CN117192266A - Online monitoring method for junction temperature of power devices in new energy vehicle inverters - Google Patents

Abstract

The invention discloses an online monitoring method for junction temperature of power devices in a new energy vehicle inverter. The method includes determining the attributes of each layer of the power module and establishing a three-dimensional model in simulation software; conducting instantaneous monitoring under different conditions for the power module and the entire bench. State thermal impedance test and analysis to obtain transient thermal impedance curves under different conditions and reference points, especially data under different coupling states; adjust the established transient thermal impedance curves in the simulation software based on the measured transient thermal impedance curves of different sets of data The parameters of the model; the calibrated three-dimensional simulation analysis model is used to construct a two-dimensional thermal network model under different coupling conditions; the junction temperature is monitored online based on the established two-dimensional thermal network model and compared with the experimental and simulation results. The invention also establishes an aging compensation device. Even if the thermal network model is modified to address the aging problem, the invention calculates the junction temperature based on the established thermal network model with low hardware requirements and fast response.

Description

Online monitoring method for junction temperature of power devices in new energy vehicle inverters

Technical field

The invention belongs to the field of power electronics technology, and specifically relates to a method for online monitoring of junction temperature of power devices in a new energy vehicle inverter.

Background technique

The main drive inverter is a key part of the electric powertrain. It is responsible for converting the DC voltage of the high-voltage battery (350-800VDC) into the AC voltage of the three-phase AC sinusoidal current to rotate the induction motor and drive the vehicle forward. Common power levels are 40kW to 250+kW. When a 400V-800V battery powers the main drive inverter, the rated voltage of the inverter components needs to be 600V-1200V, while operating current levels up to 1000A per phase. The performance of this module affects the overall energy efficiency of the vehicle, and the junction temperature has always been the most important parameter and indicator of the power device. It is also the key to the reliable operation of the entire inverter and is an industry problem. Therefore, an accurate online evaluation method of junction temperature can not only lay the foundation for the long-term reliability of the device, but also provide a methodological basis for its remaining life evaluation.

The IGBT module is an important component of the new energy vehicle inverter. The biggest factor affecting the service life of the IGBT is the junction temperature. When the junction temperature of the IGBT fluctuates greatly during operation, its failure rate will rise rapidly and its service life will also decrease. will decrease accordingly. Moreover, changes in junction temperature will cause changes in the thermal physical parameters of the IGBT module, making the modeling of IGBT complex. Therefore, monitoring the operating junction temperature of IGBT is of great significance for its life prediction and failure mechanism analysis.

Junction temperature online monitoring methods mainly include optical measurement method, physical contact method, temperature-sensitive electrical parameter method, and thermal network model method. Although the optical measurement method has high accuracy and can obtain the temperature distribution of the entire chip surface, especially the dynamic change process of the bonding wire temperature, it requires destroying the entire module package during measurement and is not practical for application. The physical contact methods are mainly thermocouples and thermistors. Although their measurement methods are simple, they have shortcomings such as slow response speed and large errors in practical applications. The temperature-sensitive electrical parameter method uses the chip or module itself as a temperature sensor, selects a device parameter affected by temperature as a characterization quantity, and establishes the corresponding relationship between the characterization quantity and junction temperature through passive heating. This process is called temperature Coefficient calibration, referred to as calibration. When the device is working, the power loss of the chip generates heat to actively heat the module. At this time, by measuring the characterization quantity and then comparing it with the calibration relationship, the operating junction temperature can be inferred. However, the current thermoelectric parameter method also has corresponding shortcomings, such as affecting the normal operation of the device, low sensitivity of the measurement parameters, and being extremely susceptible to interference, etc., which will be greatly restricted when applied to complex working conditions.

The basic principle of the thermal network model method is to construct a thermal path network from the junction temperature to the known temperature node based on the thermoelectric analogy theory, thereby achieving online monitoring and prediction of the junction temperature. As a junction temperature monitoring method recommended by device manufacturers, this method has low hardware requirements, is simple and easy to implement, and has the characteristics of low implementation cost and fast response, so it has become the most potential junction temperature extraction method. Thermal network model method can be divided into one-dimensional, two-dimensional and three-dimensional thermal network models. The one-dimensional thermal network model ignores the cross-thermal coupling phenomenon and has a large error with the actual junction temperature test. The three-dimensional thermal network model is highly dependent on the finite element method and is difficult to apply in practice.

