Converter power switch tube health monitoring method and equipment

Abstract

The invention discloses a health monitoring method and equipment for a converter power switch tube. The health monitoring method comprises the following steps: acquiring converter data uploaded by a collector; enabling the converter data to pass through a health model of a power switch tube to obtain health data of the power switch tube; and displaying the health data of the power switch tube through a user interaction interface. According to the invention, the data is collected through the collector and is intelligently analyzed, so that the purpose of accurately and efficiently analyzing the health data of the power switch tube in a real-time online and high-coverage manner is achieved, the safety of a new energy station is improved, and the service life of the converter is prolonged.

Description

Technical Field

[0001] This application relates to the field of converter technology, and in particular to a method and device for health monitoring of converter power switching transistors. Background Technology

[0002] Converters based on power semiconductor devices, such as inverters and converters, are used for power conversion and AC / DC conversion, and are widely used in new energy industries such as wind power, photovoltaics, hydrogen production, traction drives, and power supplies. The power semiconductor devices used in converters, such as insulated-gate bipolar transistors (IGBTs), feature high voltage, high current, fast switching speed, and low losses. IGBTs account for up to 15% of the cost of a converter, making them relatively expensive. The healthy operation of IGBTs is crucial for the efficient and safe operation of new energy power plant equipment; however, current new energy power plants often lack monitoring and analysis methods for IGBT health, resulting in high IGBT failure rates, impacting power generation, and incurring high replacement costs.

[0003] Publicly available literature reports methods proposed by researchers for estimating IGBT health. Method one involves estimating operating conditions in advance to calculate the expected lifespan of the IGBT. Its drawbacks are: 1. Poor accuracy: the deviation between predicted and actual operating conditions is usually large, making it difficult to accurately estimate the actual health of the converter; 2. Inability to perform continuous analysis: when operating conditions change, designers need to re-analyze; 3. Poor coverage: analysis for various operating conditions is cumbersome. Method two involves collecting historical operating data and performing retrospective analysis to obtain relatively rough health and lifespan data for the converter. Its drawbacks are: 1. Poor accuracy: the recorded data is affected by time precision, making it difficult to accurately monitor IGBT health data; 2. Inability to perform real-time continuous analysis and provide users with real-time analysis results; 3. Poor coverage: it cannot achieve large-scale IGBT monitoring and horizontal analysis in new energy power plants. Summary of the Invention

[0004] This application provides a health monitoring method and device for power switching transistors in a converter, aiming to achieve accurate, efficient, real-time, online, and high-coverage analysis of power switching transistor health data, thereby improving the safety of new energy power plants and extending the service life of the converter.

[0005] This application provides a health monitoring method for a converter power switch transistor, the health monitoring method including:

[0006] Retrieve the converter data uploaded by the collector;

[0007] After passing the converter data through the health model of the power switch, the health data of the power switch is obtained.

[0008] The health data of the power switching transistors is displayed through a user interface.

[0009] Another aspect of this application provides a health monitoring device for a converter power switch, including at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the health monitoring device for the converter power switch to perform a health monitoring method for the converter power switch.

[0010] The health monitoring method and equipment for converter power switches provided in this application collect data through a data acquisition device and perform intelligent analysis to achieve accurate, efficient, real-time, and high-coverage analysis of power switch health data, thereby improving the safety of new energy power plants and extending the service life of converters. Specifically:

[0011] 1. Real-time online monitoring of power switch transistor losses, junction temperature, and lifespan; evaluation of the health status of power switch transistors; assistance in identifying faulty power switch transistors; early maintenance measures; and improvement of the service life of power switch transistors.

[0012] 2. Monitor the aging curve of power switching transistors online to identify converters that have entered the end of their lifespan; help users understand the aging status of power switching transistors, make good spare parts plans, and avoid power generation loss due to power switching transistor failure;

[0013] 3. Collect big data on the health of power switching transistors to help new energy equipment manufacturers obtain and analyze converter operating data, improve key components and control algorithms of converters, extend the service life of converters, and thus enhance product competitiveness. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of a health monitoring method provided in an embodiment of this application;

[0015] Figure 2 This is a schematic diagram of device connection provided in an embodiment of this application;

[0016] Figure 3 This is a schematic diagram of the health monitoring process provided in an embodiment of this application;

[0017] Figure 4 This is a schematic diagram illustrating the connections between the various models provided in the embodiments of this application;

[0018] Figure 5 This is a schematic diagram of the IGBT junction temperature model provided in an embodiment of this application.

[0019] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] To make the technical problems, technical solutions, and beneficial effects to be solved by this application clearer and more understandable, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit the scope of this application.

