CN111507031B - Power module thermal network model parameter identification method based on least square method - Google Patents

Abstract

The invention discloses a power module thermal network model parameter identification method based on a least square method. The method for identifying the parameters of the heat network utilizes a plurality of groups of finite element model simulation results to calculate the heat conductance of the heat network model and identify the heat capacity of the heat network model based on the principle of a least square method. The method for identifying the parameters of the heat supply network model has simple principle, and the extracted heat supply network model has high precision and can be applied to the on-line calculation of the chip temperature and the calculation of the chip temperature in a long time scale.

Description

Parameter identification method of thermal network model of power module based on least squares method

technical field

The invention relates to a method for identifying parameters of a thermal network model of a power module, in particular to a method for identifying parameters of a thermal network model of a power module based on the least squares method.

Background technique

Industrial research has shown that the power module is the weakest component in a power electronic system. About 55% of power module failures are induced by temperature factors. The chip temperature estimation of the power module is the basis for life prediction and reliability evaluation. Thermal network models are widely used in industry for chip temperature prediction and estimation of power modules because they are easily embedded in digital signal processors and do not need to design additional hardware circuits. The finite element simulation model is often used to predict the chip temperature. The simulation results of the finite element simulation model have high accuracy but the simulation process is time-consuming. Therefore, at the expense of certain accuracy, a reduced-order thermal network model needs to be extracted from the finite element model of the power module.

At present, the extraction of the reduced-order thermal network model has become a research hotspot in the reliable operation technology of power modules. There are both in-depth theoretical analysis of this in academic papers and practical engineering methods, such as the invention patent "A method of obtaining electric power". Method and Device for Transient Temperature of Electronic Devices" (CN 102930096 B) and "Thermal Network Parameter Identification Method Based on IGBT Junction Temperature Information" (CN 105718694 B). in,

Chinese invention patent specification CN 102930096 B, published on November 29, 2014, "A Method and Device for Obtaining Transient Temperature of Power Electronic Devices", is to first obtain the initial finite element model and finite volume model of the power electronic device; The finite element volume model obtains the convective heat transfer coefficient distribution, and maps the convective heat transfer coefficient distribution to the initial finite element model to form a new finite element model; reduced order model. However, this method of obtaining a reduced-order simulation model has the following shortcomings:

1) The initial finite element model, the finite element volume model and the new finite element model are involved in the implementation process, and there are two mapping operations, and the operation of extracting the reduced-order model is complicated;

2) The initial finite element model includes the heat dissipation system, and the new finite element model uses the convective heat transfer coefficient to model the actual heat dissipation system, which will cause a large error in the extracted reduced-order model.

Chinese invention patent specification CN 105718694 B published "Thermal network parameter identification method based on IGBT junction temperature information" published on February 19, 2019. First, an equivalent thermal network model is established to change any set of measurable thermal resistances in the thermal network. and thermal capacitance, to establish two different IGBT thermal network parameter constraint equations; then measure the thermal resistance under the equivalent zero input state and the two cooling curves before and after the thermal capacitance change to obtain two sets of time constants; finally, use the IGBT thermal network parameters The constraint equation of gets the thermal network parameters of the IGBT. However, this thermal network parameter identification has the following shortcomings:

1) The extracted thermal network model does not consider the thermal coupling effect between chips, and is not suitable for temperature estimation of multi-chip power modules;

2) A set of measurable thermal resistances and thermal capacitances involved in the implementation process does not specify how to measure them.

SUMMARY OF THE INVENTION

The problem to be solved by the present invention is the deficiency of the above solution, that is, to provide a simple and accurate method for identifying parameters of a thermal network model of a power module. The present invention is achieved through the following technical solutions:

A method for identifying parameters of a thermal network model of a power module based on the least squares method, the thermal network model includes 13 temperature nodes, 4 input loss current sources, 20 thermal conductivities and 12 thermal capacitances; the 13 temperature The node is denoted as temperature node Ni, i is the serial number of temperature node, _i =1, 2...13, the temperature of temperature node Ni is denoted as temperature Ti, _i = ₁ , 2...13; the 20 The thermal conductivities are recorded as thermal conductance G _j , j is the serial number of thermal conductance, j=1, 2...20; the 12 thermal capacitances are recorded as thermal capacitance C _k , k is the serial number of thermal capacitance, k=1 , 2...12; the four input loss current sources are respectively recorded as input loss current source P ₁ , input loss current source P ₂ , input loss current source P ₃ and input loss current source P ₄ ;

The thermal network model is a three-dimensional Cauer thermal network structure, which is divided into 4 layers, the temperature node N ₁ , the temperature node N ₂ , the temperature node N ₃ and the temperature node N ₄ are in the first layer, and the temperature node N ₅ , temperature node N ₆ , temperature node N ₇ and temperature node N ₈ are in the second layer, temperature node N ₉ , temperature node N ₁₀ , temperature node N ₁₁ and temperature node N ₁₂ are in the third layer , the temperature node N ₁₃ is in the fourth layer; the temperature node N ₁ , the temperature node N ₅ and the temperature node N ₉ are aligned in the vertical direction of space, and the temperature node N ₂ , the temperature node N ₆ and the temperature node N 9 are aligned in the vertical direction of space. N ₁₀ is aligned in the vertical direction of space, temperature node N ₃ , temperature node N ₇ and temperature node N ₁₁ are aligned in the vertical direction of space, temperature node N ₄ , temperature node N ₈ and temperature node N ₁₂ Align in the vertical direction of space;

_The thermal conduction _G1 is set between the temperature node N1 and the temperature node _N5 , the thermal conduction _G2 is set between the temperature node _N2 and the temperature node _N6 , and the thermal conduction _G3 is set between the temperature node N ₃ and the temperature node _N7 , the heat conduction _G4 is set between the temperature node _N4 and the temperature node _N8 , the heat conduction _G5 is set between the temperature node _N5 and the temperature node _N8 , _The heat conduction G6 is set between the temperature node _N5 and the temperature node _N6 , the heat conduction _G7 is set between the temperature node _N6 and the temperature node _N7 , and the heat conduction _G8 is set at the temperature node N ₇ and the temperature node N ₈ , the thermal conductance G ₉ is set between the temperature node N ₅ and the temperature node N ₉ , the thermal conductance G ₁₀ is set between the temperature node N ₆ and the temperature node N ₁₀ , The heat conduction _G11 is set between the temperature node _N7 and the temperature node _N11 , the heat conduction _G12 is set between the temperature node _N8 and the temperature node _N12 , and the heat conduction _G13 is set at the temperature node N ₉ and the temperature node N ₁₂ , the thermal conductance G ₁₄ is set between the temperature node N ₉ and the temperature node N ₁₀ , the thermal conductance G ₁₅ is set between the temperature node N ₁₀ and the temperature node N ₁₁ , The thermal conduction G ₁₆ is set between the temperature node N ₁₁ and the temperature node N ₁₂ , the thermal conduction G ₁₇ is set between the temperature node N ₉ and the temperature node N ₁₃ , and the thermal conduction G ₁₈ is set between the temperature node N ₁₀ and the temperature node _N13 , the heat conduction _G19 is set between the temperature node _N11 and the temperature node _N13 , and the heat conduction _G20 is set between the temperature node _N12 and the temperature node _N13 ;

