CN111767673A - IGBT junction temperature measuring method and device, computer equipment and storage medium - Google Patents
IGBT junction temperature measuring method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111767673A CN111767673A CN202010423056.3A CN202010423056A CN111767673A CN 111767673 A CN111767673 A CN 111767673A CN 202010423056 A CN202010423056 A CN 202010423056A CN 111767673 A CN111767673 A CN 111767673A
- Authority
- CN
- China
- Prior art keywords
- junction temperature
- igbt
- igbt junction
- sample data
- random forest
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 66
- 238000003860 storage Methods 0.000 title claims abstract description 8
- 238000007637 random forest analysis Methods 0.000 claims abstract description 59
- 238000003066 decision tree Methods 0.000 claims description 43
- 238000009529 body temperature measurement Methods 0.000 claims description 23
- 238000012549 training Methods 0.000 claims description 22
- 238000004088 simulation Methods 0.000 claims description 17
- 238000000513 principal component analysis Methods 0.000 claims description 12
- 238000004590 computer program Methods 0.000 claims description 10
- 238000004140 cleaning Methods 0.000 claims description 9
- 230000002159 abnormal effect Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 238000001931 thermography Methods 0.000 abstract description 7
- 238000004519 manufacturing process Methods 0.000 abstract description 4
- 238000005259 measurement Methods 0.000 description 6
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 239000003550 marker Substances 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000010291 electrical method Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001351 cycling effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000017525 heat dissipation Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 230000035882 stress Effects 0.000 description 1
- 238000005382 thermal cycling Methods 0.000 description 1
- 230000008646 thermal stress Effects 0.000 description 1
- 230000000930 thermomechanical effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/243—Classification techniques relating to the number of classes
- G06F18/24323—Tree-organised classifiers
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Testing Of Individual Semiconductor Devices (AREA)
Abstract
The invention relates to an IGBT junction temperature measuring method, an IGBT junction temperature measuring device, computer equipment and a storage medium, wherein the method comprises the steps of obtaining characteristic parameters of IGBT junction temperature to be tested; inputting characteristic parameters of IGBT junction temperature into a preset random forest model; and determining the IGBT junction temperature corresponding to the characteristic parameters according to the output result of the random forest model. The method is mainly executed through software, complex thermal imaging equipment is not needed, and a complex measuring circuit is not needed to be installed, so that the method has the characteristics of simple hardware and contribution to industrial mass production. The IGBT junction temperature can be accurately measured and calculated through the random forest model, so that the IGBT junction temperature measuring method with high accuracy and strong industrial practicability is realized.
Description
Technical Field
The invention relates to the technical field of power electronics, in particular to an IGBT junction temperature measuring method, an IGBT junction temperature measuring device, computer equipment and a storage medium.
Background
An Insulated Gate Bipolar Transistor (IGBT) is a core device for energy conversion and transmission, and is generally directly applied to equipment such as a frequency converter and an Uninterruptible Power Supply (UPS). Due to high frequency switching, power semiconductor devices are under power cycling and thermal cycling induced thermal pulses for long periods of time, and therefore the constant changes in thermo-mechanical stress caused by junction temperature changes are very damaging to the internal structure of the IGBT, leading to device failure. Research shows that the higher the working junction temperature of the device is, the smaller the safety margin is; the larger the junction temperature fluctuation, the shorter the thermal cycle life. By monitoring the junction temperature of the IGBT, the thermal stress impact of the device can be reduced by adopting a proper control method, so that the service life of the module is prolonged, and the failure rate is reduced.
Currently, there are generally three main methods for measuring the junction temperature of a power semiconductor device: optical methods, physical contact methods, and electrical methods. The optical method is simple and easy to operate, but the thermal imaging equipment is expensive and is not easy to install, so that the thermal imaging equipment is difficult to apply to engineering; the physical contact method can measure the temperature more accurately, but the chip in the package can not be directly contacted, and the measured temperature has certain deviation; the electrical method has high measurement speed and high precision, but needs a complex measurement circuit and can influence the normal operation of the device to a certain extent. Therefore, how to provide a method for measuring the junction temperature of the IGBT with high accuracy and strong industrial practicability on the premise of not requiring a complex measurement circuit and expensive thermal imaging equipment is a problem to be solved urgently at present.
