A method and system for generating an automatic operation analysis report of a light storage and charging station based on multi-role collaborative decision-making

Abstract

This invention provides a method and system for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making. It relates to the field of intelligent operation and management technology for new energy power systems. The method includes: when executing an instruction to generate an operation analysis report, determining the acquisition parameters of the required data corresponding to each report item in the report to be generated; determining the required data from the operation data of the photovoltaic, energy storage, and charging station based on the acquisition parameters; allocating the required data to the corresponding operation agents based on their configured roles; receiving the analysis results from each operation agent; and generating the report content corresponding to each report item based on the analysis results. This invention's method and system for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making automatically generates highly readable and consistent daily operation reports, thereby improving the operation and management efficiency and decision support capabilities of photovoltaic, energy storage, and charging stations.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and management technology of new energy power systems, and in particular to a method and system for generating automated operation analysis reports for photovoltaic, energy storage and charging power stations based on multi-role collaborative decision-making. Background Technology

[0002] With the rapid deployment of photovoltaic power generation, energy storage systems, and electric vehicle charging facilities, integrated photovoltaic-energy storage-charging power stations are gradually becoming an important component of new energy infrastructure. These stations involve a variety of complex factors during operation, including fluctuations in photovoltaic output, energy storage charging and discharging strategies, changes in time-of-use electricity prices, and differences in user charging behavior, significantly increasing the difficulty of operation and management.

[0003] In existing technologies, the operation analysis of photovoltaic, energy storage, and charging stations typically relies on manual experience or statistical reports based on fixed rules. On the one hand, traditional daily reports can only reflect historical data from a single dimension and lack the ability to comprehensively analyze multi-source data. On the other hand, existing systems often separate operation analysis, problem diagnosis, and strategy recommendations, making it difficult to form a unified, continuous, and interpretable operation decision support system. This results in low response efficiency for operation personnel, highly subjective analysis conclusions, and an inability to support the needs of large-scale station operation. Summary of the Invention

[0004] One of the objectives of this invention is to provide a method and system for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making. By introducing a multi-role collaborative decision-making mechanism, the operation data of photovoltaic, energy storage, and charging stations are analyzed and processed in a clear division of labor and logical coordination, enabling intelligent identification of station operation status, operation trends, and abnormal situations, and automatically generating highly readable and consistent daily operation reports, thereby improving the operation management efficiency and decision support capabilities of photovoltaic, energy storage, and charging stations.

[0005] This invention provides a method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making, comprising: When executing the instruction to generate an operational analysis report, determine the acquisition parameters of the required data corresponding to each report item in the operational analysis report to be generated; Demand data is determined from the operational data of the photovoltaic storage and charging station based on the parameters obtained from the demand data. Based on the roles configured for each operational agent, the demand data is allocated to the corresponding operational agent; Receive analysis results from various operational intelligent agents; Based on the analysis results, generate the corresponding report content for each report item.

[0006] Preferably, the roles configured for each operational intelligent agent include one or more of the following: operational trend analysis role, revenue analysis role, anomaly diagnosis role, strategy evaluation role, and summary generation role; Among them, the operation trend analysis role is used to analyze the intraday variation trend of photovoltaic power generation, energy storage charging and discharging, and charging load; The revenue analysis role is used to analyze charging revenue at power stations, electricity price structure, and revenue changes. Anomaly diagnosis role, used to identify abnormal situations where operating indicators deviate from historical benchmarks or preset thresholds; The strategy evaluation role is used to assess the rationality of the current energy storage charging and discharging strategy and time-of-use operation strategy; The generated roles are used to synthesize and summarize the analysis results of each role, forming a unified operational conclusion.

[0007] Preferably, the steps for generating the operation analysis report generation instruction include: Receive interactive voice messages through the intelligent assistant interaction module; Semantic intent recognition for interactive voice; When the identified semantic intent is a report generation intent, generate an instruction to generate an operational analysis report corresponding to the identified intent; And / or, Monitor the operational intelligence agent. When the type of decision generated by the operational intelligence agent matches the type of decision in the pre-configured report generation trigger library, trigger the generation instruction of the corresponding operational analysis report. And / or, Based on a pre-configured report generation trigger time library, the corresponding operational analysis report generation instruction is triggered.

[0008] Preferably, the method for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making further includes: When the instruction to generate an operational analysis report is received that it is generated based on interactive voice, time verification is performed based on a pre-configured time verification analysis library. When the time verification result matches the time interval for predictive analysis configured in the time verification analysis library, predictive data is obtained by predicting operational data. Based on the generated forecast data, execute the instructions to generate an operational analysis report.

[0009] Preferably, the prediction steps for the prediction data include: Determine the ratio of the first time period corresponding to the data to be predicted to the time period corresponding to the current report as a reference parameter; When the reference parameter is less than or equal to a preset first threshold, historical operational data for a second time period prior to the current moment is obtained as analysis data; wherein, the second time period is a preset multiple of the first time period; The analysis data is sampled based on the first time length to obtain multiple sample data; Based on multiple sampled data, the predicted data is determined.

