Photovoltaic inverter IGBT module linkage defect early warning and inspection triggering method and device

Abstract

The application discloses a photovoltaic inverter IGBT module linkage defect early warning and inspection triggering method and device, the method comprising: synchronously collecting temperature field data and electrical parameters on the surface of a photovoltaic inverter IGBT module; after preprocessing the collected data, key features are extracted, the correlation degree index of temperature and electrical parameters is obtained, and the index is input into a pre-trained deep neural network model to obtain a defect probability evaluation result and a defect type; the target early warning level is determined according to the defect probability evaluation result and the defect type; in the case that the target early warning level meets preset conditions, an inspection task work order is generated; based on weather conditions, equipment operating states, inspection resources and the inspection task work order, an inspection time window and a path are planned, and an inspection instruction is generated, the inspection instruction is sent to a power station inspection system to schedule a drone or an inspection personnel to perform inspection, early fault features of the IGBT module can be captured in time, and fault misjudgment or misjudgment is avoided.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power generation equipment condition monitoring and intelligent operation and maintenance technology, and in particular to a method and device for triggering early warning and inspection of linked defects in photovoltaic inverter IGBT (Insulated Gate Bipolar Transistor) modules. Background Technology

[0002] As the core equipment of a photovoltaic power generation system, the reliability of the IGBT module in the photovoltaic inverter directly affects the power generation efficiency and operational safety of the entire system. In actual operation, due to the long-term exposure to high-frequency switching operations and high-current loads, IGBT modules are prone to defects such as overheating, aging, and poor contact. These defects are difficult to detect through routine inspections in the early stages.

[0003] Currently, photovoltaic power plants primarily employ a periodic maintenance model for inspecting photovoltaic inverters, lacking a precise early warning mechanism based on the actual operating status of the equipment. Traditional infrared temperature measurement inspections are often disconnected from equipment operating parameters, leading to delayed defect identification in photovoltaic inverters. Furthermore, fixed-cycle inspections struggle to capture early fault characteristics of IGBT modules, easily resulting in missed or incorrect fault diagnoses.

[0004] Therefore, there is an urgent need for those skilled in the art to provide an intelligent operation and maintenance method for IGBT modules that can integrate multi-source monitoring data, achieve early defect warning, and guide precise inspection. Summary of the Invention

[0005] The purpose of this invention is to provide a method and device for early warning of IGBT module linkage defects in photovoltaic inverters, which can solve at least one of the above-mentioned problems in the prior art.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for triggering early warning and inspection of IGBT module linkage defects in photovoltaic inverters, including: Simultaneously collect temperature field data on the surface of the photovoltaic inverter IGBT module and electrical parameters of the IGBT module; After preprocessing the temperature field data and the electrical parameters, key features are extracted to obtain the correlation index between the temperature and the electrical parameters; The correlation index between temperature and electrical parameters is input into a pre-trained deep neural network model to obtain the defect probability assessment result and defect type. Based on the defect probability assessment results and defect type, the target early warning level is determined, and the early warning mechanism corresponding to the target early warning level is activated; When the target warning level meets the preset conditions, an inspection task work order is generated, wherein the inspection task work order includes: the component to be inspected, the detection method, and the judgment criteria; Based on weather conditions, equipment operating status, inspection resources, and the inspection task work order, the inspection time window and path are planned and inspection instructions are generated. The inspection instructions are then sent to the power plant inspection system to dispatch drones or inspection personnel for inspection.

[0007] Optionally, the step of simultaneously acquiring temperature field data on the surface of the photovoltaic inverter IGBT module and the electrical parameters of the IGBT module includes: The infrared thermal imager array deployed inside the photovoltaic inverter cabinet acquires temperature field data in real time at key locations on the surface of the photovoltaic inverter IGBT module, including: gate drive terminal, collector-emitter connection point, and heat dissipation substrate contact surface. When collecting the temperature field data, the multi-channel data acquisition unit is controlled by a high-precision clock to synchronously acquire the on-saturation voltage drop, switching transient characteristics, gate charge characteristics, and output current ripple coefficient of the IGBT module.

[0008] Optionally, the step of preprocessing the temperature field data and the electrical parameters to extract key features and obtain the correlation index between the temperature and the electrical parameters includes: Noise filtering and non-uniformity correction are performed on the infrared image to obtain preprocessed temperature field data; wherein, the temperature field data is an infrared image; The electrical parameters are subjected to outlier removal and filtering denoising to obtain preprocessed electrical parameters; Extract temperature field data feature sequences from the preprocessed temperature field data, wherein the temperature field data feature sequences include: temperature spatial distribution features and time series features; The electrical parameter feature sequence is analyzed from the preprocessed electrical parameters. The electrical parameter feature sequence includes: drift trend, waveform distortion degree, and harmonic component changes. Based on the extracted temperature field data feature sequence and the electrical parameter feature sequence, a correlation index between temperature and electrical parameters is established.