[0021] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0022] like Figure 1 As shown, one embodiment of this application provides a health monitoring method for a converter power switch transistor, the health monitoring method including the following steps:

[0023] S11. Obtain the converter data uploaded by the collector;

[0024] S12. After passing the converter data through the health model of the power switch, the health data of the power switch is obtained.

[0025] In some examples, the health model of the power switch includes a loss model, a thermal resistance model, an integrator, and a classification model;

[0026] The step of obtaining the health data of the power switch by passing the converter data through the health model of the power switch includes:

[0027] After passing the converter data through the loss model, the loss data of the power switch is obtained;

[0028] After passing the loss data of the power switch transistor through the thermal resistance model, the junction temperature fluctuation data of the power switch transistor is obtained.

[0029] After passing the junction temperature fluctuation data of the power switch transistor through the integrator, the power cycle data and lifetime data of the power switch transistor are obtained.

[0030] After passing all the data of the power switch through the classification model, abnormal power switches are identified.

[0031] In some examples, obtaining the loss data of the power switch by passing the converter data through the loss model includes:

[0032] Based on the converter data and the pre-trained artificial intelligence model, the loss data of the power switch is obtained.

[0033] The artificial intelligence model is obtained in the following way:

[0034] Obtain at least one of the following: conduction loss, switching loss, corresponding diode reverse recovery loss data, and junction temperature fluctuation data of the power switch;

[0035] The artificial intelligence model is obtained by training the power switch based on at least one of the following: conduction loss, switching loss, corresponding diode reverse recovery loss data, and junction temperature fluctuation data: using a preset algorithm.

[0036] It should be noted that the data required for training can be obtained from the chip datasheet provided by the manufacturer; it can also be obtained from the loss and junction temperature simulation software provided by the manufacturer. By setting the operating conditions and operating the software, data such as loss, junction temperature fluctuation, and maximum junction temperature can be obtained; more accurate power switch conduction loss, switching loss, diode reverse recovery loss, and junction temperature fluctuation of the power switch can also be obtained on a test bench. For example, thermal resistance... Parameters such as thermal time constant τ, thermal resistance-PowerCycle curve, loss-IGBT junction temperature fluctuation, and junction temperature-PowerCycle curve can be obtained from IGBT manufacturer data or through aging tests performed on an IGBT test bench.

[0037] In some examples, obtaining the junction temperature fluctuation data of the power switch by passing the loss data of the power switch through the thermal resistance model includes:

[0038] Based on the power switch's loss data, preset thermal resistance and thermal capacity parameters, and preset thermal resistance-power cycle curve, the junction temperature fluctuation data of the power switch is calculated.

[0039] In some examples, obtaining the power cycle data and lifetime data of the power switch by passing the junction temperature fluctuation data of the power switch through the integrator includes:

[0040] Based on the junction temperature fluctuation data of the power switch and the preset junction temperature-power cycle curve, the power cycle data and lifetime data of the power switch are calculated.

[0041] In some examples, the process of identifying abnormal power switches after processing all data of the power switch through the classification model includes:

[0042] Clustering calculations are performed on all data of the power switch to identify abnormal power switches.

[0043] In some examples, the power cycle data of the power switch includes total power cycles, converter runtime, and used power cycles, and the lifetime data of the power switch includes lifetime percentage, total lifetime, remaining lifetime, and lifetime attrition rate; wherein,

[0044] Integrated power cycle Converter runtime Power cycling has been used Lifespan percentage Total lifespan Remaining lifespan Lifespan consumption rate ; For the calculation period, For the current period, For current period The corresponding power cycle.

[0045] S13. Display the health data of the power switch through the user interface.

[0046] The user interface provides users with data on the operating conditions, junction temperature, losses, lifespan, and health of the monitored converters. Users can independently select data from different converters and power switches, displaying operating conditions, junction temperature, losses, lifespan, and health over time as curves. The interface supports setting initial lifespan values ​​for users replacing power switches; it also supports setting initial values ​​for newly deployed systems in existing renewable energy power plants; and it supports data export.

[0047] In some examples, displaying the health data of the power switch through a user interface includes:

[0048] The health data of multiple power switching transistors are displayed in a digital twin manner.

[0049] Specifically, it can display the health status of the converter's power modules and multiple power switches in a digital twin format. It supports highlighting power modules and power switches of concern to the user or exhibiting abnormalities. It also supports panoramic display of all converters in the entire renewable energy power station.

[0050] like Figure 2 As shown, another embodiment of this application provides a health monitoring device for a converter power switch, including at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to cause the health monitoring device for the converter power switch to perform the aforementioned health monitoring method for the converter power switch.

