Intelligent and economical off-grid optical storage micro-grid system
Through the integration of high-precision environmental monitoring and advanced prediction algorithms, combined with machine learning and deep reinforcement learning, real-time photovoltaic output prediction and dynamic load management of off-grid optical storage microgrid systems are achieved, and the photovoltaic output prediction and load management of existing systems is solved. The issues of uncertainty and simplicity have improved energy utilization efficiency and system stability, and realized intelligent management and self-healing control.
Patent Information
- Application Number
- CN202411097495.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-12
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing off-grid optical storage microgrid system has uncertainty and simplicity in photovoltaic output prediction and dynamic load management, resulting in low energy utilization efficiency, insufficient system stability and reliability, lack of intelligent management methods, and difficult to respond to environmental changes. and user needs, and fault detection and response rely on manual intervention, slow response speed, lack of self-repair ability.
Adopt high-precision environmental monitoring and advanced prediction algorithms, combined with machine learning and deep reinforcement learning technology, real-time photovoltaic output prediction and dynamic load management, integrate intelligent fault diagnosis and self-healing control, promote users to participate in energy management through the blockchain platform, and optimize Resource configuration.
It realizes efficient and stable operation of the system in complex environments, maximizes energy output and utilization efficiency, reduces energy waste, improves the overall efficiency of the system, reduces operating costs, and enhances user participation and system response capabilities.
Smart Images

Figure CN119010178A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of photovoltaic power technology, and in particular to an intelligent and economical off-grid photovoltaic storage microgrid system. Background Art
[0002] In the rural areas of the northwest, due to the vast land and sparse population and the disorderly development of land resources, especially the increase in planting and breeding projects in remote mountainous areas far away from residential areas, the traditional power supply method faces huge challenges. The planting and breeding industries in these areas are scattered, and they are far away from the main residential areas, making it neither economical nor efficient to extend coverage through conventional distribution networks, especially considering the low return on investment and the possibility of equipment being abandoned in the short term. In this context, it has become an urgent need to explore an intelligent and economical off-grid photovoltaic storage microgrid system. This system can effectively connect the large power grid, distribution network and microgrid, realize the complementarity of centralized and distributed energy systems, and solve the power supply problem in remote areas at a lower cost and faster speed.
[0003] At present, although off-grid photovoltaic storage microgrid systems have shown great application potential, existing technologies still face many challenges in actual deployment. First, since solar energy resources are greatly affected by geographical location, season, and weather factors, the output of photovoltaic systems is significantly intermittent and uncertain. Traditional control systems often find it difficult to accurately predict changes in photovoltaic output, resulting in inflexible system scheduling and inability to fully match user load requirements, thereby reducing energy utilization efficiency. Secondly, although energy storage technology can alleviate the volatility of photovoltaic power generation to a certain extent, its high cost, limited cycle life, and energy density limit its widespread application in microgrid systems. In addition, the lack of effective intelligent management means makes the system slow to respond to environmental changes and user needs, and the system stability and reliability need to be improved.
[0004] In the existing technology, the off-grid photovoltaic storage microgrid system is insufficient in intelligent management and lacks efficient intelligent monitoring and prediction technology, making it difficult for the system to accurately predict photovoltaic output based on real-time environmental parameters and historical data, affecting the foresight and rationality of energy distribution. On the other hand, the existing dynamic load management strategy is relatively simple, lacks an in-depth understanding of user behavior patterns, and cannot accurately adjust energy storage and photovoltaic output, which often leads to energy waste and supply shortages. In addition, most systems rely on manual intervention in fault detection and response, have slow response speeds, and lack self-repair capabilities. Once a fault occurs, it is easy to cause power outages. Due to the lack of effective incentive mechanisms and convenient interactive platforms, users lack initiative and flexibility in the use and management of energy, which not only limits the economy of energy use, but also hinders the improvement of the overall efficiency of the energy system. Therefore, it is particularly urgent to develop an intelligent and economical off-grid photovoltaic storage microgrid system that integrates advanced artificial intelligence and optimization algorithms. Summary of the invention
[0005] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide an intelligent and economical off-grid photovoltaic storage microgrid system. It can integrate high-precision environmental monitoring and advanced prediction algorithms. The system can respond to environmental changes in real time, accurately predict future photovoltaic output, and adopt a multi-objective optimization scheduling strategy of deep reinforcement learning to maximize energy output and utilization efficiency, ensuring that the system can maintain efficient and stable operation under various complex environmental conditions.