[0010] Preferably, the prediction step for the prediction data further includes: When the reference parameter is greater than the preset first threshold and less than or equal to the preset second threshold, historical operational data of the current time period and the previous preset number of time periods are obtained as analysis data. Based on a pre-configured sampling time interval, the analysis data is sampled to obtain sampled data; Based on the sampling time interval and the first time length, the number of prediction units contained in the prediction data is determined; Each prediction unit is extracted sequentially, and based on the pre-configured data extraction rules, the sampled data and / or the predicted prediction units are extracted as reference units. Based on the reference unit, predictions are made for the prediction unit; The first threshold is less than the second threshold.

[0011] Preferably, the prediction step for the prediction data further includes: When the reference parameter is greater than the preset second threshold, the historical operation data of the current time period and the previous preset number of time periods, as well as the historical operation data of other related photovoltaic storage and charging stations and the previous preset number of time periods, are obtained as analysis data. Based on a pre-configured sampling time interval, the analysis data is sampled to obtain sampled data; Based on the sampling time interval and the first time length, the number of prediction units contained in the prediction data is determined; Each prediction unit is extracted sequentially, and based on the pre-configured data extraction rules, the sampled data and / or the predicted prediction units are extracted as reference units. Based on the reference unit, predictions are made for the prediction unit; Preferably, the method for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making further includes: When executing the instruction to generate an optimization suggestion report, an optimization suggestion collection instruction is sent to each operational agent; Based on the optimization suggestions from various operational feedback, an optimization suggestion report is generated. Among them, the optimization suggestions from each operation feedback are generated by each operation intelligence agent based on its configured role, combined with the current real-time status of the photovoltaic storage and charging station and its historical knowledge base.

[0012] Preferably, the method for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making further includes: After the operational analysis report generated by combining the forecast data is produced, a monitoring dataset is constructed based on the forecasts of the forecast units. The operation of photovoltaic, energy storage and charging stations is monitored based on the monitoring dataset; When the monitoring results meet the preset correction conditions, the operation analysis report is corrected based on the monitoring results and the pre-configured report correction library; otherwise, the operation analysis report is regenerated.

[0013] The present invention also provides a system for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making, comprising: a report generation module; The report generation module includes: a parameter determination unit, a requirement data acquisition unit, a data allocation unit, a feedback receiving unit, and a content generation unit; When the instruction to generate an operation analysis report is executed, the parameter determination unit determines the acquisition parameters of the required data corresponding to each report item in the operation analysis report to be generated; the required data acquisition unit determines the required data from the operation data of the photovoltaic storage and charging station based on the acquisition parameters of the required data; the data allocation unit allocates the required data to the corresponding operation agents based on the roles configured for each operation agent; the feedback receiving unit receives the analysis results fed back by each operation agent; and the content generation unit generates the report content corresponding to the report item based on the analysis results.

[0014] The beneficial effects of this invention are as follows: Through a multi-role collaborative decision-making mechanism, multi-dimensional and systematic analysis of photovoltaic, energy storage, and charging station operation data is achieved, significantly improving the comprehensiveness and accuracy of operation analysis. By decomposing analysis tasks into roles through an operation intelligence agent, the workload of manual analysis is reduced, and the work efficiency of operation personnel is improved. Through a standardized daily report generation mechanism, the consistency and interpretability of photovoltaic, energy storage, and charging station operation information transmission are enhanced. It has good scalability and can flexibly add or adjust analysis roles according to actual operation needs to adapt to photovoltaic, energy storage, and charging stations of different sizes and types.

[0015] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0016] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of a method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making, according to an embodiment of the present invention. Figure 2 This is a schematic diagram of an automated operation analysis report generation system for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making, as described in an embodiment of the present invention. Detailed Implementation

[0018] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0019] This invention provides a method for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making, such as... Figure 1 As shown, it includes: Step 1: When executing the command to generate the operation analysis report, determine the acquisition parameters of the required data corresponding to each report item in the operation analysis report to be generated; The parameters to be acquired include one or more of the following: data type, start time, end time, and device number that generates the required data. Step 2: Obtain parameters based on demand data. Determine demand data from the operational data of the photovoltaic storage and charging stations. The operational data of photovoltaic-storage-charging stations comprises various core business data generated during their operation. This data directly reflects the station's energy production, storage, consumption, market environment, and equipment status. Operational data includes photovoltaic power generation data, energy storage charging and discharging data, charging pile charging data, time-of-use electricity price data, and equipment status data. Photovoltaic power generation data includes electricity-related data generated by photovoltaic modules, such as power generation, output power, power generation duration, and irradiance-related data. Energy storage charging and discharging data includes charging and discharging behavior data of the energy storage system, such as charging amount, discharging amount, charging and discharging power, remaining capacity, charging and discharging duration, and peak / valley charging and discharging ratios. The data includes charging pile data for electric vehicles, such as the number of charging orders, charging volume, charging duration, charging power, single-pile charging efficiency, and user charging time distribution. Time-of-use electricity price data includes electricity market price standards set for different time periods, such as different prices during peak, flat, and off-peak hours; this is the core basis for revenue analysis and charging / discharging strategy optimization. Equipment status data includes the operating status data of the core equipment in the photovoltaic-storage-charging station, such as whether the equipment is operating normally, fault codes, operating temperature, and voltage / current parameters. The core equipment in the photovoltaic-storage-charging station includes: photovoltaic inverters, energy storage batteries, charging piles, and controllers. Step 3: Based on the roles configured for each operational agent, allocate the required data to the corresponding operational agents; Each operational intelligent agent is configured with one or more of the following roles: operational trend analysis role, revenue analysis role, anomaly diagnosis role, strategy evaluation role, and summary generation role. Specifically, the operational trend analysis role analyzes the intraday trends of photovoltaic power generation, energy storage charging and discharging, and charging load; the revenue analysis role analyzes the charging revenue, electricity price structure, and revenue changes at the power station; the anomaly diagnosis role identifies abnormal situations where operational indicators deviate from historical benchmarks or preset thresholds; the strategy evaluation role assesses the rationality of the current energy storage charging and discharging strategy and time-of-use operation strategy; and the summary generation role synthesizes and summarizes the analysis results from each role to form a unified operational conclusion. Step 4: Receive the analysis results from each operational intelligence agent; Each operational intelligence agent is configured with different roles. These roles are specialized analytical perspectives categorized according to the core needs of photovoltaic, energy storage, and charging station operation analysis. Each role corresponds to a specific operational dimension and undertakes exclusive analytical tasks. Role-level analysis conclusions are targeted analytical results output after executing the analysis tasks of the corresponding functional role. For example, the operation trend analysis role outputs that yesterday's peak photovoltaic power generation occurred between 12:00 and 13:00, a decrease of 10% compared to the historical average; the anomaly diagnosis role outputs that the charging efficiency of charging pile No. 3 is 85%, lower than the preset threshold of 90%. Step 5: Based on the analysis results, generate the report content corresponding to the report item. That is, fill the data from the analysis results into the report content area of ​​the corresponding report item.