[0051] refer to Figure 2To understand this, each converter can be equipped with one or more data acquisition devices. These devices collect data from the converter and upload it to the health monitoring equipment for the converter's power switching transistors. The health monitoring equipment then performs intelligent analysis to achieve accurate, efficient, real-time, and comprehensive analysis of the power switching transistor health data, thereby improving the safety of new energy power plants and extending the lifespan of the converters.

[0052] The following combination Figures 3-5 The following explanation uses the intelligent monitoring of IGBTs in a 100MW wind farm as an example:

[0053] (1) Overall Deployment

[0054] A 100MW wind farm has 20 5MW doubly-fed induction generator (DFIG) wind turbine converters. The converters use a three-level topology and are air-cooled. The wind farm monitoring center has a health monitoring device for the converter power switches. Each of the 20 wind turbine converters is equipped with a data acquisition unit. The data acquisition units are connected to the converters via the wind farm's local area network (LAN).

[0055] (2) Data Acquisition

[0056] The data acquisition unit collects converter data, including basic converter information (two-level, three-level, number of IGBTs, converter serial number), converter ambient temperature, switching frequency of grid-side / generator-side power modules, power module temperature, bus voltage, grid-side / generator-side current frequency, three-phase current of grid-side / generator-side modules, power factor angle of grid-side / generator-side modules, and grid-side / generator-side waveform generation mode (SVPWM, DPWM). The data acquisition unit connects to the wind farm's local area network and uploads the above data to the health monitoring equipment of the converter's power switching transistors via TCP-IP protocol.

[0057] (3) Data preparation for IGBT loss model

[0058] The data required for training the loss model is obtained using the loss and junction temperature simulation software provided by the IGBT manufacturer. By setting operating conditions (current amplitude, current frequency, switching frequency, bus voltage, ambient temperature, power factor, waveform generation method, etc.), data such as loss, junction temperature fluctuation, and maximum junction temperature are obtained. These six operating conditions need to be combined in various ways to obtain a comprehensive dataset. For example, regarding current amplitude, if the rated current of the converter module is 1200A, starting from 0A, the rated current is increased by 25% each time until 100% of the rated current, resulting in five possible input current amplitudes. Regarding current frequency, if the rated current frequency of the grid-side converter is 50Hz, starting from 45Hz, the current frequency is changed by 5Hz each time until 55Hz, resulting in three possible input current frequencies: 45Hz, 50Hz, and 55Hz. Regarding switching frequency, the converter has two possible input switching frequencies: 2000Hz and 3000Hz. Taking bus voltage as an example, the converter's bus voltage operating range is 1600V to 1900V. With a 100V bus voltage step, there are four input options: 1600V, 1700V, 1800V, and 1900V. Taking ambient temperature as another example, the converter can operate in an ambient temperature range of -10 to 40 degrees Celsius, with three input options: -10 degrees Celsius, 20 degrees Celsius, and 40 degrees Celsius. Taking power factor angle as another example, the converter has three input options: -20 degrees Celsius, 0 degrees Celsius, and 20 degrees Celsius. Taking waveform generation method as another example, there are two input options: SVPWM and DPWM. The total number of input combinations after considering these factors is 5 × 3 × 2 × 4 × 3 × 3 × 2 = 2160 combinations. Each combination yields a data point, which serves as training data for the loss model. The number of combinations can be further increased to obtain more refined and accurate results.

[0059] (4) IGBT Loss Model

[0060] Using the above loss dataset, an artificial neural network can be built and trained on machine learning platforms such as Tensorflow, based on the scikit-learn algorithm library, so that the loss model can calculate the loss of each IGBT and diode.

[0061] (5) IGBT thermal resistance model and integrator

[0062] The thermal resistance model calculates the junction temperature fluctuation of each IGBT and diode based on losses and the thermal resistance network. In other words, the thermal resistance model calculates the junction temperature fluctuation of the IGBT based on its losses, thermal resistance, and thermal capacity parameters.

[0063] Calculate the IGBT junction temperature using the following formula:

[0064] Ambient temperature Radiator temperature IGBT case temperature IGBT junction temperature Radiator thermal resistance Radiator time constant IGBT case thermal resistance IGBT case time constant IGBT core temperature thermal resistance IGBT core temperature time constant .

[0065] The thermal resistance parameter changes with power cycling, and the parameter can be obtained by aging tests on a test bench.

[0066] The integrator calculates data such as power cycle and lifetime based on preset junction temperature-power cycle curve parameters.

[0067] Calculation cycle Current cycle Current current cycle The corresponding power cycle Integrated power cycle Converter runtime Power cycling has been used Lifespan percentage Total lifespan Remaining lifespan Lifespan consumption rate .

