Igbt junction temperature prediction method, device, equipment and storage medium

Abstract

The application provides an IGBT junction temperature prediction method, device, equipment and storage medium, and particularly relates to the technical field of IGBT junction temperature prediction, and the method comprises the following steps: a processor acquires an insulated gate bipolar transistor (IGBT) collector current at the nth moment, an IGBT collector voltage at the nth moment, an IGBT working condition cumulative effect at the nth moment, a first IGBT power loss at the nth moment and a second IGBT power loss at the nth moment; the processor inputs the IGBT collector current at the nth moment, the IGBT collector voltage at the nth moment, the IGBT working condition cumulative effect at the nth moment, the first IGBT power loss at the nth moment and the second IGBT power loss at the nth moment into a trained long short-term memory network model to obtain an IGBT junction temperature value at the (n+1)th moment. The method comprehensively considers the influencing factors of the IGBT junction temperature value, significantly improves the reliability and prediction stability of the long short-term memory network model, and further improves the accuracy of the IGBT junction temperature value prediction.

Description

Technical Field

[0001] This application relates to the field of IGBT junction temperature prediction technology, and in particular to an IGBT junction temperature prediction method, apparatus, device and storage medium. Background Technology

[0002] As one of the core components of a wind power generation system, the wind power converter contains the IGBT (Insulated Gate Bipolar Transistor), a crucial part whose performance directly determines the overall efficiency and stability of the converter. Since IGBTs typically operate under high current, high load, and high temperature conditions, variations in their junction temperature have a significant impact on their performance and reliability.

[0003] Currently, wind power converters rely on temperature sensors installed externally to directly measure the IGBT junction temperature and predict the next IGBT junction temperature based on the measured value. If the predicted IGBT junction temperature exceeds a preset temperature threshold, the temperature sensor issues an alarm, prompting staff to shut down the wind power converter.

[0004] However, the influence of variable operating conditions in the actual operating environment of wind power converters on the predicted IGBT junction temperature value is ignored, resulting in inaccurate prediction results. Summary of the Invention

[0005] In view of this, the purpose of this application is to provide an IGBT junction temperature prediction method, apparatus, device and storage medium to solve the above problems and achieve accurate prediction of IGBT junction temperature values.

[0006] In a first aspect, embodiments of this application provide a method for predicting IGBT junction temperature, the method comprising: The processor acquires the collector current of the insulated gate bipolar transistor (IGBT) at time n, the collector voltage of the IGBT at time n, the cumulative effect of the IGBT operating conditions at time n, the first IGBT power loss at time n, and the second IGBT power loss at time n; wherein, the first IGBT power loss at time n is the power loss of the IGBT affected by wind speed at time n, and the second IGBT power loss at time n is the power loss of the IGBT not affected by wind speed at time n; The processor inputs the IGBT collector current, IGBT collector voltage, IGBT operating condition cumulative effect, first IGBT power loss, and second IGBT power loss at time n into the trained long short-term memory network model to obtain the IGBT junction temperature at time n+1. The Long Short-Term Memory (LSTM) network model is obtained as follows: Obtain an IGBT parameter sample set; wherein, the IGBT parameter sample set includes IGBT collector current samples, IGBT collector voltage samples, IGBT operating condition cumulative effect samples, first IGBT power loss samples, and second IGBT power loss samples. The IGBT parameter sample set is input into the original long short-term memory network model for training to obtain the long short-term memory network model.

[0007] Preferably, the method further includes: The processor acquires the IGBT junction temperature measurement values ​​and the cumulative effect of IGBT operating conditions at the first n time steps obtained from the inspection. The processor calculates the target operating condition cumulative effect correction factor at time n+1 based on the IGBT junction temperature at time n+1 and the measured IGBT junction temperature at time n. The processor calculates the correction value of the IGBT junction temperature at time n+1 based on the IGBT junction temperature at time n+1, the target operating condition cumulative effect correction factor at time n+1, and the IGBT operating condition cumulative effect at time n.

[0008] Preferably, the step of the processor calculating the corrected value of the IGBT junction temperature at time n+1 based on the IGBT junction temperature at time n+1, the target operating condition cumulative effect correction factor at time n+1, and the IGBT operating condition cumulative effect at time n includes: The formula for calculating the correction value of the IGBT junction temperature at time n+1 is as follows:

[0009] in, This is the correction value for the IGBT junction temperature at time n+1. This represents the IGBT junction temperature at time n+1. This is the correction factor for the cumulative effect of the target operating condition at time n+1. This represents the cumulative effect of IGBT operating conditions at time n.

[0010] Preferably, the step of the processor calculating the target operating condition cumulative effect correction factor at time n+1 based on the IGBT junction temperature value at time n+1 and the measured IGBT junction temperature at time n includes: The formula for calculating the cumulative effect correction factor of the target operating condition at time n+1 is as follows:

[0011] in, This is the correction factor for the cumulative effect of the target operating condition at time n+1. This represents the IGBT junction temperature at time n+1. This represents the measured junction temperature of the IGBT at time n.

