Intelligent monitoring and fault diagnosis system and method for photovoltaic power station

Abstract

The invention discloses an intelligent monitoring and fault diagnosis system and method for a photovoltaic power station, and relates to the technical field of photovoltaic power station control, and the system comprises a photovoltaic module sensor which is used for collecting the voltage, current and temperature data of a photovoltaic module; the inverter monitoring sensor is used for monitoring voltage, current and frequency data of the inverter; the data acquisition and transmission device is used for receiving data from the photovoltaic module sensor and the inverter monitoring sensor; and the central processing unit is used for carrying out multi-modal data fusion analysis on the voltage, current and temperature data of the photovoltaic module and the environment monitoring data, generating a multi-dimensional operation state diagram of the system, and carrying out probability evaluation on potential faults through a dynamic fault tree analysis algorithm. According to the invention, by integrating the multi-modal data fusion algorithm and the dynamic fault tree analysis model, potential faults in the photovoltaic power station can be accurately identified and positioned, the accuracy of fault detection is improved, and the downtime of the power station is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of photovoltaic power station control, and specifically relates to a photovoltaic power station intelligent monitoring and fault diagnosis system and method. Background Art

[0002] As the global demand for clean energy continues to grow, photovoltaic power plants have been widely used as an important form of renewable energy. However, the efficient operation of photovoltaic power plants depends on complex electrical equipment and components, which are easily affected by environmental factors, material aging, and changes in operating conditions, leading to performance degradation or failure. Therefore, how to monitor the operating status of photovoltaic power plants in real time, detect and locate faults in a timely manner, and take effective maintenance measures has become a key issue to ensure the long-term stable operation of photovoltaic power plants.

[0003] Traditional photovoltaic power station operation and maintenance usually rely on manual inspections and regular maintenance. However, this approach has many shortcomings. First, manual inspections cannot achieve real-time monitoring of power station equipment, which can easily lead to delays in the discovery and handling of faults, increasing power station downtime and maintenance costs. Secondly, regular maintenance methods often cannot accurately handle specific problems, resulting in a waste of maintenance resources, and cannot ensure effective prevention before faults occur. Finally, most existing monitoring systems rely on a single data source or simple rule judgments, lack the ability to comprehensively analyze multi-dimensional data, and have difficulty accurately identifying complex faults and potential problems.

[0004] In the existing technology, although some photovoltaic power stations have introduced automated monitoring and remote control systems, these systems still have limitations in data processing capabilities and intelligent analysis. For example, most existing fault detection systems rely on static models and cannot adapt to dynamically changing operating environments, resulting in low accuracy in fault location. In addition, although remote control systems can achieve a certain degree of operational automation, their flexibility and real-time performance still need to be improved, especially when faced with sudden faults, it is difficult to respond quickly and make optimization decisions. Summary of the invention

[0005] In view of the deficiencies of the prior art, the present invention provides a photovoltaic power station intelligent monitoring and fault diagnosis system and method, which solves the problem of poor flexibility and real-time performance in coping with faults in the prior art.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: In order to achieve the above object, the present invention adopts the following technical solution: Photovoltaic power station intelligent monitoring and fault diagnosis system, including: Photovoltaic module sensor, used to collect voltage, current and temperature data of photovoltaic modules; Inverter monitoring sensor, used to monitor the voltage, current and frequency data of the inverter; A data acquisition and transmission device for receiving data from photovoltaic module sensors and inverter monitoring sensors; The central processing unit is used to perform multi-modal data fusion analysis on the voltage, current and temperature data of the photovoltaic modules and the data of the inverter monitoring sensors, generate a multi-dimensional operating status diagram of the system, and perform probability assessment of potential faults through a dynamic fault tree analysis algorithm; The fault warning module is used to detect and identify the fault type in the system, and uses the multi-modal data fusion analysis results output by the central processing unit to perform preliminary fault detection through the support vector machine model; The fault location module is used to generate a dynamic fault tree model, calculate the probability of the fault path, locate the specific fault source, and cooperate with the central processing unit through deep data fusion, sharing of analysis results and support for dynamic decision-making; The remote control and operation module is used for remote monitoring and real-time data communication. It maintains real-time data communication with the central processing unit through WebSocket technology, and performs corresponding control operations according to the warning signals provided by the fault location module and the fault warning module. It provides rapid response and preliminary repair when a fault occurs, and continuously monitors data to support decision-making optimization of the maintenance scheduling and management modules. The maintenance scheduling and management module is used to generate maintenance tasks and optimize the scheduling paths of on-site maintenance personnel.

[0007] A further improvement of the present invention is that the photovoltaic module sensor includes a voltage sensor, a current sensor and a temperature sensor, the inverter monitoring sensor includes a voltage sensor, a current sensor and a frequency sensor, and the data acquisition and transmission device includes an analog-to-digital converter and an SPI bus or an I2C bus.

