Intelligent regulation method and device based on solar power generation prediction

Abstract

The present application provides an intelligent regulation method and device based on solar power generation prediction, which relates to the field of photovoltaic power generation technology. By connecting the equipment database of the power generation equipment, the equipment time series data set is obtained to establish an equipment status prediction model; a supply strategy set for photovoltaic power generation and energy storage is established based on load demand and weather forecast data; the equipment fitting working data set of the power generation equipment is configured with the supply strategy set, and the equipment status is predicted with the equipment working data, the equipment fitting working data set and the equipment status prediction model to establish the prediction result; multi-objective optimization analysis is performed based on the prediction result and the supply strategy set, and an adjustment plan is generated for intelligent adjustment management. The present application solves the technical problem that the prior art lacks consideration of weather data, load demand and equipment status in the process of adjusting power generation equipment, resulting in insufficient flexibility and adaptability of the supply strategy. It can quickly respond to environmental changes and load demands, and improve the overall performance of the power generation system.

Description

Technical Field

[0001] The present application relates to the field of photovoltaic power generation technology, and specifically to an intelligent regulation method and device based on solar power generation prediction. Background Art

[0002] Photovoltaic power generation, as an important part of renewable energy, is increasingly becoming a key driving force for the transformation of the global energy system. Traditionally, photovoltaic power generation systems rely on static power generation strategies, that is, the inclination and orientation of photovoltaic panels are fixed after installation, which to a certain extent limits the power generation efficiency, especially under variable climate conditions. In recent years, with the development of intelligent technology, some dynamic adjustment methods have been introduced into the field of photovoltaic power generation, such as maximum power point tracking (MPPT) technology, which can adjust the working state of photovoltaic modules in real time to obtain maximum power output.

[0003] However, existing technologies still face many challenges. On the one hand, traditional maximum power point tracking (MPPT) technology often only considers the immediate power generation efficiency, while ignoring the impact of long-term factors such as weather changes and equipment aging on power generation efficiency. On the other hand, the intermittent and instability of photovoltaic power generation poses a challenge to the stable operation of the power grid. The existing regulation methods lack accurate matching of load demand, resulting in poor power generation efficiency and economy during peak or trough load periods. In addition, the prediction and maintenance of equipment status also rely heavily on manual labor and lack of intelligent prediction models, which not only increases operation and maintenance costs, but also affects the overall performance of the system. Summary of the invention

[0004] The present application provides an intelligent regulation method and device based on solar power generation prediction, which solves the technical problems that the prior art lacks consideration of weather data, load demand and equipment status in the process of regulating power generation equipment, resulting in the inability of the regulation strategy to fully meet the working status and demand of the equipment, and the lack of flexibility and adaptability of the supply strategy. Through multi-objective balance, the technical effect of enhancing the ability of the photovoltaic power generation system to quickly respond to environmental changes and load demands and improving the overall performance of the power generation system is achieved.

[0005] In view of the above problems, on the one hand, the present application provides an intelligent regulation method based on solar power generation prediction, the method comprising: connecting to the equipment database of the power generation equipment, executing the time series data reading in the equipment database, establishing the equipment time series data set of the power generation equipment, establishing the equipment state prediction model with the equipment time series data set, the equipment state prediction model including the normal equipment prediction network and the probabilistic equipment failure prediction network; collecting and obtaining weather data, the weather data including weather forecast data, obtaining load demand, and establishing a supply strategy set for photovoltaic power generation and energy storage based on the load demand and the weather forecast data; configuring the equipment fitting working data set of the power generation equipment with the supply strategy set, collecting the equipment working data of the power generation equipment, predicting the equipment state with the equipment working data, the equipment fitting working data set and the equipment state prediction model, and establishing the prediction result; performing multi-objective optimization analysis according to the prediction result and the supply strategy set, as follows:

[0006] ;

[0007] in, is the objective function, is the weight factor of the importance of load demand satisfaction, Characterizes the load difference, the difference between the actual power generation and the load demand, , For in time The actual power generation For energy storage in time The output power, is the load demand power, is the weight factor representing the total cost, is the total cost, including power generation cost, energy storage cost and equipment abnormality cost, is the weight factor for device maintenance, Representation in time Equipment maintenance costs, is the weight factor of power generation efficiency, Representation in time power generation efficiency; generate regulation plans based on multi-objective optimization analysis results, and perform intelligent regulation management based on the regulation plans.

[0008] On the other hand, the present application also provides an intelligent regulation device based on solar power generation prediction, the device comprising: a prediction model construction module, the prediction model construction module is used to connect to the equipment database of the power generation equipment, execute the time series data reading in the equipment database, establish the equipment time series data set of the power generation equipment, and establish the equipment state prediction model with the equipment time series data set, the equipment state prediction model includes a normal equipment prediction network and a probabilistic equipment failure prediction network; a supply strategy set establishment module, the supply strategy set establishment module is used to collect and obtain weather data, the weather data includes weather forecast data, obtain load demand, and establish a supply strategy set for photovoltaic power generation and energy storage based on the load demand and the weather forecast data; an equipment state prediction module, the equipment state prediction module is used to configure the equipment fitting working data set of the power generation equipment with the supply strategy set, collect the equipment working data of the power generation equipment, and use the equipment working data, the equipment fitting working data set and the equipment state prediction model to predict the equipment state, and establish a prediction result; a multi-objective optimization analysis module, the multi-objective optimization analysis module is used to perform multi-objective optimization analysis according to the prediction result and the supply strategy set, as follows:

[0009] ;

[0010] in, is the objective function, is the weight factor of the importance of load demand satisfaction, Characterizes the load difference, the difference between the actual power generation and the load demand, , For in time The actual power generation For energy storage in time The output power, is the load demand power, is the weight factor representing the total cost, is the total cost, including power generation cost, energy storage cost and equipment abnormality cost, is the weight factor for device maintenance, Representation in time Equipment maintenance costs, is the weight factor of power generation efficiency, Representation in time power generation efficiency; an intelligent adjustment module, wherein the intelligent adjustment module is used to generate an adjustment plan based on the multi-objective optimization analysis results and perform intelligent adjustment management based on the adjustment plan.