[0020] The present invention provides a method for generating automated operation analysis reports for photovoltaic-storage-charging stations based on multi-role collaborative decision-making. This method is applied in an intelligent operation analysis system for photovoltaic-storage-charging systems based on multi-role collaborative decision-making. The intelligent operation analysis system for photovoltaic-storage-charging systems based on multi-role collaborative decision-making includes: a data acquisition module, a data preprocessing and fusion module, a multi-role collaborative decision-making module, an operation intelligence agent module, a report module, and a result output module. Taking the generation of a daily operation report as an example, the functions of each module are as follows: The data acquisition module is used to collect operational data from the photovoltaic-storage-charging station. The operational data includes at least photovoltaic power generation data, energy storage charging and discharging data, charging pile charging data, time-of-use electricity price data, and equipment status data. The data preprocessing and fusion module is used to process the collected operational data... The system performs data cleaning, time alignment, indicator normalization, and multi-source data fusion to form a unified site operation dataset. A multi-role collaborative decision-making module, based on this unified dataset, generates analysis tasks for multiple functional roles according to preset role division rules and collaboratively integrates the analysis results from each role. An operation intelligence module, acting as the execution unit of the multi-role collaborative decision-making module, executes the operation analysis tasks corresponding to each functional role and generates role-level analysis conclusions. A reporting module structures and organizes the analysis results output by the multi-role collaborative decision-making module to generate a standardized daily operation report for the photovoltaic, energy storage, and charging site. A results output module outputs the daily operation report to the operation management terminal in the form of a visual interface, structured data, or message push. The automated operation analysis report generation method for photovoltaic, energy storage, and charging sites based on multi-role collaborative decision-making in this invention is mainly completed by the reporting module. This includes cleaning, time alignment, indicator normalization, and multi-source data fusion processing of operational data to form a unified station operation dataset, including: Remove outliers, missing values, and duplicate values ​​from the running data to obtain cleaned running data; The cleaned operational data was adjusted to aligned operational data with a uniform time granularity according to the collection frequency and timestamp; According to the preset normalization method, the aligned running data with different dimensions and different value ranges are transformed into a unified numerical range to obtain normalized data; Normalized data is merged according to time and equipment association to form a unified site operation dataset.

[0021] The cleaned operational data is adjusted to a uniform time granularity according to the collection frequency and timestamp. For example, data with different collection frequencies are unified to data with a uniform collection frequency through interpolation. The time format of all data is standardized, and data with time offsets is corrected to ensure that the timestamps of multi-source data in the same time period are completely consistent. Data with time offsets includes: data with timestamp deviations due to sensor upload delays; outliers include: extreme data caused by equipment failures; and missing values ​​include: data not collected due to sensor failures. Preset normalization methods include linear normalization, suitable for indicators with relatively uniform data distribution (such as photovoltaic power generation and charging volume), with the formula: X′=(X-Xmin) / (Xmax-Xmin), where X is the original data, and Xmin is the historical minimum value of the indicator (e.g., the last 30...). The minimum value of photovoltaic power generation), Xmax is the historical maximum value, and X′ is the normalized data, ranging from [0,1]. Standardization and normalization are applicable to indicators with extreme values, including revenue, failure frequency, etc. The formula is: D=(X-μ) / σ, where μ is the indicator mean, σ is the standard deviation, and D is the normalized data, which follows a normal distribution with a mean of 0 and a variance of 1, preserving the relative fluctuation characteristics of the data. The normalized data is fused according to time association and equipment association. Specifically, the data of each dimension is first grouped by timestamp to ensure that all dimension data in the same time period are aggregated into the same group, and then the equipment-level data is associated by equipment ID. Associating equipment-level data by equipment ID includes binding the charging data of a certain charging pile with the status data of that charging pile. By cleaning, aligning, normalizing, and fusing multi-source data of the operation data, a unified station operation dataset is formed, which solves the problems of inconsistent formats, asynchronous time, noisy data, and inconsistent dimensions of multi-source data.