[0008] A further improvement of the present invention is that the central processing unit includes a multimodal data fusion algorithm processor, and the multimodal data fusion algorithm processor adopts an architecture combining a Bayesian network model and a deep learning model. The Bayesian network model is used to establish a joint probability distribution of each sensor data, and the deep learning model is used to perform classification and regression analysis on the fused data; The multimodal data fusion algorithm processor includes an adaptive weight adjustment unit, which is used to dynamically adjust the weight of each data source according to the real-time volatility of the input data. The adjustment algorithm is: ; in, For time Time-adjusted weights, is the learning rate, is the loss function, For weight gradient.

[0009] A further improvement of the present invention is that the output result of the multimodal data fusion analysis is: ; in, Indicates that the system is at time The overall status score, represents the weight of each input data source, Indicates Data sources at time The output value of The total number of data sources.

[0010] A further improvement of the present invention is that the fault warning module includes a fault detection unit and a fault type identification unit. The fault detection unit uses a support vector machine to perform preliminary fault detection based on the output result of the multimodal data fusion algorithm processor. The fault type identification unit uses a deep learning model to identify the fault type and determines the fault boundary by the following formula: ; in, is the Lagrange multiplier, is the classification label, is the kernel function, For bias.

[0011] A further improvement of the present invention is that the deep learning model of the fault type identification unit includes a convolutional neural network for extracting features from multimodal data and calculating the probability of each fault type through a Softmax layer.

[0012] A further improvement of the present invention is that the fault location module includes a fault tree generation unit and a fault probability calculation unit. The fault tree generation unit generates a dynamic fault tree model by analyzing the time series of each sensor data. The fault probability calculation unit calculates the probability of each fault path by Monte Carlo simulation technology. The total probability calculation formula of the fault occurrence is as follows: ; in, is the total probability of system failure, For the The probability of a sub-failure event.

[0013] A further improvement of the present invention is that the remote control and operation module includes a remote monitoring unit and a remote control unit. The remote monitoring unit realizes real-time data communication with the central processing unit through WebSocket technology. The remote control unit includes an API interface for remotely adjusting the operating parameters of the inverter and the angle of the photovoltaic array.

[0014] A further improvement of the present invention is that the maintenance scheduling and management module includes a task generation unit and a personnel scheduling unit. The task generation unit automatically generates maintenance tasks based on the output results of the fault warning module and the fault location module. The personnel scheduling unit optimizes the scheduling path of the on-site maintenance personnel through a genetic algorithm. The fitness function of the genetic algorithm is: ; in, For the personnel scheduling process Click to The distance of the point, For the The estimated completion time of a task, n and m are the number of scheduling points and tasks respectively.

[0015] The photovoltaic power station intelligent monitoring and fault diagnosis method comprises the following steps: Use PV module sensors to collect voltage, current and temperature data of each PV module, and use inverter monitoring sensors to monitor the voltage, current and frequency data of the inverter; After the data is transmitted to the central processing unit, the system conducts a comprehensive analysis of the data of the photovoltaic modules and the environmental monitoring data through a multimodal data fusion algorithm; the central processing unit combines the Bayesian network and the deep learning model to classify and regress the data, and dynamically adjusts the weight of each data source. Then the system uses the support vector machine model to perform preliminary fault detection and identify possible abnormal conditions. Next, the convolutional neural network is used to extract deeper features from the multimodal data and calculate the probability of various types of faults through the Softmax layer, thereby identifying the specific types of faults that may occur in the system and generating a multi-dimensional operating status diagram. After fault identification, the fault location module generates a dynamic fault tree model by analyzing the time series of each sensor data. The system uses this model to combine historical data and real-time data to locate the specific fault source, evaluate the probability of the fault, determine the severity of the fault and its possible impact range; The remote control and operation module ensures real-time communication of data for further processing through WebSocket technology and allows remote monitoring of the operating status of the photovoltaic power station. Technicians can remotely adjust the operating parameters of the inverter and the angle of the photovoltaic array through the API interface, thereby optimizing the operation of the system and reducing the impact of failures; Based on the results of fault warning and location, the system's maintenance scheduling and management module automatically generates maintenance tasks. The system uses genetic algorithms to optimize the scheduling paths of on-site maintenance personnel. The generated tasks include giving priority to tasks in high-risk areas and reasonably arranging the work order of on-site technicians. According to the optimized scheduling plan, on-site maintenance personnel perform troubleshooting and repairs according to the generated tasks. The system uploads the maintenance results through a real-time feedback mechanism, updates the system status and maintenance records, ensures that the PV power station can resume stable operation after the fault is repaired, and accumulates data for future diagnosis and optimization.

[0016] Compared with the prior art, the present invention has at least the following beneficial technical effects: By integrating multimodal data fusion algorithms and dynamic fault tree analysis models, the present invention can accurately identify and locate potential faults in photovoltaic power plants, significantly improving the accuracy of fault detection. At the same time, the system can update the fault analysis model in real time and quickly calculate the probability of fault occurrence, allowing operation and maintenance personnel to respond to problems more promptly, reducing power plant downtime and reducing operational risks.

[0017] The present invention automatically generates maintenance tasks and optimizes the dispatch path of on-site personnel through the collaborative work of the task generation unit and the personnel dispatch unit through the optimization algorithm. This not only improves the efficiency of maintenance work, but also dynamically adjusts the dispatch strategy according to the fault situation, ensures the reasonable allocation and optimal use of maintenance resources, thereby reducing maintenance costs and improving the overall operation efficiency of the power station.