[0011] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0012] Connect the equipment database of the power generation equipment, read the time series data in the equipment database, establish the equipment time series data set of the power generation equipment, establish the equipment state prediction model with the equipment time series data set, and the equipment state prediction model includes a normal equipment prediction network and a probabilistic equipment failure prediction network; collect and obtain weather data, the weather data includes weather forecast data, obtain load demand, and establish a supply strategy set for photovoltaic power generation and energy storage based on the load demand and the weather forecast data. This process provides basic data support for subsequent intelligent regulation and ensures the accuracy and real-time nature of the prediction. Configure the equipment fitting working data set of the power generation equipment with the supply strategy set, collect the equipment working data of the power generation equipment, predict the equipment state with the equipment working data, the equipment fitting working data set and the equipment state prediction model, and establish the prediction result; perform multi-objective optimization analysis based on the prediction result and the supply strategy set. This process uses a multi-objective optimization algorithm to comprehensively consider multiple objectives such as power generation efficiency, cost, and equipment maintenance, find the optimal operation strategy, solve the problem of multi-objective balancing in traditional regulation methods, and achieve the optimization of the overall performance of the power generation system. The regulation plan is generated based on the results of multi-objective optimization analysis, and intelligent regulation management is carried out based on the regulation plan, which can quickly respond to environmental changes and load demand fluctuations, ensuring the stable operation of the power grid.

[0013] In summary, this application improves the adaptability of photovoltaic power generation systems to environmental changes and load demands through data fusion and prediction of deep learning algorithms; multi-objective optimization analysis ensures the optimal balance between system efficiency, cost and equipment maintenance; and the application of real-time intelligent regulation and equipment status monitoring technology enables the power generation system to automatically adjust to maximize power generation efficiency, minimize costs, and extend equipment life, thereby significantly improving the overall performance of the power generation system as a whole, making the photovoltaic power generation system more adaptable to complex and changeable operating environments, and promoting the further development and application of renewable energy technologies.

[0014] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A schematic diagram of a process flow of an intelligent regulation method based on solar power generation prediction provided in an embodiment of the present application;

[0016] Figure 2 A schematic diagram of a process for establishing a device state prediction model in an intelligent adjustment method based on solar power generation prediction provided in an embodiment of the present application;

[0017] Figure 3 A schematic diagram of a process for establishing a device state prediction result in an intelligent adjustment method based on solar power generation prediction provided in an embodiment of the present application;

[0018] Figure 4 A schematic diagram of the structure of an intelligent regulation device based on solar power generation prediction provided in an embodiment of the present application.

[0019] Explanation of the reference numerals: prediction model building module 10 , supply strategy set building module 20 , equipment status prediction module 30 , multi-objective optimization analysis module 40 , intelligent adjustment module 50 . DETAILED DESCRIPTION

[0020] The embodiments of the present application provide an intelligent adjustment method and device based on solar power generation prediction, which solves the technical problems that the prior art lacks consideration of weather data, load demand and equipment status in the process of adjusting power generation equipment, resulting in the inability of the adjustment strategy to fully meet the working status and demand of the equipment, and the lack of flexibility and adaptability of the supply strategy. Through multi-objective balance, the technical effect of enhancing the ability of the photovoltaic power generation system to quickly respond to environmental changes and load demands and improving the overall performance of the power generation system is achieved.

[0021] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides an intelligent adjustment method based on solar power generation prediction, the method comprising:

[0022] Step S1: connect to the equipment database of the power generation equipment, execute the time series data reading in the equipment database, establish the equipment time series data set of the power generation equipment, and establish the equipment state prediction model with the equipment time series data set. The equipment state prediction model includes a normal equipment prediction network and a probabilistic equipment failure prediction network.

[0023] Specifically, the equipment database is an information system that stores data related to power generation equipment, such as the equipment's operating parameters, historical operating data, etc. Time series data is equipment operating status data recorded in chronological order, which is used to analyze the data trend of equipment operating status changing over time. The collection of extracted time series data is the equipment time series data set, which is used to analyze and predict equipment status. The equipment status prediction model is a model trained based on the equipment time series data set by combining deep learning techniques, such as long short-term memory networks (LSTMs), gated recurrent units (GRUs), etc., and is used to predict the future status of the equipment. Among them, the normal equipment prediction network is used to predict the status of the equipment under normal operating conditions; while the probabilistic equipment failure prediction network focuses on evaluating the probability of equipment failure.

[0024] The real-time monitoring data of power generation equipment is connected to the central database through a data interface or network protocol to achieve real-time synchronization and storage of data. For example, the operating parameters (such as current, voltage, and temperature) of equipment such as inverters and photovoltaic panels are transmitted to the database in real time using the OPC-UA standard protocol or Modbus protocol. The equipment operating status data recorded in chronological order, that is, time series data, are extracted from the database. The above time series data are organized into a data set to obtain the time series data set of the power generation equipment for subsequent equipment status prediction model training. The equipment status prediction model is trained using a deep learning framework such as TensorFlow or PyTorch, which includes a normal equipment prediction network and a probabilistic equipment failure prediction network. The model is trained using the time series data set of the power generation equipment, and the accuracy of the model is evaluated using the validation data set. For example, the normal equipment prediction network can be trained using a long short-term memory network (LSTM) network to predict the power generation in the next week; at the same time, the probabilistic equipment failure prediction network can be trained using a gated recurrent unit (GRU) network to predict the probability of inverter overheating.