[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: an intelligent and economical off-grid photovoltaic storage microgrid system, characterized in that the system includes the following components:
[0007] Intelligent monitoring and prediction: The system integrates high-precision meteorological sensors and advanced data analysis algorithms to monitor light intensity, temperature, wind speed and other environmental parameters in real time, and combines historical data to accurately predict photovoltaic output in the future, providing a basis for decision-making in energy allocation.
[0008] Dynamic load management: Using machine learning technology to predict system load change trends, automatically adjust the energy storage device charging and discharging strategy and photovoltaic output power to ensure supply and demand balance and reduce unnecessary energy waste;
[0009] Multi-objective optimization scheduling: Develop a multi-objective optimization model that comprehensively considers economic benefits, system stability, and resource utilization, and uses intelligent learning algorithms to achieve optimal tilt adjustment of the photovoltaic array, optimal charging and discharging strategy of the energy storage system, and intelligent distribution of the load to maximize the overall benefits of the system;
[0010] Fault diagnosis and self-healing control: The system integrates intelligent fault detection and diagnosis, monitors system status in real time, quickly locates and isolates faults, and simultaneously activates backup plans and self-healing mechanisms to ensure uninterrupted power supply;
[0011] User interaction and energy management: Build a blockchain-based energy management platform to achieve energy management between users and with the main power grid, encourage users to participate in energy management, and optimize resource allocation.
[0012] Furthermore, the specific steps of the intelligent monitoring and prediction are:
[0013] (1) Real-time monitoring of environmental parameters: The system continuously collects key parameters such as light intensity, ambient temperature, and wind speed through high-precision meteorological sensors installed at the PV-storage microgrid site, including photoresistors, temperature sensors, and anemometers;
[0014] (2) Data preprocessing: The collected data is cleaned and preprocessed to remove outliers, smooth noise, ensure data quality, and apply time series analysis techniques to process time-related continuous data and extract useful features;
[0015] (3) Historical data analysis: Use historical meteorological data and PV output data to build a database, analyze patterns and trends in historical data, and identify the nonlinear relationship between light intensity, temperature, and PV output;
[0016] (4) Model training: Select a time series prediction model, input the preprocessed data for training, learn the nonlinear relationship between environmental parameters and PV output, and thus predict future PV output;
[0017] (5) Future output prediction: Apply the trained model, input the current and future predicted environmental parameters, predict the PV array output in the next period of time, make multiple predictions and take the average and use confidence intervals to increase the accuracy of the prediction.
[0018] Furthermore, the intelligent monitoring and prediction is used to analyze and predict the patterns of data series that change over time through time series analysis methods, helping to identify the characteristics of trends, seasonality, periodicity and random fluctuations in the data, revealing the relationship between environmental parameters such as light intensity and temperature and photovoltaic output that change over time, thereby improving the accuracy of future output predictions, building a time series prediction model, using historical data to train the model, and learning the patterns and laws in the time series. The algorithm formula is: F(T+1) is the predicted value, X(i) is the actual value of the i-th period of the time series, and N is the number of averages. The trained model is used to make predictions based on the environmental parameters and features collected in real time. The prediction results include the photovoltaic output and load demand in the future. The mean square error is used to evaluate the accuracy of the prediction results. The model parameters are adjusted according to the evaluation results to optimize the prediction performance.
[0019] Furthermore, the time series prediction model is based on the mean square error MSE It is used to quantify the difference between the predicted value and the actual value, use an independent test data set to evaluate the prediction performance of the model, calculate the MSE between the predicted value and the actual value, and intuitively understand the prediction error size of the model. The algorithm formula is: y i is the actual observed value, is the model estimate, n is the number of observations, and based on the MSE evaluation results, the model parameters are adjusted and a new model structure is selected to optimize the prediction performance.