[0022] Based on a unified site operation dataset, and according to preset role division rules, analysis tasks for multiple functional roles are generated. Logical consistency checks and conflict detection are performed on the analysis results. Based on preset business priority rules, conflicting conclusions between roles are eliminated, and a comprehensive analysis result is generated, including: Receive a unified site operation dataset and index it according to data type to obtain standard multi-source data; According to the preset role division rules, standard multi-source data is intelligently allocated to five functional roles: trend analysis, profit analysis, anomaly diagnosis, strategy evaluation, and summary generation. Five functional roles of intelligent agents execute corresponding analysis tasks in parallel to generate role-level analysis conclusions; After conflict detection and resolution of the role-level analysis conclusions, a comprehensive analysis result is obtained through hierarchical integration and causal linkage. The comprehensive analysis results are then integrated and output.

[0023] The operational trend analysis agent calls upon energy flow time-series data, employs time-series analysis algorithms, and calculates in parallel the peak daily periods of photovoltaic power generation, the fluctuation range of energy storage charging and discharging loads, and the distribution characteristics of charging loads during different time periods. It outputs trend characteristic values, visualization curves, and role-level conclusions for year-on-year / month-on-month difference analysis. The revenue analysis agent calls upon economic and energy-related data, and according to preset accounting logic, performs parallel analysis of charging revenue composition, the contribution ratio of time-of-use pricing to revenue, the fluctuation range of revenue compared to the same period in history, and driving factors. It outputs revenue quantification results, key influencing factors, and role-level conclusions for fluctuation attribution. The anomaly diagnosis agent calls upon equipment status data and historical benchmark data, and analyzes indicators to determine the anomaly. The system performs several analysis steps: First, it calculates deviations and detects whether operational metrics exceed preset thresholds. It identifies abnormal device IDs, time periods, and types of anomalies, outputting an anomaly list, deviation degree, and preliminary cause predictions at the role-level. Second, it evaluates the rationality of energy storage charging / discharging threshold settings and time-of-use operation period divisions by comparing and analyzing the current strategy's revenue / efficiency with simulation results of optimized strategies. It outputs a strategy suitability score and optimization direction suggestions at the role-level. Third, it summarizes and generates data, temporarily storing the computation progress of each parallel agent. Once all role-level analysis conclusions are generated, it initiates preliminary calculations to avoid resource conflicts with other agents. The time-series analysis algorithms include sliding window methods and trend fitting models. Pre-defined calculation logic includes: charging revenue = charging volume × electricity price for the corresponding time period - energy storage charging / discharging loss cost. Operational metrics include: charging efficiency, device temperature, and energy storage charging / discharging power. The comprehensive analysis results, obtained through hierarchical integration and causal linkage, include: The core indicators and key conclusions in the role-level analysis conclusions are extracted and integrated according to four levels: operational status, economic benefits, equipment safety, and strategy optimization, to obtain a hierarchical analysis framework. Establish a logical chain based on the inherent connections between the analytical conclusions of each role level; The hierarchical analysis framework and logical chains are integrated using a natural language logical induction algorithm to obtain comprehensive analysis results. The natural language logical induction algorithm is a mature existing technology and will not be discussed further here. The logical chain is generated based on four dimensions: data, phenomenon, cause, and impact.

[0024] Hierarchical analysis frameworks, for example, integrate the trend characteristics of photovoltaics / energy storage / charging at the operational status layer, and integrate the results of revenue, cost, and profit analysis at the economic benefit layer; The steps for generating the operation analysis report include: Receive interactive voice messages through the intelligent assistant interaction module; Semantic intent recognition for interactive voice; When the identified semantic intent is a report generation intent, generate an instruction to generate an operational analysis report corresponding to the identified intent; And / or, Monitor the operational intelligence agent. When the type of decision generated by the operational intelligence agent matches the type of decision in the pre-configured report generation trigger library, trigger the generation instruction of the corresponding operational analysis report. And / or, Based on a pre-configured report generation trigger time library, the corresponding operational analysis report generation instruction is triggered.

[0025] There are three ways to generate instructions for generating operational analysis reports: The first method involves interaction between staff and an intelligent interaction module. The intelligent interaction module receives the staff's voice communication and performs semantic intent recognition on the voice. Since staff can use the intelligent interaction module not only to generate reports but also to perform other related interactive tasks, such as viewing, querying, and controlling, it is necessary to judge the results of the semantic intent recognition. Only when the recognized semantic intent is a report generation intent will a corresponding operational analysis report generation instruction be generated. Specifically, how to determine whether the recognized semantic intent is a report generation intent can be achieved by pre-configuring a classification table. That is, the classification table includes various report generation intents. When the recognized semantic intent is a semantic intent included in the classification table, it can be determined that the staff wants to generate the corresponding report. The semantic intent included in the classification table corresponds to one or more corresponding operational analysis report generation instructions, and operational analysis report generation instructions can be generated based on the classification table. The second method involves monitoring and analyzing the intelligent agent to generate an operational analysis report. Specifically, this involves using a pre-configured report generation trigger library, which specifies the decision type that triggers report generation and the corresponding report type. By monitoring the intelligent agent's decisions, the corresponding operational analysis report generation instructions are triggered. The third type is timed triggering, which means that the report is generated at the trigger time specified in the pre-configured report generation trigger time library.