[0025] Through the above steps, an equipment status prediction model was constructed, which achieved accurate prediction of the status of power generation equipment and provided data support for subsequent intelligent adjustment and maintenance optimization.

[0026] Step S2: Collect and obtain weather data, the weather data including weather forecast data, obtain load demand, and establish a supply strategy set for photovoltaic power generation and energy storage based on the load demand and the weather forecast data.

[0027] Specifically, weather forecast data refers to weather forecasts within a certain time frame in the future, which is used to predict the potential capacity of photovoltaic power generation. Load demand refers to the electricity demand of power users within a specific time. The supply strategy set is a set of strategies formulated based on weather forecast data and load demand to guide the operation of photovoltaic power generation and energy storage systems.

[0028] First, it is necessary to obtain weather data from meteorological data providers or public APIs, such as OpenWeatherMap, the Meteorological Bureau, etc., including real-time and predicted meteorological information, such as light intensity, temperature, humidity, etc. Obtain public power load demand data from government departments, power market operators, etc. Next, obtain the supply strategy set based on load demand and weather forecast data. The supply strategy set can be obtained using existing forecasting models or using time series-based ARIMA models or machine learning models such as neural networks, support vector machines, and random forests, and using big data to obtain historical load demand and historical weather data for training. Exemplarily, existing models such as "one map of adjustable load resources on the power demand side" and "one network for wind, light, storage, charging, and replacement" can be used to generate supply strategy sets based on the acquired load demand and weather forecast data. These supply strategy sets are the basis for subsequent multi-objective optimization, ensuring that subsequent multi-objective optimization has sufficient and relatively accurate data support.

[0029] Step S3: configuring the equipment fitting working data set of the power generation equipment with the supply strategy set, collecting the equipment working data of the power generation equipment, performing equipment state prediction with the equipment working data, the equipment fitting working data set and the equipment state prediction model, and establishing a prediction result.

[0030] Specifically, the equipment fitting working data set is a data set that predicts the working status of power generation equipment under different strategies, generated through simulation or historical data analysis based on the supply strategy set. Equipment working data is the working status and performance data actually collected from power generation equipment.

[0031] First, based on the supply strategy set, use historical data or simulation models, such as the Python-based simulation framework SimPy or AnyLogic, to simulate the working conditions of the equipment under specific strategies, such as different power generation or energy storage modes, and generate a device fitting working data set. Then, through a real-time data acquisition system such as a supervisory control and data acquisition system (SCADA), obtain the real-time working data of the power generation equipment, which includes power generation, equipment temperature, operating efficiency, etc. The above equipment working data and equipment fitting working data set are input into the equipment state prediction model established in step S1 to predict the equipment state and obtain the prediction result of the equipment state. This prediction result reflects the possible state of the equipment when executing a specific strategy in the supply strategy set in the current state, which can be the risk probability of failure in the future or the specific operating parameter value.

[0032] Step S4: Perform multi-objective optimization analysis based on the prediction results and the supply strategy set, as follows:

[0033] ;

[0034] in, is the objective function, is the weight factor of the importance of load demand satisfaction, Characterizes the load difference, the difference between the actual power generation and the load demand, , For in time The actual power generation For energy storage in time The output power, is the load demand power, is the weight factor representing the total cost, is the total cost, including power generation cost, energy storage cost and equipment abnormality cost, is the weight factor for device maintenance, Representation in time Equipment maintenance costs, is the weight factor of power generation efficiency, Representation in time power generation efficiency.

[0035] Specifically, step S4 uses a multi-objective optimization algorithm, such as a genetic algorithm, a particle swarm optimization algorithm, etc., to optimize the supply strategies in the supply strategy set based on the prediction results and the supply strategy set obtained in the previous steps, and takes the load demand satisfaction, total cost and equipment maintenance efficiency as optimization goals. Through multi-objective optimization analysis, decision guidance is provided for the regulation of the photovoltaic power generation system, thereby ensuring the economy and reliability of energy supply. The optimization algorithm generates multiple supply strategies, each with a different score on the objective function. By evaluating and comparing these supply strategies, a relatively good solution can be found for all objectives. Set the objective function Different supply strategies are evaluated and compared from the perspectives of meeting load demand, cost control, equipment maintenance, and power generation efficiency. The specific function expression is:

[0036] ;

[0037] in, , , , is a weight factor used to adjust the relative importance of different objectives, where is the weight factor of the importance of load demand satisfaction; is the weight factor representing the total cost; Weight factors maintained for equipment; is the weight factor for power generation efficiency. To determine the specific values ​​of these weight factors, it is necessary to first analyze the priority of each goal, understand the importance of each target system performance and business goal, and then set the corresponding weight according to the importance of each goal. This process can be completed through expert consultation and historical data analysis. Experts can judge the weights based on their experience, market trends or business strategies; historical data can reflect the criticality of these goals in past decision-making. For example, if historical data shows that equipment maintenance costs have caused system performance to decline, then the weight of equipment maintenance costs may be increased. The specific value of the weight can be dynamically adjusted according to the real-time performance of the system and changes in the external environment to adapt to the adjustment needs in various situations.

