A junction temperature detection method based on RC-IGBT

Abstract

The present invention discloses a junction temperature detection method based on RC-IGBT, which belongs to the technical field of junction temperature detection of IGBT. The present invention collects the temperature values ​​of multiple detection points of the RC-IGBT module shell, constructs a shell temperature matrix, reflects the shell temperature distribution, and then divides each temperature value in the shell temperature matrix into a high temperature part and a low temperature part, obtains a high temperature zone set and a low temperature zone set, respectively calculates the temperature rise coefficient for the high temperature zone set and the low temperature zone set, captures the temperature change of the high temperature zone and the low temperature zone, and then calculates the temperature distribution coefficient according to the temperature difference between the high temperature zone set and the low temperature zone set, reflecting the temperature diffusion. The present invention predicts the initial junction temperature according to the shell temperature matrix, and then combines the temperature diffusion of the high temperature zone and the low temperature zone, as well as the temperature rise of the high temperature zone and the low temperature zone, to compensate for the initial junction temperature, thereby improving the junction temperature detection accuracy of the RC-IGBT module.

Description

Technical Field

[0001] The present invention relates to the technical field of IGBT junction temperature detection, and in particular to a junction temperature detection method based on RC-IGBT. Background Art

[0002] In modern power electronic devices, insulated gate bipolar transistors (IGBTs) are widely used in various high-power applications such as electric vehicles, renewable energy systems, and industrial drives due to their superior switching characteristics and high efficiency. However, with the continuous increase in power density, the thermal management problem of IGBT modules has become increasingly prominent. Excessive junction temperature not only affects the performance and reliability of the IGBT, but may also cause premature failure of the device. Therefore, accurately monitoring and controlling the junction temperature of the IGBT has become a key issue that needs to be urgently addressed in the field of power electronics.

[0003] At present, the traditional junction temperature detection method mainly relies on the temperature sensor to directly measure the external temperature of the module, thereby inferring the junction temperature of the IGBT. However, the external temperature cannot accurately reflect the changes in the internal junction temperature, which may lead to the lag of temperature monitoring and cause the problem of low junction temperature detection accuracy of the IGBT. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a junction temperature detection method based on RC-IGBT to solve the problem of low junction temperature detection accuracy in the prior art.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is: a junction temperature detection method based on RC-IGBT, comprising the following steps:

[0006] S1, collect the temperature values ​​of each detection point on the RC-IGBT module housing and construct the housing temperature matrix;

[0007] S2. construct a high temperature zone set and a low temperature zone set according to each temperature value in the shell temperature matrix;

[0008] S3, respectively calculating the temperature rise coefficient for the high temperature zone set and the low temperature zone set to obtain the temperature rise coefficient for the high temperature zone and the temperature rise coefficient for the low temperature zone;

[0009] S4. Calculate the temperature distribution coefficient according to the temperature difference between the high temperature area set and the low temperature area set;

[0010] S5. The shell temperature matrix is ​​processed by using the junction temperature prediction model, and the junction temperature of the RC-IGBT module is obtained based on the compensation of the high temperature zone temperature rise coefficient, the low temperature zone temperature rise coefficient and the temperature distribution coefficient.

[0011] Further, S2 includes the following sub-steps:

[0012] S21, taking the average of each temperature value in the shell temperature matrix to obtain a temperature mean;

[0013] S22, extracting temperature values ​​greater than or equal to the temperature mean in the shell temperature matrix, and constructing a high temperature zone set;

[0014] S23, extracting temperature values ​​less than the temperature mean in the shell temperature matrix, and constructing a low temperature zone set.