Although the conventional two-dimensional thermal network model method takes into account the coupling between chips, the data fitted to the two-dimensional thermal network model established by the conventional method are obtained through simulation methods and are not calibrated experimentally. In this way, There will be a certain error between the junction temperature estimated by the established thermal network model and the actual temperature.

The main drive inverter is a key part of the electric powertrain, and the power device is an important part of the main drive inverter. The junction temperature is the most important factor affecting the reliability of the power device, so a good junction temperature monitoring Methods can lay the foundation for long-term device reliability.

Contents of the invention

In order to solve the above technical problems, the present invention provides a method for online monitoring of junction temperature of power devices in new energy vehicle inverters. Through experiments, transient thermal impedance curves under different conditions are measured, and a thermal network model is established to monitor junction temperature online. This method has low hardware requirements and fast response. The junction temperature can be calculated according to the pre-established model by simply measuring the ambient temperature.

In order to achieve the above objects, the present invention adopts the following technical solutions:

A method for online monitoring of junction temperature of power devices in new energy vehicle inverters, including the following steps:

Step 1: Establish a three-dimensional simulation analysis model of the power module based on the tested sample;

Step 2: Conduct transient thermal impedance test analysis under different conditions for the power module and the entire power device test bench;

Step 3: Calibrate the three-dimensional simulation analysis model;

Step 4: Construct a two-dimensional thermal network model under different coupling conditions based on the calibrated three-dimensional simulation analysis model;

Step 5: Verify the two-dimensional thermal network model.

Beneficial effects:

Compared with conventional methods, the present invention adopts a two-dimensional thermal network model, which takes into account the mutual coupling between chips and has greater accuracy. In addition, conventional methods for extracting transient thermal impedance are usually extracted from data sheets provided by suppliers or using finite element simulation. Compared with the experimentally extracted transient thermal impedance, these methods are not close enough to reality and have certain errors. , the junction temperature cannot be accurately estimated.

Description of the drawings

Figure 1 is the schematic diagram of temperature calibration;

Figure 2 is a schematic diagram of the saturation voltage drop under small current;

Figure 3 is a thermal network model diagram considering thermal coupling;

Figure 4 is a flow chart of the online junction temperature monitoring method of the power device in the new energy vehicle inverter of the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

As shown in Figure 4, the online junction temperature monitoring method of power devices in new energy vehicle inverters of the present invention includes the following steps:

Step 1: Establish a three-dimensional simulation analysis model of the power module based on the sample under test, including:

Step 1.1: According to the IGBT module structure, the three-dimensional simulation analysis model established in the simulation software can be divided into chip layer, upper copper layer, solder layer, ceramic layer, lower copper layer, substrate layer, TIM layer and heat sink.

Step 1.2: Based on the actual object, measure the specific dimensions of the chip layer, upper copper layer, solder layer, ceramic layer, lower copper layer, substrate layer, TIM layer and heat sink, and establish a complete three-dimensional simulation analysis model based on the measurement data.

Step 1.3: In the simulation software, select the material of the chip as Si, the material of the solder layer as Sn-Ag-Cu, the material of the ceramic layer as Al ₂ O ₃ , the material of the upper and lower copper layers as Cu, and the material of the TIM layer. The material is thermally conductive silicone grease.

Step 1.4: Input the density, thermal conductivity and constant pressure heat capacity data of Si, Sn-Ag-Cu, Al ₂ O ₃ , and thermal grease materials into the simulation software. The specific parameters are shown in Table 1.

Table 1

Step 2: Conduct transient thermal impedance test analysis under different conditions for the power module and the entire power device test bench, including:

Step 2.1: Use the small current saturation voltage drop method to measure the device junction temperature.

The testing principle of the saturation voltage drop method under low current is: because the voltage Vce and junction temperature _Tj show a linear relationship, after the load current of the device under test is cut off, a small current is passed, and the saturation at both ends of the device under test under low current can be measured. Voltage drop, the junction temperature can be obtained by comparing the relationship between the calibrated voltage Vce and the junction temperature _Tj .

Measuring junction temperature using saturation voltage drop under low current requires calibration in advance. See Figure 1 for the calibration schematic, where G is the gate, C is the collector, E is the emitter, and Isense is the measurement current. The principle is to give the gate a conduction voltage of 15v to turn on the device. When the device is turned on, the voltage Vce and the junction temperature _Tj show a linear relationship. The device is passively heated through the induction cooker, and the data of the voltage Vce and the junction temperature _Tj are recorded. Through the linear The coefficient k between Vce and T _j is obtained by fitting. The specific relationship is as follows:

_Tj =kVce+b

Among them, b is a constant.