Disclosure of Invention
Therefore, it is necessary to provide a method, an apparatus, computer equipment and a storage medium for measuring the junction temperature of the IGBT aiming at the problems of low accuracy and complex hardware of the existing IGBT junction temperature measuring method, and the IGBT junction temperature measurement with high accuracy can be realized by simple hardware equipment.
The invention provides an IGBT junction temperature measuring method which is characterized by comprising the following steps: acquiring characteristic parameters of IGBT junction temperature to be tested; inputting the characteristic parameters of the IGBT junction temperature into a preset random forest model; and determining the IGBT junction temperature corresponding to the characteristic parameters according to the output result of the random forest model.
According to the IGBT junction temperature measuring method, a random forest model of the IGBT junction temperature is obtained through pre-training, when IGBT junction temperature detection is carried out, the obtained characteristic parameters related to the IGBT junction temperature are input into the random forest model, and the IGBT junction temperature to be tested is determined according to the output result of the random forest model. The method is mainly executed through software, complex thermal imaging equipment is not needed, and a complex measuring circuit is not needed to be installed, so that the method has the characteristics of simple hardware and contribution to industrial mass production. The IGBT junction temperature can be accurately measured and calculated through the random forest model, so that the IGBT junction temperature measuring method with high accuracy and strong industrial practicability is realized.
In one embodiment, the characteristic parameters of the IGBT junction temperature include attribute parameters and operating parameters of the IGBT device.
In one embodiment, before the step of inputting the characteristic parameter of the IGBT junction temperature into a preset random forest model, the method further includes the steps of: constructing a thermal simulation model of the IGBT inverter; acquiring a plurality of groups of sample data, wherein the sample data comprises IGBT junction temperature and related parameters thereof; and training the random forest model according to the multiple groups of sample data.
In one embodiment, the step of constructing the thermal simulation model of the IGBT inverter specifically includes: according to the relation among the IGBT junction temperature, the IGBT power loss, the first thermal resistance, the second thermal resistance, the third thermal resistance and the environment temperature, a thermal simulation model of the IGBT inverter is constructed; the first thermal resistance is a junction-shell thermal resistance, the second thermal resistance is a thermal resistance from a shell to a radiator, and the third thermal resistance is a thermal resistance from the radiator to the environment.
In one embodiment, the step of acquiring multiple sets of sample data includes: selecting multiple groups of sample data which accord with preset conditions, wherein the preset conditions comprise at least one of a voltage range, a current range, a switching frequency range and an environment temperature range; and performing data cleaning on the multiple groups of sample data to filter abnormal data in the multiple groups of sample data.
In one embodiment, each set of the sample data includes a plurality of feature quantities; the step of obtaining a plurality of groups of sample data further comprises: performing principal component analysis on the cleaned sample data to determine principal characteristic quantities from the plurality of characteristic quantities; and carrying out layered sampling processing on the main characteristic quantities in the multiple groups of sample data.
In one embodiment, each set of the sample data includes a plurality of feature quantities; the step of obtaining a plurality of groups of sample data further comprises: performing principal component analysis on the cleaned sample data to determine non-relevant characteristic quantities from the plurality of characteristic quantities; and deleting the non-relevant characteristic quantities in the plurality of groups of sample data.
In one embodiment, the random forest model comprises a plurality of decision trees; the step of inputting the characteristic parameters of the IGBT junction temperature into a preset random forest model specifically comprises the following steps: inputting characteristic parameters of the IGBT junction temperature into the decision trees; the step of determining the IGBT junction temperature corresponding to the characteristic parameter according to the output result of the random forest model comprises the following steps: obtaining an output result of each decision tree; and taking the average value of the output results of the decision trees as the IGBT junction temperature corresponding to the characteristic parameter.
The invention also provides an IGBT junction temperature measuring device, which comprises: the obtaining module is used for obtaining characteristic parameters of the junction temperature of the IGBT to be tested; the input module is used for inputting the characteristic parameters of the IGBT junction temperature into a preset random forest model; and the determining module is used for determining the IGBT junction temperature corresponding to the characteristic parameters according to the output result of the random forest model.
