AI-Based Fault Detection Systems for Solar Array Monitoring

Diagnosing failures in solar cell modules using a control terminal that interacts with the power generation control device. The method involves continuously stepping the output command from 100% to 0% to acquire a partial I-V curve from the maximum power point to open circuit voltage. By comparing this partial curve to normal or known curves, failures can be diagnosed based on shape differences. This allows diagnosing failures during system operation without disconnecting.

Source

Diagnostic technique for estimating local degradation of series resistance in solar cells. It involves comparing series resistance values when the solar cell limits output due to low light conditions versus when it tracks the maximum power point. If the resistance increases more in low light, it indicates local resistance degradation, while a uniform increase indicates total resistance degradation. This allows distinguishing between localized resistance failures versus general resistance degradation in solar cells.

Source

A method for monitoring solar panel performance that separates short-term and long-term variations. The method uses a trained model to simulate average panel performance based on environmental parameters, and compares this simulated performance to actual panel output. By analyzing the difference between simulated and actual performance over time, the method can detect long-term degradation such as soiling or UV damage, and identify maintenance or weather-related events. The method can be applied to photovoltaic panels, thermal solar panels, and hybrid solar panels.

Source

Detecting solar panel system malfunctions proactively using statistical outliers. The method involves identifying solar panels that have already been flagged as malfunctioning. For those panels, it calculates estimated generation versus actual generation over time. If an accuracy ratio for a time interval is a statistical outlier compared to previous ratios, it flags that time interval as potentially indicating further malfunction. This allows proactive detection of worsening panel issues.

Source

Diagnostic system for measuring string health in solar power generation systems, comprising a control device that selectively disconnects each string from the inverter and measures the resulting power output, comparing it to a baseline measurement to determine string health and detect abnormalities.

Source

A power generation system, inverter, and short-circuit detection method for a negative branch of an inverter. The system includes a plurality of photovoltaic modules, an inverter, and a transformer, with current sensors on some or all of the negative branches. The inverter controller detects short-circuit faults on the negative branches by monitoring current values and sends fault information to facilitate repair or replacement of the faulty branch.

Source

Photovoltaic system with intelligent current scanning for improved fault diagnosis accuracy and system efficiency. The system uses a central management device to coordinate scanning of current levels in photovoltaic strings. It analyzes factors like irradiance and fluctuations to determine when to initiate scanning. This avoids frequent scans during stable conditions, reducing energy loss and power fluctuations. The central device also schedules scanning times to prevent scanning during shaded periods. This improves curve accuracy versus instant scans. The scheduled scanning can be initiated by a user.

Source

A system and method for determining degradation loss in solar panels of photovoltaic power plants using machine learning. The system obtains location, weather, and configuration data for the panels, as well as real-time operating parameters from sensors. This data is input into a pre-trained ML model to determine degradation loss for each panel, with alerts generated when losses exceed a threshold. The system can also validate sensor data and incorporate reference information from commissioning records.

Source

A photovoltaic array fault detection method that uses a signal transmission system and a filtering system to detect faults in photovoltaic arrays. The method involves sending a step signal through the array, capturing reflection signals from each module, and processing them through a filtering system to obtain target waveforms. The processed waveforms are then compared to a library of known waveforms to identify faults. The filtering system includes two high-pass filters with adjustable cutoff frequencies, which are optimized based on signal-to-noise ratio and waveform differences. The method also includes fault detection and repair procedures for the signal transmission devices and filtering system.

Source

A monitoring system for photovoltaic systems that identifies the faulty component causing reduced performance without requiring field intervention. The system comprises electronic control means that analyze data from the photovoltaic panel, inverter, and optimizer to pinpoint the source of the degradation. This enables rapid diagnosis and repair, reducing downtime and maintenance costs.

Source

A photovoltaic (PV) energy generation system with improved fault current monitoring for large-scale systems. The system comprises a power converter unit connected to a PV generator with multiple parallel-connected PV main strings. Each string pair is monitored by a differential current measuring device, circuit breaker, and controller that activates the breaker after detecting a fault exceeding a threshold value. The system features defined leakage capacitances on the DC intermediate circuit to enable ground current measurement and fault detection, even in insulated DC links.

Source

A photovoltaic power generation system and fault protection method, device, combiner box, and inverter that reduces the number of protection switches without increasing their withstand voltage level. The system connects multiple photovoltaic strings to power conversion units through fault isolation circuits, with protection switches shared between strings when power conversion units have common positive or negative connections. The control unit disconnects protection switches in response to faults, ensuring at most two strings per unit remain connected in parallel.

Source

A photovoltaic power generation system with improved fault protection, comprising a power conversion unit, a multi-pole linked switch, and N photovoltaic strings. The system includes a control portion that, in response to a fault, short-circuits the power conversion unit's positive and negative input terminals and disconnects the multi-pole linked switch, limiting parallel connections to at most three photovoltaic strings. The system also includes a parameter detection portion that monitors current and voltage parameters to detect faults.

Source

A method and system for detecting anomalies in solar power generation using a combination of inverter input/output data and hourly solar irradiance data. The system estimates expected power generation based on solar irradiance data and compares it to actual power generation based on inverter data to identify discrepancies. It performs primary and secondary verifications to determine the presence of anomalies, and further analyzes inverter and solar panel performance to diagnose specific issues.

Source

A device and method for detecting abnormalities in solar power generators based on power generation data. The device analyzes power output patterns across multiple strings to identify anomalies, using statistical techniques such as Cook's distance and leverage to determine if a string's output deviates significantly from expected behavior. The method can be applied to both string-level and inverter-level monitoring, enabling real-time detection of faults in solar power systems.

Source

Method for detecting malfunctioning MLPE devices in a photovoltaic system, comprising: controlling an MLPE device to switch its operation state; measuring the bus feature before and after the state switch; determining if the bus feature change meets a predetermined threshold; and identifying the MLPE device as faulty if the change does not meet the threshold. The method enables automated detection of faulty MLPE devices in a photovoltaic system, reducing manual troubleshooting time and costs.

Source

A photovoltaic cell detection method and apparatus that detects photovoltaic cells in a running photovoltaic system without disrupting power output. The method determines a working voltage for the cell, where the cell's output power remains stable at both voltages, and captures infrared images at both voltages to detect faults. The apparatus includes a processor and memory that execute the method, and a DC/DC conversion module that provides the working voltage.

Source

A method for determining the operating state of a photovoltaic array based on real-time output current values of photovoltaic strings. The method involves acquiring present output current values, determining a target abnormality score predicting model, and comparing the predicted abnormality score with a threshold to determine the operating state of the array. The method improves accuracy by filtering out virtually high current values and using a machine learning-based abnormality score predicting model.

Source

A photovoltaic module self-test method for string-type household solar panels, comprising determining an abnormal critical value based on the maximum power of middle three groups of panels, detecting panel abnormalities against this threshold, and recording abnormal counts. The method also includes a feature to clear previous detection data when a threshold number of exceptions is reached, to prevent false positives from temporary shading events.

Source

A method for detecting underperformance in photovoltaic (PV) plants using only electrical measurements, excluding irradiance data. The method determines performance deviation of a PV configuration from normalized power generation data of multiple similar configurations, and detects underperformance by comparing the deviation metric with a threshold. The method also determines long-term degradation of a PV configuration by analyzing trends in production data.

Source

A photovoltaic panel fault diagnosis method and apparatus that uses a simulation model to determine normal operating parameters, such as maximum power voltage and current, and then compares these values to real-time measurements to diagnose faults in the system. The method eliminates the need for a reference module and I-V curve measuring device, reducing costs and measurement errors.

Source

An operation and maintenance management platform for solar power generation systems that uses digital twin technology to predict and respond to system failures. The platform creates virtual replicas of existing solar power systems, monitors their performance, and compares it to actual system data to identify potential failures. When a failure is predicted, the platform notifies registered system managers, who can then respond quickly to prevent downtime. The platform also provides real-time monitoring and management capabilities, enabling users to check system status and performance at any time.

Source

A device and method for detecting fault currents in photovoltaic installations, particularly in transformerless inverters, where the fault current is a small resistive current amidst high capacitive leakage currents. The device uses a single current transformer to measure the difference between the summation currents of two parallel PV strings, which is small due to their similar leakage currents. This difference measurement enables reliable detection of fault currents, even at low intensities, without requiring separate measurement of the high leakage currents.

Source

A photovoltaic inverter system with automatic RSD location and fault control. The system determines RSD locations by selectively turning on RSDs and measuring voltage changes across each RSD. When an RSD fails, the system isolates the affected string while maintaining normal operation of other strings. The system's controller coordinates RSD control and fault detection across multiple inverters.

Source

A photovoltaic power generation system that protects against short-circuit faults in DC-DC converters. The system includes a protection switch with a release mechanism that disconnects the switch when a fault is detected, preventing damage to the converter and line. The release mechanism can be an electromagnetic or thermal release, and is designed to prevent triggering of external protective devices. The system can be applied to photovoltaic inverters and direct current combiner boxes, and is particularly suitable for systems with multiple DC-DC converters connected in parallel.