[0020] Furthermore, the dynamic load management comprises the following specific steps:
[0021] (1) Load data collection: The system collects real-time load data through smart meters and sensors, including power consumption and power usage pattern information of various loads;
[0022] (2) Load forecasting: Based on historical load data, environmental factors, and user behavior patterns, machine learning technology is used to predict load change trends in the future.
[0023] (3) Strategy formulation: Based on the load forecast results, combined with the current photovoltaic output and energy storage status, formulate the charging and discharging strategy of the energy storage equipment and the distribution strategy of the photovoltaic output power;
[0024] (4) Strategy execution: Automatically adjust the operating status of the photovoltaic array, the charging and discharging power of the energy storage device, and the distribution of the load to ensure the balance of system supply and demand and optimize the efficiency of energy use;
[0025] (5) Real-time monitoring and adjustment: The system monitors load changes, PV output, and energy storage status in real time, and fine-tunes strategies based on real-time data to adapt to environmental changes and fluctuations in user demand.
[0026] Furthermore, the dynamic load management collects historical load data, including historical power consumption and power consumption patterns, through the support vector machine SVM in machine learning technology, and learns the relationship between historical load data and environmental factors, that is: f(x)=sign(ω·x+b), x represents the eigenvector of the input sample, ω represents the normal vector of the hyperplane, and b represents the intercept of the hyperplane. The predicted value is calculated to obtain the predicted power load value in the future period. According to the load prediction results, combined with the current photovoltaic output and energy storage status, the charging and discharging strategy of the energy storage equipment and the distribution strategy of the photovoltaic output power are formulated. The system automatically adjusts the operating status of the photovoltaic array, the charging and discharging power of the energy storage equipment, and the distribution of the load to ensure the balance of supply and demand and optimize the energy utilization efficiency.
[0027] Furthermore, the multi-objective optimization scheduling has the following specific steps:
[0028] (1) Objective definition: Define the objective function of multi-objective optimization, which usually includes minimizing operating costs, maximizing energy utilization, maintaining system voltage and frequency stability, and improving user satisfaction;
[0029] (2) Problem modeling and algorithm design: Convert the actual problem into a mathematical model, including decision variables such as energy storage: charging and discharging strategy, photovoltaic array inclination, load distribution plan, constraints: equipment capacity limitation, energy supply and demand balance, safe operation restrictions and objective function, and select and design a multi-objective optimization algorithm based on the characteristics of the problem;
[0030] (3) Parameter setting and iterative optimization: Set parameters for the algorithm. Through the iterative process, the algorithm continuously generates a new generation of solutions and gradually approaches the optimal solution set.
[0031] (4) Optimization and interpretation of solutions: Using the decision-making method of multi-objective optimization, evaluate the solution set output by the optimization algorithm, check whether all constraints are met, and select and synthesize the final scheduling strategy;
[0032] (5) Strategy implementation and evaluation: Apply the optimized scheduling strategy to the actual system and verify the effectiveness of the strategy and the improvement of system performance through real-time monitoring and post-evaluation.
[0033] Furthermore, the multi-objective optimization scheduling can automatically adjust the inclination angle of the photovoltaic array, the charging and discharging strategy of the energy storage system and the load distribution through the deep reinforcement learning DQN control system. According to the real-time feedback and performance evaluation of the system, the parameters of the DQN algorithm are adjusted to optimize the overall performance of the system. The algorithm is used as follows: the system state s t , including multi-dimensional features of photovoltaic output, energy storage status, load demand, and external environmental parameters. These features are input into the DQN neural network for state representation, action a t Including the tilt adjustment of the photovoltaic array, the charging and discharging strategy of the energy storage system, and the intelligent distribution of the load, these actions constitute a discrete and continuous action space. In multi-objective optimization scheduling, the reward function needs to reflect multiple goals, including economic benefits, system stability, and resource utilization. Each dimension corresponds to an optimization goal, one dimension represents economic benefits, and the other dimension represents the stability of system voltage and fluctuation range. The algorithm formula is: Q(s t ,a t ) means taking action a in state s t The expected return, s t+1 is the next state. Based on the Q value learned by DQN, the system generates a multi-objective optimization strategy, which trades off between multiple objectives to find a balanced solution that meets all objective requirements. Strategy generation is achieved by selecting the action with the highest comprehensive Q value. According to the real-time feedback and performance evaluation of the system, the parameters and reward function of the DQN algorithm are further adjusted to optimize the overall performance of the system.