The invention also provides a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the method of any one of claims 1 to 8 when executing the computer program.
The invention also provides a computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 8.
Drawings
Fig. 1 is a schematic flow chart of an IGBT junction temperature measurement method according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of an IGBT junction temperature measurement method according to another embodiment of the present invention;
fig. 3 is a schematic flow chart of a method for measuring junction temperature of an IGBT according to another embodiment of the present invention;
fig. 4 is a schematic flowchart of a method for measuring junction temperature of an IGBT according to another embodiment of the present invention;
fig. 5 is a schematic block structure diagram of an IGBT junction temperature measurement apparatus according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a computer device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In one embodiment, the present invention provides an IGBT junction temperature measurement method, as shown in fig. 1, including the following steps:
and S110, acquiring characteristic parameters of the IGBT junction temperature to be tested.
The characteristic parameter of the junction temperature of the IGBT refers to a physical parameter related to the junction temperature of the IGBT. For example, the characteristic parameters of the IGBT junction temperature may include attribute parameters and operating parameters of the IGBT device.
The attribute parameters of the IGBT device are inherent attribute parameters of the IGBT device and can be obtained through a data manual of the IGBT device. For example, it may be obtained by receiving a data manual transmitted from a user input/external device. For another example, the attribute parameters of the IGBT device may be prestored, and the prestored database may be queried to obtain the parameters when detecting the IGBT junction temperature. The present application does not limit the manner of obtaining the characteristic parameter.
The operation parameters of the IGBT device are real-time parameters of the IGBT during operation, including peak current IpkModulation ratio m, power factor cos phi, switching frequency fswBus voltage VdcAnd the ambient temperature TaAnd/or the like.
And S130, inputting the characteristic parameters of the IGBT junction temperature into a preset random forest model.
The random forest model is a classifier model which is based on a machine learning algorithm and used for training and predicting samples by utilizing a plurality of trees.
In this embodiment, a random forest model is obtained by a machine learning algorithm and pre-training of a plurality of samples according to a relationship between the junction temperature of the IGBT and characteristic parameters thereof. When the IGBT junction temperature needs to be detected, the characteristic parameters of the IGBT junction temperature to be detected are input into a pre-trained random forest model, and then the corresponding junction temperature can be measured and calculated through the random forest model.
Because the random forest model is classified by adopting a plurality of decision trees, and each decision tree is trained according to a randomly extracted sample in the training process, the accuracy is very high, and a large number of input variables can be processed.
And S150, determining the IGBT junction temperature corresponding to the characteristic parameters according to the output result of the random forest model.
The random forest model comprises a plurality of decision trees, after characteristic parameters related to the junction temperature of the IGBT to be tested are input into the random forest model, each decision tree respectively carries out junction temperature measurement and calculation according to the characteristic parameters and outputs a measurement and calculation result. Optionally, the measurement result of one of the decision trees may be selected as the IGBT junction temperature corresponding to the characteristic parameter; or, taking the average value of the test results of the decision trees as the IGBT junction temperature corresponding to the characteristic parameter.
The IGBT junction temperature measuring method provided by the embodiment of the invention is characterized in that a random forest model of the IGBT junction temperature is obtained by pre-training, when the IGBT junction temperature is detected, the obtained characteristic parameters related to the IGBT junction temperature are input into the random forest model, and the IGBT junction temperature to be tested is determined according to the output result of the random forest model. The method is mainly executed through software, complex thermal imaging equipment is not needed, and a complex measuring circuit is not needed to be installed, so that the method has the characteristics of simple hardware and contribution to industrial mass production. The IGBT junction temperature can be accurately measured and calculated through the random forest model, so that the IGBT junction temperature measuring method with high accuracy and strong industrial practicability is realized.
In one embodiment, before step S130, that is, before the step of inputting the characteristic parameter of the junction temperature into the preset random forest model, the IGBT junction temperature measurement method further includes a process of training the random forest model. As shown in fig. 2, the process of training the random forest model specifically includes the following steps:
s121, constructing a thermal simulation model of the IGBT inverter;
specifically, because the temperature of the IGBT module depends on the power loss, the thermal resistance and the ambient temperature, a thermal simulation model of the IGBT inverter can be constructed according to the relationship between the IGBT junction temperature and the IGBT power loss, the first thermal resistance, the second thermal resistance, the third thermal resistance and the ambient temperature; the first thermal resistance is junction-to-shell thermal resistance, the second thermal resistance is thermal resistance from the shell to the radiator, and the third thermal resistance is thermal resistance from the radiator to the environment.