[0012] Preferably, the method further includes: The processor acquires the IGBT junction temperature measurement value at time n+1; The processor calculates the average absolute error of the junction temperature prediction based on the correction value of the IGBT junction temperature at time n+1 and the measured value of the IGBT junction temperature at time n+1. The formula for calculating the average absolute error of the junction temperature prediction is as follows:

[0013] in, The mean absolute error value for predicting the junction temperature value. This is the correction value for the IGBT junction temperature at time n+1. This represents the measured IGBT junction temperature at time n+1, where N is the number of time points. The processor calculates the mean square error of the junction temperature prediction based on the correction value of the IGBT junction temperature at time n+1 and the measured value of the IGBT junction temperature at time n+1. The formula for calculating the mean square error of the predicted junction temperature value is as follows:

[0014] in, The mean square error value for predicting the junction temperature. This is the correction value for the IGBT junction temperature at time n+1. This represents the measured IGBT junction temperature at time n+1, where N is the number of time points. If both the mean absolute error of the junction temperature prediction and the mean square error of the junction temperature prediction are within the preset standard range, then the processor determines that the long short-term memory network model meets the prediction standard.

[0015] Preferably, the method further includes: If both the mean absolute error and the mean square error of the junction temperature prediction exceed the preset standard range, the processor will retrain the long short-term memory network model.

[0016] The IGBT junction temperature prediction method provided in this application has the following beneficial effects: This application provides a method for predicting IGBT junction temperature. In this method, the processor inputs the IGBT collector current at time n, the IGBT collector voltage at time n, the cumulative effect of IGBT operating conditions at time n, the first IGBT power loss at time n, and the second IGBT power loss at time n into a trained Long Short-Term Memory (LSTM) network model to obtain the IGBT junction temperature at time n+1. This method considers the influencing factors of IGBT junction temperature more comprehensively. By training the original LSM network model with a more comprehensive training sample set (IGBT parameter sample set) corresponding to these influencing factors, the reliability and predictive stability of the LSM network model are significantly improved, thereby increasing the accuracy of IGBT junction temperature prediction.

[0017] Secondly, this application also provides an IGBT junction temperature prediction device, the device comprising: The acquisition module is used to acquire the collector current of the insulated gate bipolar transistor (IGBT) at time n, the collector voltage of the IGBT at time n, the cumulative effect of the IGBT operating conditions at time n, the first IGBT power loss at time n, and the second IGBT power loss at time n; wherein, the first IGBT power loss at time n is the power loss of the IGBT affected by wind speed at time n, and the second IGBT power loss at time n is the power loss of the IGBT not affected by wind speed at time n; The data processing module is used to input the IGBT collector current, IGBT collector voltage, IGBT operating condition cumulative effect, first IGBT power loss and second IGBT power loss at time n into the trained long short-term memory network model to obtain the IGBT junction temperature at time n+1. The Long Short-Term Memory (LSTM) network model is obtained as follows: Obtain an IGBT parameter sample set; wherein, the IGBT parameter sample set includes IGBT collector current samples, IGBT collector voltage samples, IGBT operating condition cumulative effect samples, first IGBT power loss samples, and second IGBT power loss samples. The IGBT parameter sample set is input into the original Long Short-Term Memory (LSTM) network model for training, thus obtaining the LTM network model. The IGBT junction temperature prediction device provided in this application embodiment has the same technical features as the IGBT junction temperature prediction method provided in the above embodiment, so it can also solve the same technical problems and achieve the same technical effects.

[0018] Thirdly, this application provides a computing device, including a memory and a processor; The memory stores one or more computer programs, the one or more computer programs including instructions; when the instructions are executed by the processor, the computing device performs the method as described in any one of the first aspects.

[0019] Fourthly, this application provides a computer-readable storage medium for storing a computer program for performing the method as described in any one of the first aspects.

[0020] Fifthly, this application provides a computer program product comprising one or more computer instructions, wherein when the computer instructions are executed by a computer, the computer performs the method as described in any one of the first aspects.

[0021] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing the application. The objectives and other advantages of this application are realized and obtained through the structures particularly pointed out in the description and the accompanying drawings.

[0022] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a schematic flowchart of an IGBT junction temperature prediction method provided in an embodiment of this application; Figure 2 Internal circuit diagram of the IGBT provided in the embodiments of this application; Figure 3 This is a schematic diagram of an IGBT junction temperature prediction device provided in an embodiment of this application; Figure 4 This is a schematic diagram of an electronic device structure provided in an embodiment of this application. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0026] To facilitate understanding of this embodiment, the embodiments of this application will be described in detail below.