[0038] Characterizes the load difference, that is, the difference between the actual power generation and the load demand, , For in time The actual power generation For energy storage in time The output power, The load difference reflects the supply and demand balance of the system within a specific period of time. The larger the load difference, the greater the difference between the power supply and load demand of the power generation system, that is, the greater the imbalance between supply and demand, and the load demand is not met and the power supply is in excess.

[0039] The total cost is the sum of all costs including power generation cost, energy storage cost and equipment abnormality cost, reflecting the economic efficiency of the power generation system operation. A lower total cost indicates that the power generation system operates efficiently and has good cost-effectiveness.

[0040] Representation at a specific time The cost of maintaining equipment. Usually estimated based on maintenance contracts, maintenance history, and actual costs of maintenance activities. Equipment maintenance costs reflect the investment required to maintain and continue to operate the system. Effective maintenance can reduce these costs and improve system reliability and life.

[0041] Representation at a specific time The proportion of solar energy actually converted into electrical energy. Power generation efficiency is a key indicator to measure the performance of photovoltaic power generation systems. A high power generation efficiency means that the system can more effectively utilize solar energy resources, thereby improving the overall economic benefits.

[0042] Through the multi-objective optimization analysis in step S4, the power generation system can comprehensively consider load demand satisfaction, cost control, equipment maintenance and power generation efficiency, and provide optimal strategy recommendations based on different priorities, thereby achieving intelligent, efficient and economical operation of energy management.

[0043] Step S5: Generate an adjustment plan based on the multi-objective optimization analysis results, and perform intelligent adjustment management based on the adjustment plan.

[0044] Specifically, the regulation scheme is a specific regulation strategy or operating instruction formulated for the power generation system or power generation equipment based on the results of multi-objective optimization analysis, which is used to guide the dynamic adjustment of the system in terms of power generation, energy storage and load management. First, the results of the multi-objective optimization analysis need to be converted into a specific executable regulation scheme. Key information is extracted from the optimization results, such as the power output of the power generation equipment at a specific time point, the charging and discharging strategy of the energy storage system, and the load management strategy. Subsequently, these regulation schemes are automatically implemented into the actual power generation system through intelligent control systems, such as distributed energy management systems (DERMS) or supervisory control and data acquisition systems (SCADA), to achieve precise regulation of the power generation system and power generation equipment, thereby greatly improving the operating performance of the power generation system.

[0045] Further, such as Figure 2 As shown, step S1 of the embodiment of the present application also includes:

[0046] Step S11: extracting abnormal data from the device time series data set and establishing an abnormal data identifier.

[0047] Step S12: locating the preceding abnormal window according to the abnormal data identifier, and extracting the window data set of the preceding abnormal window.

[0048] Step S13: constructing a time-sensitive probabilistic equipment failure prediction network based on the abnormal data identifier and the window data set.

[0049] Step S14: Establishing a device status prediction model based on the constructed probabilistic device failure prediction network.

[0050] Specifically, abnormal data identification is an identifier used to identify and mark abnormal data in a data set. The pre-anomaly window is a data window used to analyze the equipment status before the abnormality occurs within a period of time before the abnormal data point. The window dataset is a collection of data within the pre-anomaly window extracted from the dataset. The probabilistic equipment failure prediction network is a machine learning model used to predict the probability of equipment failure and takes into account the sensitivity of time series data.

[0051] First, the equipment time series data set is preprocessed, including data cleaning and standardization. Then, statistical methods or machine learning algorithms are used to identify abnormal data points. Exemplarily, the data in the equipment time series data set is visualized through line charts, box plots and other charts to observe the data distribution and abnormal patterns, and calculate the basic statistics of the data, such as mean, median, and standard deviation, to understand the central trend and dispersion of the data. Based on the 3σ criterion, when a data point exceeds the range of three standard deviations, it is considered an abnormal point. For example, if the temperature of the equipment suddenly rises by more than three times the standard deviation under normal operating conditions, the temperature data point will be marked as abnormal. Alternatively, use isolation forest, autoencoder, or clustering-based methods (such as K-means) to identify abnormal points, use normal data to train the anomaly detection model, and use an independent data set to verify the accuracy of the model to ensure that the model can effectively identify anomalies.

[0052] After determining the abnormal data point, an abnormal data identifier is established for it, and then the window data of the previous period of time is located as the pre-anomaly window. The size of the pre-anomaly window depends on the sensitivity of anomaly detection and the needs of actual applications. Usually, the start time of the pre-anomaly window is usually set before the time point when the abnormal point occurs, and the size of the window is determined according to the time range of the equipment state change before the abnormal point occurs. A smaller window may be used to detect rapidly changing anomalies, while a larger window may be used to detect slowly developing anomalies. The specific time point can be determined according to business needs and the purpose of data analysis. For example, if the goal is to predict failures in the short term, data from a few hours before the abnormal point occurs may be selected as window data. Extract the data in the pre-anomaly window to obtain a window data set. The window data set contains the equipment status information before the abnormality occurs. These data will be used to build a probabilistic equipment failure prediction network.

[0053] Based on the extracted window data set and abnormal data identification, a probabilistic equipment fault prediction network is built using machine learning tools such as neural networks or support vector machines. This network can learn the data pattern before the abnormality occurs and predict the probability of future equipment failure. Training with the pre-abnormal window data set can enable the model to better grasp the changing rules of equipment operation data, identify the correlation between data, and thus more accurately predict the future state of the equipment. For example, using an LSTM network, input a window data set, and output the probability of fault prediction. During the training process, LSTM can capture long-term dependencies in time series and improve the accuracy of fault prediction. Based on the trained probabilistic equipment fault prediction network, an equipment status prediction model is established.