In this embodiment, a thermal simulation model of the IGBT inverter may be constructed by professional heat dissipation simulation optimization analysis software. For example, a thermal simulation model of the inverter is constructed by ANSYS Icepak as follows:
Tj＝P×(Rthj-c+Rthc-hs+Rthhs-a)+Ta；
wherein, TjFor IGBT junction temperature, P for IGBT power loss, Rthj-cIs incrustation thermal resistance, Rthc-hsIs the thermal resistance from the case to the heat sink, Rthhs-aFrom heat sink to environmentThermal resistance, TaIs ambient temperature.
Specifically, fluid-solid coupling simulation can be performed according to boundary conditions through the thermal simulation model, so that IGBT junction temperature data can be obtained. For example, boundary conditions may be set including: the voltage range is 100-400V, the current range is 5-15A, the switching frequency range is 1-10 kHz, the ambient temperature range is 20-50 ℃, and multiple groups of junction temperature and related parameter data are collected according to the conditions and the fluid-solid coupling simulation result.
In the above equation, the power loss can be calculated, but the analysis of the inverter thermal resistance is very troublesome. Therefore, the heat conduction form of the IGBT in the inverter is analyzed by adopting ANSYS, so that IGBT junction temperature data including thermal resistance data are obtained, the obtaining process of the IGBT junction temperature data can be simplified, and the IGBT junction temperature measuring efficiency is improved.
And S123, acquiring multiple groups of sample data, wherein the sample data comprises IGBT junction temperature and relevant parameters thereof.
In this embodiment, the sample data is used to train the random forest model to obtain a model that can be finally used for predicting the IGBT junction temperature.
Because the thermal simulation model is established according to the relation between the IGBT junction temperature and the power loss of the IGBT module, the thermal resistance and the ambient temperature, each group of sample data can comprise IGBT junction temperature data, IGBT module power data, thermal resistance data and ambient temperature data.
For example, each set of sample data may include an IGBT junction temperature and a peak current I corresponding to the IGBT junction temperaturepkModulation ratio m, power factor cos phi, switching frequency fswBus voltage VdcAnd the ambient temperature Ta。
And S125, training the random forest model according to the multiple groups of sample data.
In this embodiment, a plurality of sets of acquired sample data are input into a preset random forest model, so that the random forest model is trained, and the method is suitable for predicting the junction temperature of the IGBT, and the accuracy of measuring the junction temperature of the IGBT is improved.
In an embodiment, as shown in fig. 3, the step S123 of acquiring multiple sets of sample data includes:
and S1231, selecting multiple groups of sample data meeting preset conditions. The preset condition includes at least one of a voltage range, a current range, a switching frequency range and an ambient temperature range.
The IGBT junction temperature under the steady state is mainly determined by the output voltage, the current, the switching frequency and the ambient temperature of the inverter, so that the voltage, the current, the switching frequency and the ambient temperature can be used as the characteristic attributes of the sample, the value range condition of each characteristic attribute is set, and the data of which the value meets the condition is selected as the sample data. For example, the voltage range can be set to be 100-400V, the current range is 5-15A, the switching frequency range is 1-10 kHz, and the environmental temperature range is 20-50 ℃.
And S1233, performing data cleaning on the multiple groups of sample data to filter abnormal data in the multiple groups of sample data.
The data cleaning is mainly used for filtering and screening abnormal data in the sample data so as to ensure the accuracy of the sample data. The abnormal data may include at least one of incomplete data, error data, repeated data, non-required data, and the like.
In this embodiment, the applicable cleaning rule may be predefined according to experience, and the abnormal data may be removed by the predefined cleaning rule.
In one embodiment, each set of sample data includes a plurality of feature quantities; as shown in fig. 4, after performing data cleaning on multiple sets of sample data, the step S123, namely, the step of acquiring multiple sets of sample data, further includes:
and S1235, performing principal component analysis on the cleaned sample data to determine a principal feature quantity from the plurality of feature quantities.