[0027] This application provides a method for predicting IGBT junction temperature, such as... Figure 1 As shown, Figure 1 This is a schematic flowchart of an IGBT junction temperature prediction method provided in an embodiment of this application. The method includes the following steps: S101, the processor obtains the collector current of the insulated gate bipolar transistor (IGBT) at time n, the collector voltage of the IGBT at time n, the cumulative effect of the IGBT operating conditions at time n, the power loss of the first IGBT at time n, and the power loss of the second IGBT at time n.

[0028] Wherein, the power loss of the first IGBT at time n is the power loss of the IGBT at time n affected by wind speed, and the power loss of the second IGBT at time n is the power loss of the IGBT at time n not affected by wind speed.

[0029] Specifically, such as Figure 2 As shown, Figure 2 This is an internal circuit diagram of the IGBT provided in an embodiment of this application. Rs is the P-base region resistance, RN is the modulation resistor in the IGBT base region, and the IGBT collector current is calculated according to the following formula:

[0030] in, This refers to the collector current of the IGBT. This refers to the drain current of the IGBT. This represents the base current of the IGBT.

[0031]

[0032] in, This refers to the drain current of the IGBT. This refers to the base region current of the IGBT. This refers to the emitter-collector current gain of the IGBT.

[0033] The collector voltage of the IGBT is calculated using the following formula:

[0034] in, This is the collector voltage of the IGBT. For Boltzmann constant, This is the absolute temperature of the PN junction of the IGBT. This represents the charge of electrons in the PN junction of the IGBT. The bipolar diffusion coefficient, Intrinsic carrier concentration, For cell interval, The cell width is perpendicular to the cross-sectional direction. The width of the N-type drift region. The bipolar diffusion length, This refers to the collector current of the IGBT. For the electron mobility of the inversion layer, The characteristic capacitance of the oxide layer of the IGBT. For gate bias, This refers to the collector threshold voltage of the IGBT. For IGBT channel length, This is the coefficient related to the on-state pressure drop.

[0035] Because the output power and voltage of the wind power converter are affected by changes in wind speed, the power loss of the first IGBT changes, which in turn causes a change in the IGBT junction temperature. The effect of wind speed on the power loss of the first IGBT is usually reflected in the load changes caused by wind speed variations. Therefore, the power loss of the first IGBT can be calculated using the following formula:

[0036] in, For the first IGBT power loss, The first wind speed coefficient, The second wind speed coefficient, The third wind speed coefficient, The fourth wind speed coefficient, The wind speed is represented by the first wind speed coefficient, the second wind speed coefficient, the third wind speed coefficient, and the fourth wind speed coefficient, which are used to characterize the nonlinear effect of wind speed on the power loss of the first IGBT.

[0037] The second IGBT power loss includes on-state power loss and switching power loss. On-state power loss is the power loss generated by the on-state resistance of the IGBT when it is in the on-state, and it is calculated using the following formula:

[0038] in, For on-state power loss, This refers to the collector current of the IGBT. This is the collector voltage of the IGBT.

[0039] Switching power loss refers to the power loss caused by the asynchronous switching of the IGBT collector voltage and collector current during the IGBT's transition between the on and off states. The switching power loss is calculated using the following formula:

[0040] in, For switching power loss, This refers to the collector current of the IGBT. This is the collector voltage of the IGBT. The time of IGBT on state, This refers to the time of the IGBT's off state. This represents the switching frequency.

[0041] The second IGBT power loss is calculated based on the switching loss power and conduction loss power mentioned above. The formula for calculating the second IGBT power loss is as follows:

[0042] in, For switching power loss, For on-state power loss, This refers to the power loss of the second IGBT.

[0043] During the long-term operation of wind power converters, the IGBTs are gradually fatigued due to repeated switching operations, temperature fluctuations, load changes, and other operating conditions caused by the wind power converter. This fatigue affects the accuracy of predicting the IGBT junction temperature. To accurately predict the IGBT junction temperature, the cumulative effect of IGBT operating conditions must be considered. The cumulative effect of IGBT operating conditions is calculated using the following formula:

[0044] in, This is due to the cumulative effect of IGBT operating conditions. As the initial fatigue factor, =1, This is the fatigue growth rate coefficient, used to characterize the degree of increase in fatigue factor per unit time. This refers to the operating time of the wind power converter.

[0045] S102, the processor inputs the IGBT collector current, IGBT collector voltage, IGBT operating condition cumulative effect, first IGBT power loss, and second IGBT power loss at time n into the trained long short-term memory network model to obtain the IGBT junction temperature at time n+1.