[0026] In one embodiment, the method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making further includes: When the instruction to generate an operational analysis report is received that it is generated based on interactive voice, time verification is performed based on a pre-configured time verification analysis library. When the time verification result matches the time interval for predictive analysis configured in the time verification analysis library, predictive data is obtained by predicting operational data. Based on the generated forecast data, execute the instructions to generate an operational analysis report.

[0027] For operational analysis reports generated based on interactive voice analysis, since the report generation method is initiated by staff, there may be instances where the report generation time is inappropriate. For example, company management regulations typically require daily report generation between 11 PM and 1 AM, so it's unreasonable for staff to request daily report generation at 2 PM. Therefore, using a time verification library for time validation can avoid this situation. However, while requesting daily report generation at 2 PM is unreasonable, there might still be a genuine need. This embodiment uses a configured time verification analysis library to perform verification analysis, extract this need, and conduct predictive analysis. The predicted data is then used to supplement the data required for report generation, ensuring that staff can generate reports according to their needs. The requirement is to use different display formats for reports generated entirely from forecast data, reports generated entirely from historical data within the current time period, and reports generated by combining forecast data and historical data within the current time period. The display format includes font color, font size, and font style. Furthermore, reports generated by combining forecast data and historical data within the current time period can also be generated using a combined approach. For example, the total power generation of a photovoltaic-storage-charging station includes both the total generated power and the forecasted total power. Assuming the total power generation is 3000 kWh, of which the total generated power is 1000 kWh and the forecasted total power is 2000 kWh, the generated report would be: "Total power generation of the photovoltaic-storage-charging station is 3000 (1000 + 2000) kWh."

[0028] The prediction steps for the predicted data include: Determine the ratio of the first time period corresponding to the data to be predicted to the time period corresponding to the current report as a reference parameter; When the reference parameter is less than or equal to a preset first threshold, historical operational data for a second time length prior to the current moment is obtained as analysis data; wherein, the second time length is a preset multiple of the first time length; the first threshold is any value between 0.001 and 0.05; the preset multiple is any value between 3 and 10; the sum of the second time length and the first time length is less than or equal to the time period; The analysis data is sampled based on the first time length to obtain multiple sample data; Based on multiple sampled data, the predicted data is determined.

[0029] Both the sampled data and the data to be predicted are sets of values ​​for each data item at each data point within a first time period; among them, the predicted data is determined based on multiple sampled data sets, including: The characteristic value corresponding to the sampled data is determined by the value of each data point in the sampled data. For example, for the total power generation, the sum of the values ​​of each data point is used as the characteristic value of the sampled data; for current, voltage, etc., the average value of each data point is used as the characteristic value of the sampled data. By determining the characteristic value, the sampled data is fuzzified and regarded as a whole in order to achieve the prediction of the forecast data. The predicted value of the predicted data is calculated using the characterization value corresponding to each data item. The calculation formula is as follows: ; In the formula, This indicates that the predicted data corresponds to the first... Predicted values ​​for each data item; Indicates corresponding to the first The first data item The representational value of each sampled data; This indicates the configuration assigned to the corresponding number. The first data item The weight values ​​of each sampled data; The total number of sampled data; To query the bias values ​​obtained from a pre-configured bias value table based on the time interval and data items corresponding to the predicted data; where, , Configuration is carried out by professionals based on historical data analysis of the facilities. When the reference parameter is greater than the preset first threshold and less than or equal to the preset second threshold, historical operational data of the current time period and the previous preset number of time periods are obtained as analysis data. Based on a pre-configured sampling time interval, the analysis data is sampled to obtain sampled data; wherein, the sampling time interval is an integer multiple of the time difference between adjacent data points; Based on the sampling time interval and the first time length, the number of prediction units contained in the prediction data is determined; Each prediction unit is extracted sequentially, and based on the pre-configured data extraction rules, the sampled data and / or the predicted prediction units are extracted as reference units. Based on the reference unit, predictions are made for the prediction unit; The first threshold is less than the second threshold.