[0009] Optionally, the step of inputting the correlation index between temperature and electrical parameters into a pre-trained deep neural network model to obtain the defect probability assessment result and defect type includes: The correlation index between temperature and electrical parameters is input into a pre-trained deep neural network model; wherein, the deep neural network model includes: a temperature feature extraction network branch and an electrical parameter analysis network branch; The temperature feature extraction network branch identifies temperature gradient distribution, hotspot movement trajectory, and thermal response time constant based on the temperature field data feature sequence. The electrical parameter analysis network branch extracts on-state voltage drop drift characteristics, switching loss variation trends, and gate drive anomaly information based on the electrical parameter feature sequence. By performing correlation analysis on the dual-modal data output by the temperature feature extraction network branch and the electrical parameter analysis network branch through the feature fusion layer, defect probability assessment results and defect type identification are obtained.

[0010] Optionally, the step of determining the target early warning level based on the defect probability assessment result and defect type, and activating the early warning mechanism corresponding to the target early warning level, includes: Based on the defect probability assessment results and defect type identification, obvious anomalies were found. Assess the severity and development trend of the defects; By combining rule-based reasoning, the defect probability assessment results, and the defect type, the target warning level is determined, whereby the warning levels include: attention level warning, warning level warning, and severe level warning; When the target warning level is the attention level warning, record the abnormal characteristics and start the enhanced monitoring mode; When the target warning level is a warning level or a severe level warning, an inspection task work order is generated.

[0011] Optionally, after the step of sending the inspection command to the power plant inspection system to dispatch a drone or inspection personnel for inspection, the method further includes: After the inspection is completed, inspection feedback data is sent to the system, which includes: on-site inspection results, maintenance and handling status and equipment operating status. Based on the feedback data, the system evaluates the accuracy of early warning and defect identification, and optimizes the model parameters and judgment thresholds of the deep neural network model through an incremental learning mechanism.

[0012] Optionally, before the step of acquiring temperature field data at key locations on the surface of the photovoltaic inverter IGBT module in real time using an infrared thermal imager array deployed inside the photovoltaic inverter cabinet, the method further includes: The infrared thermal imager array deployed inside the photovoltaic inverter cabinet is subjected to network calibration and parameter configuration; wherein, the network calibration includes: temperature calibration and spatial positioning of the infrared thermal imager; Range calibration of the multi-channel data acquisition unit; Under typical operating conditions of the photovoltaic inverter, normal state data of the IGBT module are collected to establish temperature distribution and electrical parameter benchmarks.

[0013] This invention also provides a photovoltaic inverter IGBT module linkage defect early warning and inspection triggering device, comprising: The acquisition module is used to simultaneously acquire temperature field data on the surface of the photovoltaic inverter IGBT module and the electrical parameters of the IGBT module; The extraction module is used to extract key features after preprocessing the temperature field data and the electrical parameters to obtain the correlation index between the temperature and the electrical parameters. The prediction module is used to input the correlation index between the temperature and electrical parameters into a pre-trained deep neural network model to obtain the defect probability assessment result and defect type. The determination module is used to determine the target warning level based on the defect probability assessment results and defect type, and to activate the warning mechanism corresponding to the target warning level; The generation module is used to generate an inspection task work order when the target warning level meets the preset conditions. The inspection task work order includes: the component to be inspected, the detection method, and the judgment criteria. The instruction generation module is used to plan the inspection time window and path and generate inspection instructions based on weather conditions, equipment operating status, inspection resources and the inspection task work order, and send the inspection instructions to the power plant inspection system to dispatch drones or inspection personnel for inspection.

[0014] Optionally, the acquisition module is specifically used for: The infrared thermal imager array deployed inside the photovoltaic inverter cabinet acquires temperature field data in real time at key locations on the surface of the photovoltaic inverter IGBT module, including: gate drive terminal, collector-emitter connection point, and heat dissipation substrate contact surface. When collecting the temperature field data, the multi-channel data acquisition unit is controlled by a high-precision clock to synchronously acquire the on-saturation voltage drop, switching transient characteristics, gate charge characteristics, and output current ripple coefficient of the IGBT module.

[0015] Optionally, the extraction module includes: The first submodule is used to perform noise filtering and non-uniformity correction on the infrared image to obtain preprocessed temperature field data; wherein, the temperature field data is the infrared image; The second submodule is used to perform outlier removal and filtering noise reduction on the electrical parameters to obtain preprocessed electrical parameters. The third submodule is used to extract temperature field data feature sequences from the preprocessed temperature field data, wherein the temperature field data feature sequences include: temperature spatial distribution features and time series features; The fourth submodule is used to analyze the electrical parameter feature sequence from the preprocessed electrical parameters. The electrical parameter feature sequence includes: drift trend, waveform distortion degree, and harmonic component changes. The fifth submodule is used to establish a correlation index between temperature and electrical parameters based on the extracted temperature field data feature sequence and the electrical parameter feature sequence.