[0015] Furthermore, S3 includes the following sub-steps:

[0016] S31. Subtract the high temperature area sets at adjacent moments to obtain the high temperature difference set, D H,t =R H,t -R H,t-1 , where D H,t is the high temperature difference set at time t, R H,t is the high temperature area set at time t, R H,t-1 is the set of high temperature areas at time t-1;

[0017] S32, calculating the temperature rise coefficient of the set according to the high temperature difference set, and obtaining the temperature rise coefficient of the high temperature zone;

[0018] S33, subtract the low temperature area sets at adjacent moments to obtain the low temperature difference set, D L,t =R L,t -R L,t-1 , where D L,t is the low temperature difference set at time t, R L,t is the low temperature zone set at time t, R L,t-1 is the set of low temperature areas at time t-1;

[0019] S34. Calculate the temperature rise coefficient of the set according to the low temperature difference set to obtain the temperature rise coefficient of the low temperature zone.

[0020] Furthermore, the formula for calculating the temperature rise coefficient of the set in S32 and S34 is: , where μ is the temperature rise coefficient of the set, e is a natural constant, and x i is the i-th element in the high temperature difference set or the low temperature difference set, N is the number of elements in the high temperature difference set or the low temperature difference set, and i is a positive integer.

[0021] Furthermore, the formula for calculating the temperature distribution coefficient in S4 is: , where θ is the temperature distribution coefficient, e is the natural constant, T H,j is the jth temperature value in the high temperature zone set, T L,j is the jth temperature value in the low temperature zone set, M H is the number of temperature values ​​in the high temperature zone set, M Lis the number of temperature values ​​in the low temperature zone set, and j is a positive integer.

[0022] Further, the junction temperature prediction model in S5 includes: a temperature prediction unit, a first compensation coefficient calculation unit, a second compensation coefficient calculation unit and a temperature compensation unit;

[0023] The temperature prediction unit is used to predict the initial junction temperature according to the housing temperature matrix;

[0024] The first compensation coefficient calculation unit is used to calculate the first compensation coefficient according to the high temperature zone temperature rise coefficient and the low temperature zone temperature rise coefficient;

[0025] The second compensation coefficient calculation unit is used to calculate the second compensation coefficient according to the temperature distribution coefficient;

[0026] The temperature compensation unit is used to compensate the initial junction temperature by using the first compensation coefficient and the second compensation coefficient to obtain the junction temperature of the RC-IGBT module.

[0027] Further, the temperature prediction unit includes: a first convolutional layer, a second convolutional layer, a third convolutional layer and a fully connected layer;

[0028] The input end of the first convolutional layer is used as the input end of the temperature prediction unit, and its output end is connected to the input end of the second convolutional layer and the input end of the third convolutional layer respectively;

[0029] The input end of the fully connected layer is connected to the output end of the second convolutional layer and the output end of the third convolutional layer respectively, and its output end serves as the output end of the temperature prediction unit;

[0030] The convolution kernel size of the first convolutional layer is , the convolution kernel size of the second convolutional layer is , the convolution kernel size of the third convolutional layer is .

[0031] Furthermore, the expression of the first compensation coefficient calculation unit is: , where γ1 is the first compensation coefficient, tanh is the hyperbolic tangent function, μ H is the temperature rise coefficient in high temperature zone, μ L is the temperature rise coefficient in the low temperature zone, ω μH is the temperature rise coefficient μ in the high temperature zone H The weight, ω μL is the temperature rise coefficient μ in the low temperature zone L The weight of b μ is the bias of the temperature rise coefficient.

[0032] Furthermore, the expression of the second compensation coefficient calculation unit is: , where γ2 is the second compensation coefficient, tanh is the hyperbolic tangent function, θ is the temperature distribution coefficient, ωθ is the weight of the temperature distribution coefficient θ, b θ is the offset of the temperature distribution coefficient.

[0033] Furthermore, the expression of the temperature compensation unit is: ,in, is the junction temperature of the RC-IGBT module, T0 is the initial junction temperature, γ1 is the first compensation coefficient, and γ2 is the second compensation coefficient.