[0030] Among them, predicting the prediction unit based on the reference unit includes: The reference units and prediction units are grouped and sorted based on the time period and the time sequence. The reference unit corresponding to the prediction unit in the group that does not contain the prediction unit is used as the verification unit, and the prediction corresponding to the verification unit is obtained based on the other reference units in the group to obtain the analysis value. Based on the verification unit and the analysis value, the deviation is determined; The predicted value is calculated using the reference cell in the prediction unit, and then corrected based on the deviation. The specific calculation formula is as follows: ; In the formula, This indicates that the prediction unit corresponds to the first... Predicted values ​​for each data item; Indicates corresponding to the first The first data item Characteristic values ​​of each reference unit; This indicates the configuration assigned to the corresponding number. The first data item The weight values ​​of each reference unit; The total number of sampled data; To query the bias values ​​obtained from the pre-configured bias value table based on the time interval and data item corresponding to the prediction unit; For the first The characterization value of the verification unit in a group that does not contain a prediction unit; No. The predicted value of the check unit in a group that does not contain a prediction unit; For the pre-configured corresponding to the first The weight values ​​of groups that do not contain prediction units; This represents the total number of groups containing the prediction unit. When the reference parameter is greater than the preset second threshold, the historical operation data of the current time period and the previous preset number of time periods, as well as the historical operation data of other related photovoltaic storage and charging stations and the previous preset number of time periods, are obtained as analysis data. Based on a pre-configured sampling time interval, the analysis data is sampled to obtain sampled data; Based on the sampling time interval and the first time length, the number of prediction units contained in the prediction data is determined; Each prediction unit is extracted sequentially, and based on the pre-configured data extraction rules, the sampled data and / or the predicted prediction units are extracted as reference units. Based on the reference unit, predictions are made for the prediction unit; The prediction of the prediction unit is based on the reference unit, and the calculation formula is as follows: ; In the formula, This indicates that the prediction unit corresponds to the first... Predicted values ​​for each data item; Indicates corresponding to the first The first data item Characteristic values ​​of each reference unit; This indicates the configuration assigned to the corresponding number. The first data item The weight values ​​of each reference unit; The total number of sampled data; To query the bias values ​​obtained from the pre-configured bias value table based on the time interval and data item corresponding to the prediction unit; For the first The characterization value of the verification unit in a group that does not contain a prediction unit; No. The predicted value of the check unit in a group that does not contain a prediction unit; For the pre-configured corresponding to the first The weight values ​​of groups that do not contain prediction units; This represents the total number of groups containing the prediction unit. , These are the first and second weights configured, respectively; The total number of other photovoltaic storage and charging stations; For the first The first of another photovoltaic storage and charging station The characterization value of the verification unit in a group that does not contain a prediction unit; For the first The first of another photovoltaic storage and charging station The predicted value of the check unit in a group that does not contain a prediction unit; For the pre-configured corresponding to the first The first of another photovoltaic storage and charging station The weight values ​​of groups that do not contain prediction units.

[0031] This embodiment performs predictive analysis based on user needs. By using predictive data, it ensures the generation of reports on customer needs, enabling users to understand the station's operational status in advance for a period of time.

[0032] To enable the generation of automated operation analysis reports for photovoltaic, energy storage, and charging power stations based on multi-role collaborative decision-making, the following methods are also included: When the instruction to generate an optimization suggestion report is executed, an optimization suggestion collection instruction is sent to each operational intelligent agent; each operational intelligent agent generates optimization suggestion content based on its configured role and in combination with the current real-time status of the optical storage and charging station and its historical knowledge base. Based on the optimization suggestions from various operational feedback, an optimization suggestion report is generated. Among them, the optimization suggestions from each operation feedback are generated by each operation intelligence agent based on its configured role, combined with the current real-time status of the photovoltaic storage and charging station and its historical knowledge base.

[0033] To maintain the validity of the generated reports, the method for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making also includes: After the operational analysis report generated by combining the forecast data is produced, a monitoring dataset is constructed based on the forecasts of the forecast units. The operation of photovoltaic, energy storage and charging stations is monitored based on the monitoring dataset; When the monitoring results meet the preset correction conditions, the operation analysis report is corrected based on the monitoring results and the pre-configured report correction library; otherwise, the operation analysis report is regenerated.

[0034] The present invention also provides a system for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making, comprising: a report generation module; Among them, such as Figure 2 As shown, the report generation module includes: parameter determination unit 1, requirement data acquisition unit 2, data allocation unit 3, feedback receiving unit 4, and content generation unit 5; When the instruction to generate the operation analysis report is executed, the parameter determination unit 1 determines the acquisition parameters of the required data corresponding to each report item in the operation analysis report to be generated; the required data acquisition unit 2 determines the required data from the operation data of the photovoltaic storage and charging station based on the acquisition parameters of the required data; the data allocation unit 3 allocates the required data to the corresponding operation agents based on the roles configured for each operation agent; the feedback receiving unit 4 receives the analysis results fed back by each operation agent; and the content generation unit 5 generates the report content corresponding to the report item based on the analysis results.

[0035] The roles configured for each operational intelligent agent include one or more of the following: operational trend analysis role, revenue analysis role, anomaly diagnosis role, strategy evaluation role, and summary generation role. Among them, the operation trend analysis role is used to analyze the intraday variation trend of photovoltaic power generation, energy storage charging and discharging, and charging load; The revenue analysis role is used to analyze charging revenue at power stations, electricity price structure, and revenue changes. Anomaly diagnosis role, used to identify abnormal situations where operating indicators deviate from historical benchmarks or preset thresholds; The strategy evaluation role is used to assess the rationality of the current energy storage charging and discharging strategy and time-of-use operation strategy; The generated roles are used to synthesize and summarize the analysis results of each role, forming a unified operational conclusion.

[0036] The steps for generating the operation analysis report include: Receive interactive voice messages through the intelligent assistant interaction module; Semantic intent recognition for interactive voice; When the identified semantic intent is a report generation intent, generate an instruction to generate an operational analysis report corresponding to the identified intent; And / or, Monitor the operational intelligence agent. When the type of decision generated by the operational intelligence agent matches the type of decision in the pre-configured report generation trigger library, trigger the generation instruction of the corresponding operational analysis report. And / or, Based on a pre-configured report generation trigger time library, the corresponding operational analysis report generation instruction is triggered.