[0034] Furthermore, the specific steps of the fault diagnosis and self-healing control are:
[0035] (1) Data judgment: Real-time collection of operating data of the photovoltaic array, energy storage equipment and inverter in the photovoltaic microgrid system, and use of sensors to monitor key parameters, including current, voltage, temperature and frequency, to ensure the normal operation of the system;
[0036] (2) Data analysis and fault detection: Process and analyze the collected data and determine whether there are potential faults and abnormal conditions based on preset thresholds and algorithms;
[0037] (3) Fault location and isolation: Once a fault or abnormality is detected, the system quickly locates the location and cause of the fault, accurately analyzes and classifies the fault, and ensures that the fault is accurately isolated;
[0038] (4) Start backup plans and self-healing mechanisms: Based on the type of fault, the system starts backup plans to ensure that the system can continue to operate stably. The self-healing mechanism includes automatically restarting the device, switching to the backup power supply, and adjusting the working parameters to restore the normal function of the system;
[0039] (5) Recording and reporting: The time, location, cause, and handling process of the fault are recorded in the log for subsequent analysis and maintenance. Fault information is reported to users and administrators through the user interface and remote monitoring platform to ensure that they are aware of the system status in a timely manner.
[0040] Furthermore, the fault diagnosis and self-healing control utilizes a threshold judgment method to detect whether a fault occurs in state monitoring and fault detection, sets one or more thresholds to determine whether the system is in a normal state and a fault occurs, and sets a threshold for the output power of the photovoltaic panel. If the actual output power is lower than the threshold, it is considered that the photovoltaic panel has a fault. For the detected fault, the system takes corresponding self-healing measures according to the type and degree of the fault, including restarting the equipment, switching to the backup power supply, and adjusting the working parameters. In this process, the threshold judgment ensures that the system's recovery operation is carried out within a safe range.
[0041] Compared with the existing technology, this intelligent and economical off-grid photovoltaic storage microgrid system has the following beneficial effects:
[0042] 1. The present invention integrates advanced artificial intelligence technology and optimization algorithms to monitor environmental parameters in real time and predict photovoltaic output, thereby automatically adjusting the charging and discharging strategy of the energy storage equipment and the photovoltaic output power, and dynamically adjusting the photovoltaic array inclination, energy storage charging and discharging strategy and load distribution according to the system status and external conditions, maximizing energy output and utilization efficiency, and ensuring that the system can maintain efficient and stable operation under various complex environmental conditions.
[0043] 2. The present invention adopts a multi-objective optimization scheduling strategy, comprehensively considers economic benefits, system stability and resource utilization, and realizes the optimal tilt adjustment of the photovoltaic array, the optimal charging and discharging strategy of the energy storage system and the intelligent distribution of the load through deep reinforcement learning. These optimization measures can maximize the overall benefits of the system, reduce energy waste, improve energy utilization efficiency, and reduce the operating costs of the system.
[0044] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0046] Figure 1 This is a process operation diagram of an intelligent and economical off-grid photovoltaic storage microgrid system.
[0047] Figure 2 This is a flow chart of an intelligent and economical off-grid photovoltaic storage microgrid system. DETAILED DESCRIPTION
[0048] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0049] Embodiment 1
[0050] This embodiment introduces in detail the specific application of an intelligent and economical off-grid photovoltaic storage microgrid system. By integrating high-precision meteorological sensors and advanced data analysis algorithms, the system can monitor environmental parameters in real time and predict photovoltaic output, thereby autonomously adjusting the energy allocation strategy to ensure the stable operation of the system. The fault diagnosis and self-healing control functions enable the system to detect and isolate faults in real time, activate backup plans and self-healing mechanisms, reduce downtime caused by faults, and improve the reliability and stability of the system.
