Method for suppressing voltage peak of silicon carbide field effect transistor

Abstract

The present application relates to the technical field of smart grid, and particularly relates to a voltage peak suppression method of silicon carbide field effect tube. Drain-source voltage, drain current, switching frequency and environment temperature of the silicon carbide field effect tube are acquired, and active miller clamp diode switching delay and off resistance are acquired; based on a neural network, the drain-source voltage is predicted to acquire a predicted drain-source voltage sequence; the silicon carbide field effect tube is simulated by using the predicted drain-source voltage to acquire an active miller clamp prediction accuracy; the predicted drain-source voltage fluctuation is acquired, and based on the active miller clamp prediction accuracy and the predicted drain-source voltage fluctuation, the active miller clamp diode switching delay is optimized and adjusted to suppress the voltage peak of the silicon carbide field effect tube. Through implementation of the present application, the voltage peak of the silicon carbide field effect tube can be precisely and adaptively suppressed.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, and in particular to a method for suppressing voltage peaks in silicon carbide field-effect transistors. Background Technology

[0002] In high-voltage, high-power scenarios such as smart grids, silicon carbide (SiC) MOSFETs are widely used due to their performance advantages. The industry commonly uses active Miller clamping technology to suppress voltage peaks during switching. The traditional approach is to preset the switching delay and turn-off resistance of the clamping diode according to the manual parameters or fixed operating conditions. However, this type of technology has significant problems: fixed parameters cannot adapt to dynamic changes in operating conditions, resulting in poor suppression effects; it also lacks voltage prediction capabilities, has a lag in response to sudden voltage increases, and the device is susceptible to high stress; relying on trial and error to optimize parameters is inefficient and makes it difficult to find the optimal combination. Summary of the Invention

[0003] This invention addresses the technical problem of poor voltage peak suppression during switching of silicon carbide field-effect transistors in the prior art by providing a method for suppressing voltage peaks in silicon carbide field-effect transistors.

[0004] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0005] This invention provides a method for suppressing voltage peaks in a silicon carbide (SiC) field-effect transistor (FET), comprising: acquiring the drain-source voltage, drain current, switching frequency, and ambient temperature of the SiC FET, and acquiring the switching delay and turn-off resistance of the active Miller clamp diode; predicting the drain-source voltage based on a neural network to obtain a predicted drain-source voltage sequence, simulating the SiC FET using the predicted drain-source voltage, and obtaining the prediction accuracy of the active Miller clamp; acquiring the predicted drain-source voltage fluctuation, and optimizing the switching delay of the active Miller clamp diode based on the prediction accuracy and the predicted drain-source voltage fluctuation to suppress the voltage peaks of the SiC FET.

[0006] Optionally, based on a neural network, the drain-source voltage is predicted to obtain the predicted drain-source voltage, including: obtaining the historical drain-source voltage, historical drain current, historical switching frequency, and historical ambient temperature of the same type of silicon carbide field-effect transistor; constructing a drain-source voltage predictor based on the neural network; training the drain-source voltage predictor with the historical drain-source voltage, historical drain current, historical switching frequency, and historical ambient temperature until convergence; and using the drain-source voltage predictor, predicting the drain-source voltage based on the obtained drain-source voltage, drain current, switching frequency, and ambient temperature to obtain the predicted drain-source voltage.

[0007] Optionally, the predicted drain-source voltage is used to simulate the silicon carbide field-effect transistor to obtain the prediction accuracy of active Miller clamping, including:

[0008] The predicted drain-source voltage, switching frequency, ambient temperature, and switching delay of the active Miller clamp diode are used to perform simulation to obtain the prediction accuracy of the active Miller clamp.

[0009] Optionally, the predicted drain-source voltage fluctuation is obtained, and based on the prediction accuracy of the active Miller clamp and the predicted drain-source voltage fluctuation, the switching delay of the active Miller clamp diode is optimized and adjusted to suppress the voltage peak of the silicon carbide field-effect transistor. This includes: determining whether the prediction accuracy of the active Miller clamp is less than or equal to a prediction accuracy threshold; determining whether the predicted drain-source voltage fluctuation is less than or equal to a fluctuation threshold; if both are satisfied, the switching delay of the active Miller clamp diode is used to suppress the voltage peak of the silicon carbide field-effect transistor; if either is not satisfied, the switching delay of the active Miller clamp diode is optimized and adjusted to obtain an optimized active Miller clamp diode switching delay for suppressing the voltage peak of the silicon carbide field-effect transistor.

