preliminary photovoltaic-storage joint operation model based on the revised initial photovoltaic output plan.

[0149] The collaborative operation strategy optimization module 206 is used to obtain real-time power supply reliability indicators based on the initial mode of photovoltaic and energy storage joint operation, iteratively optimize capacity allocation parameters based on the real-time power supply reliability indicators, and determine the collaborative operation strategy of the photovoltaic and energy storage system.

[0150] The instruction deployment and execution module 207 is used to generate operation adjustment instructions for the photovoltaic and energy storage system based on the collaborative strategy of the photovoltaic and energy storage system operation, and deploy them to the energy storage system to obtain the power supply reliability operation status.

[0151] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores static and dynamic information data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0152] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0153] In addition, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0154] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0155] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0156] This invention is not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this invention is limited only by the appended claims.

Claims

1. A distributed light storage isolated network power supply control method, characterized in that, The method comprises: acquiring photovoltaic power generation historical data and current meteorological information, combining energy storage capacity state data, using a long short-term memory network to perform trend analysis, and obtaining a photovoltaic power generation prediction error accumulation trend; based on the photovoltaic power generation prediction error accumulation trend and the energy storage capacity state data, generating an initial scheme of photovoltaic output planning; according to the photovoltaic power generation prediction error accumulation trend, acquiring night load peak demand distribution data, and classifying actual power generation gap levels through a support vector machine; based on the actual power generation gap level classification result, determining an energy storage reserved capacity shortage risk level; extracting valley recharge time period data from the energy storage reserved capacity shortage risk level, using a time series analysis algorithm to calculate a capacity deployment pressure index; and based on the capacity deployment pressure index, determining a load peak and valley recharge overlap degree quantization value; based on the load peak and valley recharge overlap degree quantization value and the distribution mismatch historical record, adjusting the energy storage capacity distribution proportion, determining a dynamic energy storage reserved increment; based on the capacity deployment pressure index and the dynamic energy storage reserved increment, calculating a capacity deployment parameter; combining the dynamic energy storage reserved increment and the photovoltaic power generation prediction error accumulation trend, correcting the initial scheme of photovoltaic output planning; based on the corrected initial scheme of photovoltaic output planning, establishing a preliminary mode of photovoltaic and energy storage joint operation; based on the preliminary mode of photovoltaic and energy storage joint operation, acquiring a real-time power supply reliability index, based on the real-time power supply reliability index, iteratively optimizing the capacity deployment parameter, and determining a collaborative strategy of photovoltaic and energy storage system operation; based on the collaborative strategy of photovoltaic and energy storage system operation, generating photovoltaic and energy storage system operation adjustment instructions, and deploying to the energy storage system to obtain a power supply reliability operation state.

2. The distributed optical storage isolated network power supply control method according to claim 1, characterized in that, The method comprises: acquiring photovoltaic power generation historical data and current meteorological information, combining energy storage capacity state data, using a long short-term memory network to perform trend analysis, and obtaining a photovoltaic power generation prediction error accumulation trend; based on the photovoltaic power generation prediction error accumulation trend and the energy storage capacity state data, generating an initial scheme of photovoltaic output planning; based on the photovoltaic power generation prediction error accumulation trend and the energy storage capacity state data, generating an initial scheme of photovoltaic output planning comprises the following steps: collecting hourly power generation and daily cumulative values in photovoltaic power generation historical data, acquiring temperature, humidity, and wind speed parameters in current meteorological information, and extracting remaining power percentage and charge-discharge power limit from energy storage capacity state data; using a long short-term memory network to extract time sequence features of the historical data and the current meteorological information, calculating based on the extracted time sequence features, and obtaining a photovoltaic power generation prediction error accumulation trend; based on the photovoltaic power generation prediction error accumulation trend and the remaining power percentage, fitting an error change slope through linear regression, calculating a correction coefficient of daily power generation prediction value based on the error change slope and the remaining power percentage, and obtaining an adjusted time period power value; generating a night load demand estimate value based on historical load data, combining the night load demand estimate value and the adjusted time period power value, determining an energy storage capacity constraint condition, and based on the constraint condition and the adjusted time period power value, generating an initial scheme of photovoltaic output planning.