[0051] In the specific implementation, the system deploys high-precision meteorological sensors at the photovoltaic microgrid site, including photoresistors, temperature sensors, and anemometers, which continuously collect key parameters such as light intensity, ambient temperature, and wind speed. The collected data undergoes a cleaning process to remove outliers and noise. Through time series analysis technology: Smooth the data to ensure data quality and extract useful features. Use historical meteorological data and photovoltaic output data to build a database, analyze data patterns and trends, identify the relationship between environmental parameters and photovoltaic output, select a time series prediction model, and use the mean square error to predict the output of photovoltaic power. MSE It is used to quantify the difference between the predicted value and the actual value, use an independent test data set to evaluate the prediction performance of the model, calculate the MSE between the predicted value and the actual value, and intuitively understand the prediction error size of the model. The algorithm formula is: y i is the actual observed value, is the model estimate, n is the number of observations. According to the MSE evaluation results, the model parameters are adjusted and a new model structure is selected to optimize the prediction performance. The model is trained using preprocessed data to learn the nonlinear relationship between environmental parameters and photovoltaic output. The trained model is used to predict the photovoltaic array output in the future. The average of multiple predictions and confidence intervals are used to improve the prediction accuracy.
[0052] Through smart meters and sensors, the power consumption and power consumption pattern information of various loads are obtained in real time. Machine learning technology is used to collect historical load data, including historical power consumption and power consumption patterns. The relationship between historical load data and environmental factors is learned, that is: f(x) = sign(ω·x+b). The predicted value is calculated to obtain the predicted power load value in the future. According to the load forecast results, combined with the current photovoltaic output and energy storage status, the charging and discharging strategy of the energy storage equipment and the allocation strategy of the photovoltaic output power are formulated. The charging and discharging strategy of the energy storage equipment and the photovoltaic output power allocation are formulated and implemented to achieve supply and demand balance. The system monitors the system status in real time and fine-tunes the strategy according to the real-time data to adapt to environmental changes and fluctuations in user demand.
[0053] Multi-objective optimization scheduling sets objective functions, such as minimizing operating costs, maximizing energy utilization, maintaining system voltage and frequency stability, and improving user satisfaction. It converts the problem into a mathematical model, including decision variables, energy storage charging and discharging strategies, photovoltaic array inclination, load distribution and constraints, equipment capacity, supply and demand balance, and safety restrictions. It sets parameters for the optimization algorithm, approximates the optimal solution set through an iterative process, evaluates the algorithm output solution set, selects a strategy that meets all constraints, applies the optimization strategy to the actual system, and monitors and evaluates the strategy effect in real time.
[0054] Collect system operation data in real time, monitor key parameters, determine whether there are potential faults, quickly locate the fault location and cause, conduct accurate analysis and classification, and activate backup plans and self-healing mechanisms according to the type of fault, such as automatically restarting equipment and switching to backup power supplies. Record fault information and report it to users and administrators through the user interface and remote monitoring platform, build a blockchain-based energy management platform, realize energy management between users and with the main power grid, encourage users to participate in energy management, and optimize resource allocation.
[0055] Embodiment 2
[0056] An intelligent and economical off-grid photovoltaic storage microgrid system in this embodiment particularly emphasizes that the system adopts a deep learning algorithm control system to automatically adjust the inclination angle of the photovoltaic array, the charging and discharging strategy of the energy storage system, and the load distribution, which effectively improves energy utilization efficiency, system reliability and user experience, and is particularly suitable for energy supply needs in remote areas or specific scenarios.
[0057] First, the system continuously collects key parameters such as light intensity, ambient temperature, and wind speed through high-precision meteorological sensors deployed at the photovoltaic microgrid site. The collected data undergoes a cleaning process to remove outliers and noise. A database is established using historical meteorological data and photovoltaic output data to analyze data patterns and trends, identify the relationship between environmental parameters and photovoltaic output, select a time series prediction model, and use the trained model to predict the photovoltaic array output in the future. The power consumption and power consumption pattern information of various loads is obtained in real time through smart meters and sensors. Machine learning technology is used to predict load change trends based on historical data, environmental factors, and user behavior patterns. According to the load prediction results and the current system status, the charging and discharging strategy of the energy storage equipment and the photovoltaic output power allocation are formulated and implemented to achieve a balance between supply and demand. The system monitors the system status in real time and fine-tunes the strategy based on real-time data to adapt to environmental changes and fluctuations in user demand.