[0034] The beneficial effects of the present invention are as follows: the present invention collects the temperature values ​​of multiple detection points of the RC-IGBT module shell, constructs a shell temperature matrix, reflects the shell temperature distribution, and then divides each temperature value in the shell temperature matrix into high temperature and low temperature parts to obtain a high temperature zone set and a low temperature zone set, respectively calculates the temperature rise coefficient for the high temperature zone set and the low temperature zone set, captures the temperature change of the high temperature zone and the low temperature zone, and then calculates the temperature distribution coefficient based on the temperature difference between the high temperature zone set and the low temperature zone set to reflect the temperature diffusion. The present invention predicts the initial junction temperature based on the shell temperature matrix, and then combines the temperature diffusion of the high temperature zone and the low temperature zone, as well as the temperature rise of the high temperature zone and the low temperature zone, to compensate for the initial junction temperature, thereby improving the junction temperature detection accuracy of the RC-IGBT module. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flow chart of a junction temperature detection method based on RC-IGBT. DETAILED DESCRIPTION

[0036] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0037] like Figure 1 As shown, a junction temperature detection method based on RC-IGBT includes the following steps:

[0038] S1, collect the temperature values ​​of each detection point on the RC-IGBT module housing and construct the housing temperature matrix;

[0039] S2. construct a high temperature zone set and a low temperature zone set according to each temperature value in the shell temperature matrix;

[0040] S3, respectively calculating the temperature rise coefficient for the high temperature zone set and the low temperature zone set to obtain the temperature rise coefficient for the high temperature zone and the temperature rise coefficient for the low temperature zone;

[0041] S4. Calculate the temperature distribution coefficient according to the temperature difference between the high temperature area set and the low temperature area set;

[0042] S5. The shell temperature matrix is ​​processed by using the junction temperature prediction model, and the junction temperature of the RC-IGBT module is obtained based on the compensation of the high temperature zone temperature rise coefficient, the low temperature zone temperature rise coefficient and the temperature distribution coefficient.

[0043] The RC-IGBT module consists of three stages: initial preheating stage: the temperature rises rapidly after power is turned on, the temperature gradient is steep, and the heating rate is high; stable working stage: the temperature tends to be flat and close to the thermal equilibrium state; heat accumulation stage: after working for a long time, the temperature continues to rise slowly, and the internal thermal resistance gradually increases.

[0044] In this embodiment, the housing of the RC-IGBT module is the package of the RC-IGBT module.

[0045] In this embodiment, the implementation process of step S1 is as follows: when the RC-IGBT module is in normal working state, an infrared thermal imager is used to vertically align the module housing, and a fixed distance (usually 10-20 cm) is maintained to collect an infrared image of the RC-IGBT module. Since the infrared image is composed of multiple pixels, each pixel has a temperature radiation intensity value, and the temperature radiation intensity value can be converted into a temperature value through Planck's radiation law, the infrared image can be used to capture the temperature value of each detection point on the RC-IGBT module housing.

[0046] In this embodiment, S2 includes the following sub-steps:

[0047] S21, taking the average of each temperature value in the shell temperature matrix to obtain a temperature mean;

[0048] S22, extracting temperature values ​​greater than or equal to the temperature mean in the shell temperature matrix, and constructing a high temperature zone set;

[0049] S23, extracting temperature values ​​less than the temperature mean in the shell temperature matrix, and constructing a low temperature zone set.

[0050] In this embodiment, S3 includes the following sub-steps:

[0051] S31. Subtract the high temperature area sets at adjacent moments to obtain the high temperature difference set, D H,t =R H,t -R H,t-1 , where D H,t is the high temperature difference set at time t, R H,t is the high temperature area set at time t, R H,t-1 is the set of high temperature areas at time t-1;

[0052] S32, calculating the temperature rise coefficient of the set according to the high temperature difference set, and obtaining the temperature rise coefficient of the high temperature zone;

[0053] S33, subtract the low temperature area sets at adjacent moments to obtain the low temperature difference set, D L,t =R L,t -R L,t-1 , where D L,t is the low temperature difference set at time t, R L,t is the low temperature zone set at time t, R L,t-1 is the set of low temperature areas at time t-1;

[0054] S34. Calculate the temperature rise coefficient of the set according to the low temperature difference set to obtain the temperature rise coefficient of the low temperature zone.