[0068] (6) Classification model

[0069] The classification model performs clustering calculations on data from all IGBTs across all converters, including junction temperature, losses, and lifetime, to identify IGBTs with abnormal data and issue warnings to users. The K-Means clustering algorithm can be used to calculate the categories of IGBTs at the same location across different converters. By calculating the cluster centroids and the Euclidean distance of each IGBT, abnormal IGBTs are identified.

[0070] (7) User interface

[0071] a. The user interface displays data such as junction temperature, losses, remaining lifetime, lifetime attrition rate, and health status based on user selection. The user interface of the converter power switch health monitoring device should include the converter number, IGBT number, and selectable drop-down menus or input boxes, supporting users to select data for different converters and different IGBTs.

[0072] The user interface displays data over time curves, including power, junction temperature, losses, and lifespan. Users can zoom in and out over time. Users can also view historical data curves and select a specific time period.

[0073] b. The user interface displays the health data of all converters in a digital twin format. The user interface uses digital twins to display the health status of converter modules and multiple IGBTs. Selected power modules and IGBTs that are malfunctioning are highlighted. It supports panoramic display of data from all converters and inverters in the entire renewable energy power plant.

[0074] c. The user interface generates analysis reports based on user needs to assist users in decision-making.

[0075] The preferred embodiments of this application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and spirit of this application shall be within the scope of the claims.

Claims

1. A method for health monitoring of power switching transistors in a converter, characterized in that, The health monitoring method includes: Retrieve the converter data uploaded by the collector; The converter data is processed through the power switch health model to obtain the power switch health data. The health data of the power switch is displayed through a user interface.

2. The health monitoring method according to claim 1, characterized in that, The health model of the power switch includes a loss model, a thermal resistance model, an integrator, and a classification model; The step of obtaining the health data of the power switch by passing the converter data through the health model of the power switch includes: After passing the converter data through the loss model, the loss data of the power switch is obtained. After passing the loss data of the power switch through the thermal resistance model, the junction temperature fluctuation data of the power switch is obtained. After passing the junction temperature fluctuation data of the power switch transistor through the integrator, the power cycle data and lifetime data of the power switch transistor are obtained. After passing all the data of the power switch through the classification model, abnormal power switches are identified.

3. The health monitoring method according to claim 2, characterized in that, The step of obtaining the loss data of the power switch by passing the converter data through the loss model includes: Based on the converter data and the pre-trained artificial intelligence model, the loss data of the power switch is obtained.

4. The health monitoring method according to claim 3, characterized in that, The artificial intelligence model was obtained in the following way: Obtain at least one of the following: conduction loss, switching loss, corresponding diode reverse recovery loss data, and junction temperature fluctuation data of the power switch; The artificial intelligence model is obtained by training the power switch based on at least one of the following: conduction loss, switching loss, corresponding diode reverse recovery loss data, and junction temperature fluctuation data: using a preset algorithm.

5. The health monitoring method according to claim 2, characterized in that, The step of obtaining the junction temperature fluctuation data of the power switch by passing the loss data of the power switch through the thermal resistance model includes: Based on the power switch's loss data, preset thermal resistance and thermal capacity parameters, and preset thermal resistance-power cycle curve, the junction temperature fluctuation data of the power switch is calculated.

6. The health monitoring method according to claim 2, characterized in that, The process of passing the junction temperature fluctuation data of the power switch transistor through the integrator to obtain the power cycle data and lifetime data of the power switch transistor includes: Based on the junction temperature fluctuation data of the power switch and the preset junction temperature-power cycle curve, the power cycle data and lifetime data of the power switch are calculated.

7. The health monitoring method according to claim 2, characterized in that, The step of identifying abnormal power switches after processing all data of the power switch through the classification model includes: Clustering calculations are performed on all data of the power switch to identify abnormal power switches.

8. The health monitoring method according to claim 2, characterized in that, The power cycle data of the power switch includes the total power cycle, converter operating time, and used power cycles; the lifetime data of the power switch includes the percentage of lifetime used, total lifetime, remaining lifetime, and lifetime consumption rate. Integrated power cycle Converter runtime Power cycling has been used Lifespan percentage Total lifespan Remaining lifespan Lifespan consumption rate ; For the calculation period, For the current period, For current period The corresponding power cycle.

9. The health monitoring method according to claim 1, characterized in that, The method of displaying the health data of the power switch through a user interface includes: The health data of multiple power switching transistors are displayed in a digital twin manner.

10. A health monitoring device for a converter power switching transistor, characterized in that, The device includes at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, cause the health monitoring device of the converter power switch to perform the health monitoring method of the converter power switch according to any one of claims 1-9.

Images