Source

Fault detection system for photovoltaic modules and strings that uses machine learning to analyze the performance of photovoltaic arrays. The system measures array performance from both the array and module perspectives, calculates a performance metric based on environmental conditions, and identifies faults by comparing measured performance against estimated performance. The system employs an AI algorithm to analyze the characteristic curves of both array and module levels, enabling accurate fault detection even when the array is composed of multiple modules.

Source

A photovoltaic panel system with automated tripping and restoration capabilities that enables real-time monitoring and intervention of malfunctioning solar panels. The system comprises a network of sensors that transmit electrical data from each panel, a central control unit that receives and processes this data, and tripping/restoration devices that automatically intervene based on panel performance metrics. The system incorporates intelligent cleaning mechanisms, including a roll-up/roll-down cover that covers the photovoltaic surface while the panel is being monitored, and a brush-based cleaning system that removes debris from the panel's surface. The system achieves rapid intervention and restoration through its automated control architecture, eliminating the need for site personnel and reducing maintenance downtime.

Source

Photovoltaic system for efficient power measurement and management through a single monitoring device. The system integrates a single monitoring device that measures both module voltage and current from individual solar cells in series strings, eliminating the need for separate voltage and current measurement devices. The device transmits measured data to a current measuring device connected to the solar cells, enabling direct power calculation and total system monitoring. This approach simplifies system architecture while maintaining accurate power measurement capabilities.

Source

A fault protection apparatus for photovoltaic power generation systems that prevents overvoltage damage to half-bus capacitors during inverter bridge arm faults. The apparatus includes a primary circuit breaker with two switch transistors connected in series between the capacitor bridge arm's reference output port and the inverter bridge arm's reference input port. The circuit breaker is controlled to turn on and off based on voltage or current deviations between the reference ports, allowing the connection between the ports to be dynamically adjusted to prevent capacitor overvoltage.

Source

A deterioration diagnosis system for solar cell strings and photovoltaic power systems that enables precise location of deteriorated solar cells within the string. The system measures the cumulative resistance of each solar cell in the string over time, calculating a normalized resistance value for each cell. By comparing the resistance values to a predetermined threshold, the system identifies cells that have exceeded this threshold, indicating deterioration. This approach eliminates the need for manual inspection of individual solar cells, significantly reducing maintenance time and labor costs.

Source

Abnormality determination system for solar cell modules that detects defects in the parallel resistance component of connected solar cells by measuring changes in output current when modulated light is applied under different illumination conditions. The system controls light irradiation to two states, one with reduced and one with increased incident light, and calculates a ratio of phase detection outputs to determine if the parallel resistance component is normal or defective.

Source

A power generation state determination device for photovoltaic systems that accurately determines system performance by grouping panels based on output measurements, calculating a reference value from a selected group, and comparing it to the output of the target group.

Source

Method for detecting defects in solar cells of photovoltaic plants through infrared imaging. The method uses two infrared images captured under different operating conditions - one capturing ambient reflections when the solar cell is in a short-circuit state and another capturing emitted radiation when the solar cell operates normally. The images are combined to create a third image that suppresses ambient noise while highlighting defects. This approach eliminates the need for artificial lighting and enables comprehensive solar cell detection through natural illumination.

Source

Apparatus for diagnosing photovoltaic devices that improves accuracy of state diagnosis by comparing power generation differences between grouped devices through analysis of historical power generation trends, rather than comparing predicted and actual power generation. The apparatus processes power generation data from photovoltaic devices based on failure history, maintenance records, and environmental factors to enable precise grouping and minimize error. It also provides a power balancing device to optimize power generation from abnormal devices before maintenance.

Source

Fault detection method, device, and photovoltaic power generation system that automates photovoltaic module inspection through coordinated control of inverters and image acquisition terminals. The system controls photovoltaic modules to a target state through inverters, then captures images using mobile terminals to detect faults. The process is repeated in a predetermined sequence, enabling efficient and automated fault detection across multiple modules.

Source

A protection apparatus for photovoltaic power generation systems that prevents short circuits and overcurrents by monitoring branch currents and disconnecting faulty units. The apparatus connects multiple photovoltaic units in parallel and includes a protection switch that can isolate up to three units from the DC bus. When a fault is detected, the controller opens the protection switch to prevent damage to the system. The apparatus eliminates the need for fuses, reducing power loss and cable costs.

Source

A method for determining detection conditions for photovoltaic power generation systems that eliminates the need for dedicated environmental monitoring equipment. The method uses real-time monitoring of working parameters, such as irradiance and current, to determine when environmental conditions are suitable for detection. When conditions are met, the system initiates detection cycles, monitoring the target unit's performance over a predetermined time period. The detection results are then analyzed to determine the unit's operational status.