[0018] The present invention uses a remote control and operation module, which enables operators to monitor the operating status of the photovoltaic power station in real time while away from the site, and to make flexible remote adjustments through the API interface. Whether it is the parameter setting of the inverter or the angle adjustment of the photovoltaic array, it can be completed through remote operation, thereby greatly reducing the need for on-site operation and improving the flexibility of system management and the ability to respond to emergencies. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic diagram of the system framework of the present invention; Figure 2 It is a schematic diagram of the photovoltaic module sensor framework of the present invention; Figure 3 This is a schematic diagram of the inverter monitoring sensor framework of the present invention; Figure 4 It is a schematic diagram of the central processing unit framework of the present invention; Figure 5 This is a schematic diagram of the fault warning module framework of the present invention; Figure 6 This is a schematic diagram of the fault location module framework of the present invention; Figure 7 This is a schematic diagram of the maintenance scheduling and management module framework of the present invention; Figure 8 It is a schematic diagram of the process steps of the present invention. DETAILED DESCRIPTION

[0020] In the following, only some exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the present invention. Therefore, the drawings and descriptions are considered to be exemplary and non-restrictive in nature.

[0021] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0022] It should also be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0023] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0024] Various structural schematic diagrams of the embodiments disclosed in the present invention are shown in the accompanying drawings. These figures are not drawn to scale, and some details are magnified and some details may be omitted for the purpose of clear expression. The shapes of various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are only exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.

[0025] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0026] Embodiment 1: The embodiment of the present invention provides a photovoltaic power station intelligent monitoring and fault diagnosis system, including: Photovoltaic module sensor, used to collect voltage, current and temperature data of photovoltaic modules; Inverter monitoring sensor, used to monitor the voltage, current and frequency data of the inverter; A data acquisition and transmission device for receiving data from photovoltaic module sensors and inverter monitoring sensors; The central processing unit is used to perform multi-modal data fusion analysis on the voltage, current and temperature data of the photovoltaic modules and the data of the inverter monitoring sensors, generate a multi-dimensional operating status diagram of the system, and perform probability assessment of potential faults through a dynamic fault tree analysis algorithm; The fault warning module is used to detect and identify the fault type in the system, and uses the multi-modal data fusion analysis results output by the central processing unit to perform preliminary fault detection through the support vector machine model; The fault location module is used to generate a dynamic fault tree model, calculate the probability of the fault path, locate the specific fault source, and cooperate with the central processing unit through deep data fusion, sharing of analysis results and support for dynamic decision-making; The remote control and operation module is used for remote monitoring and real-time data communication. It maintains real-time data communication with the central processing unit through WebSocket technology, and performs corresponding control operations according to the warning signals provided by the fault location module and the fault warning module. It provides rapid response and preliminary repair when a fault occurs, and continuously monitors data to support decision-making optimization of the maintenance scheduling and management modules. The maintenance scheduling and management module is used to generate maintenance tasks and optimize the scheduling paths of on-site maintenance personnel.

[0027] Embodiment 2: The embodiment of the present invention provides a photovoltaic power station intelligent monitoring and fault diagnosis system, including: Please see attached Figure 1 -Attached Figure 2 , photovoltaic module sensor, used to collect voltage, current and temperature data of photovoltaic modules; including voltage sensor, current sensor and temperature sensor; Specifically, it is used to accurately collect the voltage, current and temperature data of photovoltaic modules to support the operation of the overall monitoring and diagnosis system of photovoltaic power plants. This module includes three core sensing devices: voltage sensor, current sensor and temperature sensor. Each sensor is responsible for data collection in different dimensions. The voltage sensor is used to measure the output voltage of each photovoltaic module in real time to ensure that the system can grasp the working status of the module and detect possible voltage anomalies in time. The current sensor is used to monitor the output current of the photovoltaic module, and combines the voltage data to calculate the power output and detect potential current instability problems.

[0028] Temperature sensors are used to monitor the surface temperature of photovoltaic modules, help evaluate the impact of temperature on power station efficiency, and prevent module aging or failure due to overheating. These three sensors are connected to the data acquisition and transmission device through a standardized data interface, usually using SPI or I2C bus to achieve high-speed and stable data transmission. After preliminary processing locally, the collected data is transmitted to the central processing unit in real time. In cooperation with other modules, these data are not only the basic input for the system's multimodal data fusion analysis, but also provide the necessary parameters for the fault warning module, ensuring the accuracy of fault detection and positioning. Through fusion analysis with environmental monitoring data, the data of photovoltaic module sensors can support the generation of a multi-dimensional operating status diagram of the system, providing a key basis for overall fault diagnosis and system optimization.

[0029] Please see attached Figure 3 ,Inverter monitoring sensors, used to monitor the voltage, current and frequency data of the inverter, including voltage sensors, current sensors and frequency sensors; Specifically, the module consists of three parts: voltage sensor, current sensor and frequency sensor, which are responsible for different types of key parameter monitoring. The voltage sensor is used to accurately measure the inverter output voltage to help identify voltage fluctuations or instability, thereby preventing possible losses or equipment damage during power transmission. The current sensor is responsible for monitoring the inverter output current to ensure the continuity and stability of the current, and at the same time, combined with the voltage data, to provide a complete power output analysis. The frequency sensor is used to measure the output frequency of the inverter to ensure that the power output meets the grid frequency standards and avoid system incompatibility or power quality problems caused by frequency deviations.