3. The distributed power supply control method for an optical storage isolated network according to claim 1, wherein, The method comprises the following steps: According to the cumulative trend of photovoltaic power generation prediction error, the peak load demand distribution data at night is obtained, and the actual power generation gap level is classified by support vector machine; the risk level of insufficient reserved capacity of energy storage is determined based on the classification result of actual power generation gap level. According to the cumulative trend of photovoltaic power generation prediction error, the daily error peak value and error occurrence time are extracted, the peak load demand distribution data at night is obtained, and the distribution data includes hourly peak load, average load and load fluctuation range; Based on the error peak value and the peak load demand distribution data at night, the number of overlapping hours of the load peak period and the error exceeding period is calculated, and the feature vector including the number of overlapping hours and the percentage of remaining capacity of energy storage is constructed; 4. The distributed optical storage micro-grid power supply control method according to claim 1, characterized in that, The feature vector is classified by using support vector machine, and the actual power generation gap level is determined based on the classification result; the risk level of insufficient reserved capacity of energy storage is determined according to the actual power generation gap level and the percentage of remaining capacity of energy storage. The method comprises the following steps: According to the cumulative trend of photovoltaic power generation prediction error, the peak load demand distribution data at night is obtained, and the actual power generation gap level is classified by support vector machine; the risk level of insufficient reserved capacity of energy storage is determined based on the classification result of actual power generation gap level. Based on the capacity allocation pressure index and the dynamic energy storage reserved increment, the capacity allocation parameter is calculated, comprising the following steps: According to the cumulative trend of photovoltaic power generation prediction error, the peak load demand distribution data at night is obtained, and the actual power generation gap level is classified by support vector machine; the risk level of insufficient reserved capacity of energy storage is determined based on the classification result of actual power generation gap level.

5. The distributed power supply control method for an optical storage isolated network according to claim 1, wherein, Based on the dynamic energy storage reserved increment and the capacity allocation pressure index, the capacity allocation proportion adjustment value, the reserved capacity increment value and the recharge time optimization value are calculated, and the capacity allocation parameter is integrated. The method comprises the following steps: According to the cumulative trend of photovoltaic power generation prediction error, the peak load demand distribution data at night is obtained, and the actual power generation gap level is classified by support vector machine; the risk level of insufficient reserved capacity of energy storage is determined based on the classification result of actual power generation gap level. Based on the dynamic energy storage reserved increment and the cumulative trend of photovoltaic power generation prediction error, the initial scheme of photovoltaic output plan is corrected; based on the corrected initial scheme of photovoltaic output plan, the preliminary mode of photovoltaic and energy storage combined operation is established, comprising the following steps:

6. The distributed optical storage microgrid power supply control method according to claim 1, wherein, According to the proportion of dynamic energy storage reserved increment in total energy storage capacity, the correction coefficient is determined, and the power generation prediction value of each period is adjusted based on the correction coefficient to obtain the corrected period power sequence; Based on the corrected period power sequence, the photovoltaic main supply period and the energy storage power supply period are divided, and the photovoltaic and energy storage power distribution scheme is determined; Based on the photovoltaic and energy storage power distribution scheme, the parameters of photovoltaic inverter and energy storage converter are set, and the preliminary mode of photovoltaic and energy storage combined operation is established. The method comprises the following steps: According to the preliminary mode of photovoltaic and energy storage combined operation, the real-time power supply reliability index is obtained, and the capacity allocation parameter is iteratively optimized based on the real-time power supply reliability index to determine the collaborative strategy of photovoltaic and energy storage system operation. According to the preliminary mode of photovoltaic and energy storage combined operation, the real-time power supply reliability index is obtained, and the comprehensive reliability score is calculated; The capacity allocation parameter is adjusted by using gradient descent method, the photovoltaic output adjustment scheme and the energy storage charging and discharging plan are formulated based on the optimized capacity allocation parameter, the load demand response mechanism is established, and the collaborative strategy of photovoltaic and energy storage system operation is integrated.

7. The distributed power supply control method for optical storage isolated network according to claim 1, wherein, The cooperative strategy based on the operation of the light storage system generates an operation adjustment instruction of the light storage system and deploys the operation adjustment instruction to the energy storage system, so as to obtain a power supply reliability operation state, including the following steps: The cooperative strategy based on the operation of the light storage system generates an adjustment instruction set containing an execution time and a parameter value; The adjustment instruction set is deployed to the energy storage system through a data transmission protocol, and a power supply reliability operation state is determined based on response state data after execution.

8. A distributed optical storage isolated network power supply control system, characterized in that, The system comprises: A power generation prediction and initial scheme generation module is configured to obtain photovoltaic power generation historical data and current meteorological information, combine energy storage capacity state data, perform trend analysis by using a long short-term memory network, and obtain a photovoltaic power generation prediction error accumulation trend; based on the photovoltaic power generation prediction error accumulation trend and the energy storage capacity state data, an initial photovoltaic output plan is generated; An energy storage risk level evaluation module is configured to obtain night load peak demand distribution data according to the photovoltaic power generation prediction error accumulation trend, and classify actual power generation gap levels by using a support vector machine; based on the classification result of the actual power generation gap level, an energy storage reserved capacity deficiency risk level is determined; A load and recharge overlap analysis module is configured to extract low-valley recharge time period data from the energy storage reserved capacity deficiency risk level, calculate a capacity allocation pressure index by using a time series analysis algorithm, and determine a load peak and low-valley recharge overlap degree quantitative value based on the capacity allocation pressure index; A dynamic capacity allocation calculation module is configured to adjust an energy storage capacity allocation proportion according to the load peak and low-valley recharge overlap degree quantitative value and a distribution mismatch historical record, determine a dynamic energy storage reserved increment, and calculate a capacity allocation parameter based on the capacity allocation pressure index and the dynamic energy storage reserved increment; A light storage combined operation modeling module is configured to combine the dynamic energy storage reserved increment and the photovoltaic power generation prediction error accumulation trend, correct the initial photovoltaic output plan, and establish a light storage combined operation preliminary mode based on the corrected initial photovoltaic output plan; A cooperative operation strategy optimization module is configured to obtain real-time power supply reliability indexes according to the light storage combined operation preliminary mode, iteratively optimize the capacity allocation parameter based on the real-time power supply reliability indexes, and determine a cooperative strategy of the light storage system operation; An instruction deployment and execution module is configured to generate an operation adjustment instruction of the light storage system based on the cooperative strategy of the light storage system operation, deploy the operation adjustment instruction to the energy storage system, and obtain a power supply reliability operation state.

9. The distributed optical storage microgrid power supply control system according to claim 8, characterized in that, The photovoltaic power generation historical data and the current meteorological information are obtained, the energy storage capacity state data is combined, a long short-term memory network is used for trend analysis, and a photovoltaic power generation prediction error accumulation trend is obtained; Based on the photovoltaic power generation prediction error accumulation trend and the energy storage capacity state data, an initial photovoltaic output plan is generated, including the following steps: Hourly power generation and daily cumulative values in the photovoltaic power generation historical data are collected, temperature, humidity, and wind speed parameters in the current meteorological information are obtained, and a remaining power percentage and a charge-discharge power limit value are extracted from the energy storage capacity state data; A long short-term memory network is used to extract time sequence features of the historical data and the current meteorological information, and a photovoltaic power generation prediction error accumulation trend is obtained based on the extracted time sequence features. Based on the cumulative trend of photovoltaic power generation prediction error and the percentage of remaining power, the error change slope is obtained by linear regression fitting, and the correction coefficient of daytime power generation prediction value is calculated based on the error change slope and the percentage of remaining power, to obtain the adjusted time-sharing generation power value; According to the historical load data, the night load demand estimation value is generated, and the adjusted time-sharing generation power value is combined with the night load demand estimation value to determine the energy storage capacity constraint condition. Based on the constraint condition and the adjusted time-sharing generation power value, the photovoltaic output plan initial scheme is generated.