[0037] To ensure the effective execution of report generation instructions based on interactive voice, the automated operation analysis report generation system for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making also includes: a verification module; The verification module performs the following operations: When the instruction to generate an operational analysis report is received that it is generated based on interactive voice, time verification is performed based on a pre-configured time verification analysis library. When the time verification result matches the time interval for predictive analysis configured in the time verification analysis library, predictive data is obtained by predicting operational data. Based on the generated forecast data, execute the instructions to generate an operational analysis report.

[0038] The prediction steps for the predicted data include: Determine the ratio of the first time period corresponding to the data to be predicted to the time period corresponding to the current report as a reference parameter; When the reference parameter is less than or equal to a preset first threshold, historical operational data for a second time period prior to the current moment is obtained as analysis data; wherein, the second time period is a preset multiple of the first time period; The analysis data is sampled based on the first time length to obtain multiple sample data; Based on multiple sampled data, the predicted data is determined.

[0039] When the reference parameter is greater than the preset first threshold and less than or equal to the preset second threshold, historical operational data of the current time period and the previous preset number of time periods are obtained as analysis data. Based on a pre-configured sampling time interval, the analysis data is sampled to obtain sampled data; Based on the sampling time interval and the first time length, the number of prediction units contained in the prediction data is determined; Each prediction unit is extracted sequentially, and based on the pre-configured data extraction rules, the sampled data and / or the predicted prediction units are extracted as reference units. Based on the reference unit, predictions are made for the prediction unit; The first threshold is less than the second threshold.

[0040] When the reference parameter is greater than the preset second threshold, the historical operation data of the current time period and the previous preset number of time periods, as well as the historical operation data of other related photovoltaic storage and charging stations for the time periods and the previous preset number of time periods, are obtained as analysis data; the other related photovoltaic storage and charging stations are other photovoltaic storage and charging stations in the association table that the staff has built in advance for the station. Based on a pre-configured sampling time interval, the analysis data is sampled to obtain sampled data; Based on the sampling time interval and the first time length, the number of prediction units contained in the prediction data is determined; Each prediction unit is extracted sequentially, and based on the pre-configured data extraction rules, the sampled data and / or the predicted prediction units are extracted as reference units. Based on the reference unit, predictions are made for the prediction unit; To ensure the provision of optimization suggestions, the automated operation analysis report generation system for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making also includes: an intelligent assistant interaction module; The intelligent assistant interaction module performs the following operations: When the instruction to generate an optimization suggestion report is executed, an optimization suggestion collection instruction is sent to each operational intelligent agent; each operational intelligent agent generates optimization suggestion content based on its configured role and in combination with the current real-time status of the optical storage and charging station and its historical knowledge base. Based on the optimization suggestions from various operational feedback, an optimization suggestion report is generated. Among them, the optimization suggestions from each operation feedback are generated by each operation intelligence agent based on its configured role, combined with the current real-time status of the photovoltaic storage and charging station and its historical knowledge base.

[0041] The intelligent assistant interaction module enables natural language interaction between operators and a multi-role intelligent agent in the backend. Based on Large Language Model (LLM) or semantic analysis algorithms, this module has the following functions: Semantic intent recognition: It can parse the natural language questions from operators and convert them into structured query instructions; examples include: "Why did energy storage revenue decline yesterday?" or "What were the specific causes of yesterday's fault?"; On-demand role invocation: Based on the recognized intent, it dynamically schedules specific roles in the "multi-role collaborative decision-making module" to perform in-depth attribution analysis, retrieving the raw data and logical chain behind the daily reports; Dynamic strategy suggestions: Combining the current real-time status of the site with historical knowledge base, it generates targeted optimization suggestions for the user's questions, realizing a shift from "passive daily report push" to "proactive decision-making assistance." Optimization suggestions include: "Suggest adjusting peak and off-peak charging and discharging thresholds" or "Suggest checking the inverter's heat dissipation module."

[0042] To ensure the continued effectiveness of reports generated based on predictive data, the automated operation analysis report generation system for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making also includes: a tracking and monitoring module; The tracking and monitoring module performs the following operations: After the operational analysis report generated by combining the forecast data is produced, a monitoring dataset is constructed based on the forecasts of the forecast units. The operation of photovoltaic, energy storage and charging stations is monitored based on the monitoring dataset; When the monitoring results meet the preset correction conditions, the operation analysis report is corrected based on the monitoring results and the pre-configured report correction library; otherwise, the operation analysis report is regenerated.

[0043] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making, characterized in that, include: When executing the instruction to generate an operational analysis report, determine the acquisition parameters of the required data corresponding to each report item in the operational analysis report to be generated; Demand data is determined from the operational data of the photovoltaic storage and charging station based on the parameters obtained from the demand data. Based on the roles configured for each operational agent, the demand data is allocated to the corresponding operational agent; Receive analysis results from various operational intelligent agents; Based on the analysis results, generate the corresponding report content for each report item.