[0016] Optionally, the prediction module is specifically used for: The correlation index between temperature and electrical parameters is input into a pre-trained deep neural network model; wherein, the deep neural network model includes: a temperature feature extraction network branch and an electrical parameter analysis network branch; The temperature feature extraction network branch identifies temperature gradient distribution, hotspot movement trajectory, and thermal response time constant based on the temperature field data feature sequence. The electrical parameter analysis network branch extracts on-state voltage drop drift characteristics, switching loss variation trends, and gate drive anomaly information based on the electrical parameter feature sequence. By performing correlation analysis on the dual-modal data output by the temperature feature extraction network branch and the electrical parameter analysis network branch through the feature fusion layer, defect probability assessment results and defect type identification are obtained.

[0017] Optionally, the determining module includes: Based on the defect probability assessment results and defect type identification, obvious anomalies were found. Assess the severity and development trend of the defects; By combining rule-based reasoning, the defect probability assessment results, and the defect type, the target warning level is determined, whereby the warning levels include: attention level warning, warning level warning, and severe level warning; When the target warning level is the attention level warning, record the abnormal characteristics and start the enhanced monitoring mode; When the target warning level is a warning level or a severe level warning, an inspection task work order is generated.

[0018] Optionally, the device further includes: The feedback module is used to send the inspection instruction to the power plant inspection system by the instruction generation module to dispatch drones or inspection personnel to carry out the inspection, and then send inspection feedback data to the system after the inspection is completed. The inspection feedback data includes: on-site inspection results, maintenance and handling status and equipment operating status. The system evaluates the accuracy of early warning and defect identification based on the feedback data, and optimizes the model parameters and judgment thresholds of the deep neural network model through an incremental learning mechanism.

[0019] Optionally, the device further includes: The calibration module is used to perform network calibration and parameter configuration on the infrared thermal imager array deployed inside the photovoltaic inverter cabinet before the acquisition module acquires temperature field data of key locations on the surface of the photovoltaic inverter IGBT module in real time through the infrared thermal imager array deployed inside the photovoltaic inverter cabinet; wherein, the network calibration includes: temperature calibration and spatial positioning of the infrared thermal imager. The calibration module is used to perform range calibration on the multi-channel data acquisition unit. The benchmark establishment module is used to collect normal state data of the IGBT module under typical operating conditions of the photovoltaic inverter and establish temperature distribution benchmark and electrical parameter benchmark.

[0020] The photovoltaic inverter IGBT module linkage defect early warning and inspection triggering scheme provided in this embodiment of the invention synchronously collects temperature field data on the surface of the photovoltaic inverter IGBT module and the electrical parameters of the IGBT module; after preprocessing the temperature field data and electrical parameters, key features are extracted to obtain the correlation index between temperature and electrical parameters; the correlation index between temperature and electrical parameters is input into a pre-trained deep neural network model to obtain defect probability assessment results and defect types; based on the defect probability assessment results and defect types, the target early warning level is determined, and the early warning mechanism corresponding to the target early warning level is activated; when the target early warning level meets preset conditions, an inspection task work order is generated; based on weather conditions, equipment operating status, inspection resources, and the inspection task work order, an inspection time window and path are planned and an inspection instruction is generated, which is sent to the power plant inspection system to dispatch drones or inspection personnel for inspection. The solution provided by this invention has the following advantages: First, it can establish a correlation between temperature anomalies and changes in electrical parameters before IGBT module failure, thus enabling timely identification of IGBT module defects; second, it can capture early fault characteristics of IGBT modules in a timely manner, avoiding missed or false fault detection; third, it integrates multi-source monitoring data for early defect warning, making the warning results more reliable; and fourth, it can guide more intelligent operation and maintenance of photovoltaic inverters during precise inspections. Attached Figure Description

[0021] The accompanying drawings, which form part of this specification, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart illustrating a photovoltaic inverter IGBT module linkage defect early warning and inspection triggering method according to an embodiment of this application; Figure 2 This is a structural block diagram illustrating a photovoltaic inverter IGBT module linkage defect early warning and inspection triggering device according to an embodiment of this application. Detailed Implementation

[0022] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described herein can be combined with each other.

[0023] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this invention is for describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0024] This invention aims to solve the problems of inefficient task scheduling, low equipment coordination, and unreasonable path planning in the inspection of new energy power plants. It achieves intelligent scheduling and globally optimal path planning based on real-time data, thereby improving inspection efficiency and equipment utilization. Specifically, it relates to an intelligent scheduling and path optimization scheme for energy plant inspection tasks based on digital twins, applicable to the collaborative management and efficient operation of various types of inspection equipment such as drones, robots, and cameras in new energy power plants such as photovoltaic power plants and wind farms.

[0025] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0026] The following description, in conjunction with the accompanying drawings, details the digital twin-based new energy power station inspection task scheduling scheme provided in this application through specific embodiments and application scenarios.