[0010] The acquisition of the predicted drain-source voltage fluctuation includes: calculating the standard deviation and mean of the predicted drain-source voltage; and calculating the ratio of the standard deviation and mean of the predicted drain-source voltage to obtain the predicted drain-source voltage fluctuation.

[0011] If any one of these conditions is not met, the turn-off resistor is optimized to obtain an optimized active Miller clamp diode switching delay for voltage peak suppression of the silicon carbide field-effect transistor (MOSFET). This includes: adjusting the turn-off resistor to optimize the active Miller clamp diode switching delay, obtaining an optimized turn-off resistor and an optimized active Miller clamp diode switching delay; using the optimized turn-off resistor to suppress the voltage peak of the MOSFET and obtaining the optimized drain-source voltage fluctuation after voltage peak suppression; when the optimized drain-source voltage fluctuation is greater than a fluctuation threshold, the turn-off resistor is further optimized until the optimized drain-source voltage fluctuation is less than or equal to the fluctuation threshold, and then the optimized turn-off resistor is used to obtain an optimized active Miller clamp diode switching delay for voltage peak suppression of the MOSFET.

[0012] The process of adjusting the turn-off resistor to optimize the switching delay of the active Miller clamp diode, and obtaining the optimized turn-off resistor and the optimized active Miller clamp diode switching delay, includes: obtaining the turn-off resistor adjustment range; obtaining the turn-off resistor adjustment step size based on the ratio of the predicted drain-source voltage fluctuation to the fluctuation threshold; obtaining the corrected turn-off resistor adjustment step size based on the ratio of the active Miller clamp prediction accuracy to the fluctuation threshold and the turn-off resistor adjustment range; iteratively optimizing the turn-off resistor based on the corrected turn-off resistor adjustment step size and the turn-off resistor adjustment range to obtain the optimized turn-off resistor; and obtaining the optimized active Miller clamp diode switching delay based on the optimized turn-off resistor.

[0013] By implementing this invention, it is possible to obtain the drain-source voltage, drain current, switching frequency, and ambient temperature of a silicon carbide field-effect transistor, as well as the switching delay and turn-off resistance of an active Miller clamp diode. The collected parameters cover the key characteristics of the device's operating status, external environment, and control structure, and can comprehensively reflect the working scenario of the silicon carbide field-effect transistor, laying a data foundation for subsequent accurate suppression of voltage peaks.

[0014] Based on neural networks, the drain-source voltage is predicted to obtain a predicted drain-source voltage sequence. The predicted drain-source voltage is then used to simulate a silicon carbide field-effect transistor to obtain the prediction accuracy of active Miller clamping. By obtaining the prediction accuracy of active Miller clamping through simulation, the effectiveness of the current clamping strategy can be evaluated in advance, avoiding voltage peak exceeding the limit due to strategy failure when directly applied to actual devices, thus reducing the risk of device damage.

[0015] The predicted drain-source voltage fluctuation is obtained, and the switching delay of the active Miller clamp diode is optimized and adjusted based on the prediction accuracy and the predicted drain-source voltage fluctuation of the active Miller clamp. This is used to suppress the voltage peak of the silicon carbide field-effect transistor. The switching delay is dynamically optimized based on dual indicators. Compared with fixed parameter control, it can adapt to different operating conditions and ensure the adaptability and continuity of the voltage peak suppression effect.

[0016] In summary, by implementing this invention, precise and adaptive suppression of voltage peak values ​​in silicon carbide field-effect transistors can be achieved. Attached Figure Description

[0017] Figure 1 A schematic flowchart of a voltage peak suppression method for a silicon carbide field-effect transistor provided by the present invention;

[0018] Figure 2 This is a schematic diagram of the process for obtaining the predicted drain-source voltage in a voltage peak suppression method for a silicon carbide field-effect transistor provided by the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0021] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0022] Example 1, as Figure 1 As shown, this embodiment of the invention provides a method for suppressing voltage peaks in a silicon carbide field-effect transistor, comprising:

[0023] S100: Obtain the drain-source voltage, drain current, switching frequency, and ambient temperature of the silicon carbide field-effect transistor, and obtain the switching delay and turn-off resistance of the active Miller clamp diode.