The characteristic attributes of the sample data include voltage, current, switching frequency and ambient temperature, correspondingly, the characteristic quantities included in the sample data are respectively a voltage value, a current value, a switching frequency value and an ambient temperature value, and each group of sample data includes the above four characteristic quantities.
For the sample data after data cleaning, normalization processing may be performed on the above four feature quantities, and then the variance contribution ratio of each feature quantity may be analyzed by using a Principal Component Analysis (PCA) method. After the principal component analysis, the feature quantity having the largest influence on the sample marker can be used as the principal feature quantity. For example, in the present embodiment, the current value is used as the main feature amount because the main component analysis method finds that the variance contribution rate of the phase current is the largest, that is, the current has the largest influence on the sample mark.
And S1237, performing layered sampling processing on the main characteristic quantity in the multiple groups of sample data.
Specifically, the determined main characteristic quantity is layered in a unit of a certain range, and a preset amount of data is collected for each layer until the total number of the collected sample data reaches a preset total number. For example, taking the main characteristic quantity as a current value as an example, the current data is processed by using a hierarchical sampling method, specifically: the current values are layered by taking 1A as a unit, and 20 groups of data are collected in each layer, namely 20 groups of data are collected in each 1A, and 200 groups of data are collected in total.
The collected sample data can be used as a training set for training the model. The collected sample data can also be divided into a training set and a test set, the training set is used for model training, and the test set is used for model performance verification. For example, for 200 sets of data collected, the data may be collected as 8: the scale of 2 is divided into a training set and a test set. The training set needs to be preprocessed and used for training, and the test set is used for prediction to reveal the performance of the model.
In one embodiment, each set of sample data includes a plurality of feature quantities; as shown in fig. 5, after performing data cleaning on multiple sets of sample data, the step of acquiring multiple sets of sample data further includes:
and S1239, performing principal component analysis on the cleaned sample data to determine an uncorrelated feature quantity from the plurality of feature quantities.
For a specific implementation of principal component analysis on the cleaned sample data, reference may be made to the foregoing embodiment, which is not described herein again.
In this embodiment, through principal component analysis, not only the principal feature quantity having a large influence on the sample marker among the four feature quantities but also the uncorrelated feature quantity having a small influence on the sample marker among the four feature quantities can be determined. For example, through principal component analysis, it is found that the variance contribution rate of the voltage value is minimal, that is, the influence of the voltage on the sample mark is minimal, so the voltage value can be used as the uncorrelated feature quantity.
And S1240, deleting the non-relevant characteristic quantity in the multiple groups of sample data.
Since the influence of the non-relevant feature quantity on the sample marking is minimum, the non-relevant feature quantity is determined and then deleted from the sample data in the embodiment, so that the non-relevant data in the sample data is removed, and the rationality of the sample data is improved.
In this embodiment, if the sample mark is found to be possibly wrong, the simulation model may be run again for verification.
In one embodiment, the random forest model includes a plurality of decision trees; the step of inputting the characteristic parameters of the junction temperature into the preset random forest model specifically comprises the following steps: inputting characteristic parameters of junction temperature into a plurality of decision trees; the step of determining the IGBT junction temperature corresponding to the characteristic parameters according to the output result of the random forest model comprises the following steps: obtaining an output result of each decision tree; and taking the average value of the output results of the decision trees as the IGBT junction temperature corresponding to the characteristic parameter.
The random forest algorithm is an algorithm which combines a plurality of weak learners (namely, a base decision tree) into a strong learner based on an integrated learning idea. In one embodiment, the regression problem takes the average of the sum of all base decision tree results for the final output result, and the classification problem is chosen by "minority-majority-compliant".
The establishment of the random forest model is described below by way of a specific example.
In this embodiment, 10 base decision trees may be selected during the random forest model training, the maximum feature number is selected to be 3 when a single base decision tree is constructed, and the minimum value of the sample included in the leaf node is set to be 1 as a default value. The establishment of the random forest model comprises the following two steps of generation of a decision tree and model combination.