Heat capacity C ₁ , heat capacity C ₂ , heat capacity C ₃ , heat capacity C ₄ , heat capacity C ₅ , heat capacity C ₆ , heat capacity C ₇ , heat capacity C ₈ , heat capacity C ₉ , heat capacity C ₁₀ , One end of the heat capacity C ₁₁ and the heat capacity C ₁₂ are respectively corresponding to the temperature node N ₁ , the temperature node N ₂ , the temperature node N ₃ , the temperature node N ₄ , the temperature node N ₅ , and the temperature node N ₆ . , temperature node N ₇ , temperature node N ₈ , temperature node N ₉ , temperature node N ₁₀ , temperature node N ₁₁ , temperature node N ₁₂ are connected, and the other end is grounded; input loss current source P ₁ , One end of the input loss current source P ₂ , the input loss current source P ₃ , and the input loss current source P ₄ are respectively corresponding to the temperature node N ₁ , the temperature node N ₂ , the temperature node N ₃ , and the temperature node N ₄ . connected, the other end is grounded;

The steps of the identification method of the present invention are as follows:

Step 1, build the finite element simulation model of the power module

The physical size of the power module is sampled, and a finite element simulation model S of the power module including the liquid cooling system is built according to the physical size; the finite element simulation model S includes 8 layers, and the order from top to bottom is chip layer, chip solder layer, upper copper layer, ceramic layer, lower copper layer, substrate solder layer, heat dissipation substrate layer and liquid cooling layer; the chip layer is composed of 2 IGBT chips and 2 Diode chips, and the 2 IGBT chips are respectively recorded as chip X ₁ and chip X ₃ , 2 Diode chips are respectively recorded as chip X ₂ and chip X ₄ ;

A total of 12 temperature monitoring points are set in the finite element simulation model S, and any one of them is recorded as the temperature monitoring point M _m , where m is the serial number of the temperature monitoring point, m=1, 2...12; among them, the temperature monitoring point M ₁ is located in the center _of the chip X1, the temperature monitoring point _M2 is located in the center of the chip _X2 , the temperature monitoring point M3 is located in the center _of the chip _X3 , and the temperature monitoring point _M4 is located in the center of the chip _X4 Location; temperature monitoring point M ₅ , temperature monitoring point M ₆ , temperature monitoring point M ₇ and temperature monitoring point M ₈ are located on the upper copper layer, temperature monitoring point M ₉ , temperature monitoring point M ₁₀ , temperature monitoring point M ₁₁ and temperature monitoring point M 11 The monitoring point M ₁₂ is set on the solder layer of the substrate; the temperature monitoring point M ₅ and the temperature monitoring point M ₉ are aligned with the temperature monitoring point M ₁ in the vertical direction of space, and the temperature monitoring point M ₆ and the temperature monitoring point M ₁₀ are in the vertical direction of space Aligned with the temperature monitoring point M ₂ , the temperature monitoring point M ₇ and the temperature monitoring point M ₁₁ are aligned with the temperature monitoring point M ₃ in the vertical direction of space, and the temperature monitoring point M ₈ and the temperature monitoring point M 12 are aligned with the temperature monitoring point M ₁₂ in the vertical direction of space. The monitoring point M ₄ is aligned;

Set the liquid cooling layer flow velocity V, the liquid cooling layer temperature Y, the simulation termination time T _sim and the logarithmic uniform simulation step size T _step ;

Step 2, finite element simulation model simulation experiment

记录温度监测点M_m的稳态温度

m＝1,2...12，得到仿真实验1的稳态温度矩阵

和仿真实验1对应的热网络模型的损耗矩阵U¹，

和U¹的表达式如下：Simulation experiment 1, set the loss input P _X1 of chip X ₁ as the step power of amplitude A, the loss input P _X2 of chip X ₂ is 0, the loss input P _X3 of chip X ₃ is 0, and the loss input of chip X ₄ P _X4 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

m=1,2...12, get the steady-state temperature matrix of simulation experiment 1

The loss matrix U ¹ of the thermal network model corresponding to the simulation experiment 1,

and U ¹ are expressed as follows:

U ¹ =[A 0 0 0 0 0 0 0 0 0 0 0 -A] ^T

记录温度监测点M_m的稳态温度

m＝1,2...12，得到仿真实验2的稳态温度矩阵

和仿真实验2对应的热网络模型的损耗矩阵U²，

和U²的表达式如下：Simulation experiment 2, set the loss input P _X2 of chip X ₂ as the step power of amplitude A, the loss input P _X1 of chip X ₁ is 0, the loss input P _X3 of chip X ₃ is 0, and the loss input of chip X ₄ P _X4 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

m=1,2...12, the steady-state temperature matrix of simulation experiment 2 is obtained

The loss matrix U ² of the thermal network model corresponding to simulation experiment 2,

and U ² are expressed as follows:

U ² =[0 A 0 0 0 0 0 0 0 0 0 0 -A] ^T

记录温度监测点M_m的稳态温度

m＝1,2...12，得到仿真实验3的稳态温度矩阵

和仿真实验3对应的热网络模型的损耗矩阵U³，

和U³的表达式如下：Simulation experiment 3, set the loss input P _X3 of chip X ₃ as the step power of amplitude A, the loss input P _X1 of chip X ₁ is 0, the loss input P _X2 of chip X ₂ is 0, and the loss input of chip X ₄ P _X4 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

m=1,2...12, get the steady-state temperature matrix of simulation experiment 3

The loss matrix U ³ of the thermal network model corresponding to simulation experiment 3,

and the expression for ^U3 is as follows:

U ³ =[0 0 A 0 0 0 0 0 0 0 0 0 -A] ^T

记录温度监测点M_m的稳态温度

m＝1,2...12，得到仿真实验4的稳态温度矩阵

和仿真实验4对应的热网络模型的损耗矩阵U⁴，

和U⁴的表达式如下：Simulation experiment 4, set the loss input P _X4 of chip X ₄ as the step power of amplitude A, the loss input P _X1 of chip X ₁ is 0, the loss input P _X2 of chip X ₂ is 0, and the loss input of chip X ₃ P _X3 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

m=1,2...12, the steady-state temperature matrix of simulation experiment 4 is obtained

The loss matrix U ⁴ of the thermal network model corresponding to simulation experiment 4,

and the expression for U ⁴ is as follows:

U ¹ =[0 0 0 A 0 0 0 0 0 0 0 0 -A] ^T

Among them, the loss matrix U ¹ of the thermal network model corresponding to the simulation experiment 1, the loss matrix U ² of the thermal network model corresponding to the simulation experiment 2, the loss matrix U ³ of the thermal network model corresponding to the simulation experiment 3, and the thermal network model corresponding to the simulation experiment 4 . Each element in the loss matrix U ⁴ of the network model corresponds to the temperature T _i of the temperature node Ni of the thermal network model one-to-one, _i =1, 2, . . . 13;

Step 3, Calculate the thermal conductivity of the thermal network model

The thermal network model is described by a system of ordinary differential equations with the following expressions:

in,

T is a temperature matrix composed of 13 temperatures, and its expression is:

T=[T ₁ T ₂ T ₃ T ₄ T ₅ T ₆ T ₇ T ₈ T ₉ T ₁₀ T ₁₁ T ₁₂ T ₁₃ ] ^T

P is the loss matrix of the thermal network model, and its expression is:

P=[P ₁ P ₂ P ₃ P ₄ 0 0 0 0 0 0 0 0 -(P ₁ +P ₂ +P ₃ +P ₄ )] ^T ;

C is a heat capacity matrix composed of 12 heat capacities, and its expression is:

G is the thermal conductance matrix corresponding to 20 thermal conductivities, and its expression is:

When the temperature _Ti of the temperature node _Ni of the thermal network model reaches a steady state, equation (1) is simplified to:

GT=P (2)

仿真实验1对应的热网络模型的损耗矩阵U¹、仿真实验2的稳态温度矩阵

仿真实验2对应的热网络模型的损耗矩阵U²、仿真实验3的稳态温度矩阵

仿真实验3对应的热网络模型的损耗矩阵U³、仿真实验4的稳态温度矩阵

和仿真实验4对应的热网络模型的损耗矩阵U⁴满足以下超定方程组：Let the steady-state temperature matrix of simulation experiment 1 be

The loss matrix U ¹ of the thermal network model corresponding to simulation experiment 1 and the steady-state temperature matrix of simulation experiment 2

The loss matrix U ² of the thermal network model corresponding to the simulation experiment 2 and the steady-state temperature matrix of the simulation experiment 3

The loss matrix U ³ of the thermal network model corresponding to simulation experiment 3 and the steady-state temperature matrix of simulation experiment 4

The loss matrix U ⁴ of the thermal network model corresponding to simulation experiment 4 satisfies the following overdetermined equations:

Use the principle of least squares to solve the overdetermined equation system (3), and obtain the thermal conductance G _j of the thermal network model, j=1,2,...20;

Step 3. Identify the heat capacity of the thermal network model

(1) Identify the heat capacity C ₁

For the temperature node N ₁ , let it satisfy the following relation:

赋值给温度结点N₁的温度T₁、温度监测点M₅的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₅的温度T₅，并令P₁＝Au(t)，其中u(t)为单位阶跃函数；The numerical solution _of the transient temperature of the temperature monitoring point M1 obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₁ assigned to the temperature node N ₁ and the transient temperature of the temperature monitoring point M ₅ with the simulation time t

Assign the temperature T ₅ to the temperature node N ₅ , and let P ₁ =Au(t), where u(t) is the unit step function;

The heat capacity C ₁ of the relation (4) is identified based on the least squares method;

(2) Identify the heat capacity C ₂

For the temperature node N ₂ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₂的温度T₂、温度监测点M₆的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₆的温度T₆，并令P₂＝Au(t)，其中u(t)为单位阶跃函数；The numerical solution of the transient temperature of the temperature monitoring point M ₂ obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₂ assigned to the temperature node N ₂ and the transient temperature of the temperature monitoring point M ₆ with the simulation time t

Assign the temperature T ₆ to the temperature node N ₆ , and let P ₂ =Au(t), where u(t) is the unit step function;

The heat capacity C ₂ of the relation (5) is identified based on the least squares method;

(3) Identify the heat capacity C ₅

For the temperature node N ₅ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₁的温度T₁、温度监测点M₅的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₅的温度T₅、温度监测点M₆的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₆的温度T₆、温度监测点M₈的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₈的温度T₈、温度监测点M₉的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₉的温度T₉；The numerical solution _of the transient temperature of the temperature monitoring point M1 obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₁ assigned to the temperature node N ₁ and the transient temperature of the temperature monitoring point M ₅ with the simulation time t

Numerical solution of the temperature T ₅ assigned to the temperature node N ₅ and the transient temperature of the temperature monitoring point M ₆ as a function of the simulation time t

Numerical solution of the temperature T ₆ assigned to the temperature node N ₆ and the transient temperature of the temperature monitoring point M ₈ with the simulation time t

Numerical solution of the temperature T ₈ assigned to the temperature node N ₈ and the transient temperature of the temperature monitoring point M ₉ as a function of the simulation time t

The temperature T _{9 assigned to the temperature node N 9 ;}

The heat capacity C ₅ of the relation (6) is identified based on the least squares method;

(4) Identify the heat capacity C ₆

For the temperature node N ₆ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₂的温度T₂、温度监测点M₅的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₅的温度T₅、温度监测点M₆的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₆的温度T₆、温度监测点M₇的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₇的温度T₇、温度监测点M₁₀的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₀的温度T₁₀；The numerical solution of the transient temperature of the temperature monitoring point M ₂ obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₂ assigned to the temperature node N ₂ and the transient temperature of the temperature monitoring point M ₅ with the simulation time t

Numerical solution of the temperature T ₅ assigned to the temperature node N ₅ and the transient temperature of the temperature monitoring point M ₆ as a function of the simulation time t

Numerical solution of the temperature T ₆ assigned to the temperature node N ₆ and the transient temperature of the temperature monitoring point M ₇ with the simulation time t

Numerical solutions of the temperature T ₇ assigned to the temperature node N ₇ and the transient temperature of the temperature monitoring point M ₁₀ with the simulation time t

The temperature T _{10 assigned to the temperature node N 10 ;}

The heat capacity C ₆ of the relation (7) is identified based on the least square method;

(5) Identify the heat capacity C ₉

For the temperature node N ₉ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₅的温度T₅、温度监测点M₉的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₉的温度T₉、温度监测点M₁₀的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₀的温度T₁₀、温度监测点M₁₂的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₂的温度T₁₂；The numerical solution of the transient temperature of the temperature monitoring point M ₅ obtained by the simulation in step 2 as a function of the simulation time t

Numerical solution of the temperature T ₅ assigned to the temperature node N ₅ and the transient temperature of the temperature monitoring point M ₉ with the simulation time t

Numerical solution of the temperature T ₉ assigned to the temperature node N ₉ and the transient temperature of the temperature monitoring point M ₁₀ with the simulation time t

Numerical solution of the temperature T ₁₀ assigned to the temperature node N ₁₀ and the transient temperature of the temperature monitoring point M ₁₂ with the simulation time t

The temperature T _{12 assigned to the temperature node N 12 ;}

The heat capacity C ₉ of the relation (8) is identified based on the least squares method;

(6) Identify the heat capacity C ₁₀

For the temperature node N ₁₀ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₆的温度T₆、温度监测点M₉的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₉的温度T₉、温度监测点M₁₀的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₀的温度T₁₀、温度监测点M₁₁的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₁的温度T₁₁；The numerical solution of the transient temperature of the temperature monitoring point _M6 obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₆ assigned to the temperature node N ₆ and the transient temperature of the temperature monitoring point M ₉ as a function of the simulation time t

Numerical solution of the temperature T ₉ assigned to the temperature node N ₉ and the transient temperature of the temperature monitoring point M ₁₀ with the simulation time t

Numerical solution of the temperature T ₁₀ assigned to the temperature node N ₁₀ and the transient temperature of the temperature monitoring point M ₁₁ with the simulation time t

The temperature T _{11 assigned to the temperature node N 11 ;}

The heat capacity C ₁₀ of the relation (9) is identified based on the least squares method;

Let C3 _{= C1} , _C4 = _C2 , C7= _C5 , _C8 = _C6 , _C11 = _C9 , _C12 = _C10 , identify ₁₂ heat capacities.

The least square method based on the power module thermal network model parameter identification method disclosed by the invention extracts a thermal network model of a multi-chip power module, and realizes the rapid acquisition of the transient temperature of the chip. Compared with the prior art, the present invention has The beneficial effects are reflected in:

1) It is suitable for extracting the thermal network model of multi-chip power modules, taking into account the thermal coupling effect between chips;

2) The extracted thermal network model is suitable for on-line calculation of chip temperature, and it is easy to be embedded in a circuit simulator for long-term chip temperature calculation;

3) The finite element model is not simplified by the equivalent heat transfer coefficient, and the extracted thermal network model has high accuracy.

Description of drawings

FIG. 1 is a flowchart of a method for identifying parameters of a thermal network model of a power module based on the least squares method according to the present invention.