[0058] Then, the multi-objective optimization scheduling uses the deep reinforcement learning DQN control system to automatically adjust the tilt angle of the photovoltaic array, the charging and discharging strategy of the energy storage system, and the load distribution. According to the real-time feedback and performance evaluation of the system, the parameters of the DQN algorithm are adjusted to optimize the overall performance of the system. The algorithm is used as follows: the system state s t , including multi-dimensional features of photovoltaic output, energy storage status, load demand, and external environmental parameters. These features are input into the DQN neural network for state representation, action a tIncluding the tilt adjustment of the photovoltaic array, the charging and discharging strategy of the energy storage system, and the intelligent distribution of the load, these actions constitute a discrete and continuous action space. In multi-objective optimization scheduling, the reward function needs to reflect multiple goals, including economic benefits, system stability, and resource utilization. Each dimension corresponds to an optimization goal, one dimension represents economic benefits, and the other dimension represents the stability of system voltage and fluctuation range. The algorithm formula is: Q(s t ,a t ) means taking action a in state s t The expected return, s t+1 is the next state. Based on the Q value learned by DQN, the system generates a multi-objective optimization strategy, which trades off between multiple objectives to find a balanced solution that meets all objective requirements. Strategy generation is achieved by selecting the action with the highest comprehensive Q value. According to the real-time feedback and performance evaluation of the system, the parameters and reward function of the DQN algorithm are further adjusted to optimize the overall performance of the system.
[0059] Finally, the system operation data is collected in real time, key parameters are monitored, and it is determined whether there are potential faults. If a system failure occurs, the system will quickly locate and isolate the fault, and at the same time activate the backup plan and self-healing mechanism to ensure the continuity and stability of the power supply and uninterrupted power supply. Through the blockchain-based energy management platform, users can view their energy usage, participate in energy management, and optimize resource allocation.
[0060] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.
Claims
1. An intelligent and economical off-grid photovoltaic storage microgrid system, characterized in that: The system consists of the following components: Intelligent monitoring and prediction: The system integrates high-precision meteorological sensors and advanced data analysis algorithms to monitor light intensity, temperature, wind speed and other environmental parameters in real time, and combines historical data to accurately predict photovoltaic output in the future, providing a basis for decision-making in energy allocation. Dynamic load management: Using machine learning technology to predict system load change trends, automatically adjust the energy storage device charging and discharging strategy and photovoltaic output power to ensure supply and demand balance and reduce unnecessary energy waste; Multi-objective optimization scheduling: Develop a multi-objective optimization model that comprehensively considers economic benefits, system stability, and resource utilization, and uses intelligent learning algorithms to achieve optimal tilt adjustment of the photovoltaic array, optimal charging and discharging strategy of the energy storage system, and intelligent distribution of the load to maximize the overall benefits of the system; Fault diagnosis and self-healing control: The system integrates intelligent fault detection and diagnosis, monitors system status in real time, quickly locates and isolates faults, and simultaneously activates backup plans and self-healing mechanisms to ensure uninterrupted power supply; User interaction and energy management: Build a blockchain-based energy management platform to achieve energy management between users and with the main power grid, encourage users to participate in energy management, and optimize resource allocation.
2. According to claim 1, a smart and economical off-grid photovoltaic storage microgrid system is characterized in that: The specific steps of intelligent monitoring and prediction are as follows: (1) Real-time monitoring of environmental parameters: The system continuously collects key parameters such as light intensity, ambient temperature, and wind speed through high-precision meteorological sensors installed at the PV-storage microgrid site, including photoresistors, temperature sensors, and anemometers; (2) Data preprocessing: The collected data is cleaned and preprocessed to remove outliers, smooth noise, ensure data quality, and apply time series analysis techniques to process time-related continuous data and extract useful features; (3) Historical data analysis: Use historical meteorological data and PV output data to build a database, analyze patterns and trends in historical data, and identify the nonlinear relationship between light intensity, temperature, and PV output; (4) Model training: Select a time series prediction model and input preprocessed data for training. Learn the nonlinear relationship between environmental parameters and PV output to predict future PV output; (5) Future output prediction: Apply the trained model, input the current and future predicted environmental parameters, predict the PV array output in the next period of time, make multiple predictions and take the average and use confidence intervals to increase the accuracy of the prediction.
3. According to claim 1, a smart and economical off-grid photovoltaic storage microgrid system is characterized in that: The intelligent monitoring and prediction uses a time series analysis method to analyze and predict the patterns of data series that change over time, help identify the characteristics of trends, seasonality, periodicity, and random fluctuations in the data, and reveal the relationship between environmental parameters such as light intensity and temperature and photovoltaic output that change over time, thereby improving the accuracy of future output predictions, building a time series prediction model, using historical data to train the model, and learning the patterns and laws in the time series. The algorithm formula is: F(T+1) is the predicted value, X(i) is the actual value of the i-th period of the time series, and N is the number of averages. The trained model is used to make predictions based on the environmental parameters and features collected in real time. The prediction results include the photovoltaic output and load demand in the future. The mean square error is used to evaluate the accuracy of the prediction results. The model parameters are adjusted according to the evaluation results to optimize the prediction performance.
4. The intelligent and economical off-grid photovoltaic storage microgrid system according to claim 3, characterized in that: The time series prediction model is based on the mean square error MSE It is used to quantify the difference between the predicted value and the actual value, use an independent test data set to evaluate the prediction performance of the model, calculate the MSE between the predicted value and the actual value, and intuitively understand the prediction error size of the model. The algorithm formula is: y i is the actual observed value, is the model estimate, n is the number of observations, and based on the MSE evaluation results, the model parameters are adjusted and a new model structure is selected to optimize the prediction performance.
5. The intelligent and economical off-grid photovoltaic storage microgrid system according to claim 1, characterized in that: The specific steps of dynamic load management are as follows: (1) Load data collection: The system collects real-time load data through smart meters and sensors, including power consumption and power usage pattern information of various loads; (2) Load forecasting: Based on historical load data, environmental factors, and user behavior patterns, machine learning technology is used to predict the load change trend in the future. (3) Strategy formulation: Based on the load forecast results, combined with the current photovoltaic output and energy storage status, formulate the charging and discharging strategy of the energy storage equipment and the distribution strategy of the photovoltaic output power; (4) Strategy execution: Automatically adjust the operating status of the photovoltaic array, the charging and discharging power of the energy storage device, and the distribution of the load to ensure the balance of system supply and demand and optimize the efficiency of energy use; (5) Real-time monitoring and adjustment: The system monitors load changes, PV output, and energy storage status in real time, and fine-tunes strategies based on real-time data to adapt to environmental changes and fluctuations in user demand.
6. The intelligent and economical off-grid photovoltaic storage microgrid system according to claim 5, characterized in that: The dynamic load management collects historical load data, including historical power consumption and power consumption patterns, through the support vector machine SVM in machine learning technology, and learns the relationship between historical load data and environmental factors, that is, f(x)=sign(ω·x+b), where x represents the eigenvector of the input sample, ω represents the normal vector of the hyperplane, and b represents the intercept of the hyperplane. The predicted value is calculated to obtain the predicted power load value in the future period. According to the load prediction result, combined with the current photovoltaic output and energy storage status, the charging and discharging strategy of the energy storage equipment and the distribution strategy of the photovoltaic output power are formulated. The system automatically adjusts the operating status of the photovoltaic array, the charging and discharging power of the energy storage equipment, and the distribution of the load to ensure the balance of supply and demand and optimize the energy utilization efficiency.
7. The intelligent and economical off-grid photovoltaic storage microgrid system according to claim 6, characterized in that: The specific steps of the multi-objective optimization scheduling are as follows: (1) Objective definition: Define the objective function of multi-objective optimization, which usually includes minimizing operating costs, maximizing energy utilization, maintaining system voltage and frequency stability, and improving user satisfaction; (2) Problem modeling and algorithm design: Convert the actual problem into a mathematical model, including decision variables such as energy storage: charging and discharging strategy, photovoltaic array inclination, load distribution plan, constraints: equipment capacity limitation, energy supply and demand balance, safe operation restrictions and objective function, and select and design a multi-objective optimization algorithm based on the characteristics of the problem; (3) Parameter setting and iterative optimization: Set parameters for the algorithm. Through the iterative process, the algorithm continuously generates a new generation of solutions and gradually approaches the optimal solution set. (4) Optimization and interpretation of solutions: Using the decision-making method of multi-objective optimization, evaluate the solution set output by the optimization algorithm, check whether all constraints are met, and select and synthesize the final scheduling strategy; (5) Strategy implementation and evaluation: Apply the optimized scheduling strategy to the actual system and verify the effectiveness of the strategy and the improvement of system performance through real-time monitoring and post-evaluation.