[0024] S200: Based on a neural network, the drain-source voltage is predicted, the predicted drain-source voltage sequence is obtained, and the predicted drain-source voltage is used to simulate the silicon carbide field-effect transistor to obtain the prediction accuracy of active Miller clamp.

[0025] S300: Obtain the predicted drain-source voltage fluctuation, and based on the prediction accuracy of the active Miller clamp and the predicted drain-source voltage fluctuation, optimize and adjust the switching delay of the active Miller clamp diode to suppress the voltage peak of the silicon carbide field-effect transistor.

[0026] In step S100 of this application embodiment, it is necessary to obtain the drain-source voltage, drain current, switching frequency, and ambient temperature of the silicon carbide field-effect transistor, and to obtain the switching delay and turn-off resistance of the active Miller clamp diode.

[0027] Step S100 is the basic data acquisition step for the entire silicon carbide field-effect transistor voltage peak suppression method. Its core purpose is to provide comprehensive and accurate raw data support for subsequent drain-source voltage prediction, active Miller clamp simulation verification and parameter optimization.

[0028] To achieve the above objectives, it is first necessary to clarify the scope and source of the parameters to be collected. Specifically, the parameters to be collected are divided into two categories: "silicon carbide field-effect transistor (SiC) operating parameters" and "active Miller clamp circuit parameters." Specifically, SiC operating parameters include drain-source voltage, drain current, switching frequency, and ambient temperature; active Miller clamp circuit parameters include the switching delay and turn-off resistance of the active Miller clamp diode.

[0029] In this embodiment, the drain-source voltage (Vds) reflects the voltage state between the drain and source of the transistor and is the core monitoring object for voltage peak suppression; the drain current (Id) reflects the current load of the drain of the transistor and affects the dynamic change of the drain-source voltage; the switching frequency (f) determines the switching rate of the transistor and is an important cause of drain-source voltage fluctuations; the ambient temperature (T) affects the on-resistance, carrier mobility and other electrical characteristics of the transistor, indirectly changing the drain-source voltage; the active Miller clamp diode switching delay (td) determines the response speed of the clamp diode to turn on / off and directly affects the clamping effect; the turn-off resistor (Roff) controls the current change rate during the turn-off process of the transistor and is the core adjustment parameter for optimizing the clamp switching delay.

[0030] In actual data acquisition, the "operating parameters of silicon carbide field-effect transistors," namely drain-source voltage, drain current, switching frequency, and ambient temperature, can be acquired in real time through hardware sensors or detection circuits. For example, a voltage sensor can be used to measure the drain-source voltage, a current sensor to measure the drain current, a temperature sensor to measure the ambient temperature, and the switching frequency can be directly read from the output signal of the control chip.

[0031] For the "Active Miller Clamp Circuit Parameters," the switching delay of the active Miller clamp diode can be obtained by measuring the time difference between the clamp diode control signal and the actual turn-on / turn-off signal using an oscilloscope, or by directly using the typical delay parameter in the diode device datasheet as the initial value. For the turn-off resistor, the nominal value of the turn-off resistor in the circuit design can be directly read as the initial value.

[0032] like Figure 2 As shown, in step S200 of this embodiment, the drain-source voltage is predicted based on a neural network to obtain the predicted drain-source voltage, including:

[0033] Obtain historical drain-source voltage, historical drain current, historical switching frequency, and historical ambient temperature of the same type of silicon carbide field-effect transistor;

[0034] A drain-source voltage predictor is constructed based on a neural network.

[0035] The drain-source voltage predictor is trained using the historical drain-source voltage, historical drain current, historical switching frequency, and historical ambient temperature until convergence.

[0036] The drain-source voltage predictor is used to predict the drain-source voltage based on the acquired drain-source voltage, drain current, switching frequency, and ambient temperature, thereby obtaining the predicted drain-source voltage.

[0037] In this embodiment of the application, the above-mentioned subdivision steps in step S200 are the core modeling and reasoning links for predicting the drain-source voltage of silicon carbide field-effect transistors based on neural networks. The core purpose is to build a high-precision drain-source voltage predictor through a data-driven approach, so as to provide a reliable prediction basis for subsequent simulation verification and parameter optimization of active Miller clamping.

[0038] To achieve the above objectives, it is first necessary to obtain the historical drain-source voltage, historical drain current, historical switching frequency, and historical ambient temperature of the same type of silicon carbide field-effect transistor (MOSFET). This data can be obtained through experimental testing of the same type of device under different operating conditions, long-term operation logs, or simulation data accumulation. It needs to cover a wide range of operating conditions, such as different load currents, switching frequencies, and temperature conditions, to ensure the representativeness of the data. After cleaning, normalizing, and timing-aligning the raw data, it can be used as training data for the drain-source voltage predictor.

[0039] Next, a drain-source voltage predictor needs to be constructed based on a neural network, and trained using the historical drain-source voltage, historical drain current, historical switching frequency, and historical ambient temperature until convergence. Depending on the task type of the drain-source voltage predictor, a Long Short-Term Memory (LSTM) network can be used to build it. This network can effectively capture the dynamic changes of operating parameters over time and adapt to the timing correlation characteristics between the drain-source voltage and drain current, switching frequency, and ambient temperature of the silicon carbide field-effect transistor.

[0040] The drain-source voltage predictor consists of an input layer, a hidden layer, and an output layer. The input layer contains four neurons, corresponding to the input features of drain-source voltage, drain current, switching frequency, and ambient temperature. The hidden layer comprises two LSTM units: the first LSTM unit contains 64 neurons using the ReLU activation function, and the second LSTM unit contains 32 neurons using the ReLU activation function. The output layer consists of one neuron, which outputs the predicted drain-source voltage using a linear activation function.

[0041] In terms of parameter settings for the drain-source voltage predictor, the batch size is 32; the learning rate is 0.001; the time step is 10, which means using historical data from 10 consecutive time steps to predict the drain-source voltage at the next time step; the Dropout coefficient is 0.2; and the L2 regularization coefficient is 0.0001.

[0042] The training samples for the drain-source voltage predictor are derived from the historical operating data of the aforementioned silicon carbide field-effect transistors (MOSFETs) under different operating conditions. Each training sample includes historical drain-source voltage, historical drain current, historical switching frequency, and historical ambient temperature. At least 10,000 training sample data sets are collected, covering various operating conditions, and divided into a validation set and a training set at a 2:8 ratio. Each training sample contains time-series data for 10 consecutive time points. During training, the maximum number of training rounds is set to 1000. Training can be terminated if the convergence criterion is met early. When the mean square error (MSE) of the validation set does not decrease for 20 consecutive rounds, and the MSE value is below 0.01, corresponding to a voltage prediction error of less than 1%, the model is considered converged, training is stopped, and the drain-source voltage predictor is obtained.

[0043] In step S200 of this application embodiment, the predicted drain-source voltage is used to simulate the silicon carbide field-effect transistor to obtain the prediction accuracy of active Miller clamping, including:

[0044] The predicted drain-source voltage, switching frequency, ambient temperature, and switching delay of the active Miller clamp diode are used to perform simulation to obtain the prediction accuracy of the active Miller clamp.

[0045] In this embodiment of the application, the above-mentioned subdivision steps in step S200 are key verification links connecting drain-source voltage prediction and subsequent parameter optimization. The core purpose is to evaluate the suppression effect of the active Miller clamp circuit on the predicted drain-source voltage through simulation, and to provide a quantitative basis for determining whether the circuit parameters need to be adjusted.

[0046] To achieve the above objectives, it is first necessary to define the simulation input parameters. In this embodiment, the simulation input parameters are the aforementioned predicted drain-source voltage, switching frequency, ambient temperature, and active Miller clamp diode switching delay. These parameters together constitute the boundary conditions of the simulation, ensuring that the simulated scenario is consistent with the actual operating conditions.

[0047] Next, a joint simulation model of the silicon carbide field-effect transistor and the active Miller clamping circuit needs to be built using power electronics simulation software such as PSpice and Saber. The model needs to include the switching characteristics of the silicon carbide field-effect transistor, the on / off behavior of the active Miller clamping diode, and the external circuit topology.

[0048] Furthermore, the above input parameters need to be imported into the co-simulation model to simulate the dynamic process of the active Miller clamp diode operating according to its switching delay characteristics under the predicted drain-source voltage conditions of the silicon carbide field-effect transistor, with a focus on recording the actual drain-source voltage waveform after clamping.

[0049] The degree of agreement between the simulated actual drain-source voltage after clamping and the "target suppression value of the predicted drain-source voltage" is used as the metric. The target suppression value of the predicted drain-source voltage can be a preset safe voltage threshold. Specifically, the active Miller clamp prediction accuracy can be calculated as: Active Miller clamp prediction accuracy = (1 - |actual drain-source voltage after clamping - target suppression value| / target suppression value) × 100%, where the target suppression value is preset according to the safe operating range of the silicon carbide field-effect transistor. Optionally, the average of the active Miller clamp prediction accuracies from multiple simulations can be used as the final Miller clamp prediction accuracy.

[0050] In step S300 of this application embodiment, the predicted drain-source voltage fluctuation is obtained, and based on the prediction accuracy of the active Miller clamp and the predicted drain-source voltage fluctuation, the switching delay of the active Miller clamp diode is optimized and adjusted to suppress the voltage peak of the silicon carbide field-effect transistor, including:

[0051] Determine whether the prediction accuracy of the active Miller clamp is less than or equal to the prediction accuracy threshold;

[0052] Determine whether the predicted drain-source voltage fluctuation is less than or equal to the fluctuation threshold;

[0053] If both conditions are met, then an active Miller clamp diode switching delay is used to suppress the voltage peak of the silicon carbide field-effect transistor.

[0054] If any one of these conditions is not met, the switching delay of the active Miller clamp diode is optimized and adjusted to obtain an optimized active Miller clamp diode switching delay, thereby suppressing the voltage peak of the silicon carbide field-effect transistor.

[0055] In this embodiment of the application, the purpose of step S300 is to realize the core decision-making link of active Miller clamp parameter optimization and voltage peak suppression. Its core purpose is to dynamically optimize the switching delay of the active Miller clamp diode by judging the dual thresholds of active Miller clamp prediction accuracy and predicted drain-source voltage fluctuation, so as to ensure that the drain-source voltage peak of the silicon carbide field-effect transistor is stably suppressed within a safe range.

[0056] To achieve the above objectives, it is first necessary to obtain the predicted drain-source voltage fluctuation.

[0057] In step S300 of this application embodiment, the acquisition of the predicted drain-source voltage fluctuation includes:

[0058] Calculate the standard deviation and mean of the predicted drain-source voltage;

[0059] Calculate the ratio of the predicted drain-source voltage standard deviation to the mean to obtain the predicted drain-source voltage fluctuation.

[0060] Specifically, the first step is to take the "predicted drain-source voltage" output by the drain-source voltage predictor in step S200 as the calculation object, select the predicted drain-source voltage sequence of multiple consecutive time points, such as 100 sampling time points, to ensure that the data has temporal representativeness, then calculate the average value of the sequence using the arithmetic mean formula, and then calculate the dispersion of the sequence using the statistical standard deviation formula.

[0061] Furthermore, it is necessary to calculate the ratio of the predicted drain-source voltage standard deviation to the mean to obtain the predicted drain-source voltage fluctuation. Specifically, the method of "predicted drain-source voltage fluctuation = standard deviation ÷ mean × 100%" can be directly used to convert the ratio of standard deviation to mean into a percentage form, eliminating the influence of dimensions and making the fluctuation a relative indicator that can be compared across operating conditions. The smaller the predicted drain-source voltage fluctuation value, the more stable the predicted drain-source voltage changes in time; the larger the value, the more severe the voltage fluctuation, which may exceed the stable operating range of the silicon carbide field-effect transistor, requiring improvement through optimization of the active Miller clamp diode switching delay.

[0062] Furthermore, it is necessary to determine whether the prediction accuracy of the active Miller clamp is less than or equal to the prediction accuracy threshold; and whether the predicted drain-source voltage fluctuation is less than or equal to the fluctuation threshold. If both conditions are met, then the voltage peak of the silicon carbide field-effect transistor is suppressed by using the active Miller clamp diode switching delay.

[0063] First, a "prediction accuracy threshold" needs to be preset, such as 95%, which needs to be set according to the device safety requirements.

[0064] Next, check if the "active Miller clamp prediction accuracy is less than or equal to the prediction accuracy threshold". If the active Miller clamp prediction accuracy is less than or equal to the prediction accuracy threshold, it indicates insufficient clamping accuracy, which needs optimization; if it is higher than the prediction accuracy threshold, it indicates that the accuracy meets the standard. Then, check if the "predicted drain-source voltage fluctuation is less than or equal to the fluctuation threshold". If the predicted drain-source voltage fluctuation is higher than the fluctuation threshold, it indicates poor voltage stability, which needs optimization; if it is less than or equal to the fluctuation threshold, it indicates that the stability meets the standard.

[0065] If both conditions are met—that is, the prediction accuracy of the active Miller clamp is less than or equal to the prediction accuracy threshold, and the predicted drain-source voltage fluctuation is less than or equal to the fluctuation threshold—then no parameter adjustment is needed. The current "active Miller clamp diode switching delay" can be used to control the clamping circuit, and drain-source voltage peak suppression can be achieved through the timely switching on / off of the diode.

[0066] In step S300 of this application embodiment, if any one of the conditions is not met, the turn-off resistor is optimized and adjusted to obtain an optimized active Miller clamp diode switching delay, and voltage peak suppression of the silicon carbide field-effect transistor is performed, including:

[0067] Adjust the turn-off resistor to optimize the switching delay of the active Miller clamp diode, and obtain the optimized turn-off resistor and the optimized switching delay of the active Miller clamp diode.

[0068] The optimized turn-off resistor is used to suppress the voltage peak of the silicon carbide field-effect transistor, and the optimized drain-source voltage fluctuation after voltage peak suppression is obtained.

[0069] When the optimized drain-source voltage fluctuation is greater than the fluctuation threshold, the turn-off resistor is further optimized until the optimized drain-source voltage fluctuation is less than or equal to the fluctuation threshold. The optimized turn-off resistor is then used to obtain the optimized active Miller clamp diode switching delay for voltage peak suppression of the silicon carbide field-effect transistor.

[0070] If either of the following conditions is not met: the prediction accuracy of the active Miller clamp is greater than the prediction accuracy threshold, or the predicted drain-source voltage fluctuation is greater than the fluctuation threshold, then the turn-off resistor needs to be adjusted first to optimize the switching delay of the active Miller clamp diode, thereby obtaining the optimized turn-off resistor and the optimized switching delay of the active Miller clamp diode.

[0071] In step S300 of this application embodiment, adjusting the turn-off resistor optimizes the switching delay of the active Miller clamp diode, obtaining the optimized turn-off resistor and the optimized active Miller clamp diode switching delay, including:

[0072] Obtain the adjustment range of the turn-off resistor;

[0073] Based on the ratio of the predicted drain-source voltage fluctuation to the fluctuation threshold, the turn-off resistor adjustment step size is obtained.

[0074] Based on the ratio of the active Miller clamp prediction accuracy to the volatility threshold and the turn-off resistor adjustment range, the corrected turn-off resistor adjustment step size is obtained.

[0075] Based on the modified turn-off resistor adjustment step size and turn-off resistor adjustment range, the turn-off resistor is iteratively optimized to obtain the optimized turn-off resistor.

[0076] Based on the optimized turn-off resistance, the optimized active Miller clamp diode switching delay is obtained.

[0077] In this embodiment of the application, the purpose of the above-mentioned subdivision steps in step S300 is to indirectly optimize the switching delay of the active Miller clamp diode by dynamically adjusting the turn-off resistor, and finally reduce the drain-source voltage fluctuation to within the threshold, thereby achieving stable suppression of the peak voltage of the silicon carbide field-effect transistor.