FIG. 2 is a structural diagram of a thermal network model involved in the identification method of the present invention.

Figure 3 is a schematic diagram of the chip plane layout of the finite element simulation model.

Figure 4 is a schematic cross-sectional view of the finite element simulation model.

FIG. 5 is a graph showing the change of the transient temperature of the temperature monitoring point M ₁ , the temperature monitoring point M ₅ and the temperature monitoring point M ₉ in the simulation experiment 1 with the simulation time t.

FIG. 6 is the transient temperature curve of the temperature node N ₁ of the thermal network model and the temperature monitoring point M ₁ of the finite element model obtained by the simulation experiment.

Detailed ways

The present invention will be further described in detail below with reference to the accompanying drawings.

FIG. 2 is a structural diagram of the thermal network model involved in the identification method of the present invention. It can be seen from FIG. 2 that the thermal network model in the present invention includes 13 temperature nodes, 4 input loss current sources, 20 thermal conductivity and 12 thermal The 13 temperature nodes are denoted as temperature nodes Ni, i is the serial number of temperature nodes, _i =1, 2...13, the temperature of temperature nodes _Ni is denoted as temperature Ti, _i =1, 2...13; the 20 thermal conductivities are recorded as thermal conductance G _j , j is the serial number of the thermal conductance, j=1, 2...20; the 12 thermal capacitances are recorded as thermal capacitance C _k , k is the thermal conductance The serial number of the capacitor, k=1, 2...12; the four input loss current sources are respectively recorded as input loss current source P ₁ , input loss current source P ₂ , input loss current source P ₃ and input loss current source P ₄ .

The thermal network model is a three-dimensional Cauer thermal network structure, which is divided into 4 layers, the temperature node N ₁ , the temperature node N ₂ , the temperature node N ₃ and the temperature node N ₄ are in the first layer, and the temperature node N ₅ , temperature node N ₆ , temperature node N ₇ and temperature node N ₈ are in the second layer, temperature node N ₉ , temperature node N ₁₀ , temperature node N ₁₁ and temperature node N ₁₂ are in the third layer , the temperature node N ₁₃ is in the fourth layer; the temperature node N ₁ , the temperature node N ₅ and the temperature node N ₉ are aligned in the vertical direction of space, and the temperature node N ₂ , the temperature node N ₆ and the temperature node N 9 are aligned in the vertical direction of space. N ₁₀ is aligned in the vertical direction of space, temperature node N ₃ , temperature node N ₇ and temperature node N ₁₁ are aligned in the vertical direction of space, temperature node N ₄ , temperature node N ₈ and temperature node N ₁₂ Align vertically in space.

_The thermal conduction _G1 is set between the temperature node N1 and the temperature node _N5 , the thermal conduction _G2 is set between the temperature node _N2 and the temperature node _N6 , and the thermal conduction _G3 is set between the temperature node N ₃ and the temperature node _N7 , the heat conduction _G4 is set between the temperature node _N4 and the temperature node _N8 , the heat conduction _G5 is set between the temperature node _N5 and the temperature node _N8 , _The heat conduction G6 is set between the temperature node _N5 and the temperature node _N6 , the heat conduction _G7 is set between the temperature node _N6 and the temperature node _N7 , and the heat conduction _G8 is set at the temperature node N ₇ and the temperature node N ₈ , the thermal conductance G ₉ is set between the temperature node N ₅ and the temperature node N ₉ , the thermal conductance G ₁₀ is set between the temperature node N ₆ and the temperature node N ₁₀ , The heat conduction _G11 is set between the temperature node _N7 and the temperature node _N11 , the heat conduction _G12 is set between the temperature node _N8 and the temperature node _N12 , and the heat conduction _G13 is set at the temperature node N ₉ and the temperature node N ₁₂ , the thermal conductance G ₁₄ is set between the temperature node N ₉ and the temperature node N ₁₀ , the thermal conductance G ₁₅ is set between the temperature node N ₁₀ and the temperature node N ₁₁ , The thermal conduction G ₁₆ is set between the temperature node N ₁₁ and the temperature node N ₁₂ , the thermal conduction G ₁₇ is set between the temperature node N ₉ and the temperature node N ₁₃ , and the thermal conduction G ₁₈ is set between the temperature node N ₁₀ and the temperature node N ₁₃ , the heat conduction G ₁₉ is set between the temperature node N ₁₁ and the temperature node N ₁₃ , and the heat conduction G ₂₀ is set between the temperature node N ₁₂ and the temperature node N ₁₃ .

Heat capacity C ₁ , heat capacity C ₂ , heat capacity C ₃ , heat capacity C ₄ , heat capacity C ₅ , heat capacity C ₆ , heat capacity C ₇ , heat capacity C ₈ , heat capacity C ₉ , heat capacity C ₁₀ , One end of the heat capacity C ₁₁ and the heat capacity C ₁₂ are respectively corresponding to the temperature node N ₁ , the temperature node N ₂ , the temperature node N ₃ , the temperature node N ₄ , the temperature node N ₅ , and the temperature node N ₆ . , temperature node N ₇ , temperature node N ₈ , temperature node N ₉ , temperature node N ₁₀ , temperature node N ₁₁ , temperature node N ₁₂ are connected, and the other end is grounded; input loss current source P ₁ , One end of the input loss current source P ₂ , the input loss current source P ₃ , and the input loss current source P ₄ are respectively corresponding to the temperature node N ₁ , the temperature node N ₂ , the temperature node N ₃ , and the temperature node N ₄ . connected, and the other end is grounded.

Fig. 1 is the flow chart of the power module thermal network model parameter identification method based on the least squares method of the present invention, as can be seen from Fig. 1, the parameter identification method of the present invention comprises the following steps:

Step 1, build the finite element simulation model of the power module

The physical size of the power module is sampled, and a finite element simulation model S of the power module including the liquid cooling system is built according to the physical size; the finite element simulation model S includes 8 layers, and the order from top to bottom is chip layer, chip solder layer, upper copper layer, ceramic layer, lower copper layer, substrate solder layer, heat dissipation substrate layer and liquid cooling layer; the chip layer is composed of 2 IGBT chips and 2 Diode chips, and the 2 IGBT chips are respectively recorded as chip X ₁ and chip X ₃ , 2 Diode chips are respectively recorded as chip X ₂ and chip X ₄ .

FIG. 3 is a schematic diagram of a chip plane layout of the finite element simulation model S, and FIG. 4 is a schematic cross-sectional diagram of the finite element simulation model S. As shown in FIG.

A total of 12 temperature monitoring points are set in the finite element simulation model S, and any one of them is recorded as the temperature monitoring point M _m , where m is the serial number of the temperature monitoring point, m=1, 2...12; among them, the temperature monitoring point M ₁ is located in the center _of the chip X1, the temperature monitoring point _M2 is located in the center of the chip _X2 , the temperature monitoring point M3 is located in the center _of the chip _X3 , and the temperature monitoring point _M4 is located in the center of the chip _X4 Location; temperature monitoring point M ₅ , temperature monitoring point M ₆ , temperature monitoring point M ₇ and temperature monitoring point M ₈ are located on the upper copper layer, temperature monitoring point M ₉ , temperature monitoring point M ₁₀ , temperature monitoring point M ₁₁ and temperature monitoring point M 11 The monitoring point M ₁₂ is set on the solder layer of the substrate; the temperature monitoring point M ₅ and the temperature monitoring point M ₉ are aligned with the temperature monitoring point M ₁ in the vertical direction of space, and the temperature monitoring point M ₆ and the temperature monitoring point M ₁₀ are in the vertical direction of space Aligned with the temperature monitoring point M ₂ , the temperature monitoring point M ₇ and the temperature monitoring point M ₁₁ are aligned with the temperature monitoring point M ₃ in the vertical direction of space, and the temperature monitoring point M ₈ and the temperature monitoring point M 12 are aligned with the temperature monitoring point M ₁₂ in the vertical direction of space. The monitoring point M ₄ is aligned.