[0055] The temperature rise coefficient in the high temperature zone is the temperature rise coefficient corresponding to the high temperature difference set, and the temperature rise coefficient in the low temperature zone is the temperature rise coefficient corresponding to the low temperature difference set.

[0056] The present invention divides the shell temperature matrix into a high-temperature part and a low-temperature part, and captures the temperature rise coefficients of the high-temperature part and the low-temperature part respectively. The high-temperature part is the area that mainly emits heat, and the low-temperature part is the area affected by the heat. Therefore, capturing the temperature changes of these two parts can reflect the temperature transfer from the inside of the RC-IGBT module to the outside.

[0057] In this embodiment, the formula for calculating the temperature rise coefficient of the set in S32 and S34 is: , where μ is the temperature rise coefficient of the set, e is a natural constant, and x i is the i-th element in the high temperature difference set or the low temperature difference set, N is the number of elements in the high temperature difference set or the low temperature difference set, and i is a positive integer.

[0058] In this embodiment, the formula for calculating the temperature distribution coefficient in S4 is: , where θ is the temperature distribution coefficient, e is the natural constant, T H,j is the jth temperature value in the high temperature zone set, T L,j is the jth temperature value in the low temperature zone set, M H is the number of temperature values ​​in the high temperature zone set, M L is the number of temperature values ​​in the low temperature zone set, and j is a positive integer.

[0059] The present invention reflects the temperature distribution of the shell through the temperature difference between the high-temperature part and the low-temperature part. When the temperature distribution coefficient is large, the temperature distribution gradient is large, which belongs to the initial preheating stage. When the temperature distribution gradient is small, it belongs to the heat accumulation stage. When the temperature distribution coefficient is close to 0, it belongs to the stable working stage.

[0060] In this embodiment, the junction temperature prediction model in S5 includes: a temperature prediction unit, a first compensation coefficient calculation unit, a second compensation coefficient calculation unit and a temperature compensation unit;

[0061] The temperature prediction unit is used to predict the initial junction temperature according to the housing temperature matrix;

[0062] The first compensation coefficient calculation unit is used to calculate the first compensation coefficient according to the high temperature zone temperature rise coefficient and the low temperature zone temperature rise coefficient;

[0063] The second compensation coefficient calculation unit is used to calculate the second compensation coefficient according to the temperature distribution coefficient;

[0064] The temperature compensation unit is used to compensate the initial junction temperature by using the first compensation coefficient and the second compensation coefficient to obtain the junction temperature of the RC-IGBT module.

[0065] When predicting the junction temperature, the shell temperature matrix, the high temperature zone temperature rise coefficient, the low temperature zone temperature rise coefficient and the temperature distribution coefficient are the data at the same time.

[0066] In this embodiment, the temperature prediction unit includes: a first convolutional layer, a second convolutional layer, a third convolutional layer and a fully connected layer;

[0067] The input end of the first convolutional layer is used as the input end of the temperature prediction unit, and its output end is connected to the input end of the second convolutional layer and the input end of the third convolutional layer respectively;

[0068] The input end of the fully connected layer is connected to the output end of the second convolutional layer and the output end of the third convolutional layer respectively, and its output end serves as the output end of the temperature prediction unit;

[0069] The convolution kernel size of the first convolutional layer is , the convolution kernel size of the second convolutional layer is , the convolution kernel size of the third convolutional layer is .

[0070] In the present invention, the first convolution layer is used to extract the shell temperature matrix to obtain temperature features, and then two convolution layers of different sizes are used to extract features of different scales. Finally, a fully connected layer is used to combine the features of two different scales to predict the initial junction temperature. Since the external temperature cannot accurately reflect the changes in the internal junction temperature, it may cause the hysteresis of temperature monitoring. Therefore, the present invention extracts the temperature rise coefficient and the temperature distribution coefficient to reflect the temperature dynamics, and further reflects the internal junction temperature changes through the external dynamics.