[0078] First, the adjustment range of the turn-off resistor needs to be determined. This means defining the adjustable range of the turn-off resistor, which should be set based on the safe operating parameters of the silicon carbide field-effect transistor and the design specifications of the active Miller clamp circuit, such as 10Ω to 100Ω. Specifically, the lower limit of the turn-off resistor adjustment range must ensure that the turn-off speed does not exceed the device's tolerance to avoid voltage overshoot, while the upper limit must ensure that the clamping response speed meets the suppression requirements to avoid excessive delay leading to suppression failure, thus providing boundary constraints for subsequent iterative optimization.

[0079] Next, the turn-off resistor adjustment step size needs to be obtained based on the ratio of the predicted drain-source voltage fluctuation to the fluctuation threshold. For example, if the current predicted drain-source voltage fluctuation is twice the fluctuation threshold, and the ratio of the predicted drain-source voltage fluctuation to the fluctuation threshold is 2, then the turn-off resistor adjustment step size can be set to 1 / 5 of the total length of the turn-off resistor adjustment range as the base step size. The larger the ratio of the predicted drain-source voltage fluctuation to the fluctuation threshold, the larger the set base step size.

[0080] Then, based on the ratio of the active Miller clamp prediction accuracy to the volatility threshold and the turn-off resistor adjustment range, a corrected turn-off resistor adjustment step size needs to be obtained. That is, the base step size is corrected based on the ratio of the active Miller clamp prediction accuracy to the volatility threshold. For example, if the active Miller clamp prediction accuracy is only 80% of the volatility threshold (i.e., the ratio of active Miller clamp prediction accuracy to volatility threshold = 0.8), then the base step size is increased by 20%. In other words, the lower the active Miller clamp prediction accuracy, the larger the step size correction, to enhance the adjustment strength, ultimately obtaining the corrected turn-off resistor adjustment step size.

[0081] Furthermore, based on the modified turn-off resistor adjustment step size and the turn-off resistor adjustment range, iterative optimization of the turn-off resistor is needed to obtain an optimized turn-off resistor. That is, starting with the initial turn-off resistor obtained in step S100, the resistance value is gradually adjusted within the adjustment range according to the modified turn-off resistor adjustment step size. For example, if the current turn-off resistor is 50Ω and the modified turn-off resistor adjustment step size is 10Ω, then 60Ω, 70Ω, etc., are tried sequentially. Each time the turn-off resistor is adjusted, the corresponding "optimized active Miller clamp diode switching delay" is calculated through simulation, and the active Miller clamp prediction accuracy under this optimized active Miller clamp diode switching delay is evaluated simultaneously.

[0082] Furthermore, the optimized turn-off resistor needs to be used to suppress the voltage peak of the silicon carbide field-effect transistor, and the optimized drain-source voltage fluctuation after voltage peak suppression needs to be obtained. That is, after using the optimized turn-off resistor, the "optimized drain-source voltage fluctuation" after voltage peak suppression needs to be recalculated, and the specific calculation method is the same as the method for obtaining the drain-source voltage fluctuation in the aforementioned steps.

[0083] If the optimized drain-source voltage fluctuation is less than or equal to the fluctuation threshold, and the prediction accuracy of the active Miller clamp meets the standard, then the iteration stops, and the current turn-off resistor is the "optimized turn-off resistor", and the corresponding delay is the "optimized active Miller clamp diode switching delay". If the optimized drain-source voltage fluctuation is greater than the fluctuation threshold, then the adjustment continues according to the corrected turn-off resistor adjustment step size until the condition is met.

[0084] Finally, an optimized turn-off resistor is used to control the active Miller clamping circuit, and the optimized active Miller clamping diode switching delay is used to achieve precise clamping and suppress the voltage peak of the silicon carbide field-effect transistor.