[0027] As attached Figure 1 As shown, the photovoltaic inverter IGBT module linkage defect early warning and inspection triggering method according to this application includes the following steps: Step 101: Synchronously collect temperature field data on the surface of the photovoltaic inverter IGBT module and electrical parameters of the IGBT module.

[0028] IGBT, or Insulated Gate Bipolar Transistor, is a core component in power electronic equipment. Functionally, an IGBT is a circuit switch—not a mechanical switch, but a computer-controlled one. Its advantages include voltage control, low saturation voltage drop, and high withstand voltage. IGBTs are used in high-voltage applications with voltage levels ranging from tens to thousands of volts and current levels ranging from tens to hundreds of amperes, making them crucial components for energy conversion and transmission. This application's embodiments utilize a multi-source data fusion monitoring network and intelligent diagnostic system to achieve early identification and accurate warning of IGBT module defects.

[0029] In one optional embodiment, the method for simultaneously acquiring temperature field data on the surface of the photovoltaic inverter IGBT module and electrical parameters of the IGBT module can be as follows: The infrared thermal imager array deployed inside the photovoltaic inverter cabinet acquires temperature field data at key locations on the surface of the photovoltaic inverter IGBT module in real time. While acquiring temperature field data, a high-precision clock controls a multi-channel data acquisition unit to simultaneously acquire the IGBT module's on-state saturation voltage drop, switching transient characteristics, gate charge characteristics, and output current ripple coefficient.

[0030] Key locations include the gate drive terminal, the collector-emitter connection point, and the heat sink substrate contact surface. A fixed array of infrared thermal imagers is deployed at key locations inside the photovoltaic inverter cabinet. Each imager is equipped with a dual-mode lens (wide-angle and macro) to achieve comprehensive monitoring of the IGBT module surface temperature field. The system employs a high-precision clock synchronization module to ensure the time synchronization accuracy of temperature field data and electrical parameter acquisition, thereby establishing a multi-source database with strict time-series correspondence. In actual implementation, the infrared thermal imagers acquire IGBT module surface temperature field data at a set frequency, while the electrical parameter acquisition system, i.e., the multi-channel data acquisition unit, synchronously records operating parameters such as on-state voltage drop and switching characteristics. All data included in the acquired temperature field data and electrical parameters are appended with precise timestamps and transmitted to the central processing unit via a data bus to ensure data consistency and integrity.

[0031] In an optional embodiment, before acquiring temperature field data of key locations on the surface of the photovoltaic inverter IGBT module in real time using an infrared thermal imager array deployed inside the photovoltaic inverter cabinet, the method further includes: S1: Perform network calibration and parameter configuration for the infrared thermal imager array deployed inside the photovoltaic inverter cabinet; the network calibration includes: temperature calibration and spatial positioning of the infrared thermal imager. S2: Perform range verification on the multi-channel data acquisition unit; S3: Under typical operating conditions of photovoltaic inverters, collect normal state data of IGBT modules to establish temperature distribution benchmarks and electrical parameter benchmarks.

[0032] In this alternative embodiment, calibrating the acquisition unit and infrared thermal imager array before acquiring temperature field data and electrical parameters can improve the reliability of subsequent data acquisition.

[0033] Step 102: After preprocessing the temperature field data and electrical parameters, extract key features to obtain the correlation index between temperature and electrical parameters.

[0034] This step involves quality assessment and preprocessing of the collected multi-source data, followed by extraction of key features from the preprocessed data. Temperature field data analysis includes spatial distribution and time-series characteristics; electrical parameter analysis focuses on parameter drift trends, waveform distortion, and harmonic component variations. Finally, the correlation index between temperature and electrical parameters is calculated to provide input data for IGBT module condition assessment.

[0035] In an optional embodiment, the method of extracting key features after preprocessing the temperature field data and electrical parameters to obtain the correlation index between temperature and electrical parameters may include the following sub-steps: Sub-step 1: Perform noise filtering and non-uniformity correction on the infrared image to obtain preprocessed temperature field data.

[0036] The temperature field data is presented as infrared images.

[0037] Sub-step 2: Perform outlier removal and filtering noise reduction on the electrical parameters to obtain the preprocessed electrical parameters.

[0038] Sub-step 3: Extract the temperature field data feature sequence from the preprocessed temperature field data.

[0039] Among them, the temperature field data feature sequence includes: temperature spatial distribution features and time series features.

[0040] Sub-step 4: Analyze the characteristic sequence of electrical parameters from the preprocessed electrical parameters.

[0041] The electrical parameter characteristic sequence includes: drift trend, waveform distortion degree, and harmonic component changes.

[0042] Sub-step 5: Based on the extracted temperature field data feature sequence and electrical parameter feature sequence, establish a correlation index between temperature and electrical parameters.

[0043] In this alternative embodiment, preprocessing the electrical parameters and temperature field data can effectively filter out interfering data and improve the reliability of the established correlation index between temperature and electrical parameters.

[0044] Step 103: Input the correlation index between temperature and electrical parameters into the pre-trained deep neural network model to obtain the defect probability assessment results and defect types.

[0045] The deep neural network model is built based on historical operating data. It consists of a temperature feature extraction network branch and an electrical parameter analysis network branch. The temperature feature extraction network branch processes infrared thermographic sequences to identify features such as temperature gradient distribution, hotspot movement trajectory, and thermal response time constant. The electrical parameter analysis network branch extracts parameters such as on-state voltage drop drift characteristics, switching loss variation trends, and gate drive anomalies. A feature fusion layer enables correlation analysis of the dual-modal data. The deep neural network model outputs, including but not limited to, defect probability assessment, defect type identification, and severity classification. The model employs an online learning mechanism, continuously optimizing network parameters based on newly acquired data.

[0046] In an optional embodiment, the method of inputting the correlation index between temperature and electrical parameters into a pre-trained deep neural network model to obtain the defect probability assessment result and defect type may include the following sub-steps: Sub-step 1: Input the correlation index between temperature and electrical parameters into the pre-trained deep neural network model; Sub-step 2: The temperature feature extraction network branch identifies the temperature gradient distribution, hotspot movement trajectory, and thermal response time constant based on the temperature field data feature sequence; Sub-step 3: The electrical parameter analysis network branch extracts on-state voltage drop drift characteristics, switching loss variation trends, and gate drive anomaly information based on the electrical parameter feature sequence; Sub-step 4: Perform correlation analysis on the dual-modal data output by the temperature feature extraction network branch and the electrical parameter analysis network branch through the feature fusion layer to obtain the defect probability assessment result and defect type identification.

[0047] Step 104: Determine the target warning level based on the defect probability assessment results and defect type, and activate the warning mechanism corresponding to the target warning level.

[0048] In one optional embodiment, the method for determining the target warning level based on the defect probability assessment result and the defect type, and activating the warning mechanism corresponding to the target warning level, can be as follows: First, based on the defect probability assessment results and defect type identification, obvious anomalies were found; Secondly, assess the severity and development trend of the defects; Finally, the target warning level is determined by combining rule-based reasoning, defect probability assessment results, and defect type. In addition to determining the target warning level, detailed anomaly descriptions and possible cause analyses can also be generated. The warning levels include: Attention Level Warning, Warning Level Warning, and Severe Level Warning.

[0049] The warning mechanism corresponding to each warning level can be set as follows: when the target warning level is the attention level warning, record the abnormal characteristics and start the enhanced monitoring mode; when the target warning level is the warning level warning or the severe level warning, generate an inspection task work order.

[0050] Step 105: If the target warning level meets the preset conditions, generate an inspection task work order.

[0051] The inspection task order includes: the component to be inspected, the testing method, and the judgment criteria. The component to be inspected may include, but is not limited to: cooling fans and wiring terminals.

[0052] Step 106: Based on weather conditions, equipment operating status, inspection resources, and inspection task work orders, plan the inspection time window and route, generate inspection instructions, and send the inspection instructions to the power plant inspection system to dispatch drones or inspection personnel for inspection.

[0053] In practice, the inspection task can clearly specify a list of components that require key checks, including the operating status of cooling fans, the tightness of wiring terminals, and the working status of drive circuits. The system provides an inspection guide, detailing the testing methods and judgment criteria to ensure the standardization and effectiveness of the inspection work.

[0054] This application embodiment constructs a complete defect analysis system, including defect feature extraction, cause analysis, development trend prediction, and maintenance suggestion generation. The system employs multi-dimensional data visualization technology to display information such as temperature distribution cloud maps, electrical parameter change curves, and defect development trajectories. Based on historical maintenance data and an expert experience database, it provides targeted handling suggestions and preventative measures to support maintenance personnel in making informed decisions.

[0055] In one optional embodiment, to form a closed-loop optimization process from monitoring, early warning, inspection to feedback, and to continuously improve system performance and reliability, a feedback mechanism is also provided after the inspection is completed. After sending the inspection command to the power plant inspection system to dispatch drones or inspection personnel for inspection, inspection feedback data is sent to the system after the inspection is completed. Based on the feedback data, the system evaluates the accuracy of early warning and the accuracy of defect identification, and optimizes the model parameters and judgment thresholds of the deep neural network model through an incremental learning mechanism.

[0056] The inspection feedback data includes: on-site inspection results, maintenance and handling status, and equipment operating status.