[0046] Specifically, the trained Long Short-Term Memory (LSTM) network model can optimize the IGBT collector current, IGBT collector voltage, IGBT operating condition cumulative effect, first IGBT power loss, and second IGBT power loss at time n, yielding the following target formula:

[0047] in, This represents the IGBT junction temperature at time n+1. This is the initial junction temperature of the IGBT. The scaling factor for converting IGBT power loss into temperature increment. The coefficient is used to determine the nonlinear relationship between the power loss and temperature increment of the IGBT. This is the collector voltage of the IGBT. This refers to the collector current of the IGBT. The correction factor is used to characterize the impact of the first and second IGBT power losses on the IGBT junction temperature and the impact of wind speed on the heat dissipation of the IGBT PN junction. For the power loss of the second IGBT, For the first IGBT power loss, This is the coefficient representing the influence of the fatigue factor on the IGBT junction temperature. This is the cumulative effect of IGBT operating conditions.

[0048] The Long Short-Term Memory (LSTM) network model is obtained as follows: an IGBT parameter sample set is acquired; wherein, the IGBT parameter sample set includes IGBT collector current samples, IGBT collector voltage samples, IGBT operating condition cumulative effect samples, first IGBT power loss samples, and second IGBT power loss samples; the IGBT parameter sample set is input into the original LSM network model for training to obtain the LSM network model.

[0049] Specifically, to eliminate the influence of different dimensions and magnitudes on the training of the Long Short-Term Memory (LSTM) network model, the IGBT parameter sample set needs to be normalized, that is, converted into a [0,1] distribution. The IGBT parameter sample set serves as the temporal input of the model, forming the input layer of the original LSM network model, while the target formula forms the output layer. The original LSM network model learns the temporal relationships of the input IGBT parameter sample set, and the original LSM network model... , , , Make corrections (optimizations) to ensure the prediction accuracy of the trained Long Short-Term Memory network model. , , , The correction (optimization) process is as follows: (1) Divide the IGBT parameter sample set into training set, validation set and test set according to the proportions of 70%, 15% and 15% respectively; wherein, the training set is used to train the model and the validation set is used to adjust the model. , , , The test set is used to evaluate the performance of the Long Short-Term Memory network model.

[0050] (2) The output data of the forget gate is calculated according to the following formula:

[0051] in, Output data for the forget gate. [0,1], This is an activation function used to control the proportion of historical information forgotten. Here is the weight matrix for the forget gate. This represents the hidden state at time n-1. Let n be the IGBT parameter sample set at time n. This is a bias term for the Long Short-Term Memory (LSTM) network model, used to adjust the retention strength of cell state information during different time steps in the dynamic learning of the LTM network model.

[0052] (3) The input gate is calculated according to the following formula:

[0053] in, For input gate, Here is the weight matrix of the input gate. This represents the hidden state at time n-1. Let n be the IGBT parameter sample set at time n. For the bias term of the input gate, The bias term for candidate memory states. Let n be the candidate memory state at time n.

[0054] (4) The updated memory unit is calculated according to the following formula:

[0055] (5) Update hidden state.

[0056] (6) Based on the dynamic characteristics of the IGBT parameter sample set (time series) at time n-1, the optimized result is obtained. , , , .

[0057] This application provides an IGBT junction temperature prediction method. In this method, the processor inputs the IGBT collector current at time n, the IGBT collector voltage at time n, the IGBT operating condition cumulative effect at time n, the first IGBT power loss at time n, and the second IGBT power loss at time n into a trained Long Short-Term Memory (LSTM) network model to obtain the IGBT junction temperature at time n+1. Compared with the prior art, which directly measures the IGBT junction temperature using a temperature sensor installed outside the IGBT and predicts the IGBT junction temperature at the next time step based on the measured value, this method considers the influencing factors of the IGBT junction temperature more comprehensively. By training the original LSM network model with a more comprehensive training sample set (IGBT parameter sample set) corresponding to the influencing factors, the reliability and predictive stability of the LSM network model are significantly improved, thereby improving the accuracy of IGBT junction temperature prediction. This method enhances the adaptability of the Long Short-Term Memory (LSTM) network model, enabling the trained LTM model to better adapt to the complex and variable operating environment of IGBTs. It achieves real-time prediction of IGBT junction temperature, allowing operators to take timely protective measures against the IGBTs and preventing damage to the wind power converter due to excessively high junction temperatures. This ensures the safe operation of the wind power converter and improves its safety and reliability. Through long-term monitoring, data collection, and iterative optimization of IGBT collector current samples, IGBT collector voltage samples, IGBT cumulative effect samples, and power loss samples of the first and second IGBTs, this method ensures the effectiveness of the LTM network model throughout the entire operating cycle of the wind power converter, providing reliable technical support for the intelligent operation of wind power converters.