[0085] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0086] Those skilled in the art will understand that embodiments of the present invention can provide methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0087] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0088] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0089] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0090] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0091] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for suppressing voltage peaks in a silicon carbide field-effect transistor, characterized in that, include: Obtain the drain-source voltage, drain current, switching frequency, and ambient temperature of the silicon carbide field-effect transistor, and obtain the switching delay and turn-off resistance of the active Miller clamp diode. Based on a neural network, the drain-source voltage is predicted to obtain a predicted drain-source voltage sequence. The predicted drain-source voltage is then used to simulate a silicon carbide field-effect transistor (MOSFET) to obtain the prediction accuracy of active Miller clamping, including: Simulations were performed using the predicted drain-source voltage, switching frequency, ambient temperature, and the switching delay of the active Miller clamp diode to obtain the prediction accuracy of the active Miller clamp. The formula for calculating the Miller clamp prediction accuracy is as follows: Miller clamp prediction accuracy = (1 - |actual drain-source voltage after clamping - target suppression value| / target suppression value) × 100%; The predicted drain-source voltage fluctuation is obtained, and based on the prediction accuracy and predicted drain-source voltage fluctuation of the active Miller clamp, the switching delay of the active Miller clamp diode is optimized and adjusted to suppress the voltage peak of the silicon carbide field-effect transistor, including: Determine whether the prediction accuracy of the active Miller clamp is less than or equal to the prediction accuracy threshold; Determining whether the predicted drain-source voltage fluctuation is less than or equal to a fluctuation threshold, wherein obtaining the predicted drain-source voltage fluctuation includes: Calculate the standard deviation and mean of the predicted drain-source voltage; Calculate the ratio of the predicted drain-source voltage standard deviation to the mean to obtain the predicted drain-source voltage fluctuation. If both conditions are met, then an active Miller clamp diode switching delay is used to suppress the voltage peak of the silicon carbide field-effect transistor. If any one of these conditions is not met, the switching delay of the active Miller clamp diode is optimized and adjusted to obtain an optimized active Miller clamp diode switching delay, thereby suppressing the voltage peak of the silicon carbide field-effect transistor.

2. The voltage peak suppression method for a silicon carbide field-effect transistor according to claim 1, characterized in that, Based on neural networks, the drain-source voltage is predicted, and the predicted drain-source voltage is obtained, including: Obtain historical drain-source voltage, historical drain current, historical switching frequency, and historical ambient temperature of the same type of silicon carbide field-effect transistor; A drain-source voltage predictor is constructed based on a neural network. The drain-source voltage predictor is trained using the historical drain-source voltage, historical drain current, historical switching frequency, and historical ambient temperature until convergence. The drain-source voltage predictor is used to predict the drain-source voltage based on the acquired drain-source voltage, drain current, switching frequency, and ambient temperature, thereby obtaining the predicted drain-source voltage.

3. The voltage peak suppression method for a silicon carbide field-effect transistor according to claim 1, characterized in that, If any one of these conditions is not met, the turn-off resistor is optimized and adjusted to obtain an optimized active Miller clamp diode switching delay, and voltage peak suppression of the silicon carbide field-effect transistor is performed, including: Adjust the turn-off resistor to optimize the switching delay of the active Miller clamp diode, and obtain the optimized turn-off resistor and the optimized switching delay of the active Miller clamp diode. The optimized turn-off resistor is used to suppress the voltage peak of the silicon carbide field-effect transistor, and the optimized drain-source voltage fluctuation after voltage peak suppression is obtained. When the optimized drain-source voltage fluctuation is greater than the fluctuation threshold, the turn-off resistor is further optimized until the optimized drain-source voltage fluctuation is less than or equal to the fluctuation threshold. The optimized turn-off resistor is then used to obtain the optimized active Miller clamp diode switching delay for voltage peak suppression of the silicon carbide field-effect transistor.

4. The voltage peak suppression method for a silicon carbide field-effect transistor according to claim 3, characterized in that, Adjusting the turn-off resistor optimizes the switching delay of the active Miller clamp diode, obtaining the optimized turn-off resistor and the optimized switching delay of the active Miller clamp diode, including: Obtain the adjustment range of the turn-off resistor; Based on the ratio of the predicted drain-source voltage fluctuation to the fluctuation threshold, the turn-off resistor adjustment step size is obtained. Based on the ratio of the active Miller clamp prediction accuracy to the volatility threshold and the turn-off resistor adjustment range, the corrected turn-off resistor adjustment step size is obtained. Based on the modified turn-off resistor adjustment step size and turn-off resistor adjustment range, the turn-off resistor is iteratively optimized to obtain the optimized turn-off resistor. Based on the optimized turn-off resistance, the optimized active Miller clamp diode switching delay is obtained.

Images