When the decision tree is generated, assuming that the feature space has D features in total, in each round of decision tree generation, D features (D < D) are randomly selected from the D features to form a new feature set, and the decision tree is generated by using the new feature set. And generating k decision trees through k rounds of decision tree generation. Since the k decision trees are random in both the selection of the training set and the selection of the features, the k decision trees are independent of each other.
When model combination is performed, because the generated k decision trees are independent from each other, and the importance of each decision tree is equal, the decision trees can be set to have the same weight, and the average value output in all decision making is used as the final output result.
For example, the feature space has 3 features, and in each round of generating the decision tree, 1 feature of the 3 features is randomly selected to form a new feature set, and the new feature set is used to generate the decision tree. Through 10 rounds of decision tree generation processes, 10 decision trees are generated together. Since the 10 decision trees are random in both the selection of the training set and the selection of the features, the generated 10 decision trees are independent of each other. The 10 decision tree models are then combined together. Since the generated 10 decision trees are independent of each other, and the importance of each decision tree is equal, the same weight value is set for each decision tree, and the average value of the outputs of all decisions is used as the final output result.
In order to overcome the defects of low industrial practicability, low measurement precision and poor real-time performance of the conventional IGBT junction temperature measurement method, the IGBT junction temperature measurement method provides a method for obtaining IGBT junction temperature data by using inverter thermal simulation and then predicting the IGBT junction temperature based on a random forest algorithm model. The method is simple to implement, can be widely applied to various different power devices and different working conditions, and is high in prediction precision and strong in real-time performance.
In another aspect of the present invention, an IGBT junction temperature measuring apparatus is further provided, as shown in fig. 5, the apparatus includes an obtaining module 501, an inputting module 503, and a determining module 507. The obtaining module 501 is configured to obtain a characteristic parameter of the junction temperature of the IGBT to be tested; the input module 503 is configured to input the characteristic parameter of the IGBT junction temperature into a preset random forest model; the determining module 507 is configured to determine an IGBT junction temperature corresponding to the characteristic parameter according to an output result of the random forest model.
The IGBT junction temperature measuring device is characterized in that a random forest model of IGBT junction temperature is obtained through pre-training, when IGBT junction temperature detection is carried out, the obtained characteristic parameters related to the IGBT junction temperature are input into the random forest model, and the IGBT junction temperature to be tested is determined according to the output result of the random forest model. The method is mainly executed through software, complex thermal imaging equipment is not needed, and a complex measuring circuit is not needed to be installed, so that the method has the characteristics of simple hardware and contribution to industrial mass production. The IGBT junction temperature can be accurately measured and calculated through the random forest model, so that the IGBT junction temperature measuring method with high accuracy and strong industrial practicability is realized.
In an embodiment, the present application further provides an IGBT junction temperature measurement device, which includes a functional module corresponding to the IGBT junction temperature measurement method described in any of the above embodiments, and is configured to implement each step of the IGBT junction temperature measurement method.
In another aspect, the present invention further provides a computer device, as shown in fig. 6, the computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, the steps of the IGBT junction temperature measurement method according to any of the above embodiments are implemented.
Yet another aspect of the embodiments of the present invention further provides a readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements the steps of the IGBT junction temperature measurement method according to any of the embodiments described above.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only show some embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (11)
1. An IGBT junction temperature measurement method, characterized in that the method comprises the following steps:
acquiring characteristic parameters of IGBT junction temperature to be tested;
inputting the characteristic parameters of the IGBT junction temperature into a preset random forest model;
and determining the IGBT junction temperature corresponding to the characteristic parameters according to the output result of the random forest model.
2. The IGBT junction temperature measurement method of claim 1,
the characteristic parameters of the IGBT junction temperature comprise attribute parameters and operation parameters of the IGBT device.
3. The IGBT junction temperature measurement method according to claim 1, wherein before the step of inputting the characteristic parameters of the IGBT junction temperature into a preset random forest model, the method further comprises the steps of:
constructing a thermal simulation model of the IGBT inverter;
acquiring a plurality of groups of sample data, wherein the sample data comprises IGBT junction temperature and related parameters thereof;
and training the random forest model according to the multiple groups of sample data.