After obtaining the relationship between voltage Vce and junction temperature _Tj , the junction temperature data when the device is cooled is measured through the circuit diagram shown in Figure 2. Among them, K is the switch and I _load is the load current.

Step 2.2: Calculate the transient thermal impedance curve under cooling conditions:

The transient thermal impedance curve can be calculated according to the following formula:

Among them, Z _th is the transient thermal impedance, Ta _is the ambient temperature, which can be changed by adjusting the temperature of the water-cooling plate, and P is the power loss;

Step 2.3: Calculate the data under different coupling states. The first step is to measure the junction temperature of a single IGBT in the power module when it is working, and calculate the respective transient thermal impedance curves. Then measure the chip occurrence when two IGBTs are working. The junction temperature of thermal coupling and the new transient thermal impedance curve, up to the junction temperature and respective transient thermal impedance curves of the chip when 6 IGBTs work simultaneously.

Step 2.4: Obtain transient thermal impedance curves for different conditions and reference points.

Change the ambient temperature, that is, adjust the temperature and heat dissipation conditions of the water-cooling plate, and repeat step 2.3 to obtain new transient thermal impedance data.

Step 3: Calibrate the 3D simulation analysis model, including:

Step 3.1: Set the current input and output terminals of the three-dimensional simulation analysis model, the current flow path, and the initial value of the current. Set the bottom of the heat sink as a boundary condition and give a heat transfer coefficient. Set the thermal expansion and electromagnetic heat in multiphysics as coupled and add a boundary probe at the chip.

Step 3.2: Use transient simulation to conduct research, read the data of the chip cooling in the three-dimensional simulation analysis model, and calculate the simulated transient thermal impedance curve Z _{th_t} according to the formula. Compare the two sets of transient thermal impedance curves obtained from the experiment and simulation. Comparative analysis is performed. If there is a difference in the transient thermal impedance curve data before 0.01s, the area and thickness of the chip layer of the three-dimensional simulation analysis model are adjusted. The transient thermal impedance after 0.01s is adjusted for Si, Cu, and Al ₂ O. The thermal conductivity and heat capacity of ₃ are modified.

Step 4: Construct a two-dimensional thermal network model under different coupling conditions based on the calibrated three-dimensional simulation analysis model, including:

Step 4.1: Obtain the calibrated transient thermal impedance curve Z _{th_1} , Z _{th_2} , Z _{th_3} , Z _{th_4} , Z _{th_5} , Z _{th_6} and establish a self-impedance matrix Z _self whose expression is as follows;

Since the two-dimensional thermal network model takes thermal coupling into account, the thermal impedance of mutual coupling between chips is also required. When a loss is applied to chip n, the coupling thermal impedance between the temperature reference point and chip m is defined as follows

Among them, _Tm is the junction temperature of chip m, and _Pn is the power loss of chip n.

According to the calculated coupling thermal impedance of each chip, a coupling thermal impedance matrix Z _couple can be established. The expression of Z _couple is as follows

According to the superposition principle, the chip junction temperature calculation expression when taking into account the influence of thermal coupling is:

T _j =(Z _couple +Z _self )P+T _a

After discretizing Z _th in the thermal impedance matrix formula according to the actual sampling time, the discretized thermal network for online extraction of junction temperature can be obtained. Based on this, the coupled thermal impedance network model structure is shown in Figure 3.

In Figure 3, taking the first branch as an example, the power loss P _{loss_1} is applied to the device to obtain the junction temperature T _j1 of the device. The heat is transferred downward along the branch, and the ambient temperature Ta.Zth_ha is obtained through the energy loss of _{Zth_ch1} and Zth_ha. is the thermal impedance of the radiator. T _j1 -T _jn is the junction temperature of each chip, P _loss1 -P _lossn-1 is the power loss of each chip, Z _{th_ch1} -Z _{th_chn} is the transient thermal impedance between the shell of each chip and the radiator.

Step 4.2 In the Foster thermal network model, the thermal resistance and heat capacity parameter values are fitted by the thermal impedance curve. Generally, multi-order exponential curve fitting can be used to obtain the thermal impedance parameter Z _{th_m} (t) between the two, that is :

Among them, t is time, R _i is thermal resistance, n is fitting order, and i is a natural number;

The time constant τ is determined by the thermal resistance and heat capacity of each order, and the expression is:

τ _i =R _i C _i

Among them, C _i is the thermal resistance.

Self-heating impedance and coupled thermal impedance are usually fitted using the nonlinear fitting function lsqcurvefit. The higher the fitting order, the more accurate the junction temperature calculation. However, too high a fitting order will also cause the problem of too long calculation time. . The self-heating impedance of the present invention is fitted using a fourth-order RC to obtain the values of R _i and C _i . The coupling thermal impedance is fitted using a first-order RC to obtain the respective coupling thermal resistance R _{(m, n)} and coupling heat capacity. C _(m,n) .

Step 5: Verify the two-dimensional thermal network model, including:

Step 5.1: Use experimental methods to verify. Place the device in the circuit diagram shown in Figure 2, intermittently turn on the current I _load cycle to heat the device, and turn off the current I _load to cool down the device. The turn-on and turn-off times are t _on and t _off respectively, and the ratio of t _on and t _off is 1:2. When I _load is turned off, measure the voltage Vce across the device under the applied current I _sense . According to the calibration relationship between Vce and T _j , obtain and record the experimentally measured junction temperature T _{j_e} .

The ambient temperature T _a is measured, and the value of T _a is brought into the established thermal network model to verify whether it is consistent with the junction temperature fluctuation measured experimentally.

Step 5.2: Use simulation method to verify, set the current value I _load applied to the device in the simulation software, determine the values of t _on and t _off , heat the device under test during t _on time, and measure the chip cooling during t _off time. When the temperature T _{j_t} is measured, compare the measured T _{j_t} with the value calculated by the two-dimensional thermal network model to verify whether the junction temperature fluctuation trend between the two is similar and whether the error is small.

It is easy for those skilled in the art to understand that the above descriptions are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements, etc., made within the spirit and principles of the present invention, All should be included in the protection scope of the present invention.

Claims (6)

1. The on-line junction temperature monitoring method for the power devices in the new energy automobile inverter is characterized by comprising the following steps of:

step 1: establishing a three-dimensional simulation analysis model of the power module according to the measured sample;

step 2: carrying out transient thermal impedance test analysis under different conditions aiming at the power module and the whole power device test bench;

step 3: calibrating a three-dimensional simulation analysis model;

step 4: constructing a two-dimensional thermal network model under different coupling conditions according to the calibrated three-dimensional simulation analysis model;

step 5: and (5) performing two-dimensional thermal network model verification.

2. The method for on-line monitoring the junction temperature of a power device in an inverter of a new energy automobile according to claim 1, wherein the step 1 comprises:

step 1.1: according to the IGBT module structure, the established three-dimensional simulation analysis is divided into a chip layer, an upper copper layer, a solder layer, a ceramic layer, a lower copper layer, a substrate layer, a TIM layer and a radiator;

step 1.2: according to the physical object, measuring the specific dimensions of the chip layer, the upper copper layer, the solder layer, the ceramic layer, the lower copper layer, the substrate layer, the TIM layer and the radiator, and establishing a complete three-dimensional simulation analysis model according to the measured data;

step 1.3: the material of the selected chip is Si, the material of the solder layer is Sn-Ag-Cu, and the material of the ceramic layer is Al ₂ O ₃ The upper copper layer and the lower copper layer are made of Cu, and the TIM layer is made of heat-conducting silicone grease;

step 1.4: inputting Si, sn-Ag-Cu and Al ₂ O ₃ Data of density, thermal conductivity and constant pressure heat capacity of the thermally conductive silicone grease material.

3. The method for on-line monitoring the junction temperature of a power device in an inverter of a new energy automobile according to claim 2, wherein the step 2 comprises:

step 2.1: measuring the junction temperature of the device by adopting a small current saturation voltage drop method, switching on a small current after the load current of the device to be measured is cut off, measuring to obtain the saturation voltage drop at two ends of the device to be measured under the small current, and comparing the calibrated voltage Vce with the junction temperature T _j Obtaining junction temperature according to the relation of the temperature;

the junction temperature is measured by saturation voltage drop under small current for calibration in advance, the tested device is passively heated by an electromagnetic oven, and the voltage Vce and the junction temperature T are recorded _j Data, vce and T are obtained by linear fitting _j The coefficient k between the two is expressed as follows:

T _j ＝kVce+b

wherein b is a constant;

obtaining the voltage Vce and the junction temperature T _j After the relation, junction temperature data of the device under the condition of cooling is measured;

step 2.2: calculating a transient thermal impedance curve under the condition of temperature reduction:

the transient thermal impedance curve is calculated according to the following formula:

wherein Z is _th T is transient thermal impedance _a The temperature of the water cooling plate is adjusted to be the ambient temperature, and P is the power loss;

step 2.3: calculating data under different coupling states, firstly measuring the junction temperature of each IGBT in a power module when the single IGBT works, calculating to obtain respective transient thermal impedance curves, and then measuring the junction temperature of the chip and new transient thermal impedance curves when the two IGBTs work until the junction temperature of the chip and the respective transient thermal impedance curves under the condition that the 6 IGBTs work and are coupled simultaneously;

step 2.4: obtaining transient thermal impedance curves of different conditions and reference points, namely changing the ambient temperature, namely adjusting the temperature of the water cooling plate and the heat dissipation condition, and repeating the step 2.3 to obtain new transient thermal impedance data.

4. The method for on-line monitoring the junction temperature of a power device in an inverter of a new energy automobile according to claim 3, wherein the step 3 comprises:

step 3.1: setting a current input/output end, a current flow path and an initial value of the current of the three-dimensional simulation analysis model; setting the bottom of the radiator as a boundary condition, setting a heat transfer coefficient, setting thermal expansion and electromagnetic heat in multiple physical fields as coupling, and adding a boundary probe at the chip;

step 3.2: the transient simulation is used for research, data of a three-dimensional simulation analysis model during chip cooling is read, and a simulated transient thermal impedance curve Z is obtained according to formula calculation _{th_t} Comparing and analyzing two groups of transient thermal impedance curves obtained by experiments and simulation, if the transient thermal impedance curve data are different before 0.01s, adjusting the area and the thickness of a chip layer of a three-dimensional simulation analysis model, and adjusting Si, cu and Al according to the transient thermal impedance after 0.01s ₂ O ₃ Is modified by the thermal conductivity and heat capacity of the material.

5. The method for on-line monitoring the junction temperature of a power device in an inverter of a new energy automobile according to claim 4, wherein the step 4 comprises:

step 4.1: obtaining a calibrated transient thermal impedance curve Z _{th_1} 、Z _{th_2} 、Z _{th_3} 、Z _{th_4} 、Z _{th_5} 、Z _{th_6} Establishing a self-impedance matrix Z _self The expression is as follows;

when a loss is applied to chip n, the coupling thermal impedance between the temperature reference point and chip m is defined as follows:

wherein T is _m For junction temperature of chip m, P _n Power loss for chip n;

establishing a coupling thermal impedance matrix Z according to the calculated coupling thermal impedance of each chip _couple ，Z _couple The expression of (2) is as follows:

the chip junction temperature calculation expression when the thermal coupling influence is taken into account according to the superposition principle is as follows:

T _j ＝(Z _couple +Z _self )P+T _a

z in the thermal resistance matrix _th Discretizing according to the actual sampling time to obtain a discretized thermal network for online junction temperature extraction;

step 4.2 in the Foster thermal network model, the thermal impedance parameter Z between the two is obtained by adopting multi-order exponential curve fitting _{th_m} (t), namely:

wherein t is time, R _i Is thermal resistance, n is fitting order, i is natural number i;

the time constant τ is determined by the thermal resistance and heat capacity of each stage, and the expression is:

τ _i ＝R _i C _i

wherein C is _i Is heat capacity;

fitting the self-heating impedance by adopting fourth-order RC to obtain R _i 、C _i The coupling thermal impedance is fitted by adopting a first-order RC to obtain respective coupling thermal resistance R _(m，n) And coupled heat capacity C _(m，n) 。

6. The method for on-line monitoring the junction temperature of a power device in an inverter of a new energy automobile according to claim 5, wherein the step 5 comprises:

step 5.1: verifying by experimental method, and intermittently turning on current I _load Cycling the device to heat, turning off current I _load Cooling the device; the on and off times are t respectively _on And t _off ，t _on And t _off The ratio of (2) is 1:2; at I _load Measuring applied current I when turned off _sense The voltage Vce across the lower device, according to Vce and T _j Obtained and recorded the experimentally measured junction temperature T _{j_e} ；

Measuring and obtaining the ambient temperature T _a Will T _a The values of (2) are brought into the established thermal network model, and whether the values are consistent with junction temperature fluctuation measured by experiments or not is verified;

step 5.2: verification by simulation, setting the current value I applied to the device _load Determining t _on 、t _off At t _on Time, heat the device under test, t _off Temperature T of chip during temperature reduction _{j_t} To be measured T _{j_t} And comparing the calculated values with the two-dimensional thermal network model, and verifying whether the junction temperature fluctuation trend between the two is similar or not and whether the error is within 5%.