[0054] Preferably, step S14 of the embodiment of the present application further includes:

[0055] The abnormal data identifier and the window data set are used to remove data from the equipment time series data set to generate a normal time series data set; the normal time series data set is subjected to data time series verification to establish a time series verification result; a normal equipment prediction network is built according to the time series verification result, and an equipment state prediction model is established based on the normal equipment prediction network and the probabilistic equipment failure prediction network.

[0056] Specifically, use abnormal data identification to remove abnormal data points from the equipment time series data set, generate a normal time series data set containing only normal data, and reduce the impact of abnormal data on model training. Perform time series verification on the generated normal time series data set to check the continuity and rationality of the data and ensure data quality. For example, use the autocorrelation function (ACF) of the time series to check the autocorrelation of the data, and use the sliding window method to check the continuity of the data. Record the problems found in the time series verification and their processing results, such as missing values ​​of the data, outlier processing methods, etc. Based on the normal time series data set after time series verification, use time series prediction models (such as ARIMA, Prophet) or machine learning models (such as decision trees, random forests) to train the normal equipment prediction network to predict the behavior of the equipment under normal conditions. Combine the normal equipment prediction network and the probabilistic equipment fault prediction network to establish an equipment status prediction model, which can simultaneously consider the normal operating status and potential fault probability of the equipment and provide a comprehensive equipment status prediction.

[0057] Further, such as Figure 3 As shown, step S3 of the embodiment of the present application also includes:

[0058] Step S31: The equipment working data is used as the first round of input data and input into the equipment state prediction model. The state is identified through the normal equipment prediction network and the probabilistic equipment failure prediction network respectively. The output results are synchronized to the fusion analysis layer, anomaly verification is performed, and basic anomaly prediction results are generated.

[0059] Step S32: Use the equipment fitting working data set as the second round of input data, input it into the equipment state prediction model, perform state identification through the normal equipment prediction network and the probabilistic equipment fault prediction network respectively, synchronize the output results to the fusion analysis layer, perform anomaly verification, and generate additional anomaly prediction results.

[0060] Step S33: Perform superimposed anomaly analysis based on the basic anomaly prediction result and the additional anomaly prediction result, and output the prediction result.

[0061] Specifically, the basic anomaly prediction result is the anomaly prediction result obtained by the device state prediction model based on the device working data. The additional anomaly prediction result is the anomaly prediction result obtained by the device state prediction model based on the device fitting working data set. The fusion analysis layer is a comprehensive analysis platform for integrating and analyzing data and prediction results from different sources.

[0062] First, the equipment working data is used as the first round of input data and input into the normal equipment prediction network and the probabilistic equipment fault prediction network in the equipment state prediction model to predict the normal state and fault probability of the equipment respectively. The output results are synchronized to the fusion analysis layer to perform anomaly verification to identify and confirm abnormal conditions and generate basic anomaly prediction results. Then, the equipment fitting working data set is used as the second round of input data and input into the same equipment state prediction model to perform state identification again to predict the abnormal conditions of the equipment under different operation strategies. The output results are also synchronized to the fusion analysis layer to perform anomaly verification and generate additional anomaly prediction results. Finally, through superimposed anomaly analysis, the basic anomaly prediction results and the additional anomaly prediction results are fused and analyzed to more comprehensively understand the abnormal state and potential problems of the equipment and output the final prediction results. Through this process, not only can anomaly detection be performed based on actual operation data, but also anomaly prediction can be further verified and supplemented by simulation data, thereby improving the accuracy and comprehensiveness of predictions and providing stronger support for equipment maintenance and optimized operation.

[0063] Furthermore, step S2 of the embodiment of the present application further includes:

[0064] Step S21: performing maximum output power fitting of the power generation equipment based on the weather forecast data, and establishing power generation constraints based on the maximum output power fitting result.

[0065] Step S22: Obtain the remaining energy of the energy storage, and establish an energy storage capacity constraint based on the remaining energy.

[0066] Step S23: Establishing a supply strategy set according to the power generation constraint and energy storage capacity constraint.

[0067] Specifically, by using weather forecast data, the maximum output power of the power generation equipment under specific weather conditions is predicted by establishing a performance model of the power generation equipment, such as a prediction model based on machine learning, including but not limited to linear regression, support vector machine, neural network, etc. Based on the maximum output power fitting result, the power output upper limit of the power generation equipment at different time points, that is, the power generation constraint, is determined to guide the power generation scheduling. For example, for solar power generation, the maximum output power can be predicted based on light intensity and temperature. Then, by monitoring the current state of the energy storage system, the remaining energy of the energy storage, that is, the energy that the energy storage system can currently provide, is obtained, which is usually achieved through a battery management system (BMS). According to the remaining energy, the charging and discharging capacity limit of the energy storage system at different time points, that is, the energy storage capacity constraint, is set. Integrate the power generation constraint and the energy storage capacity constraint, and establish a supply strategy set through optimization algorithms such as linear programming and mixed integer programming to guide the scheduling of the energy system. The supply strategy set generated by the above process contains power generation and energy storage operation plans at different time points, which limits the selection range for subsequent multi-objective optimization, ensuring that the generated strategy meets the equipment performance constraints and actual load requirements while maximizing the system's operating efficiency and economy.