[0030] These sensors are connected to the data acquisition and transmission device through the standard Modbus RTU or Modbus TCP protocol to achieve efficient transmission of real-time data. The collected voltage, current and frequency data will be transmitted to the central processing unit in real time as an important input for multimodal data fusion analysis. In collaboration with other modules, the data provided by the inverter monitoring sensor not only directly affects the real-time evaluation of the system's operating status, but also provides basic data support for the fault warning and fault location modules, enabling the system to accurately identify and locate possible faults in the inverter. In addition, these data are also used in the remote control and operation module to help technicians remotely debug and adjust parameters of the inverter to ensure the overall stability and efficient operation of the system. Through the monitoring of these precision sensors, the system can adjust the operating strategy in real time to respond to potential anomalies or optimize power output, ultimately ensuring the continued efficient operation of the photovoltaic power station.

[0031] Please see attached Figure 1, a data acquisition and transmission device for receiving data from a photovoltaic module sensor and an inverter monitoring sensor, including an analog-to-digital converter and an SPI bus or an I2C bus; Specifically, the main function of the analog-to-digital converter is to convert the analog signals collected by the sensor into digital signals so that the central processing unit can directly read and process these data. Since the operating environment of photovoltaic panels and inverters may have a wide range of voltage and current changes, the accuracy and response speed of the analog-to-digital converter are crucial to ensure the accuracy and real-time nature of the data.

[0032] In terms of data transmission, the data acquisition and transmission device is connected to each sensor module through the SPI bus or the I2C bus. The SPI bus is usually used in scenarios that require high-speed data transmission. It can transmit a large amount of data in a short time and is suitable for high-frequency data acquisition needs. The I2C bus provides a simpler multi-master and multi-slave communication mechanism, which is suitable for application scenarios with a large number of sensors and high data transmission requirements. The choice of the two bus methods depends on the specific system requirements and design.

[0033] In terms of connection and working coordination with other modules, the data acquisition and transmission device acts as a bridge between the sensor module and the central processing unit. It is not only responsible for the aggregation and transmission of sensor data, but also plays a role in preliminary data filtering to avoid the impact of noise and outliers on subsequent data processing. After these processed data are transmitted to the central processing unit, they will become the basic input for multimodal data fusion analysis, thereby supporting the system's fault warning, fault location, and remote control functions. In addition, the device also provides real-time data support for remote monitoring, so that the operating status of the photovoltaic power station can be continuously monitored and adjusted in time to ensure the efficiency and reliability of the system.

[0034] Please refer to the attached Figure 4 The central processing unit includes a multimodal data fusion algorithm processor, which adopts an architecture that combines a Bayesian network model and a deep learning model. The Bayesian network model is used to establish a joint probability distribution of each sensor data, and the deep learning model is used to perform classification and regression analysis on the fused data.

[0035] It is used to perform multi-modal data fusion analysis on the voltage, current, temperature data of photovoltaic modules and environmental monitoring data, generate a multi-dimensional operating status diagram of the system, and perform probability assessment of potential faults through a dynamic fault tree analysis algorithm. The output results of the multi-modal data fusion analysis are: ; in, Indicates that the system is at time The overall status score, represents the weight of each input data source, Indicates Data sources at time The output value of is the total number of data sources; The multimodal data fusion algorithm processor includes an adaptive weight adjustment unit, which is used to dynamically adjust the weight of each data source according to the real-time volatility of the input data. The adjustment algorithm is: ; in, For time Time-adjusted weights, is the learning rate, is the loss function, For weight The gradient of Specifically, the main role of the Bayesian network model in the central processing unit is to model and analyze the relationship between different data sources by establishing the joint probability distribution of each sensor data. This probability distribution can effectively handle the complex relationship between the voltage, current, temperature data of the photovoltaic module and the environmental monitoring data, providing a solid foundation for the overall status assessment of the system. Through the Bayesian network, the system can calculate the probability distribution of each data source under different operating conditions, thereby providing more accurate input for the subsequent deep learning model.

[0036] The deep learning model is used to perform deeper classification and regression analysis on the fused data. It combines the probability distribution of the Bayesian network output to predict and evaluate the overall state of the system and generate a multi-dimensional operating state diagram of the system. This state diagram intuitively reflects the real-time operating status of the photovoltaic power station and provides an important reference for fault warning and system optimization.

[0037] During the data fusion and analysis process, the multimodal data fusion algorithm processor also includes an adaptive weight adjustment unit. This unit is used to dynamically adjust the weights of different data sources according to the real-time volatility of each input data, thereby optimizing the overall analysis results. This weight adjustment mechanism enables the system to adaptively respond to changes in different sensor data to ensure the accuracy and real-time nature of the analysis results. Specifically, the adaptive weight adjustment unit will adjust the weights of each data source in real time according to changes in input data, so that data sources with higher importance gain greater influence in the analysis, thereby improving the system's fault prediction accuracy.

[0038] In the workflow of the entire system, the central processing unit integrates and analyzes data from photovoltaic modules, inverters, and environmental monitoring sensors. First, the data is transmitted to the central processing unit through analog-to-digital converters and bus systems. Then, the central processing unit uses Bayesian networks and deep learning models to perform multimodal fusion analysis on the data to generate a multidimensional image reflecting the operating status of the system. Finally, the system uses a dynamic fault tree analysis algorithm to perform a probability assessment of potential faults, determine which components may have problems, and provide corresponding early warning information.

[0039] The central processing unit works closely with the fault warning module and the fault location module to provide real-time data support and analysis results. This integration enables the entire system to identify potential faults at an early stage and prevent them by adjusting strategies in real time. At the same time, the central processing unit also provides data support for the remote control and operation module, allowing technicians to remotely monitor and adjust the system operation status to ensure the stability and efficiency of the photovoltaic power station.

[0040] Please refer to the attached Figure 5 , a fault warning module is used to detect and identify the fault type in the system; it includes a fault detection unit and a fault type identification unit. The fault detection unit uses a support vector machine to perform preliminary fault detection based on the output result of the multimodal data fusion algorithm. The fault type identification unit uses a deep learning model to identify the fault type and determines the fault boundary through the following formula: ; in, is the Lagrange multiplier, is the classification label, is the kernel function, is bias; The deep learning model of the fault type identification unit includes a convolutional neural network to extract features from multimodal data and calculate the probability of each fault type through a Softmax layer; Specifically, the fault detection unit uses the support vector machine (SVM) to perform preliminary fault detection by using the data processed by the multimodal data fusion algorithm. SVM is a supervised learning model widely used in classification problems and is suitable for processing high-dimensional data features. The fault detection unit uses the SVM model to determine whether the data point is in a normal operating state or has deviated from the normal range. Based on the output results of the multimodal data fusion algorithm, the unit inputs the real-time data provided by each sensor into the SVM and identifies the anomaly by determining the fault boundary. The SVM model uses kernel functions to map data to a higher-dimensional feature space, so that it can handle nonlinear problems. Under the synergistic effect of Lagrange multipliers, classification labels and kernel functions, the SVM model can efficiently detect possible early signs of faults in the system and generate preliminary fault alarm information.

[0041] After detecting a possible fault, the fault warning module passes the data to the fault type identification unit to further identify the specific fault type. This unit uses deep learning models, especially convolutional neural networks (CNNs), to extract deeper features from multimodal data. CNNs are good at processing data with spatial structures and can identify complex patterns and abnormal changes. In this system, CNN can not only extract important features from multimodal data, but also calculate the probability distribution of each fault type through its internal Softmax layer. The Softmax layer standardizes the probability of the fault type so that the output probability value can be directly used to determine the specific type of system fault.

[0042] The output results of the fault type identification unit will be used to further guide the system's fault location module and maintenance scheduling module. By clarifying the fault type, the system can more accurately locate the fault location and formulate corresponding maintenance strategies. This module is closely connected to the central processing unit and relies on the results of multimodal data fusion analysis to perform fault warnings and pass the identified fault information to other parts of the system to ensure the efficient and safe operation of the photovoltaic power station. Through this set of mechanisms, the fault warning module can not only detect potential problems in the system in a timely manner, but also provide strong data support for subsequent processing steps through accurate fault type identification.

[0043] Please see attached Figure 6 , a fault location module is used to generate a dynamic fault tree model, calculate the probability of the fault path, and locate the specific fault source; it includes a fault tree generation unit and a fault probability calculation unit. The fault tree generation unit generates a dynamic fault tree model by analyzing the time series of each sensor data. The fault probability calculation unit calculates the probability of each fault path through Monte Carlo simulation technology. The total probability of fault occurrence is calculated as follows: ; in, is the total probability of system failure, For the The probability of a sub-failure event; Specifically, the main task of the fault tree generation unit is to build a dynamic fault tree model by analyzing the data time series of each sensor in the system. The fault tree is a systematic analysis tool used to graphically represent the fault events in the system and the logical relationship between the events. In a photovoltaic power station, each sensor continuously monitors specific operating parameters such as voltage, current, temperature, etc. The fault tree generation unit collects the time series data of these sensors, identifies the correlations between different components, and maps these correlations into the fault tree. This dynamic fault tree model can reflect the operating status of the system in real time and dynamically update to reflect new data input or environmental changes. It not only describes the various basic events that may cause system failures, but also clarifies the logical relationship between these events, thereby providing a structured framework for subsequent fault analysis.

[0044] The fault probability calculation unit is responsible for quantitative analysis of the dynamic fault tree model constructed by the fault tree generation unit. Specifically, it uses Monte Carlo simulation technology to perform probability calculations on each possible fault path in the fault tree. Monte Carlo simulation is a method that uses random sampling and statistical methods to simulate and analyze complex systems, and is particularly suitable for handling system failure analysis with uncertainty and complexity. In a photovoltaic power station, the fault probability calculation unit calculates the probability of occurrence of each basic event and path in the fault tree through a large number of random sample simulations. Through these simulation results, the system can derive the probability of occurrence of each fault path and finally calculate the total probability of failure of the entire system. These calculation results can not only quantify the risk level of the system, but also clearly point out the key paths and nodes that are most likely to cause system failures, thereby helping operation and maintenance personnel to concentrate resources for targeted maintenance.

[0045] The fault location module works closely with the fault warning module and the central processing unit to analyze sensor data in real time and provide feedback. Through the combined action of the fault tree generation unit and the fault probability calculation unit, the module can quickly locate the source of the fault before or just after the fault occurs and provide a detailed fault report. These reports can be directly used in the maintenance scheduling and management module to develop an efficient maintenance plan. In addition, the fault location module also has the ability to learn and self-improve. Over time, it will continuously optimize the fault tree model based on historical data and new data to improve the accuracy and efficiency of fault location.

[0046] Please see attached Figure 1, a remote control and operation module, used for remote monitoring and real-time data communication, includes a remote monitoring unit and a remote control unit. The remote monitoring unit realizes real-time data communication with the central processing unit through WebSocket technology. The remote control unit includes an API interface for remotely adjusting the operating parameters of the inverter and the angle of the photovoltaic array.

[0047] Specifically, the remote monitoring unit establishes a real-time data communication connection with the central processing unit through WebSocket technology. WebSocket is a full-duplex communication protocol that allows instant and continuous data exchange between the server and the client, and is very suitable for application scenarios that require low latency and high real-time performance. In the monitoring of photovoltaic power plants, the remote monitoring unit uses WebSocket to maintain a connection with the central processing unit and continuously receives real-time data updates from the central processing unit, including the voltage, current, and temperature information of the photovoltaic modules, as well as the operating status of the inverter and environmental monitoring data. This real-time communication capability enables operators to obtain the latest operating status of the power station at any time. Whether in the control center or through mobile devices, they can intuitively view the health status and key parameters of the power station through the monitoring interface.

[0048] The remote control unit provides an API interface that allows operators to remotely adjust the operating parameters of the inverter and the angle of the photovoltaic array when an abnormality is found or the operation needs to be optimized. The design of the API interface makes the remote operation of the system highly flexible and scalable. Operators can send instructions through the API interface based on the real-time monitored operating data to adjust the output voltage and frequency of the inverter, or change the angle of the photovoltaic array to optimize light capture and power generation efficiency. For example, when the system detects that the inverter output is unstable, the operator can immediately adjust the setting parameters of the inverter through the API interface to avoid greater impact on power output. Similarly, when environmental monitoring data indicates that the light angle has changed, the operator can also adjust the angle of the photovoltaic array through the remote control unit to maximize power output.

[0049] The remote control and operation module works closely with the central processing unit and the fault warning module to ensure that all adjustment operations are based on real-time and accurate data analysis results. This real-time monitoring and control capability greatly improves the operating efficiency of the photovoltaic power station and its ability to respond to emergencies, allowing the power station to operate safely and efficiently without human supervision or with only a small number of on-site personnel.

[0050] Please refer to the attached Figure 7, a maintenance scheduling and management module is used to generate maintenance tasks and optimize the scheduling path of on-site maintenance personnel; the maintenance scheduling and management module includes a task generation unit and a personnel scheduling unit, the task generation unit automatically generates maintenance tasks based on the output results of the fault warning module and the fault location module, and the personnel scheduling unit optimizes the scheduling path of on-site maintenance personnel through a genetic algorithm, and the fitness function of the genetic algorithm is: ; in, For the personnel scheduling process Click to The distance of the point, For the The estimated completion time of a task, n and m are the number of scheduling points and tasks respectively.

[0051] Specifically, the task generation unit automatically generates specific maintenance tasks by monitoring the operating status of the system in real time and combining the output results of the fault warning module and the fault location module. When the system detects a potential fault or an abnormality that has already occurred, the task generation unit will automatically formulate a detailed maintenance task list based on information such as the type, location, severity, and scope of impact of the fault. These tasks may include operations such as component replacement, inverter debugging, and cable inspection. The task generation unit can not only generate maintenance tasks for a single fault, but also integrate multiple fault information to generate a series of prioritized tasks to optimize the efficiency of on-site maintenance. This process is automated and can ensure that maintenance tasks are generated as soon as possible after a fault occurs, and the task information is passed to the personnel scheduling unit for further processing.

[0052] The personnel scheduling unit is responsible for optimizing the scheduling path of on-site maintenance personnel based on the generated maintenance tasks using genetic algorithms. Genetic algorithms are intelligent optimization algorithms that simulate natural selection and genetic mechanisms, and are very suitable for solving complex scheduling and path optimization problems. The unit first obtains the current location information, task urgency, equipment status and other data of on-site maintenance personnel, and then calculates an optimal path through genetic algorithms to ensure that maintenance personnel can complete all tasks with the lowest movement cost and time. The fitness function of the genetic algorithm is designed to measure the pros and cons of each path, specifically considering the distance between each task point and the estimated completion time of each task. Through repeated iterations, the genetic algorithm can gradually optimize a path that minimizes the overall task completion time, thereby maximizing maintenance efficiency.

[0053] In actual operation, when the task generation unit creates a new maintenance task, the personnel scheduling unit will immediately start calculating the optimal path, and after the scheduling is completed, send task notifications and action plans to relevant maintenance personnel. Maintenance personnel will perform on-site operations according to the assigned task list and optimized path to ensure that each task can be completed efficiently and orderly. At the same time, the maintenance scheduling and management module will also monitor the progress of the task in real time. If a new fault is detected or an unexpected situation occurs during the maintenance process, the scheduling plan can be adjusted immediately and the path can be re-optimized to ensure that the maintenance work can proceed smoothly.

[0054] This module works closely with the central processing unit, fault warning module and fault location module to form a highly intelligent maintenance management system. Through automated task generation and intelligent personnel scheduling, the maintenance scheduling and management module not only greatly reduces the need for manual intervention, but also significantly improves the overall operational efficiency of the photovoltaic power station and reduces the economic losses caused by failures.

[0055] Embodiment 3: Please see attached Figure 8 The embodiment of the present invention provides a photovoltaic power station intelligent monitoring and fault diagnosis method, comprising the following steps: Use PV module sensors to collect voltage, current and temperature data of each PV module, and use inverter monitoring sensors to monitor the voltage, current and frequency data of the inverter; After the data is transmitted to the central processing unit, the system conducts a comprehensive analysis of the data of the photovoltaic modules and the environmental monitoring data through a multimodal data fusion algorithm; the central processing unit combines the Bayesian network and the deep learning model to classify and regress the data, and dynamically adjusts the weight of each data source. Then the system uses the support vector machine model to perform preliminary fault detection and identify possible abnormal conditions. Next, the convolutional neural network is used to extract deeper features from the multimodal data and calculate the probability of various types of faults through the Softmax layer, thereby identifying the specific types of faults that may occur in the system and generating a multi-dimensional operating status diagram. After fault identification, the fault location module generates a dynamic fault tree model by analyzing the time series of each sensor data. The system uses this model to combine historical data and real-time data to locate the specific fault source, evaluate the probability of the fault, determine the severity of the fault and its possible impact range; The remote control and operation module ensures real-time communication of data for further processing through WebSocket technology and allows remote monitoring of the operating status of the photovoltaic power station. Technicians can remotely adjust the operating parameters of the inverter and the angle of the photovoltaic array through the API interface, thereby optimizing the operation of the system and reducing the impact of failures; Based on the results of fault warning and location, the system's maintenance scheduling and management module automatically generates maintenance tasks. The system uses genetic algorithms to optimize the scheduling paths of on-site maintenance personnel. The generated tasks include giving priority to tasks in high-risk areas and reasonably arranging the work order of on-site technicians. According to the optimized scheduling plan, on-site maintenance personnel perform troubleshooting and repairs according to the generated tasks. The system uploads the maintenance results through a real-time feedback mechanism, updates the system status and maintenance records, ensures that the PV power station can resume stable operation after the fault is repaired, and accumulates data for future diagnosis and optimization.

[0056] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any figure mark in the claims should not be regarded as limiting the claims involved.

[0057] In addition, it should be understood that although this specification is described in accordance with the implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation modes that can be understood by those skilled in the art. The above content is only to illustrate the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution according to the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. Photovoltaic power station intelligent monitoring and fault diagnosis system, characterized by: include: Photovoltaic module sensor, used to collect voltage, current and temperature data of photovoltaic modules; Inverter monitoring sensor, used to monitor the voltage, current and frequency data of the inverter; A data acquisition and transmission device for receiving data from photovoltaic module sensors and inverter monitoring sensors; The central processing unit is used to perform multi-modal data fusion analysis on the voltage, current and temperature data of the photovoltaic modules and the data of the inverter monitoring sensors, generate a multi-dimensional operating status diagram of the system, and perform probability assessment of potential faults through a dynamic fault tree analysis algorithm; The fault warning module is used to detect and identify the fault type in the system, and uses the multi-modal data fusion analysis results output by the central processing unit to perform preliminary fault detection through the support vector machine model; The fault location module is used to generate a dynamic fault tree model, calculate the probability of the fault path, locate the specific fault source, and cooperate with the central processing unit through deep data fusion, sharing of analysis results and support for dynamic decision-making; The remote control and operation module is used for remote monitoring and real-time data communication. It maintains real-time data communication with the central processing unit through WebSocket technology, and performs corresponding control operations according to the warning signals provided by the fault location module and the fault warning module. It provides rapid response and preliminary repair when a fault occurs, and continuously monitors data to support decision-making optimization of the maintenance scheduling and management modules. The maintenance scheduling and management module is used to generate maintenance tasks and optimize the scheduling paths of on-site maintenance personnel.

2. The photovoltaic power station intelligent monitoring and fault diagnosis system according to claim 1 is characterized in that: Photovoltaic module sensors include voltage sensors, current sensors and temperature sensors, inverter monitoring sensors include voltage sensors, current sensors and frequency sensors, and data acquisition and transmission devices include analog-to-digital converters and SPI buses or I2C buses.

3. The photovoltaic power station intelligent monitoring and fault diagnosis system according to claim 1 is characterized in that: The central processing unit includes a multimodal data fusion algorithm processor, which adopts an architecture combining a Bayesian network model and a deep learning model. The Bayesian network model is used to establish the joint probability distribution of each sensor data, and the deep learning model is used to perform classification and regression analysis on the fused data. The multimodal data fusion algorithm processor includes an adaptive weight adjustment unit, which is used to dynamically adjust the weight of each data source according to the real-time volatility of the input data. The adjustment algorithm is: ； in, For time Time-adjusted weights, is the learning rate, is the loss function, For weight gradient.

4. The photovoltaic power station intelligent monitoring and fault diagnosis system according to claim 3 is characterized in that: The output results of multimodal data fusion analysis are: ； in, Indicates that the system is at time The overall status score, represents the weight of each input data source, Indicates Data sources at time The output value of The total number of data sources.

5. The photovoltaic power station intelligent monitoring and fault diagnosis system according to claim 3 is characterized in that: The fault warning module includes a fault detection unit and a fault type identification unit. The fault detection unit uses a support vector machine to perform preliminary fault detection based on the output results of the multimodal data fusion algorithm processor. The fault type identification unit uses a deep learning model to identify the fault type and determines the fault boundary using the following formula: ； in, is the Lagrange multiplier, is the classification label, is the kernel function, For bias.

6. The photovoltaic power station intelligent monitoring and fault diagnosis system according to claim 5, characterized in that: The deep learning model of the fault type identification unit includes a convolutional neural network to extract features from multimodal data and calculate the probability of each fault type through a Softmax layer.

7. The photovoltaic power station intelligent monitoring and fault diagnosis system according to claim 1, characterized in that: The fault location module includes a fault tree generation unit and a fault probability calculation unit. The fault tree generation unit generates a dynamic fault tree model by analyzing the time series of each sensor data. The fault probability calculation unit calculates the probability of each fault path through Monte Carlo simulation technology. The total probability calculation formula of the fault is as follows: ； in, is the total probability of system failure, For the The probability of a sub-failure event.

8. The photovoltaic power station intelligent monitoring and fault diagnosis system according to claim 1, characterized in that: The remote control and operation module includes a remote monitoring unit and a remote control unit. The remote monitoring unit realizes real-time data communication with the central processing unit through WebSocket technology. The remote control unit includes an API interface for remotely adjusting the operating parameters of the inverter and the angle of the photovoltaic array.

9. The photovoltaic power station intelligent monitoring and fault diagnosis system according to claim 1, characterized in that: The maintenance scheduling and management module includes a task generation unit and a personnel scheduling unit. The task generation unit automatically generates maintenance tasks based on the output results of the fault warning module and the fault location module. The personnel scheduling unit optimizes the scheduling path of on-site maintenance personnel through a genetic algorithm. The fitness function of the genetic algorithm is: ； in, For the personnel scheduling process Click to The distance of the point, For the The estimated completion time of a task, n and m are the number of scheduling points and tasks respectively.

10. Photovoltaic power station intelligent monitoring and fault diagnosis method, characterized in that: The method is based on the photovoltaic power station intelligent monitoring and fault diagnosis system according to any one of claims 1 to 9, and comprises the following steps: Use PV module sensors to collect voltage, current and temperature data of each PV module, and use inverter monitoring sensors to monitor the voltage, current and frequency data of the inverter; After the data is transmitted to the central processing unit, the system uses a multimodal data fusion algorithm to conduct a comprehensive analysis of the data of the photovoltaic modules and the environmental monitoring data. The central processing unit combines the Bayesian network and the deep learning model to classify and regress the data, and dynamically adjust the weights of each data source. Then the system uses the support vector machine model to perform preliminary fault detection and identify possible abnormal conditions. Next, the convolutional neural network is used to extract deeper features from the multimodal data and calculate the probability of various types of faults through the Softmax layer, thereby identifying the specific types of faults that may occur in the system and generating a multi-dimensional operating status diagram. After fault identification, the fault location module generates a dynamic fault tree model by analyzing the time series of each sensor data. The system uses this model to combine historical data and real-time data to locate the specific fault source, evaluate the probability of the fault, determine the severity of the fault and its possible impact range; The remote control and operation module ensures real-time communication of data for further processing through WebSocket technology and allows remote monitoring of the operating status of the photovoltaic power station. Technicians can remotely adjust the operating parameters of the inverter and the angle of the photovoltaic array through the API interface, thereby optimizing the operation of the system and reducing the impact of failures; Based on the results of fault warning and location, the system's maintenance scheduling and management module automatically generates maintenance tasks. The system uses genetic algorithms to optimize the scheduling paths of on-site maintenance personnel. The generated tasks include giving priority to tasks in high-risk areas and reasonably arranging the work order of on-site technicians. According to the optimized scheduling plan, on-site maintenance personnel perform troubleshooting and repairs according to the generated tasks. The system uploads the maintenance results through a real-time feedback mechanism, updates the system status and maintenance records, ensures that the PV power station can resume stable operation after the fault is repaired, and accumulates data for future diagnosis and optimization.

Images