[0058] In one implementation, the processor acquires the IGBT junction temperature measurement values ​​and the IGBT operating condition cumulative effect at the first n time steps obtained from the inspection; and calculates the target operating condition cumulative effect correction factor at the (n+1)th time step based on the IGBT junction temperature value at the (n)th time step and the IGBT junction temperature measurement value at the (n)th time step. The formula for calculating the cumulative effect correction factor of the target operating condition at time n+1 is as follows:

[0059] in, This is the correction factor for the cumulative effect of the target operating condition at time n+1. This represents the IGBT junction temperature at time n+1. The measured value of the IGBT junction temperature at time n; The processor calculates the correction value of the IGBT junction temperature at time n+1 based on the IGBT junction temperature at time n+1, the target operating condition cumulative effect correction factor at time n+1, and the IGBT operating condition cumulative effect at time n. The formula for calculating the correction value of the IGBT junction temperature at time n+1 is as follows:

[0060] in, This is the correction value for the IGBT junction temperature at time n+1. This represents the IGBT junction temperature at time n+1. This is the correction factor for the cumulative effect of the target operating condition at time n+1. This represents the cumulative effect of IGBT operating conditions at time n.

[0061] Specifically, since the cumulative effect of IGBT operating conditions changes with the IGBT collector current, IGBT collector voltage, first IGBT power loss, and second IGBT power loss, the cumulative effect of IGBT operating conditions at time n+1 is updated (optimized) by the correction factor of the target cumulative effect at time n calculated from the IGBT junction temperature at time n and the measured IGBT junction temperature at time n. This corrects the IGBT junction temperature at time n+1, resulting in the corrected value of the IGBT junction temperature at time n+1.

[0062] This method combines measured data (IGBT junction temperature measurement) with predicted IGBT junction temperature (IGBT junction temperature at time n+1). By calculating the cumulative effect correction factor under the target operating condition, the IGBT junction temperature at time n+1 is further optimized. This makes the prediction results closer to the actual situation of the IGBT. By calculating the cumulative effect correction factor under the target operating condition, the impact of the cumulative effect of the IGBT operating condition on the IGBT junction temperature can be determined more accurately, thereby improving the accuracy of the IGBT junction temperature.

[0063] In one implementation, the processor acquires the IGBT junction temperature measurement value at time n+1; and calculates the average absolute error value for junction temperature prediction based on the correction value of the IGBT junction temperature value at time n+1 and the IGBT junction temperature measurement value at time n+1. The formula for calculating the mean absolute error of junction temperature prediction is as follows:

[0064] in, The mean absolute error value for predicting junction temperature. This is the correction value for the IGBT junction temperature at time n+1. The value is the IGBT junction temperature measurement at time n+1, where N is the number of time points. The processor calculates the mean square error of the junction temperature prediction based on the correction value of the IGBT junction temperature at time n+1 and the measured value of the IGBT junction temperature at time n+1. The formula for calculating the mean square error of the junction temperature prediction is as follows:

[0065] in, The mean square error value for predicting the junction temperature. This is the correction value for the IGBT junction temperature at time n+1. This represents the measured IGBT junction temperature at time n+1, where N is the number of time points. If both the mean absolute error and the mean square error of the junction temperature prediction are within the preset standard range, the processor determines that the long short-term memory network model meets the prediction standard; if both the mean absolute error and the mean square error of the junction temperature prediction exceed the preset standard range, the long short-term memory network model is retrained.

[0066] Specifically, the IGBT junction temperature measurement at time n+1 is an actual measurement with high accuracy. If the mean absolute error and mean square error of the junction temperature prediction are both within the preset standard range, it means that the correction value of the IGBT junction temperature at time n+1 is close to the measured value of the IGBT junction temperature at time n+1. The prediction accuracy of the trained long short-term memory network model is high, which means it meets the prediction standard.

[0067] If both the mean absolute error and the mean square error of the junction temperature prediction exceed the preset standard range, it indicates that the corrected value of the IGBT junction temperature at time n+1 deviates significantly from the measured value at time n+1. This means the trained Long Short-Term Memory (LSTM) network model is underfitting or has excessive noise, thus failing to meet the prediction standard. In this case, the LTM network model needs to be retrained.

[0068] In this method, the mean absolute error of the junction temperature prediction amplifies the error between the correction value of the IGBT junction temperature at time n+1 and the measured value of the IGBT junction temperature at time n+1. This helps to identify IGBT junction temperatures (measured values) that are significantly deviated from the actual values, thereby helping staff to judge the accuracy of the trained long short-term memory network model.