8. The intelligent and economical off-grid photovoltaic storage microgrid system according to claim 7, characterized in that:The multi-objective optimization scheduling realizes automatic adjustment of the inclination angle of the photovoltaic array, the charging and discharging strategy of the energy storage system and the load distribution through the deep reinforcement learning DQN control system. According to the real-time feedback and performance evaluation of the system, the parameters of the DQN algorithm are adjusted to optimize the overall performance of the system. The algorithm is used as follows: the system state s t , including multi-dimensional features of photovoltaic output, energy storage status, load demand, and external environmental parameters. These features are input into the DQN neural network for state representation, action a t Including the tilt adjustment of the photovoltaic array, the charging and discharging strategy of the energy storage system, and the intelligent distribution of the load, these actions constitute a discrete and continuous action space. In multi-objective optimization scheduling, the reward function needs to reflect multiple goals, including economic benefits, system stability, and resource utilization. Each dimension corresponds to an optimization goal, one dimension represents economic benefits, and the other dimension represents the stability of system voltage and fluctuation range. The algorithm formula is: Q(s t ,a t ) means taking action a in state s t The expected return, s t+1 is the next state. Based on the Q value learned by DQN, the system generates a multi-objective optimization strategy, which trades off between multiple objectives to find a balanced solution that meets all objective requirements. Strategy generation is achieved by selecting the action with the highest comprehensive Q value. According to the real-time feedback and performance evaluation of the system, the parameters and reward function of the DQN algorithm are further adjusted to optimize the overall performance of the system.
9. The intelligent and economical off-grid photovoltaic storage microgrid system according to claim 1, characterized in that: The specific steps of the fault diagnosis and self-healing control are as follows: (1) Data judgment: Real-time collection of operating data of the photovoltaic array, energy storage equipment and inverter in the photovoltaic microgrid system, and use of sensors to monitor key parameters, including current, voltage, temperature and frequency, to ensure the normal operation of the system; (2) Data analysis and fault detection: Process and analyze the collected data and determine whether there are potential faults and abnormal conditions based on preset thresholds and algorithms; (3) Fault location and isolation: Once a fault or abnormality is detected, the system quickly locates the location and cause of the fault, accurately analyzes and classifies the fault, and ensures that the fault is accurately isolated; (4) Start backup plans and self-healing mechanisms: Based on the type of fault, the system starts backup plans to ensure that the system can continue to operate stably. The self-healing mechanism includes automatically restarting the device, switching to the backup power supply, and adjusting the working parameters to restore the normal function of the system; (5) Recording and reporting: The time, location, cause, and handling process of the fault are recorded in the log for subsequent analysis and maintenance. Fault information is reported to users and administrators through the user interface and remote monitoring platform to ensure that they are aware of the system status in a timely manner.
10. The intelligent and economical off-grid photovoltaic storage microgrid system according to claim 1, characterized in that: The fault diagnosis and self-healing control uses a threshold judgment method to detect whether a fault occurs in state monitoring and fault detection, sets one or more thresholds to judge whether the system is in a normal state and a fault occurs, and sets a threshold for the output power of the photovoltaic panel. If the actual output power is lower than the threshold, it is considered that the photovoltaic panel has a fault. For the detected fault, the system takes corresponding self-healing measures according to the type and degree of the fault, including restarting the equipment, switching to the backup power supply, and adjusting the working parameters. In this process, the threshold judgment ensures that the system's recovery operation is carried out within a safe range.
Citation Information
Cited By
Energy management method of photovoltaic energy storage system
CN119231653A
Distributed photovoltaic power generation grid-connected anti-countercurrent control system and control method
CN119362451A
Self-adaptive power balance control method of wind-solar hybrid power generation system
CN119482745A
Grid-connected and off-grid switching method for hydrogen fuel cell grid-connected controller
CN119518939A
Method for on-off grid switching of hydrogen fuel cell grid tied controller
CN119518939B