2. The method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making as described in claim 1, characterized in that, The roles configured for each operational intelligent agent include: one or more of the following: operational trend analysis role, revenue analysis role, anomaly diagnosis role, strategy evaluation role, and summary generation role. Among them, the operation trend analysis role is used to analyze the intraday variation trend of photovoltaic power generation, energy storage charging and discharging, and charging load; The revenue analysis role is used to analyze charging revenue at power stations, electricity price structure, and revenue changes. Anomaly diagnosis role, used to identify abnormal situations where operating indicators deviate from historical benchmarks or preset thresholds; The strategy evaluation role is used to assess the rationality of the current energy storage charging and discharging strategy and time-of-use operation strategy; The generated roles are used to synthesize and summarize the analysis results of each role, forming a unified operational conclusion.

3. The method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making as described in claim 1, characterized in that, The steps for generating an operational analysis report include: Receive interactive voice messages through the intelligent assistant interaction module; Semantic intent recognition for interactive voice; When the identified semantic intent is a report generation intent, generate an instruction to generate an operational analysis report corresponding to the identified intent; And / or, Monitor the operational intelligence agent. When the type of decision generated by the operational intelligence agent matches the type of decision in the pre-configured report generation trigger library, trigger the generation instruction of the corresponding operational analysis report. And / or, Based on a pre-configured report generation trigger time library, the corresponding operational analysis report generation instruction is triggered.

4. The method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making as described in claim 3, characterized in that, Also includes: When the instruction to generate an operational analysis report is received that it is generated based on interactive voice, time verification is performed based on a pre-configured time verification analysis library. When the time verification result matches the time interval for predictive analysis configured in the time verification analysis library, predictive data is obtained by predicting operational data. Based on the generated forecast data, execute the instructions to generate an operational analysis report.

5. The method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making as described in claim 4, characterized in that, The prediction steps for the predicted data include: Determine the ratio of the first time period corresponding to the data to be predicted to the time period corresponding to the current report as a reference parameter; When the reference parameter is less than or equal to a preset first threshold, historical operational data for a second time period prior to the current moment is obtained as analysis data; wherein, the second time period is a preset multiple of the first time period; The analysis data is sampled based on the first time length to obtain multiple sample data; Based on multiple sampled data, the predicted data is determined.

6. The method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making as described in claim 5, characterized in that, The forecasting steps for the forecast data also include: When the reference parameter is greater than the preset first threshold and less than or equal to the preset second threshold, historical operational data of the current time period and the previous preset number of time periods are obtained as analysis data. Based on a pre-configured sampling time interval, the analysis data is sampled to obtain sampled data; Based on the sampling time interval and the first time length, the number of prediction units contained in the prediction data is determined; Each prediction unit is extracted sequentially, and based on the pre-configured data extraction rules, the sampled data and / or the predicted prediction units are extracted as reference units. Based on the reference unit, predictions are made for the prediction unit; The first threshold is less than the second threshold.

7. The method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making as described in claim 6, characterized in that, The forecasting steps for the forecast data also include: When the reference parameter is greater than the preset second threshold, the historical operation data of the current time period and the previous preset number of time periods, as well as the historical operation data of other related photovoltaic storage and charging stations and the previous preset number of time periods, are obtained as analysis data. Based on a pre-configured sampling time interval, the analysis data is sampled to obtain sampled data; Based on the sampling time interval and the first time length, the number of prediction units contained in the prediction data is determined; Each prediction unit is extracted sequentially, and based on the pre-configured data extraction rules, the sampled data and / or the predicted prediction units are extracted as reference units. Based on the reference unit, predictions are made for the prediction unit.

8. The method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making as described in claim 6, characterized in that, Also includes: When executing the instruction to generate an optimization suggestion report, an optimization suggestion collection instruction is sent to each operational agent; Based on the optimization suggestions from various operational feedback, an optimization suggestion report is generated. Among them, the optimization suggestions from each operation feedback are generated by each operation intelligence agent based on its configured role, combined with the current real-time status of the photovoltaic storage and charging station and its historical knowledge base.

9. The method for generating an automated operation analysis report for a photovoltaic, energy storage, and charging station based on multi-role collaborative decision-making as described in claim 6, characterized in that, Also includes: After the operational analysis report generated by combining the forecast data is produced, a monitoring dataset is constructed based on the forecasts of the forecast units. The operation of photovoltaic, energy storage and charging stations is monitored based on the monitoring dataset; When the monitoring results meet the preset correction conditions, the operation analysis report is corrected based on the monitoring results and the pre-configured report correction library; Otherwise, the operational analysis report will be generated again.

10. A system for generating automated operation analysis reports for photovoltaic, energy storage, and charging stations based on multi-role collaborative decision-making, characterized in that: include: Report generation module; The report generation module includes: a parameter determination unit, a requirement data acquisition unit, a data allocation unit, a feedback receiving unit, and a content generation unit; When the instruction to generate an operation analysis report is executed, the parameter determination unit determines the acquisition parameters of the required data corresponding to each report item in the operation analysis report to be generated; the required data acquisition unit determines the required data from the operation data of the photovoltaic storage and charging station based on the acquisition parameters of the required data; the data allocation unit allocates the required data to the corresponding operation agents based on the roles configured for each operation agent; the feedback receiving unit receives the analysis results fed back by each operation agent; and the content generation unit generates the report content corresponding to the report item based on the analysis results.

Images