[0071] In this embodiment, the expression of the first compensation coefficient calculation unit is: , where γ1 is the first compensation coefficient, tanh is the hyperbolic tangent function, μ H is the temperature rise coefficient in high temperature zone, μL is the temperature rise coefficient in the low temperature zone, ω μH is the temperature rise coefficient μ in the high temperature zone H The weight, ω μL is the temperature rise coefficient μ in the low temperature zone L The weight of b μ is the bias of the temperature rise coefficient.

[0072] In this embodiment, the expression of the second compensation coefficient calculation unit is: , where γ2 is the second compensation coefficient, tanh is the hyperbolic tangent function, θ is the temperature distribution coefficient, ω θ is the weight of the temperature distribution coefficient θ, b θ is the offset of the temperature distribution coefficient.

[0073] In this embodiment, the expression of the temperature compensation unit is: ,in, is the junction temperature of the RC-IGBT module, T0 is the initial junction temperature, γ1 is the first compensation coefficient, and γ2 is the second compensation coefficient.

[0074] The present invention adopts a first compensation coefficient calculation unit and a second compensation coefficient calculation unit to respectively calculate the ratio of junction temperature compensation. When the initial junction temperature is predicted based on the shell temperature matrix, compensation is performed according to the shell temperature dynamics to improve the accuracy of junction temperature detection.

[0075] In this embodiment, the weights and biases in the junction temperature prediction model can be obtained using the existing gradient descent method and GA genetic algorithm, and the actual junction temperature of the RC-IGBT module in the sample can be obtained by installing a thermistor inside the RC-IGBT module.

[0076] The present invention collects the temperature values ​​of multiple detection points of the RC-IGBT module shell, constructs a shell temperature matrix, reflects the shell temperature distribution, and then divides each temperature value in the shell temperature matrix into high temperature and low temperature parts to obtain a high temperature zone set and a low temperature zone set, respectively calculates the temperature rise coefficient for the high temperature zone set and the low temperature zone set, captures the temperature change of the high temperature zone and the low temperature zone, and then calculates the temperature distribution coefficient according to the temperature difference between the high temperature zone set and the low temperature zone set to reflect the temperature diffusion. The present invention predicts the initial junction temperature according to the shell temperature matrix, and then combines the temperature diffusion of the high temperature zone and the low temperature zone, as well as the temperature rise of the high temperature zone and the low temperature zone, to compensate for the initial junction temperature, thereby improving the junction temperature detection accuracy of the RC-IGBT module.

[0077] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A junction temperature detection method based on RC-IGBT, characterized in that: The following steps are involved: S1, collect the temperature values ​​of each detection point on the RC-IGBT module housing and construct the housing temperature matrix; S2. construct a high temperature zone set and a low temperature zone set according to each temperature value in the shell temperature matrix; S3, respectively calculating the temperature rise coefficient for the high temperature zone set and the low temperature zone set to obtain the temperature rise coefficient for the high temperature zone and the temperature rise coefficient for the low temperature zone; S4. Calculate the temperature distribution coefficient according to the temperature difference between the high temperature area set and the low temperature area set; S5. Using the junction temperature prediction model to process the shell temperature matrix, based on the compensation of the high temperature zone temperature rise coefficient, the low temperature zone temperature rise coefficient and the temperature distribution coefficient, the junction temperature of the RC-IGBT module is obtained; The S3 comprises the following sub-steps: S31. Subtract the high temperature area sets at adjacent moments to obtain the high temperature difference set, D H,t =R H,t -R H,t-1 , where D H,t is the high temperature difference set at time t, R H,t is the high temperature area set at time t, R H,t-1 is the set of high temperature areas at time t-1; S32, calculating the temperature rise coefficient of the set according to the high temperature difference set, and obtaining the temperature rise coefficient of the high temperature zone; S33, subtract the low temperature area sets at adjacent moments to obtain the low temperature difference set, D L,t =R L,t -R L,t-1 , where D L,t is the low temperature difference set at time t, R L,t is the low temperature zone set at time t, R L,t-1 is the set of low temperature areas at time t-1; S34, calculating the temperature rise coefficient of the set according to the low temperature difference set, and obtaining the temperature rise coefficient of the low temperature zone; The formula for calculating the temperature distribution coefficient in S4 is: , where θ is the temperature distribution coefficient, e is the natural constant, T H,j is the jth temperature value in the high temperature zone set, T L,j is the jth temperature value in the low temperature zone set, M H is the number of temperature values ​​in the high temperature zone set, M L is the number of temperature values ​​in the low temperature zone set, j is a positive integer; The junction temperature prediction model in S5 includes: a temperature prediction unit, a first compensation coefficient calculation unit, a second compensation coefficient calculation unit and a temperature compensation unit; The temperature prediction unit is used to predict the initial junction temperature according to the housing temperature matrix; The first compensation coefficient calculation unit is used to calculate the first compensation coefficient according to the high temperature zone temperature rise coefficient and the low temperature zone temperature rise coefficient; The second compensation coefficient calculation unit is used to calculate the second compensation coefficient according to the temperature distribution coefficient; The temperature compensation unit is used to compensate the initial junction temperature by using the first compensation coefficient and the second compensation coefficient to obtain the junction temperature of the RC-IGBT module; The expression of the first compensation coefficient calculation unit is: , where γ1 is the first compensation coefficient, tanh is the hyperbolic tangent function, μ H is the temperature rise coefficient in high temperature zone, μ L is the temperature rise coefficient in the low temperature zone, ω μH is the temperature rise coefficient μ in the high temperature zone H The weight, ω μL is the temperature rise coefficient μ in the low temperature zone L The weight of b μ is the bias of the temperature rise coefficient; The expression of the second compensation coefficient calculation unit is: , where γ2 is the second compensation coefficient, tanh is the hyperbolic tangent function, θ is the temperature distribution coefficient, ω θ is the weight of the temperature distribution coefficient θ, b θ is the bias of the temperature distribution coefficient; The expression of the temperature compensation unit is: ,in, is the junction temperature of the RC-IGBT module, T0 is the initial junction temperature, γ1 is the first compensation coefficient, and γ2 is the second compensation coefficient.

2. The junction temperature detection method based on RC-IGBT according to claim 1, characterized in that: The S2 comprises the following sub-steps: S21, taking the average of each temperature value in the shell temperature matrix to obtain a temperature mean; S22, extracting temperature values ​​greater than or equal to the temperature mean in the shell temperature matrix, and constructing a high temperature zone set; S23, extracting temperature values ​​less than the temperature mean in the shell temperature matrix, and constructing a low temperature zone set.

3. The junction temperature detection method based on RC-IGBT according to claim 1, characterized in that: The formula for calculating the temperature rise coefficient of the set in S32 and S34 is: , where μ is the temperature rise coefficient of the set, e is a natural constant, and x i is the i-th element in the high temperature difference set or the low temperature difference set, N is the number of elements in the high temperature difference set or the low temperature difference set, and i is a positive integer.

4. The junction temperature detection method based on RC-IGBT according to claim 1, characterized in that: The temperature prediction unit includes: a first convolutional layer, a second convolutional layer, a third convolutional layer and a fully connected layer; The input end of the first convolutional layer is used as the input end of the temperature prediction unit, and the output end thereof is connected to the input end of the second convolutional layer and the input end of the third convolutional layer respectively; The input end of the fully connected layer is connected to the output end of the second convolutional layer and the output end of the third convolutional layer respectively, and the output end thereof serves as the output end of the temperature prediction unit; The convolution kernel size of the first convolutional layer is , the convolution kernel size of the second convolutional layer is , the convolution kernel size of the third convolutional layer is .

Images