[0057] The photovoltaic inverter IGBT module linkage defect early warning and inspection triggering method provided in this application embodiment synchronously collects temperature field data of the photovoltaic inverter IGBT module surface and electrical parameters of the IGBT module; after preprocessing the temperature field data and electrical parameters, key features are extracted to obtain the correlation index between temperature and electrical parameters; the correlation index between temperature and electrical parameters is input into a pre-trained deep neural network model to obtain defect probability assessment results and defect types; based on the defect probability assessment results and defect types, the target early warning level is determined, and the early warning mechanism corresponding to the target early warning level is activated; when the target early warning level meets preset conditions, an inspection task work order is generated; based on weather conditions, equipment operating status, inspection resources, and the inspection task work order, an inspection time window and path are planned and an inspection instruction is generated, which is sent to the power plant inspection system to dispatch drones or inspection personnel for inspection. The method provided in this embodiment of the invention has the following advantages: First, it can establish a correlation between temperature anomalies and changes in electrical parameters before IGBT module failure, thus enabling timely identification of IGBT module defects; second, it can capture early fault characteristics of IGBT modules in a timely manner, avoiding missed or false fault detection; third, it integrates multi-source monitoring data for early defect warning, making the warning results more reliable; and fourth, it can guide more intelligent operation and maintenance of photovoltaic inverters during precise inspections.

[0058] Figure 2 The structural block diagram of the photovoltaic inverter IGBT module linkage defect early warning and inspection triggering device in this embodiment of the application is shown.

[0059] The photovoltaic inverter IGBT module linkage defect early warning and inspection triggering device in this application embodiment includes the following functional modules: The acquisition module 201 is used to simultaneously acquire temperature field data on the surface of the photovoltaic inverter IGBT module and electrical parameters of the IGBT module; Extraction module 202 is used to extract key features after preprocessing the temperature field data and the electrical parameters to obtain the correlation index between the temperature and the electrical parameters; The prediction module 203 is used to input the correlation index between the temperature and electrical parameters into a pre-trained deep neural network model to obtain the defect probability assessment result and defect type. The determination module 204 is used to determine the target warning level based on the defect probability assessment result and the defect type, and to activate the warning mechanism corresponding to the target warning level; The generation module 205 is used to generate an inspection task work order when the target warning level meets the preset conditions. The inspection task work order includes: the component to be inspected, the detection method, and the judgment criteria. The instruction generation module 206 is used to plan the inspection time window and path and generate inspection instructions based on weather conditions, equipment operating status, inspection resources and the inspection task work order, and send the inspection instructions to the power plant inspection system to dispatch drones or inspection personnel for inspection.

[0060] Optionally, the acquisition module is specifically used for: The infrared thermal imager array deployed inside the photovoltaic inverter cabinet acquires temperature field data in real time at key locations on the surface of the photovoltaic inverter IGBT module, including: gate drive terminal, collector-emitter connection point, and heat dissipation substrate contact surface. When collecting the temperature field data, the multi-channel data acquisition unit is controlled by a high-precision clock to synchronously acquire the on-saturation voltage drop, switching transient characteristics, gate charge characteristics, and output current ripple coefficient of the IGBT module.

[0061] Optionally, the extraction module includes: The first submodule is used to perform noise filtering and non-uniformity correction on the infrared image to obtain preprocessed temperature field data; wherein, the temperature field data is the infrared image; The second submodule is used to perform outlier removal and filtering noise reduction on the electrical parameters to obtain preprocessed electrical parameters. The third submodule is used to extract temperature field data feature sequences from the preprocessed temperature field data, wherein the temperature field data feature sequences include: temperature spatial distribution features and time series features; The fourth submodule is used to analyze the electrical parameter feature sequence from the preprocessed electrical parameters. The electrical parameter feature sequence includes: drift trend, waveform distortion degree, and harmonic component changes. The fifth submodule is used to establish a correlation index between temperature and electrical parameters based on the extracted temperature field data feature sequence and the electrical parameter feature sequence.

[0062] Optionally, the prediction module is specifically used for: The correlation index between temperature and electrical parameters is input into a pre-trained deep neural network model; wherein, the deep neural network model includes: a temperature feature extraction network branch and an electrical parameter analysis network branch; The temperature feature extraction network branch identifies temperature gradient distribution, hotspot movement trajectory, and thermal response time constant based on the temperature field data feature sequence. The electrical parameter analysis network branch extracts on-state voltage drop drift characteristics, switching loss variation trends, and gate drive anomaly information based on the electrical parameter feature sequence. By performing correlation analysis on the dual-modal data output by the temperature feature extraction network branch and the electrical parameter analysis network branch through the feature fusion layer, defect probability assessment results and defect type identification are obtained.

[0063] Optionally, the determining module includes: Based on the defect probability assessment results and defect type identification, obvious anomalies were found. Assess the severity and development trend of the defects; By combining rule-based reasoning, the defect probability assessment results, and the defect type, the target warning level is determined, whereby the warning levels include: attention level warning, warning level warning, and severe level warning; When the target warning level is the attention level warning, record the abnormal characteristics and start the enhanced monitoring mode; When the target warning level is a warning level or a severe level warning, an inspection task work order is generated.