[0068] Preferably, step S22 of the embodiment of the present application further includes:

[0069] Establish a predicted load demand, which is a predicted demand after a predetermined time period; evaluate the energy storage supply demand based on the predicted load demand and generate an evaluation result; configure energy reservation of the remaining energy according to the evaluation result, and establish energy storage energy constraints based on the energy reservation.

[0070] Specifically, load demand forecasting refers to the prediction of power demand for a power grid or device after a predetermined time period in the future, which can be 24 hours, a week, etc. Energy storage energy constraints are restrictions on the charging and discharging operations of energy storage systems based on energy conservation, ensuring that the energy storage system maintains a healthy operating state while meeting load demand.

[0071] First, obtain the predicted load demand for a period of time in the future from the government website or power supplier or power market. Based on the predicted load demand, calculate the energy that the energy storage system needs to provide to meet the load demand, and then evaluate the supply capacity of the energy storage system within a predetermined time period to generate an evaluation result. This evaluation result is a specific quantitative indicator for the evaluation of energy storage supply demand, reflecting the ability of the energy storage system to meet future load demands. Exemplarily, the energy demand ratio of the energy storage system can be calculated as an evaluation result. Based on the evaluation results, in order to ensure that the energy storage system meets the load demand, determine the energy that needs to be retained to avoid excessive discharge. According to the energy reservation, set the upper and lower limits of the energy storage system's charge and discharge to ensure that the energy storage system can meet the load demand without damaging its long-term operation health, thereby improving the overall efficiency and stability of the energy system.

[0072] Furthermore, the method described in the embodiment of the present application also includes:

[0073] Establishing equipment load thresholds, accumulating the continuous triggering time of equipment load thresholds of power generation equipment with multi-objective optimization analysis results, and establishing cumulative warnings; and compensating for multi-objective optimization analysis results with the cumulative warnings.

[0074] Specifically, the equipment load threshold is the maximum load or load ratio allowed for the equipment within the safe operating range. Exceeding this threshold may affect the life of the equipment or cause a malfunction. Determine the equipment load threshold based on the parameters recommended by the equipment manufacturer and historical operating data. Accumulate the duration of the equipment load exceeding the threshold state in the multi-objective analysis results obtained in step S4. Set a safety threshold. When the cumulative triggering time of the equipment load threshold reaches or exceeds the threshold, the system automatically triggers an early warning to alert the operator. Based on the cumulative early warning, adjust the results of the multi-objective optimization analysis, re-run the optimization algorithm, consider the risk of equipment overload, and find a new balance point. For example, reduce the equipment load, increase energy storage discharge, etc. to reduce the risk of equipment overload.

[0075] Through the above steps, the equipment load can be effectively monitored to avoid the equipment being in a high-risk state for a long time. At the same time, by adjusting the multi-objective optimization analysis results, the safe and efficient operation of the power generation system can be guaranteed.

[0076] In summary, the intelligent adjustment method based on solar power generation prediction provided by the embodiment of the present application has the following technical effects:

[0077] Connect the equipment database of the power generation equipment, read the time series data in the equipment database, establish the equipment time series data set of the power generation equipment, establish the equipment state prediction model with the equipment time series data set, and the equipment state prediction model includes a normal equipment prediction network and a probabilistic equipment failure prediction network; collect and obtain weather data, the weather data includes weather forecast data, obtain load demand, and establish a supply strategy set for photovoltaic power generation and energy storage based on the load demand and the weather forecast data. This process provides basic data support for subsequent intelligent regulation and ensures the accuracy and real-time nature of the prediction. Configure the equipment fitting working data set of the power generation equipment with the supply strategy set, collect the equipment working data of the power generation equipment, predict the equipment state with the equipment working data, the equipment fitting working data set and the equipment state prediction model, and establish the prediction result; perform multi-objective optimization analysis based on the prediction result and the supply strategy set. This process uses a multi-objective optimization algorithm to comprehensively consider multiple objectives such as power generation efficiency, cost, and equipment maintenance, find the optimal operation strategy, solve the problem of multi-objective balancing in traditional regulation methods, and achieve the optimization of the overall performance of the power generation system. In addition, the equipment load threshold is established, and the continuous triggering time of the equipment load threshold of the power generation equipment is accumulated based on the multi-objective optimization analysis results to establish a cumulative warning; the multi-objective optimization analysis results are compensated through the cumulative warning to avoid the power generation equipment being in a high-risk state for a long time, so as to ensure the stable operation of the power generation equipment and extend the service life of the equipment. The adjustment plan is generated based on the multi-objective optimization analysis results, and intelligent adjustment management is performed based on the adjustment plan, which can quickly respond to environmental changes and load demand fluctuations, ensuring the stable operation of the power grid.

[0078] In general, the embodiments of the present application improve the adaptability of the photovoltaic power generation system to environmental changes and load demands through data fusion and prediction of deep learning algorithms; multi-objective optimization analysis ensures the optimal balance between system efficiency, cost and equipment maintenance; and the application of real-time intelligent regulation and equipment status monitoring technology enables the power generation system to automatically adjust to maximize power generation efficiency, minimize costs, and extend equipment life, thereby significantly improving the overall performance of the power generation system as a whole, making the photovoltaic power generation system more adaptable to complex and changeable operating environments, and promoting the further development and application of renewable energy technologies.

[0079] Embodiment 2, as Figure 4 As shown, the embodiment of the present application provides an intelligent adjustment device based on solar power generation prediction, the device comprising:

[0080] The prediction model building module 10 is used to connect to the equipment database of the power generation equipment, execute the time series data reading in the equipment database, establish the equipment time series data set of the power generation equipment, and establish the equipment state prediction model with the equipment time series data set. The equipment state prediction model includes a normal equipment prediction network and a probabilistic equipment failure prediction network.