Set the liquid cooling layer flow velocity V, the liquid cooling layer temperature Y, the simulation termination time T _sim and the logarithmic uniform simulation step size T _step .

其中t为仿真时间。In this example, the liquid cooling layer flow rate V=8L/min, the liquid cooling layer Y=65℃, the simulation termination time _Tsim =100s, and the logarithmic uniform simulation step size

where t is the simulation time.

Step 2, finite element simulation model simulation experiment

记录温度监测点M_m的稳态温度

m＝1,2...12，得到仿真实验1的稳态温度矩阵

和仿真实验1对应的热网络模型的损耗矩阵U¹，

和U¹的表达式如下：Simulation experiment 1, set the loss input P _X1 of chip X ₁ as the step power of amplitude A, the loss input P _X2 of chip X ₂ is 0, the loss input P _X3 of chip X ₃ is 0, and the loss input of chip X ₄ P _X4 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

m=1,2...12, get the steady-state temperature matrix of simulation experiment 1

The loss matrix U ¹ of the thermal network model corresponding to the simulation experiment 1,

and U ¹ are expressed as follows:

U ¹ =[A 0 0 0 0 0 0 0 0 0 0 0 -A] ^T

FIG. 5 is a curve of the transient curves of the temperature monitoring point M ₁ , the temperature monitoring point M ₅ and the temperature monitoring point M ₉ in the simulation experiment 1 changing with the simulation time t.

记录温度监测点M_m的稳态温度

m＝1,2...12，得到仿真实验2的稳态温度矩阵

和仿真实验2对应的热网络模型的损耗矩阵U²，

和U²的表达式如下：Simulation experiment 2, set the loss input P _X2 of chip X ₂ as the step power of amplitude A, the loss input P _X1 of chip X ₁ is 0, the loss input P _X3 of chip X ₃ is 0, and the loss input of chip X ₄ P _X4 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

m=1,2...12, the steady-state temperature matrix of simulation experiment 2 is obtained

The loss matrix U ² of the thermal network model corresponding to simulation experiment 2,

and U ² are expressed as follows:

U ² =[0 A 0 0 0 0 0 0 0 0 0 0 -A] ^T

记录温度监测点M_m的稳态温度

m＝1,2...12，得到仿真实验3的稳态温度矩阵

和仿真实验3对应的热网络模型的损耗矩阵U³，

和U³的表达式如下：Simulation experiment 3, set the loss input P _X3 of chip X ₃ as the step power of amplitude A, the loss input P _X1 of chip X ₁ is 0, the loss input P _X2 of chip X ₂ is 0, and the loss input of chip X ₄ P _X4 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

m=1,2...12, get the steady-state temperature matrix of simulation experiment 3

The loss matrix U ³ of the thermal network model corresponding to simulation experiment 3,

and the expression for ^U3 is as follows:

U ³ =[0 0 A 0 0 0 0 0 0 0 0 0 -A] ^T

记录温度监测点M_m的稳态温度

m＝1,2...12，得到仿真实验4的稳态温度矩阵

和仿真实验4对应的热网络模型的损耗矩阵U⁴，

和U⁴的表达式如下：Simulation experiment 4, set the loss input P _X4 of chip X ₄ as the step power of amplitude A, the loss input P _X1 of chip X ₁ is 0, the loss input P _X2 of chip X ₂ is 0, and the loss input of chip X ₃ P _X3 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

m=1,2...12, the steady-state temperature matrix of simulation experiment 4 is obtained

The loss matrix U ⁴ of the thermal network model corresponding to simulation experiment 4,

and the expression for U ⁴ is as follows:

U ¹ =[0 0 0 A 0 0 0 0 0 0 0 0 -A] ^T

Among them, the loss matrix U ¹ of the thermal network model corresponding to the simulation experiment 1, the loss matrix U ² of the thermal network model corresponding to the simulation experiment 2, the loss matrix U ³ of the thermal network model corresponding to the simulation experiment 3, and the thermal network model corresponding to the simulation experiment 4 . Each element in the loss matrix U ⁴ of the network model has a one-to-one correspondence with the temperature Ti of the temperature node _Ni of the thermal network model, _i =1, 2, . . . 13.

In this example, the amplitude A=100.

Step 3, Calculate the thermal conductivity of the thermal network model

The thermal network model is described by a system of ordinary differential equations with the following expressions:

in,

T is a temperature matrix composed of 13 temperatures, and its expression is:

T=[T ₁ T ₂ T ₃ T ₄ T ₅ T ₆ T ₇ T ₈ T ₉ T ₁₀ T ₁₁ T ₁₂ T ₁₃ ] ^T

P is the loss matrix of the thermal network model, and its expression is:

P=[P ₁ P ₂ P ₃ P ₄ 0 0 0 0 0 0 0 0 -(P ₁ +P ₂ +P ₃ +P ₄ )] ^T ;

C is a heat capacity matrix composed of 12 heat capacities, and its expression is:

G is the thermal conductance matrix corresponding to 20 thermal conductivities, and its expression is:

When the temperature _Ti of the temperature node _Ni of the thermal network model reaches a steady state, equation (1) is simplified to:

GT=P (2)

仿真实验1对应的热网络模型的损耗矩阵U¹、仿真实验2的稳态温度矩阵

仿真实验2对应的热网络模型的损耗矩阵U²、仿真实验3的稳态温度矩阵

仿真实验3对应的热网络模型的损耗矩阵U³、仿真实验4的稳态温度矩阵

和仿真实验4对应的热网络模型的损耗矩阵U⁴满足以下超定方程组：Let the steady-state temperature matrix of simulation experiment 1 be

The loss matrix U ¹ of the thermal network model corresponding to simulation experiment 1 and the steady-state temperature matrix of simulation experiment 2

The loss matrix U ² of the thermal network model corresponding to the simulation experiment 2 and the steady-state temperature matrix of the simulation experiment 3

The loss matrix U ³ of the thermal network model corresponding to simulation experiment 3 and the steady-state temperature matrix of simulation experiment 4

The loss matrix U ⁴ of the thermal network model corresponding to simulation experiment 4 satisfies the following overdetermined equations:

Using the principle of least squares to solve the overdetermined equations (3), the thermal conductance G _j of the thermal network model is obtained, j=1,2,...20.