[0069] Based on the above embodiments of the IGBT junction temperature prediction method, this application also provides an IGBT junction temperature prediction device, such as... Figure 3 As shown, Figure 3 This is a schematic diagram of an IGBT junction temperature prediction device provided in an embodiment of this application. The device includes an acquisition module 31 and a data processing module 32. The functions of each module are as follows: The acquisition module 31 is used to acquire the collector current of the insulated gate bipolar transistor (IGBT) at time n, the collector voltage of the IGBT at time n, the cumulative effect of the IGBT operating conditions at time n, the first IGBT power loss at time n, and the second IGBT power loss at time n; wherein, the first IGBT power loss at time n is the power loss of the IGBT affected by wind speed at time n, and the second IGBT power loss at time n is the power loss of the IGBT not affected by wind speed at time n; The data processing module 32 is used to input the IGBT collector current, IGBT collector voltage, IGBT operating condition cumulative effect, first IGBT power loss and second IGBT power loss at time n into the trained long short-term memory network model to obtain the IGBT junction temperature value at time n+1. The Long Short-Term Memory (LSTM) network model is obtained as follows: Obtain an IGBT parameter sample set; wherein, the IGBT parameter sample set includes IGBT collector current sample, IGBT collector voltage sample, IGBT operating condition cumulative effect sample, first IGBT power loss sample and second IGBT power loss sample; The IGBT parameter sample set is input into the original long short-term memory network model for training, thus obtaining the long short-term memory network model.

[0070] Preferably, the acquisition module 31 is further configured to acquire the measured values ​​of IGBT junction temperature for the first n time moments and the cumulative effect of IGBT operating conditions for the first n time moments. The data processing module 32 is also used to calculate the target operating condition cumulative effect correction factor at time n+1 based on the IGBT junction temperature value at time n+1 and the measured IGBT junction temperature value at time n; and to calculate the correction value of the IGBT junction temperature value at time n+1 based on the IGBT junction temperature value at time n+1, the target operating condition cumulative effect correction factor at time n+1, and the IGBT operating condition cumulative effect at time n.

[0071] Preferably, the data processing module 32 is specifically used to calculate the correction value of the IGBT junction temperature at time n+1 according to the following formula:

[0072] in, This is the correction value for the IGBT junction temperature at time n+1. This represents the IGBT junction temperature at time n+1. This is the correction factor for the cumulative effect of the target operating condition at time n+1. This represents the cumulative effect of IGBT operating conditions at time n.

[0073] Preferably, the data processing module 32 is specifically used to calculate the target operating condition cumulative effect correction factor at time n+1 according to the following formula:

[0074] in, This is the correction factor for the cumulative effect of the target operating condition at time n+1. This represents the IGBT junction temperature at time n+1. This represents the measured junction temperature of the IGBT at time n.

[0075] Preferably, the acquisition module 31 is further configured to acquire the IGBT junction temperature measurement value at time n+1; The data processing module 32 is also used to calculate the average absolute error value of the junction temperature prediction based on the correction value of the IGBT junction temperature value at time n+1 and the measured value of the IGBT junction temperature at time n+1. The formula for calculating the mean absolute error of junction temperature prediction is as follows:

[0076] in, The mean absolute error value for predicting junction temperature. This is the correction value for the IGBT junction temperature at time n+1. Let N be the IGBT junction temperature measurement value at time n+1, where N is the number of time points. Based on the correction value of the IGBT junction temperature at time n+1 and the IGBT junction temperature measurement value at time n+1, the mean square error of the junction temperature prediction is calculated. The formula for calculating the mean square error of the junction temperature prediction is as follows:

[0077] in, The mean square error value for predicting the junction temperature. This is the correction value for the IGBT junction temperature at time n+1. Let N be the IGBT junction temperature measurement value at time n+1, where N is the number of time points. If both the average absolute error and the mean square error of the junction temperature prediction are within the preset standard range, then the long short-term memory network model is determined to meet the prediction standard.

[0078] Preferably, the data processing module 32 is further configured to retrain the long short-term memory network model if both the average absolute error value and the mean square error value of the junction temperature prediction exceed the preset standard range.

[0079] The IGBT junction temperature prediction device provided in this application embodiment has the same technical features as the IGBT junction temperature prediction method provided in the above embodiment, so it can also solve the same technical problems and achieve the same technical effects.

[0080] This application also provides a computing device. For example... Figure 4 As shown in the figure, this is a schematic diagram of a computing device provided in an embodiment of this application. The computing device 400 includes a bus 401, a processor 402, a communication interface 403, and a memory 404. The processor 402, the memory 404, and the communication interface 403 communicate with each other via the bus 401.

[0081] Bus 401 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0082] Processor 402 can be any one or more of the following processors: central processing unit (CPU), graphics processing unit (GPU), microprocessor (MP), or digital signal processor (DSP).

[0083] Communication interface 403 is used for communication with external devices. Memory 404 may include volatile memory, such as random access memory (RAM). Memory 404 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0084] The memory 404 stores executable code, and the processor 402 executes the executable code to perform the aforementioned IGBT junction temperature prediction method.

[0085] This application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium capable of being stored by a computing device, or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive). The computer-readable storage medium includes instructions that instruct the computing device to perform the above-described method.

[0086] This application also provides a computer program product comprising one or more computer instructions. When the computer instructions are loaded and executed on a computing device, all or part of the processes or functions described in this application are generated.