[0064] Optionally, the device further includes: The feedback module is used to send the inspection instruction to the power plant inspection system by the instruction generation module to dispatch drones or inspection personnel to carry out the inspection, and then send inspection feedback data to the system after the inspection is completed. The inspection feedback data includes: on-site inspection results, maintenance and handling status and equipment operating status. The system evaluates the accuracy of early warning and defect identification based on the feedback data, and optimizes the model parameters and judgment thresholds of the deep neural network model through an incremental learning mechanism.

[0065] Optionally, the device further includes: The calibration module is used to perform network calibration and parameter configuration on the infrared thermal imager array deployed inside the photovoltaic inverter cabinet before the acquisition module acquires temperature field data of key locations on the surface of the photovoltaic inverter IGBT module in real time through the infrared thermal imager array deployed inside the photovoltaic inverter cabinet; wherein, the network calibration includes: temperature calibration and spatial positioning of the infrared thermal imager. The calibration module is used to perform range calibration on the multi-channel data acquisition unit. The benchmark establishment module is used to collect normal state data of the IGBT module under typical operating conditions of the photovoltaic inverter and establish temperature distribution benchmark and electrical parameter benchmark.

[0066] The photovoltaic inverter IGBT module linkage defect early warning and inspection triggering device provided in this application embodiment has the following advantages: First, it can establish a correlation between temperature anomalies and changes in electrical parameters before IGBT module failure, thus enabling timely identification of IGBT module defects; second, it can capture early fault characteristics of IGBT modules in a timely manner, avoiding missed or false fault diagnosis; third, it integrates multi-source monitoring data for early defect early warning, making the warning results more reliable; and fourth, it can guide more intelligent operation and maintenance of photovoltaic inverters during precise inspections.

[0067] The embodiments provided in this application Figure 2 The photovoltaic inverter IGBT module linkage defect early warning and inspection triggering device shown can achieve Figure 1 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0068] This application also provides an electronic device, which includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus.

[0069] The memory is used to store computer programs; the processor is used to execute the programs stored in the memory to implement the digital twin-based new energy power station inspection task scheduling process in the above embodiments.

[0070] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. The communication interface is used for communication between the aforementioned terminal and other devices.

[0071] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0072] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0073] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0074] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for early warning and inspection triggering of IGBT module linkage defects in a photovoltaic inverter, characterized in that, include: Simultaneously collect temperature field data on the surface of the photovoltaic inverter IGBT module and electrical parameters of the IGBT module; After preprocessing the temperature field data and the electrical parameters, key features are extracted to obtain the correlation index between the temperature and the electrical parameters; The correlation index between temperature and electrical parameters is input into a pre-trained deep neural network model to obtain the defect probability assessment result and defect type. Based on the defect probability assessment results and defect type, the target early warning level is determined, and the early warning mechanism corresponding to the target early warning level is activated; When the target warning level meets the preset conditions, an inspection task work order is generated, wherein the inspection task work order includes: the component to be inspected, the detection method, and the judgment criteria; Based on weather conditions, equipment operating status, inspection resources, and the inspection task work order, the inspection time window and path are planned and inspection instructions are generated. The inspection instructions are then sent to the power plant inspection system to dispatch drones or inspection personnel for inspection.

2. The method according to claim 1, characterized in that, The steps for synchronously acquiring temperature field data on the surface of the photovoltaic inverter IGBT module and the electrical parameters of the IGBT module include: The infrared thermal imager array deployed inside the photovoltaic inverter cabinet acquires temperature field data in real time at key locations on the surface of the photovoltaic inverter IGBT module, including: gate drive terminal, collector-emitter connection point, and heat dissipation substrate contact surface. When collecting the temperature field data, the multi-channel data acquisition unit is controlled by a high-precision clock to synchronously acquire the on-saturation voltage drop, switching transient characteristics, gate charge characteristics, and output current ripple coefficient of the IGBT module.

3. The method according to claim 1, characterized in that, The step of preprocessing the temperature field data and the electrical parameters to extract key features and obtain the correlation index between the temperature and the electrical parameters includes: Noise filtering and non-uniformity correction are performed on the infrared image to obtain preprocessed temperature field data; wherein, the temperature field data is an infrared image; The electrical parameters are subjected to outlier removal and filtering denoising to obtain preprocessed electrical parameters; Extract temperature field data feature sequences from the preprocessed temperature field data, wherein the temperature field data feature sequences include: temperature spatial distribution features and time series features; The electrical parameter feature sequence is analyzed from the preprocessed electrical parameters. The electrical parameter feature sequence includes: drift trend, waveform distortion degree, and harmonic component changes. Based on the extracted temperature field data feature sequence and the electrical parameter feature sequence, a correlation index between temperature and electrical parameters is established.