Step 3. Identify the heat capacity of the thermal network model

(1) Identify the heat capacity C ₁

For the temperature node N ₁ , let it satisfy the following relation:

赋值给温度结点N₁的温度T₁、温度监测点M₅的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₅的温度T₅，并令P₁＝Au(t)，其中u(t)为单位阶跃函数；The numerical solution _of the transient temperature of the temperature monitoring point M1 obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₁ assigned to the temperature node N ₁ and the transient temperature of the temperature monitoring point M ₅ with the simulation time t

Assign the temperature T ₅ to the temperature node N ₅ , and let P ₁ =Au(t), where u(t) is the unit step function;

The heat capacity C ₁ of the relation (4) is identified based on the least squares method;

(2) Identify the heat capacity C ₂

For the temperature node N ₂ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₂的温度T₂、温度监测点M₆的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₆的温度T₆，并令P₂＝Au(t)，其中u(t)为单位阶跃函数；The numerical solution of the transient temperature of the temperature monitoring point M ₂ obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₂ assigned to the temperature node N ₂ and the transient temperature of the temperature monitoring point M ₆ with the simulation time t

Assign the temperature T ₆ to the temperature node N ₆ , and let P ₂ =Au(t), where u(t) is the unit step function;

The heat capacity C ₂ of the relation (5) is identified based on the least squares method;

(3) Identify the heat capacity C ₅

For the temperature node N ₅ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₁的温度T₁、温度监测点M₅的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₅的温度T₅、温度监测点M₆的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₆的温度T₆、温度监测点M₈的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₈的温度T₈、温度监测点M₉的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₉的温度T₉；The numerical solution _of the transient temperature of the temperature monitoring point M1 obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₁ assigned to the temperature node N ₁ and the transient temperature of the temperature monitoring point M ₅ with the simulation time t

Numerical solution of the temperature T ₅ assigned to the temperature node N ₅ and the transient temperature of the temperature monitoring point M ₆ as a function of the simulation time t

Numerical solution of the temperature T ₆ assigned to the temperature node N ₆ and the transient temperature of the temperature monitoring point M ₈ with the simulation time t

Numerical solution of the temperature T ₈ assigned to the temperature node N ₈ and the transient temperature of the temperature monitoring point M ₉ as a function of the simulation time t

The temperature T _{9 assigned to the temperature node N 9 ;}

The heat capacity C ₅ of the relation (6) is identified based on the least squares method;

(4) Identify the heat capacity C ₆

For the temperature node N ₆ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₂的温度T₂、温度监测点M₅的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₅的温度T₅、温度监测点M₆的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₆的温度T₆、温度监测点M₇的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₇的温度T₇、温度监测点M₁₀的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₀的温度T₁₀；The numerical solution of the transient temperature of the temperature monitoring point M ₂ obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₂ assigned to the temperature node N ₂ and the transient temperature of the temperature monitoring point M ₅ with the simulation time t

Numerical solution of the temperature T ₅ assigned to the temperature node N ₅ and the transient temperature of the temperature monitoring point M ₆ as a function of the simulation time t

Numerical solution of the temperature T ₆ assigned to the temperature node N ₆ and the transient temperature of the temperature monitoring point M ₇ with the simulation time t

Numerical solutions of the temperature T ₇ assigned to the temperature node N ₇ and the transient temperature of the temperature monitoring point M ₁₀ with the simulation time t

The temperature T _{10 assigned to the temperature node N 10 ;}

The heat capacity C ₆ of the relation (7) is identified based on the least square method;

(5) Identify the heat capacity C ₉

For the temperature node N ₉ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₅的温度T₅、温度监测点M₉的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₉的温度T₉、温度监测点M₁₀的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₀的温度T₁₀、温度监测点M₁₂的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₂的温度T₁₂；The numerical solution of the transient temperature of the temperature monitoring point M ₅ obtained by the simulation in step 2 as a function of the simulation time t

Numerical solution of the temperature T ₅ assigned to the temperature node N ₅ and the transient temperature of the temperature monitoring point M ₉ with the simulation time t

Numerical solution of the temperature T ₉ assigned to the temperature node N ₉ and the transient temperature of the temperature monitoring point M ₁₀ with the simulation time t

Numerical solution of the temperature T ₁₀ assigned to the temperature node N ₁₀ and the transient temperature of the temperature monitoring point M ₁₂ with the simulation time t

The temperature T _{12 assigned to the temperature node N 12 ;}

The heat capacity C ₉ of the relation (8) is identified based on the least squares method;

(6) Identify the heat capacity C ₁₀

For the temperature node N ₁₀ of the thermal network model, let it satisfy the following relation:

赋值给温度结点N₆的温度T₆、温度监测点M₉的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₉的温度T₉、温度监测点M₁₀的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₀的温度T₁₀、温度监测点M₁₁的瞬态温度随仿真时间t变化的数值解

赋值给温度结点N₁₁的温度T₁₁；The numerical solution of the transient temperature of the temperature monitoring point _M6 obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₆ assigned to the temperature node N ₆ and the transient temperature of the temperature monitoring point M ₉ as a function of the simulation time t

Numerical solution of the temperature T ₉ assigned to the temperature node N ₉ and the transient temperature of the temperature monitoring point M ₁₀ with the simulation time t

Numerical solution of the temperature T ₁₀ assigned to the temperature node N ₁₀ and the transient temperature of the temperature monitoring point M ₁₁ with the simulation time t

The temperature T _{11 assigned to the temperature node N 11 ;}

The heat capacity C ₁₀ of the relation (9) is identified based on the least squares method;

Let C3 _{= C1} , _C4 = _C2 , C7= _C5 , _C8 = _C6 , _C11 = _C9 , _C12 = _C10 , identify ₁₂ heat capacities.

In order to verify the effectiveness of the present invention, the present invention is simulated and verified.

Set the loss input of chip X1 as P _{X1 =} -50sin(2πt)+|50sin( _2πt )|, the loss input of chip X2 as P _X2 =50sin(2πt)+|50sin(2πt)|, the loss of chip _X3 Input P _X3 =50sin(2πt)+|50sin(2πt)|, and input P _X4 =-50sin(2πt)+|50sin(2πt)| for the loss of chip _X4 .

Build the thermal network model in MATLAB/Simulink, and set the input loss current source P ₁ =-50sin(2πt)+|50sin(2πt)|, the input loss current source P ₂ =50sin(2πt)+|50sin(2πt)| , the input loss current source P ₃ =50sin(2πt)+|50sin(2πt)|, and the input loss current source P4= _−50sin (2πt)+|50sin(2πt)|.

Figure 6 shows the transient temperature curve of the temperature node N1 _of the thermal network model obtained from the simulation experiment and the transient temperature curve of the temperature monitoring point M1 _of the finite element model, and the transient temperature change of the temperature node N1 _of the thermal network model It is basically consistent with the transient temperature change of the temperature monitoring point M1 _of the finite element model, and the maximum error does not exceed 1%.

Claims (1)