[0081] The supply strategy set establishment module 20 is used to collect and obtain weather data, wherein the weather data includes weather forecast data, obtain load demand, and establish a supply strategy set for photovoltaic power generation and energy storage based on the load demand and the weather forecast data.

[0082] The equipment state prediction module 30 is used to configure the equipment fitting working data set of the power generation equipment with the supply strategy set, collect the equipment working data of the power generation equipment, predict the equipment state with the equipment working data, the equipment fitting working data set and the equipment state prediction model, and establish a prediction result.

[0083] The multi-objective optimization analysis module 40 is used to perform multi-objective optimization analysis according to the prediction results and the supply strategy set, as follows:

[0084] ;

[0085] in, is the objective function, is the weight factor of the importance of load demand satisfaction, Characterizes the load difference, the difference between the actual power generation and the load demand, , For in time The actual power generation For energy storage in time The output power, is the load demand power, is the weight factor representing the total cost, is the total cost, including power generation cost, energy storage cost and equipment abnormality cost, is the weight factor for device maintenance, Representation in time Equipment maintenance costs, is the weight factor of power generation efficiency, Representation in time power generation efficiency.

[0086] The intelligent adjustment module 50 is used to generate an adjustment plan based on the multi-objective optimization analysis results and perform intelligent adjustment management based on the adjustment plan.

[0087] Furthermore, the prediction model building module 10 of the embodiment of the present application is also used to perform the following steps:

[0088] Abnormal data is extracted from the equipment time series data set, and an abnormal data identifier is established; the preceding abnormal window is located according to the abnormal data identifier, and the window data set of the preceding abnormal window is extracted; a time-series sensitive probabilistic equipment fault prediction network is constructed according to the abnormal data identifier and the window data set; and an equipment state prediction model is established according to the constructed probabilistic equipment fault prediction network.

[0089] Preferably, the prediction model building module 10 of the embodiment of the present application is also used to perform the following steps:

[0090] The abnormal data identifier and the window data set are used to remove data from the equipment time series data set to generate a normal time series data set; the normal time series data set is subjected to data time series verification to establish a time series verification result; a normal equipment prediction network is built according to the time series verification result, and an equipment state prediction model is established based on the normal equipment prediction network and the probabilistic equipment failure prediction network.

[0091] Furthermore, the device state prediction module 30 of the embodiment of the present application is also used to perform the following steps:

[0092] The equipment working data is used as the first round of input data, and is input into the equipment state prediction model. The state identification is performed respectively through the normal equipment prediction network and the probabilistic equipment fault prediction network. The output result is synchronized to the fusion analysis layer, and anomaly verification is performed to generate a basic anomaly prediction result. The equipment fitting working data set is used as the second round of input data, and is input into the equipment state prediction model. The state identification is performed respectively through the normal equipment prediction network and the probabilistic equipment fault prediction network. The output result is synchronized to the fusion analysis layer, and anomaly verification is performed to generate an additional anomaly prediction result. The basic anomaly prediction result and the additional anomaly prediction result are used to perform superimposed anomaly analysis, and the prediction result is output.

[0093] Furthermore, the provisioning strategy set establishing module 20 of the embodiment of the present application is also used to perform the following steps:

[0094] Based on the weather forecast data, the maximum output power of the power generation equipment is fitted, and the power generation power constraint is established based on the maximum output power fitting result; the remaining energy of the energy storage is obtained, and the energy storage capacity constraint is established based on the remaining energy; and a supply strategy set is established according to the power generation constraint and the energy storage capacity constraint.

[0095] Preferably, the provisioning strategy set establishing module 20 of the embodiment of the present application is further configured to perform the following steps:

[0096] Establish a predicted load demand, which is a predicted demand after a predetermined time period; evaluate the energy storage supply demand based on the predicted load demand and generate an evaluation result; configure energy reservation of the remaining energy according to the evaluation result, and establish energy storage energy constraints based on the energy reservation.

[0097] Furthermore, the device described in the embodiment of the present application is also used to perform the following steps:

[0098] Establishing equipment load thresholds, accumulating the continuous triggering time of equipment load thresholds of power generation equipment with multi-objective optimization analysis results, and establishing cumulative warnings; and compensating for multi-objective optimization analysis results with the cumulative warnings.

[0099] Through the above detailed description of the intelligent regulation method based on solar power generation prediction in this specification, those skilled in the art can clearly understand the intelligent regulation device based on solar power generation prediction in this embodiment. For the device disclosed in Example 2, since it corresponds to the method disclosed in Example 1 and has corresponding functional modules and beneficial effects, the relevant parts can be referred to the description of the method part.

[0100] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent regulation method based on solar power generation prediction, characterized in that: The method comprises: Connecting to an equipment database of a power generation device, reading time series data in the equipment database, establishing an equipment time series data set of the power generation device, and establishing an equipment state prediction model with the equipment time series data set, wherein the equipment state prediction model includes a normal equipment prediction network and a probabilistic equipment failure prediction network; Collect and obtain weather data, the weather data including weather forecast data, obtain load demand, and establish a supply strategy set for photovoltaic power generation and energy storage based on the load demand and the weather forecast data; configuring an equipment fitting working data set of a power generation equipment with the supply strategy set, collecting equipment working data of the power generation equipment, performing equipment state prediction with the equipment working data, the equipment fitting working data set and the equipment state prediction model, and establishing a prediction result; Based on the forecast results and supply strategy set, a multi-objective optimization analysis is performed as follows: ； in, is the objective function, is the weight factor of the importance of load demand satisfaction, Characterizes the load difference, the difference between the actual power generation and the load demand, , For in time The actual power generation For energy storage in time The output power, is the load demand power, is the weight factor representing the total cost, is the total cost, including power generation cost, energy storage cost and equipment abnormality cost, is the weight factor for device maintenance, Representation in time Equipment maintenance costs, is the weight factor of power generation efficiency, Representation in time Power generation efficiency; Generate adjustment plans based on multi-objective optimization analysis results, and perform intelligent adjustment management based on the adjustment plans.