[0087] The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, or data center to another website, computer, or data center via wired (e.g., coaxial cable, fiber optic) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0088] When the computer program product is executed by a computer, the computer performs any of the aforementioned IGBT junction temperature prediction methods. The computer program product can be a software installation package; when any of the aforementioned IGBT junction temperature prediction methods needs to be used, the computer program product can be downloaded and executed on the computer.

[0089] The descriptions of the processes or structures corresponding to the above figures each have their own emphasis. For parts of a process or structure that are not described in detail, please refer to the relevant descriptions of other processes or structures.

[0090] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be covered within the scope of protection of this application.

Claims

1. A method for predicting IGBT junction temperature, characterized in that, The method includes: The processor acquires the collector current of the Insulated Gate Bipolar Transistor (IGBT) at time n, the collector voltage of the IGBT at time n, the cumulative effect of the IGBT operating conditions at time n, the power loss of the first IGBT at time n, and the power loss of the second IGBT at time n; wherein, the power loss of the first IGBT at time n is the power loss of the IGBT affected by wind speed at time n, calculated according to the following formula: in, For the first IGBT power loss, The first wind speed coefficient, The second wind speed coefficient, The third wind speed coefficient, The fourth wind speed coefficient, Let the wind speed be the second IGBT power loss at time n. The second IGBT power loss at time n is the power loss of the IGBT before the wind speed affects it. The second IGBT power loss includes conduction loss power and switching loss power. The conduction loss power is calculated according to the following formula: in, For on-state power loss, This is the collector current of the IGBT. The IGBT collector voltage is used; the switching loss power is calculated according to the following formula: in, For switching power loss, The time of IGBT on state, This refers to the time of the IGBT's off state. The switching frequency is given; the power loss of the second IGBT is calculated using the following formula: in, For the power loss of the second IGBT; The cumulative effect of IGBT operating conditions is calculated according to the following formula: in, This is due to the cumulative effect of IGBT operating conditions. As the initial fatigue factor, =1, This is the fatigue growth rate coefficient, used to characterize the degree of increase in fatigue factor per unit time. This refers to the operating time of the wind power converter. The processor inputs the IGBT collector current, IGBT collector voltage, IGBT operating condition cumulative effect, first IGBT power loss, and second IGBT power loss at time n into the trained long short-term memory network model to obtain the IGBT junction temperature at time n+1. The Long Short-Term Memory (LSTM) network model uses the following objective formula as its output layer: in, This represents the IGBT junction temperature at time n+1. This is the initial junction temperature of the IGBT. The scaling factor for converting IGBT power loss into temperature increment. The coefficient is used to determine the nonlinear relationship between the power loss and temperature increment of the IGBT. The correction factor is used to characterize the impact of the first and second IGBT power losses on the IGBT junction temperature and the impact of wind speed on the heat dissipation of the IGBT PN junction. This is the coefficient representing the influence of the fatigue factor on the IGBT junction temperature. The Long Short-Term Memory (LSTM) network model learns the temporal relationship of the input IGBT parameter sample set and performs... , , , Make corrections; The Long Short-Term Memory (LSTM) network model is obtained as follows: Obtain an IGBT parameter sample set; wherein, the IGBT parameter sample set includes IGBT collector current samples, IGBT collector voltage samples, IGBT operating condition cumulative effect samples, first IGBT power loss samples, and second IGBT power loss samples. The IGBT parameter sample set is input into the original long short-term memory network model for training to obtain the long short-term memory network model.

2. The IGBT junction temperature prediction method according to claim 1, characterized in that, The method further includes: The processor acquires the IGBT junction temperature measurement values ​​and the cumulative effect of IGBT operating conditions at the first n time steps obtained from the inspection. The processor calculates the target operating condition cumulative effect correction factor at time n+1 based on the IGBT junction temperature at time n+1 and the measured IGBT junction temperature at time n. The processor calculates the correction value of the IGBT junction temperature at time n+1 based on the IGBT junction temperature at time n+1, the target operating condition cumulative effect correction factor at time n+1, and the IGBT operating condition cumulative effect at time n.

3. The IGBT junction temperature prediction method according to claim 2, characterized in that, The step of the processor calculating the correction value of the IGBT junction temperature at time n+1 based on the IGBT junction temperature at time n+1, the target operating condition cumulative effect correction factor at time n+1, and the IGBT operating condition cumulative effect at time n includes: The formula for calculating the correction value of the IGBT junction temperature at time n+1 is as follows: in, This is the correction value for the IGBT junction temperature at time n+1. This represents the IGBT junction temperature at time n+1. This is the cumulative effect correction factor for the target operating condition at time n+1.