Source

A method and system for fault detection, diagnosis, and localization in solar panel networks. The system generates a network simulation model based on the solar panel network's topology and parameters, and then uses a convolutional neural network (CNN) to classify fault scenarios from simulated datasets. The CNN is trained on a combination of fault and no-fault simulated datasets, allowing it to accurately detect and diagnose faults in the solar panel network.

Source

A machine-learning-based photovoltaic power generation control system that collects real-time voltage and power data from photovoltaic modules, learns the data through a machine learning platform, and controls the modules based on extracted control information. The system analyzes connected devices in real-time to optimize power generation, and enables modeling of various service functions such as device recognition, protocol modeling, and power forecasting.

Source

Fault detection method for photovoltaic modules using image analysis. The method captures images of the module during operation, specifically when it emits light, and analyzes the captured images to detect faults. When the image quality is compromised due to environmental factors, the method automatically adjusts the power supply parameters to restore optimal image quality. This approach enables continuous monitoring of module performance while maintaining image quality, thereby improving overall system reliability.

Source

A solar panel array communication system that enables bi-directional communication between solar panels and a central converter. The system uses toroidal transformers to inject and sense signals across the panel bridges, allowing for monitoring of panel performance, temperature, current, and voltage. The system also enables panel isolation and replacement through switch control.

Source

Automated fault detection and classification in utility-scale solar arrays through machine learning-based neural networks. The approach leverages a combination of feature extraction and neural network architecture to accurately identify and classify faults in solar panels. The feature set includes key parameters such as open circuit voltage, short circuit current, temperature, irradiance, and power output, which are processed through a neural network trained using dropout regularization. The trained network enables the system to classify faults into specific categories like electrical faults, shading-related issues, and soiling conditions, providing real-time monitoring and control capabilities.

Source

Fault classification for photovoltaic arrays using graph signal processing. The method employs graph theory to develop a classification system that automatically identifies PV array faults beyond traditional diagnostic capabilities. It analyzes data from the PV array's power output, irradiance, and environmental conditions to classify faults into four distinct categories: standard test conditions, shading, degraded modules, and short circuits. The classification is achieved through a graph-based approach that leverages the properties of graphs to represent the PV array's behavior, enabling accurate fault detection and diagnosis.

Source

System for detecting and diagnosing performance issues in solar inverters across multiple solar power plants. The system receives data from inverters via a network, compares actual output to expected output, and applies machine learning models to identify underperformance, classify causes, and provide alerts to operators.

Source

A photovoltaic (PV) sensor for early detection of potential induced degradation (PID) in PV systems. The sensor is specifically designed to monitor PV modules at the negative voltage end of a string, where degradation is most pronounced. The sensor operates with low power currents (typically below 10 mA) and measures shunt resistance within the PV module. It isolates the PV module during interrogation to maintain system integrity. The sensor provides real-time monitoring of shunt resistance and PV module power loss, enabling early detection of PID before significant degradation occurs.

Source

A method for controlling photovoltaic power generation that enables rapid adjustment of output voltage reference values in response to blocking or component failure. The method involves a control device periodically sending output voltage reference value update instructions to optimizers in a photovoltaic string, which automatically update their reference values based on an adjustment coefficient. This enables optimizers to rapidly transition between maximum power point tracking and power limiting states, minimizing power loss during system faults.

Source

Inverter short-circuit detection method, device, and inverter for photovoltaic systems. The method detects short-circuits in boost circuits of an inverter by monitoring overcurrent protection states when the inverter is in a non-operating state. When a boost circuit's overcurrent protector is disconnected, the method identifies the short-circuited circuit and takes protective measures to prevent damage to the photovoltaic string connected to it. The inverter device includes multiple boost circuits, a common bus, and a DC-AC inverter module, with each boost circuit's switch tube connected in series with an overcurrent protector.

Source

A method for monitoring a solar panel power generation system with multiple modules connected in parallel, enabling prediction of individual module power generation states using a single voltage/current meter. The method employs a factorial hidden Markov model (FHMM) to analyze hourly operation states of each module series, aided by weather information, and compares predicted values against historical data to identify abnormal modules.

Source

A photovoltaic array fault diagnosis method based on composite information, comprising: collecting and preprocessing composite information data of photovoltaic array working state, including image data and text data; training a deep convolutional neural network model using the image data to obtain an image fault classification model; training a support vector machine model using the text data to obtain a text fault classification model; and fusing the image fault classification model and the text fault classification model to obtain a photovoltaic array fault diagnosis model based on composite information.

Source

A photovoltaic panel fault diagnosis system based on infrared imaging technology for solar power plants. The system uses infrared cameras to detect temperature anomalies in photovoltaic panels, automatically identifying faulty panels through image recognition. The system connects multiple cameras via a network interface, enabling remote monitoring and real-time fault detection.

Source