10. The distributed optical storage microgrid power supply control system of claim 8, wherein, According to the cumulative trend of photovoltaic power generation prediction error, the night load peak demand distribution data is obtained, and the actual generation gap level is classified by support vector machine; the risk level of insufficient reserved energy storage capacity is determined based on the classification result of the actual generation gap level, including the following steps: According to the cumulative trend of photovoltaic power generation prediction error, the daily error peak value and the error occurrence time are extracted, the night load peak demand distribution data is obtained, and the distribution data includes hourly load peak value, average load and load fluctuation range; Based on the error peak value and the night load peak demand distribution data, the number of overlapping hours of the load peak period and the error exceeding period is calculated, and the feature vector including the number of overlapping hours and the percentage of remaining energy storage capacity is constructed; The feature vector is classified by using support vector machine, and the actual generation gap level is determined based on the classification result; the risk level of insufficient reserved energy storage capacity is determined according to the actual generation gap level and the percentage of remaining energy storage capacity.

11. The distributed optical storage microgrid power supply control system of claim 8, wherein, The dynamic energy storage reserved increment is determined by adjusting the energy storage capacity allocation ratio according to the load peak value and the low valley recharge overlap degree quantitative value and the allocation mismatch historical record; The capacity allocation parameter is calculated based on the capacity allocation pressure index and the dynamic energy storage reserved increment, including the following steps: According to the load peak value and the low valley recharge overlap degree quantitative value, the allocation mismatch historical record is queried, and the deviation feature value is calculated based on the allocation mismatch historical record; The allocation failure probability is calculated by using Bayesian inference algorithm, the energy storage capacity allocation ratio is adjusted based on the allocation failure probability, and the dynamic energy storage reserved increment is determined; Based on the dynamic energy storage reserved increment and the capacity allocation pressure index, the capacity allocation ratio adjustment value, the reserved capacity increment value and the recharge time optimization value are calculated, and the capacity allocation parameter is integrated.

12. The distributed optical storage microgrid power supply control system of claim 8, wherein, The photovoltaic output plan initial scheme is corrected by combining the dynamic energy storage reserved increment and the cumulative trend of photovoltaic power generation prediction error; the preliminary mode of photovoltaic storage combined operation is established based on the corrected photovoltaic output plan initial scheme, including the following steps: The correction coefficient is determined according to the proportion of dynamic energy storage reserved increment in total energy storage capacity, and the generation prediction value of each period is adjusted based on the correction coefficient to obtain the corrected time-sharing generation power sequence; Based on the corrected time-sharing generation power sequence, the photovoltaic main supply period and the energy storage power supply period are divided to determine the photovoltaic storage power allocation scheme; Based on the photovoltaic storage power allocation scheme, the photovoltaic inverter and energy storage converter parameters are set to establish the preliminary mode of photovoltaic storage combined operation.

13. The distributed optical storage microgrid power supply control system of claim 8, wherein, The real-time power supply reliability index is obtained according to the preliminary mode of the combined operation of the light storage, the capacity allocation parameter is iteratively optimized based on the real-time power supply reliability index, and the cooperative strategy of the operation of the light storage system is determined, and the steps include the following: According to the preliminary mode of the combined operation of the light storage, the real-time power supply reliability index is obtained, and the comprehensive reliability score is calculated; The capacity allocation parameter is adjusted by using the gradient descent method, the light output adjustment scheme and the energy storage charging and discharging plan are formulated based on the optimized capacity allocation parameter, the load demand response mechanism is established, and the cooperative strategy of the operation of the light storage system is integrated.

14. The distributed optical storage microgrid power supply control system of claim 8, wherein, The cooperative strategy of the operation of the light storage system is used to generate the operation adjustment instruction of the light storage system, and is deployed to the energy storage system to obtain the power supply reliability operation state, and the steps include the following: Based on the cooperative strategy of the operation of the light storage system, an adjustment instruction set containing execution time and parameter value is generated; The adjustment instruction set is deployed to the energy storage system through a data transmission protocol, and the power supply reliability operation state is determined based on the response state data after execution.

15. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that, The processor executes the computer program to realize the steps of the method in any one of claims 1 to 7.

16. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method in any one of claims 1 to 7.

Images