4. The method according to claim 1, characterized in that, The steps of inputting the correlation index between temperature and electrical parameters into a pre-trained deep neural network model to obtain the defect probability assessment result and defect type include: The correlation index between temperature and electrical parameters is input into a pre-trained deep neural network model; wherein, the deep neural network model includes: a temperature feature extraction network branch and an electrical parameter analysis network branch; The temperature feature extraction network branch identifies temperature gradient distribution, hotspot movement trajectory, and thermal response time constant based on the temperature field data feature sequence. The electrical parameter analysis network branch extracts on-state voltage drop drift characteristics, switching loss variation trends, and gate drive anomaly information based on the electrical parameter feature sequence. By performing correlation analysis on the dual-modal data output by the temperature feature extraction network branch and the electrical parameter analysis network branch through the feature fusion layer, defect probability assessment results and defect type identification are obtained.

5. The method according to claim 1, characterized in that, The steps of determining the target early warning level based on the defect probability assessment results and defect type, and activating the early warning mechanism corresponding to the target early warning level, include: Based on the defect probability assessment results and defect type identification, obvious anomalies were found. Assess the severity and development trend of the defects; By combining rule-based reasoning, the defect probability assessment results, and the defect type, the target warning level is determined, whereby the warning levels include: attention level warning, warning level warning, and severe level warning; When the target warning level is the attention level warning, record the abnormal characteristics and start the enhanced monitoring mode; When the target warning level is a warning level or a severe level warning, an inspection task work order is generated.

6. The method according to claim 1, characterized in that, After the step of sending the inspection command to the power plant inspection system to dispatch drones or inspection personnel for inspection, the method further includes: After the inspection is completed, inspection feedback data is sent to the system, which includes: on-site inspection results, maintenance and handling status and equipment operating status. Based on the feedback data, the system evaluates the accuracy of early warning and defect identification, and optimizes the model parameters and judgment thresholds of the deep neural network model through an incremental learning mechanism.

7. The method according to claim 2, characterized in that, Before the step of acquiring real-time temperature field data at key locations on the surface of the photovoltaic inverter IGBT module using an infrared thermal imager array deployed inside the photovoltaic inverter cabinet, the method further includes: The infrared thermal imager array deployed inside the photovoltaic inverter cabinet is subjected to network calibration and parameter configuration; wherein, the network calibration includes: temperature calibration and spatial positioning of the infrared thermal imager; Range calibration of the multi-channel data acquisition unit; Under typical operating conditions of the photovoltaic inverter, normal state data of the IGBT module are collected to establish temperature distribution and electrical parameter benchmarks.

8. A photovoltaic inverter IGBT module linkage defect early warning and inspection triggering device, characterized in that, include: The acquisition module is used to simultaneously acquire temperature field data on the surface of the photovoltaic inverter IGBT module and the electrical parameters of the IGBT module; The extraction module is used to extract key features after preprocessing the temperature field data and the electrical parameters to obtain the correlation index between the temperature and the electrical parameters. The prediction module is used to input the correlation index between the temperature and electrical parameters into a pre-trained deep neural network model to obtain the defect probability assessment result and defect type. The determination module is used to determine the target warning level based on the defect probability assessment results and defect type, and to activate the warning mechanism corresponding to the target warning level; The generation module is used to generate an inspection task work order when the target warning level meets the preset conditions. The inspection task work order includes: the component to be inspected, the detection method, and the judgment criteria. The instruction generation module is used to plan the inspection time window and path and generate inspection instructions based on weather conditions, equipment operating status, inspection resources and the inspection task work order, and send the inspection instructions to the power plant inspection system to dispatch drones or inspection personnel for inspection.

9. The apparatus according to claim 8, characterized in that, The acquisition module is specifically used for: The infrared thermal imager array deployed inside the photovoltaic inverter cabinet acquires temperature field data in real time at key locations on the surface of the photovoltaic inverter IGBT module, including: gate drive terminal, collector-emitter connection point, and heat dissipation substrate contact surface. When collecting the temperature field data, the multi-channel data acquisition unit is controlled by a high-precision clock to synchronously acquire the on-saturation voltage drop, switching transient characteristics, gate charge characteristics, and output current ripple coefficient of the IGBT module.

10. The apparatus according to claim 8, characterized in that, The extraction module includes: The first submodule is used to perform noise filtering and non-uniformity correction on the infrared image to obtain preprocessed temperature field data; wherein, the temperature field data is the infrared image; The second submodule is used to perform outlier removal and filtering noise reduction on the electrical parameters to obtain preprocessed electrical parameters. The third submodule is used to extract temperature field data feature sequences from the preprocessed temperature field data, wherein the temperature field data feature sequences include: temperature spatial distribution features and time series features; The fourth submodule is used to analyze the electrical parameter feature sequence from the preprocessed electrical parameters. The electrical parameter feature sequence includes: drift trend, waveform distortion degree, and harmonic component changes. The fifth submodule is used to establish a correlation index between temperature and electrical parameters based on the extracted temperature field data feature sequence and the electrical parameter feature sequence.

Images