2. The intelligent regulation method based on solar power generation prediction according to claim 1, characterized in that: The establishing of the equipment state prediction model based on the equipment time series data set also includes: Extracting abnormal data from the device time series data set and establishing abnormal data identification; Locating a preceding abnormal window according to the abnormal data identifier, and extracting a window data set of the preceding abnormal window; Building a time-sensitive probabilistic equipment failure prediction network based on the abnormal data identifier and window data set; Establish an equipment status prediction model based on the completed probabilistic equipment failure prediction network.

3. The intelligent regulation method based on solar power generation prediction according to claim 2, characterized in that: The step of establishing a device state prediction model based on the constructed probabilistic device failure prediction network also includes: Using the abnormal data identifier and the window data set to remove data from the device time series data set, to generate a normal time series data set; Performing data time series verification on the normal time series data set to establish a time series verification result; A normal equipment prediction network is built according to the timing verification results, and an equipment state prediction model is established based on the normal equipment prediction network and the probabilistic equipment failure prediction network.

4. The intelligent regulation method based on solar power generation prediction according to claim 3, characterized in that: The performing of equipment state prediction based on the equipment working data, the equipment fitting working data set and the equipment state prediction model to establish a prediction result also includes: The equipment working data is used as the first round of input data and input into the equipment state prediction model, and the state is identified through the normal equipment prediction network and the probabilistic equipment failure prediction network respectively, and the output results are synchronized to the fusion analysis layer, and anomaly verification is performed to generate basic anomaly prediction results; The equipment fitting working data set is used as the second round of input data, and is input into the equipment state prediction model, and the state recognition is performed respectively through the normal equipment prediction network and the probabilistic equipment fault prediction network, and the output results are synchronized to the fusion analysis layer, and anomaly verification is performed to generate additional anomaly prediction results; The basic anomaly prediction result and the additional anomaly prediction result are used to perform superimposed anomaly analysis, and the prediction result is output.

5. The intelligent regulation method based on solar power generation prediction according to claim 1, characterized in that: The step of establishing a photovoltaic power generation and energy storage supply strategy set based on the load demand and the weather forecast data further includes: Performing maximum output power fitting of the power generation equipment based on the weather forecast data, and establishing a power generation power constraint based on the maximum output power fitting result; Acquiring remaining energy of energy storage, and establishing energy storage capacity constraints based on the remaining energy; A supply strategy set is established according to the power generation constraint and the energy storage capacity constraint.

6. The intelligent regulation method based on solar power generation prediction according to claim 5, characterized in that: The establishing of energy storage capacity constraint based on the remaining energy further includes: Establishing a predicted load demand, wherein the predicted load demand is a predicted demand after a predetermined time period; Evaluate the energy storage supply demand based on the predicted load demand and generate an evaluation result; Energy reservation of the remaining energy is configured according to the evaluation result, and energy storage energy constraints are established based on the energy reservation.

7. The intelligent regulation method based on solar power generation prediction according to claim 1, characterized in that: The method further comprises: Establish equipment load thresholds, accumulate the continuous triggering time of equipment load thresholds of power generation equipment based on multi-objective optimization analysis results, and establish cumulative warnings; The multi-objective optimization analysis result compensation is performed through the cumulative early warning.

8. An intelligent regulating device based on solar power generation prediction, characterized in that: The device is used to execute the intelligent regulation method based on solar power generation prediction according to any one of claims 1 to 7, and the device comprises: A prediction model building module, the prediction model building module is used to connect to the equipment database of the power generation equipment, execute the time series data reading in the equipment database, establish the equipment time series data set of the power generation equipment, and establish the equipment state prediction model with the equipment time series data set, the equipment state prediction model includes a normal equipment prediction network and a probabilistic equipment failure prediction network; A supply strategy set establishment module, the supply strategy set establishment module is used to collect and obtain weather data, the weather data includes weather forecast data, obtain load demand, and establish a supply strategy set for photovoltaic power generation and energy storage based on the load demand and the weather forecast data; An equipment state prediction module, the equipment state prediction module is used to configure an equipment fitting working data set of a power generation equipment with the supply strategy set, collect equipment working data of the power generation equipment, perform equipment state prediction with the equipment working data, the equipment fitting working data set and the equipment state prediction model, and establish a prediction result; The multi-objective optimization analysis module is used to perform multi-objective optimization analysis based on the prediction results and the supply strategy set, as follows: ; in, is the objective function, is the weight factor of the importance of load demand satisfaction, Characterizes the load difference, the difference between the actual power generation and the load demand, , For in time The actual power generation For energy storage in time The output power, is the load demand power, is the weight factor representing the total cost, is the total cost, including power generation cost, energy storage cost and equipment abnormality cost, is the weight factor for device maintenance, Representation in time Equipment maintenance costs, is the weight factor of power generation efficiency, Representation in time power generation efficiency; An intelligent adjustment module is used to generate an adjustment plan based on the multi-objective optimization analysis results and perform intelligent adjustment management based on the adjustment plan.

Images