4. The IGBT junction temperature prediction method according to claim 2, characterized in that, The step of the processor calculating the target operating condition cumulative effect correction factor at time n+1 based on the IGBT junction temperature at time n+1 and the measured IGBT junction temperature at time n includes: The formula for calculating the cumulative effect correction factor of the target operating condition at time n+1 is as follows: in, This is the correction factor for the cumulative effect of the target operating condition at time n+1. This represents the IGBT junction temperature at time n+1. The measured value of the IGBT junction temperature at time n; The number of moments.

5. The IGBT junction temperature prediction method according to claim 2, characterized in that, The method further includes: The processor acquires the IGBT junction temperature measurement value at time n+1; The processor calculates the average absolute error of the junction temperature prediction based on the correction value of the IGBT junction temperature at time n+1 and the measured value of the IGBT junction temperature at time n+1. The formula for calculating the average absolute error of the predicted junction temperature value is as follows: in, The mean absolute error value for predicting the junction temperature value. This is the correction value for the IGBT junction temperature at time n+1. This represents the measured IGBT junction temperature at time n+1, where N is the number of time points. The processor calculates the mean square error of the junction temperature prediction based on the correction value of the IGBT junction temperature at time n+1 and the measured value of the IGBT junction temperature at time n+1. The formula for calculating the mean square error of the predicted junction temperature value is as follows: in, The mean square error value for predicting the junction temperature; If both the mean absolute error of the junction temperature prediction and the mean square error of the junction temperature prediction are within the preset standard range, then the processor determines that the long short-term memory network model meets the prediction standard.

6. The IGBT junction temperature prediction method according to claim 5, characterized in that, The method further includes: If both the mean absolute error and the mean square error of the junction temperature prediction exceed the preset standard range, the processor will retrain the long short-term memory network model.

7. An IGBT junction temperature prediction device, characterized in that, The device includes: The acquisition module is used to acquire the collector current of the Insulated Gate Bipolar Transistor (IGBT) at time n, the collector voltage of the IGBT at time n, the cumulative effect of the IGBT operating conditions at time n, the power loss of the first IGBT at time n, and the power loss of the second IGBT at time n; wherein, the power loss of the first IGBT at time n is the power loss of the IGBT affected by wind speed at time n, which is calculated according to the following formula: in, For the first IGBT power loss, The first wind speed coefficient, The second wind speed coefficient, The third wind speed coefficient, The fourth wind speed coefficient, Let the wind speed be the second IGBT power loss at time n. The second IGBT power loss at time n is the power loss of the IGBT before the wind speed affects it. The second IGBT power loss includes conduction loss power and switching loss power. The conduction loss power is calculated according to the following formula: in, For on-state power loss, This is the collector current of the IGBT. The IGBT collector voltage is used; the switching loss power is calculated according to the following formula: in, For switching power loss, The time of IGBT on state, This refers to the time of the IGBT's off state. The switching frequency is given; the power loss of the second IGBT is calculated using the following formula: in, For the power loss of the second IGBT; The cumulative effect of IGBT operating conditions is calculated according to the following formula: in, This is due to the cumulative effect of IGBT operating conditions. As the initial fatigue factor, =1, This is the fatigue growth rate coefficient, used to characterize the degree of increase in fatigue factor per unit time. This refers to the operating time of the wind power converter. The data processing module is used to input the IGBT collector current, IGBT collector voltage, IGBT operating condition cumulative effect, first IGBT power loss and second IGBT power loss at time n into the trained long short-term memory network model to obtain the IGBT junction temperature at time n+1. The Long Short-Term Memory (LSTM) network model uses the following objective formula as its output layer: in, This represents the IGBT junction temperature at time n+1. This is the initial junction temperature of the IGBT. The scaling factor for converting IGBT power loss into temperature increment. The coefficient is used to determine the nonlinear relationship between the power loss and temperature increment of the IGBT. The correction factor is used to characterize the impact of the first and second IGBT power losses on the IGBT junction temperature and the impact of wind speed on the heat dissipation of the IGBT PN junction. This is the coefficient representing the influence of the fatigue factor on the IGBT junction temperature. The Long Short-Term Memory (LSTM) network model learns the temporal relationship of the input IGBT parameter sample set and performs... , , , Make corrections; The Long Short-Term Memory (LSTM) network model is obtained as follows: Obtain an IGBT parameter sample set; wherein, the IGBT parameter sample set includes IGBT collector current samples, IGBT collector voltage samples, IGBT operating condition cumulative effect samples, first IGBT power loss samples, and second IGBT power loss samples. The IGBT parameter sample set is input into the original long short-term memory network model for training to obtain the long short-term memory network model.

8. A computing device, characterized in that, Including memory and processor; The memory stores one or more computer programs, the one or more computer programs including instructions; when the instructions are executed by the processor, the computing device performs the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program for performing the method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, The computer program product includes one or more computer instructions that, when executed by a computer, perform the method as described in any one of claims 1 to 6.

Images