4. The IGBT junction temperature measurement method according to claim 3, wherein the step of constructing the thermal simulation model of the IGBT inverter specifically includes:
according to the relation among the IGBT junction temperature, the IGBT power loss, the first thermal resistance, the second thermal resistance, the third thermal resistance and the environment temperature, a thermal simulation model of the IGBT inverter is constructed;
the first thermal resistance is a junction-shell thermal resistance, the second thermal resistance is a thermal resistance from a shell to a radiator, and the third thermal resistance is a thermal resistance from the radiator to the environment.
5. The IGBT junction temperature measurement method of claim 3, wherein the step of acquiring multiple sets of sample data comprises:
selecting multiple groups of sample data which accord with preset conditions, wherein the preset conditions comprise at least one of a voltage range, a current range, a switching frequency range and an environment temperature range;
and performing data cleaning on the multiple groups of sample data to filter abnormal data in the multiple groups of sample data.
6. The IGBT junction temperature measurement method according to claim 5, wherein each set of the sample data comprises a plurality of characteristic quantities; the step of obtaining a plurality of groups of sample data further comprises:
performing principal component analysis on the cleaned sample data to determine principal characteristic quantities from the plurality of characteristic quantities;
and carrying out layered sampling processing on the main characteristic quantities in the multiple groups of sample data.
7. The IGBT junction temperature measurement method according to claim 5, wherein each set of the sample data comprises a plurality of characteristic quantities; the step of obtaining a plurality of groups of sample data further comprises:
performing principal component analysis on the cleaned sample data to determine non-relevant characteristic quantities from the plurality of characteristic quantities;
and deleting the non-relevant characteristic quantities in the plurality of groups of sample data.
8. The IGBT junction temperature measurement method according to any one of claims 1-7, wherein the random forest model comprises a plurality of decision trees; the step of inputting the characteristic parameters of the IGBT junction temperature into a preset random forest model specifically comprises the following steps: inputting characteristic parameters of the IGBT junction temperature into the decision trees;
the step of determining the IGBT junction temperature corresponding to the characteristic parameter according to the output result of the random forest model comprises the following steps:
obtaining an output result of each decision tree;
and taking the average value of the output results of the decision trees as the IGBT junction temperature corresponding to the characteristic parameter.
9. An IGBT junction temperature measurement device, characterized by comprising:
the obtaining module is used for obtaining characteristic parameters of the junction temperature of the IGBT to be tested;
the input module is used for inputting the characteristic parameters of the IGBT junction temperature into a preset random forest model;
and the determining module is used for determining the IGBT junction temperature corresponding to the characteristic parameters according to the output result of the random forest model.
10. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1 to 8 are implemented when the computer program is executed by the processor.
11. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010423056.3A CN111767673A (en) | 2020-05-19 | 2020-05-19 | IGBT junction temperature measuring method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010423056.3A CN111767673A (en) | 2020-05-19 | 2020-05-19 | IGBT junction temperature measuring method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111767673A true CN111767673A (en) | 2020-10-13 |
Family
ID=72719521
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010423056.3A Pending CN111767673A (en) | 2020-05-19 | 2020-05-19 | IGBT junction temperature measuring method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111767673A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101567256B1 (en) * | 2014-11-06 | 2015-11-13 | 현대자동차주식회사 | Apparatus and method for predicting junction temperature for inverter |
| CN105785788A (en) * | 2015-11-26 | 2016-07-20 | 华中科技大学 | Rapid three-phase voltage phase-locked loop method and dynamic response performance analyzing method thereof |
| CN109145965A (en) * | 2018-08-02 | 2019-01-04 | 深圳辉煌耀强科技有限公司 | Cell recognition method and device based on random forest disaggregated model |