1. A power module thermal network model parameter identification method based on the least squares method, the thermal network model comprises 13 temperature nodes, 4 input loss current sources, 20 thermal conductivities and 12 thermal capacitances; the 13 Each temperature node is recorded as temperature node Ni, i is the serial number of temperature node, _i =1, 2...13, the temperature of temperature node Ni is recorded as temperature Ti, _i = ₁ , 2...13; The 20 thermal conductivities are recorded as thermal conductance G _j , j is the serial number of thermal conductance, j=1,2...20; the 12 thermal capacitances are recorded as thermal capacitance C _k , k is the serial number of thermal capacitance, k =1, 2...12; the four input loss current sources are respectively denoted as input loss current source P ₁ , input loss current source P ₂ , input loss current source P ₃ and input loss current source P ₄ ; The thermal network model is a three-dimensional Cauer thermal network structure, which is divided into 4 layers, the temperature node N ₁ , the temperature node N ₂ , the temperature node N ₃ and the temperature node N ₄ are in the first layer, and the temperature node N ₅ , temperature node N ₆ , temperature node N ₇ and temperature node N ₈ are in the second layer, temperature node N ₉ , temperature node N ₁₀ , temperature node N ₁₁ and temperature node N ₁₂ are in the third layer , the temperature node N ₁₃ is in the fourth layer; the temperature node N ₁ , the temperature node N ₅ and the temperature node N ₉ are aligned in the vertical direction of space, and the temperature node N ₂ , the temperature node N ₆ and the temperature node N 9 are aligned in the vertical direction of space. N ₁₀ is aligned in the vertical direction of space, temperature node N ₃ , temperature node N ₇ and temperature node N ₁₁ are aligned in the vertical direction of space, temperature node N ₄ , temperature node N ₈ and temperature node N ₁₂ Align in the vertical direction of space; _The thermal conduction _G1 is set between the temperature node N1 and the temperature node _N5 , the thermal conduction _G2 is set between the temperature node _N2 and the temperature node _N6 , and the thermal conduction _G3 is set between the temperature node N ₃ and the temperature node _N7 , the heat conduction _G4 is set between the temperature node _N4 and the temperature node _N8 , the heat conduction _G5 is set between the temperature node _N5 and the temperature node _N8 , _The heat conduction G6 is set between the temperature node _N5 and the temperature node _N6 , the heat conduction _G7 is set between the temperature node _N6 and the temperature node _N7 , and the heat conduction _G8 is set at the temperature node N ₇ and the temperature node N ₈ , the thermal conductance G ₉ is set between the temperature node N ₅ and the temperature node N ₉ , the thermal conductance G ₁₀ is set between the temperature node N ₆ and the temperature node N ₁₀ , The heat conduction _G11 is set between the temperature node _N7 and the temperature node _N11 , the heat conduction _G12 is set between the temperature node _N8 and the temperature node _N12 , and the heat conduction _G13 is set at the temperature node N ₉ and the temperature node N ₁₂ , the thermal conductance G ₁₄ is set between the temperature node N ₉ and the temperature node N ₁₀ , the thermal conductance G ₁₅ is set between the temperature node N ₁₀ and the temperature node N ₁₁ , The thermal conduction G ₁₆ is set between the temperature node N ₁₁ and the temperature node N ₁₂ , the thermal conduction G ₁₇ is set between the temperature node N ₉ and the temperature node N ₁₃ , and the thermal conduction G ₁₈ is set between the temperature node N ₁₀ and the temperature node _N13 , the heat conduction _G19 is set between the temperature node _N11 and the temperature node _N13 , and the heat conduction _G20 is set between the temperature node _N12 and the temperature node _N13 ; Heat capacity C ₁ , heat capacity C ₂ , heat capacity C ₃ , heat capacity C ₄ , heat capacity C ₅ , heat capacity C ₆ , heat capacity C ₇ , heat capacity C ₈ , heat capacity C ₉ , heat capacity C ₁₀ , One end of the heat capacity C ₁₁ and the heat capacity C ₁₂ are respectively corresponding to the temperature node N ₁ , the temperature node N ₂ , the temperature node N ₃ , the temperature node N ₄ , the temperature node N ₅ , and the temperature node N ₆ . , temperature node N ₇ , temperature node N ₈ , temperature node N ₉ , temperature node N ₁₀ , temperature node N ₁₁ , temperature node N ₁₂ are connected, and the other end is grounded; input loss current source P ₁ , One end of the input loss current source P ₂ , the input loss current source P ₃ , and the input loss current source P ₄ are respectively corresponding to the temperature node N ₁ , the temperature node N ₂ , the temperature node N ₃ , and the temperature node N ₄ . connected, the other end is grounded; It is characterized in that, the steps of described identification method are as follows: Step 1, build the finite element simulation model of the power module The physical size of the power module is sampled, and a finite element simulation model S of the power module including the liquid cooling system is built according to the physical size; the finite element simulation model S includes 8 layers, and the order from top to bottom is chip layer, chip solder layer, upper copper layer, ceramic layer, lower copper layer, substrate solder layer, heat dissipation substrate layer and liquid cooling layer; the chip layer is composed of 2 IGBT chips and 2 Diode chips, and the 2 IGBT chips are respectively recorded as chip X ₁ and chip X ₃ , 2 Diode chips are respectively recorded as chip X ₂ and chip X ₄ ; A total of 12 temperature monitoring points are set in the finite element simulation model S, and any one of them is recorded as the temperature monitoring point M _m , where m is the serial number of the temperature monitoring point, m=1, 2...12; among them, the temperature monitoring point M ₁ is located in the center _of the chip X1, the temperature monitoring point _M2 is located in the center of the chip _X2 , the temperature monitoring point M3 is located in the center _of the chip _X3 , and the temperature monitoring point _M4 is located in the center of the chip _X4 Location; temperature monitoring point M ₅ , temperature monitoring point M ₆ , temperature monitoring point M ₇ and temperature monitoring point M ₈ are located on the upper copper layer, temperature monitoring point M ₉ , temperature monitoring point M ₁₀ , temperature monitoring point M ₁₁ and temperature monitoring point M 11 The monitoring point M ₁₂ is set on the solder layer of the substrate; the temperature monitoring point M ₅ and the temperature monitoring point M ₉ are aligned with the temperature monitoring point M ₁ in the vertical direction of space, and the temperature monitoring point M ₆ and the temperature monitoring point M ₁₀ are in the vertical direction of space Aligned with the temperature monitoring point M ₂ , the temperature monitoring point M ₇ and the temperature monitoring point M ₁₁ are aligned with the temperature monitoring point M ₃ in the vertical direction of space, and the temperature monitoring point M ₈ and the temperature monitoring point M 12 are aligned with the temperature monitoring point M ₁₂ in the vertical direction of space. The monitoring point M ₄ is aligned; Set the liquid cooling layer flow velocity V, the liquid cooling layer temperature Y, the simulation termination time T _sim and the logarithmic uniform simulation step size T _step ; Step 2, finite element simulation model simulation experiment Simulation experiment 1, set the loss input P _X1 of chip X ₁ as the step power of amplitude A, the loss input P _X2 of chip X ₂ is 0, the loss input P _X3 of chip X ₃ is 0, and the loss input of chip X ₄ P _X4 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

Obtain the steady-state temperature matrix of simulation experiment 1

The loss matrix U ¹ of the thermal network model corresponding to the simulation experiment 1,

and U ¹ are expressed as follows:

U ¹ =[A 0 0 0 0 0 0 0 0 0 0 0 -A] ^T Simulation experiment 2, set the loss input P _X2 of chip X ₂ as the step power of amplitude A, the loss input P _X1 of chip X ₁ is 0, the loss input P _X3 of chip X ₃ is 0, and the loss input of chip X ₄ P _X4 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

Obtain the steady-state temperature matrix of simulation experiment 2

The loss matrix U ² of the thermal network model corresponding to simulation experiment 2,

and U ² are expressed as follows:

U ² =[0 A 0 0 0 0 0 0 0 0 0 0 -A] ^T Simulation experiment 3, set the loss input P _X3 of chip X ₃ as the step power of amplitude A, the loss input P _X1 of chip X ₁ is 0, the loss input P _X2 of chip X ₂ is 0, and the loss input of chip X ₄ P _X4 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

Obtain the steady-state temperature matrix of simulation experiment 3

The loss matrix U ³ of the thermal network model corresponding to simulation experiment 3,

and the expression for ^U3 is as follows:

U ³ =[0 0 A 0 0 0 0 0 0 0 0 0 -A] ^T Simulation experiment 4, set the loss input P _X4 of chip X ₄ as the step power of amplitude A, the loss input P _X1 of chip X ₁ is 0, the loss input P _X2 of chip X ₂ is 0, and the loss input of chip X ₃ P _X3 is 0, and the simulation obtains the numerical solution of the transient temperature of the temperature monitoring point M _m changing with the simulation time t

Record the steady-state temperature of the temperature monitoring point M _m

Obtain the steady-state temperature matrix of simulation experiment 4

The loss matrix U ⁴ of the thermal network model corresponding to simulation experiment 4,

and the expression for U ⁴ is as follows:

U ¹ =[0 0 0 A 0 0 0 0 0 0 0 0 -A] ^T Among them, the loss matrix U ¹ of the thermal network model corresponding to simulation experiment 1, the loss matrix U ² of the thermal network model corresponding to simulation experiment 2, the loss matrix U ³ of the thermal network model corresponding to simulation experiment 3, and the thermal network model corresponding to simulation experiment 4 Each element in the loss matrix U ⁴ of the network model corresponds to the temperature T _i of the temperature node Ni of the thermal network model one-to-one, _i =1, 2, . . . 13; Step 3, Calculate the thermal conductivity of the thermal network model The thermal network model is described by a system of ordinary differential equations with the following expressions:

in, T is a temperature matrix composed of 13 temperatures, and its expression is: T=[T ₁ T ₂ T ₃ T ₄ T ₅ T ₆ T ₇ T ₈ T ₉ T ₁₀ T ₁₁ T ₁₂ T ₁₃ ] ^T P is the loss matrix of the thermal network model, and its expression is: P=[P ₁ P ₂ P ₃ P ₄ 0 0 0 0 0 0 0 0 -(P ₁ +P ₂ +P ₃ +P ₄ )] ^T ; C is a heat capacity matrix composed of 12 heat capacities, and its expression is:

G is the thermal conductance matrix corresponding to 20 thermal conductivities, and its expression is:

When the temperature _Ti of the temperature node _Ni of the thermal network model reaches a steady state, equation (1) is simplified to: GT=P (2) Let the steady-state temperature matrix of simulation experiment 1 be

The loss matrix U ¹ of the thermal network model corresponding to simulation experiment 1 and the steady-state temperature matrix of simulation experiment 2

The loss matrix U ² of the thermal network model corresponding to the simulation experiment 2 and the steady-state temperature matrix of the simulation experiment 3

The loss matrix U ³ of the thermal network model corresponding to simulation experiment 3 and the steady-state temperature matrix of simulation experiment 4

The loss matrix U ⁴ of the thermal network model corresponding to simulation experiment 4 satisfies the following overdetermined equations:

Use the principle of least squares to solve the overdetermined equation system (3), and obtain the thermal conductance G _j of the thermal network model, j=1,2,...20; Step 4: Identify the heat capacity of the thermal network model (1) Identify the heat capacity C ₁ For the temperature node N ₁ , let it satisfy the following relation:

Assign the numerical solution T ₁ ¹ (t) of the transient temperature of the temperature monitoring point M ₁ obtained by the simulation in step 2 to the simulation time t to the temperature T ₁ of the temperature node N ₁ and the transient temperature of the temperature monitoring point M ₅ Numerical solution as a function of simulation time t

Assign the temperature T ₅ to the temperature node N ₅ , and let P ₁ =Au(t), where u(t) is the unit step function; The heat capacity C ₁ of the relation (4) is identified based on the least squares method; (2) Identify the heat capacity C ₂ For the temperature node N ₂ of the thermal network model, let it satisfy the following relation:

The numerical solution of the transient temperature of the temperature monitoring point M ₂ obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₂ assigned to the temperature node N ₂ and the transient temperature of the temperature monitoring point M ₆ with the simulation time t

Assign the temperature T ₆ to the temperature node N ₆ , and let P ₂ =Au(t), where u(t) is the unit step function; The heat capacity C ₂ of the relation (5) is identified based on the least squares method; (3) Identify the heat capacity C ₅ For the temperature node N ₅ of the thermal network model, let it satisfy the following relation:

Assign the numerical solution T ₁ ¹ (t) of the transient temperature of the temperature monitoring point M ₁ obtained by the simulation in step 2 to the simulation time t to the temperature T ₁ of the temperature node N ₁ and the transient temperature of the temperature monitoring point M ₅ Numerical solution as a function of simulation time t

Numerical solution of the temperature T ₅ assigned to the temperature node N ₅ and the transient temperature of the temperature monitoring point M ₆ as a function of the simulation time t

Numerical solution of the temperature T ₆ assigned to the temperature node N ₆ and the transient temperature of the temperature monitoring point M ₈ with the simulation time t

Numerical solution of the temperature T ₈ assigned to the temperature node N ₈ and the transient temperature of the temperature monitoring point M ₉ as a function of the simulation time t

The temperature T _{9 assigned to the temperature node N 9 ;} The heat capacity C ₅ of the relation (6) is identified based on the least squares method; (4) Identify the heat capacity C ₆ For the temperature node N ₆ of the thermal network model, let it satisfy the following relation:

The numerical solution of the transient temperature of the temperature monitoring point M ₂ obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₂ assigned to the temperature node N ₂ and the transient temperature of the temperature monitoring point M ₅ with the simulation time t

Numerical solution of the temperature T ₅ assigned to the temperature node N ₅ and the transient temperature of the temperature monitoring point M ₆ as a function of the simulation time t

Numerical solution of the temperature T ₆ assigned to the temperature node N ₆ and the transient temperature of the temperature monitoring point M ₇ with the simulation time t

Numerical solutions of the temperature T ₇ assigned to the temperature node N ₇ and the transient temperature of the temperature monitoring point M ₁₀ with the simulation time t

The temperature T _{10 assigned to the temperature node N 10 ;} The heat capacity C ₆ of the relation (7) is identified based on the least square method; (5) Identify the heat capacity C ₉ For the temperature node N ₉ of the thermal network model, let it satisfy the following relation:

The numerical solution of the transient temperature of the temperature monitoring point M ₅ obtained by the simulation in step 2 as a function of the simulation time t

Numerical solution of the temperature T ₅ assigned to the temperature node N ₅ and the transient temperature of the temperature monitoring point M ₉ with the simulation time t

Numerical solution of the temperature T ₉ assigned to the temperature node N ₉ and the transient temperature of the temperature monitoring point M ₁₀ with the simulation time t

Numerical solution of the temperature T ₁₀ assigned to the temperature node N ₁₀ and the transient temperature of the temperature monitoring point M ₁₂ with the simulation time t

The temperature T _{12 assigned to the temperature node N 12 ;} The heat capacity C ₉ of the relation (8) is identified based on the least squares method; (6) Identify the heat capacity C ₁₀ For the temperature node N ₁₀ of the thermal network model, let it satisfy the following relation:

The numerical solution of the transient temperature of the temperature monitoring point _M6 obtained by the simulation in step 2 as a function of the simulation time t

Numerical solutions of the temperature T ₆ assigned to the temperature node N ₆ and the transient temperature of the temperature monitoring point M ₉ as a function of the simulation time t

Numerical solution of the temperature T ₉ assigned to the temperature node N ₉ and the transient temperature of the temperature monitoring point M ₁₀ with the simulation time t

Numerical solution of the temperature T ₁₀ assigned to the temperature node N ₁₀ and the transient temperature of the temperature monitoring point M ₁₁ with the simulation time t

The temperature T _{11 assigned to the temperature node N 11 ;} The heat capacity C ₁₀ of the relation (9) is identified based on the least squares method; Let C3 _{= C1} , _C4 = _C2 , C7= _C5 , _C8 = _C6 , _C11 = _C9 , _C12 = _C10 , identify ₁₂ heat capacities.

Images