| CN110133466A (en) * | 2019-05-16 | 2019-08-16 | 上海金脉电子科技有限公司 | The junction temperature calculation method and system of IGBT module |
| CN110286279A (en) * | 2019-06-05 | 2019-09-27 | 武汉大学 | Based on extreme random forest and the sparse Fault Diagnosis of Power Electronic Circuits method from encryption algorithm of stacking-type |
| US10885259B2 (en) * | 2019-08-30 | 2021-01-05 | Intel Corporation | Random forest model for prediction of chip layout attributes |
-
2020
- 2020-05-19 CN CN202010423056.3A patent/CN111767673A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101567256B1 (en) * | 2014-11-06 | 2015-11-13 | 현대자동차주식회사 | Apparatus and method for predicting junction temperature for inverter |
| CN105785788A (en) * | 2015-11-26 | 2016-07-20 | 华中科技大学 | Rapid three-phase voltage phase-locked loop method and dynamic response performance analyzing method thereof |
| CN109145965A (en) * | 2018-08-02 | 2019-01-04 | 深圳辉煌耀强科技有限公司 | Cell recognition method and device based on random forest disaggregated model |
| CN110133466A (en) * | 2019-05-16 | 2019-08-16 | 上海金脉电子科技有限公司 | The junction temperature calculation method and system of IGBT module |
| CN110133466B (en) * | 2019-05-16 | 2021-08-17 | 上海金脉电子科技有限公司 | Junction temperature calculation method and system of IGBT module |
| CN110286279A (en) * | 2019-06-05 | 2019-09-27 | 武汉大学 | Based on extreme random forest and the sparse Fault Diagnosis of Power Electronic Circuits method from encryption algorithm of stacking-type |
| US10885259B2 (en) * | 2019-08-30 | 2021-01-05 | Intel Corporation | Random forest model for prediction of chip layout attributes |
Non-Patent Citations (3)
| Title |
|---|
| H. NIU;: "Sensing IGBT junction temperature using gate drive output transient properties", 2015 IEEE APPLIED POWER ELECTRONICS CONFERENCE AND EXPOSITION (APEC), pages 2492 - 2499 * |
| 姚芳;王少杰;李志刚;: "逆变器中IGBT功率模块的电热联合仿真模型", 半导体技术, no. 06 * |
| 李昊宇: "基于智能信息处理技术的IGBT可靠性分析方法研究", 中国优秀硕士学位论文全文数据库工程科技Ⅱ辑, no. 2, pages 042 - 581 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111460617B (en) | IGBT junction temperature prediction method based on neural network | |
| CN109143094B (en) | Abnormal data detection method and device for power battery | |
| CN103649858B (en) | For the method and apparatus of prediction component or system status | |
| CN110838075A (en) | Training and predicting method and device for prediction model of transient stability of power grid system | |
| CN113759225B (en) | IGBT residual life prediction and state evaluation realization method | |
| CN109284543B (en) | IGBT residual life prediction method based on optimal scale Gaussian process model | |
| KR102648764B1 (en) | Battery performance evaluation method and battery performance evaluation device | |
| EP3696693A1 (en) | Method and apparatus for monitoring state of device in process industry and medium | |
| Lee et al. | Assessing the lifetime performance index of exponential products with step-stress accelerated life-testing data | |
| CN113484813B (en) | Intelligent ammeter fault rate prediction method and system under multi-environment stress | |
| US11976984B1 (en) | Method and system for predicting junction temperature of power semiconductor module in full life cycle, and terminal | |
| TW202246793A (en) | Battery state determination method, and battery state determination apparatus | |
| CN113447828A (en) | Lithium battery temperature estimation method and system based on Bayesian neural network | |
| Bai et al. | Prognostics using an adaptive self-cognizant dynamic system approach | |
| CN113125928A (en) | IGBT module aging characterization method and system based on Miner theory | |
| CN111260113A (en) | SiC MOSFET module full life cycle junction temperature online prediction method | |
| CN109643484A (en) | External action is considered in early warning system to learn the method and system for the method for latest data | |
| CN112560328B (en) | IGBT bonding lead fault diagnosis method based on surface micro-strain signal | |
| CN111767673A (en) | IGBT junction temperature measuring method and device, computer equipment and storage medium | |
| CN110489790B (en) | IGBT junction temperature prediction method based on improved ABC-SVR | |
| Zhang et al. | Thermal network parameters identifying during the cooling procedure of IGBT module | |
| CN113627621B (en) | Active learning method for optical network transmission quality regression estimation | |
| CN113191075B (en) | Photovoltaic array fault diagnosis method based on improved goblet sea squirt group algorithm | |
| CN111831862B (en) | High-quality insulation evaluation system | |
| CN110197012B (en) | Support capacitor life evaluation method considering fault influence of traction transmission system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |