CN107622167B - Collector current soft measurement method for grid control device - Google Patents
Collector current soft measurement method for grid control device Download PDFInfo
- Publication number
- CN107622167B CN107622167B CN201710879802.8A CN201710879802A CN107622167B CN 107622167 B CN107622167 B CN 107622167B CN 201710879802 A CN201710879802 A CN 201710879802A CN 107622167 B CN107622167 B CN 107622167B
- Authority
- CN
- China
- Prior art keywords
- control device
- grid control
- current
- bipolar transistor
- collector current
- 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.)
- Active
Links
Images
Landscapes
- Testing Of Individual Semiconductor Devices (AREA)
Abstract
A collector current soft measurement method for a grid control device belongs to the technical field of power electronics. Firstly, establishing a steady-state current analytic model and a transient current analytic model of a grid control device, establishing a global current analytic model of the grid control device according to the steady-state current analytic model and the transient current analytic model, and screening key parameters of the grid control device; then collecting a plurality of groups of test data in an application circuit of the grid control device, wherein each group of test data comprises key parameters of the grid control device screened out by analyzing soft measurement modeling and collector current of the corresponding grid control device; establishing a statistical data model by using the collected multiple groups of test data, namely establishing a relation between the key parameter and the collector current of the corresponding grid control device by using a machine learning algorithm; and acquiring key parameters in real time during actual measurement, and substituting the acquired key parameters into a statistical data model established by empirical data soft measurement modeling for calculation to obtain indirectly measured collector current of the gate control device. The invention has the advantages of high speed and low cost.
Description
Technical Field
The invention belongs to the technical field of power electronics, and particularly relates to a collector current soft measurement method for a grid control device.
Background
In the field of power electronics, a gate control device is a voltage-controlled power gate control device and comprises a VDMOS (vertical double-diffused metal oxide semiconductor), an LDMOS (laterally diffused metal oxide semiconductor), an IGBT (insulated gate bipolar transistor), an MCT (multi-channel transistor) and the like, wherein the IGBT is widely applied as a novel power device, and the figure of the IGBT can be seen from power generation, power transmission to power transformation. With the development of power electronic technology, in a power electronic system using an IGBT as a core power device, measurement of a collector current of the IGBT is indispensable, regardless of power control, IGBT drive control, and the like.
The existing grid control device collector current measuring technology comprises a current measuring transformer, an LEM sensor, a Hall sensor, a resistor, a Rogowski coil and the like, but with the continuous improvement of the performance of a grid control device, the switching speed of the grid control device is faster and faster, and the situation that a plurality of grid control devices are applied in parallel is more and more.
Disclosure of Invention
The invention aims to provide a collector current soft measuring method for a grid control device, aiming at the application and the requirement, which can meet the requirements on speed and cost.
The technical scheme of the invention is as follows:
a collector current soft measurement method for a grid control device comprises the stages of analytic soft measurement modeling, empirical data soft measurement modeling and measurement,
the analytical soft measurement modeling comprises the following steps:
1.1 establishing a steady-state current analytic model of the grid control device;
1.2 establishing a transient current analytic model of the grid control device;
1.3, establishing a full local current analytic model of the grid control device according to a steady-state current analytic model and a transient current analytic model of the grid control device, and screening key parameters of the grid control device;
the empirical data soft measurement modeling comprises the following steps:
2.1, establishing an application circuit of the grid control device, and setting a measurement interface of the key parameter screened in the step 1.3 on the application circuit;
2.2 collecting a plurality of groups of test data in the circuit established in the step 2.1, wherein the test data comprise the key parameters of the grid control device screened out by the analytic soft measurement modeling and the collector current of the corresponding grid control device;
2.3, establishing a statistical data model according to the plurality of groups of test data acquired in the step 2.2, namely establishing a relation between the key parameters and the collector current of the corresponding grid control device by using a machine learning algorithm;
the measuring stage is as follows: and acquiring the key parameters in real time in an application circuit of the grid control device, and bringing the acquired key parameters into a statistical data model established by empirical data soft measurement modeling for calculation to obtain the collector current of the grid control device.
In particular, the key parameters include process parameters that can be directly measured and on-line application parameters derived from the process parameters.
Specifically, when the gate control device is an insulated gate bipolar transistor, the process parameter includes a gate current I of the insulated gate bipolar transistorGVoltage V between collector and emitter of insulated gate bipolar transistorCEVoltage V between gate and emitter of insulated gate bipolar transistorGEAnd the temperature T of the insulated gate bipolar transistor, wherein the online application parameter comprises the gate charge Q of the insulated gate bipolar transistorGAnd threshold voltage V of the insulated gate bipolar transistorTHGate charge Q of said insulated gate bipolar transistorGBy means of a gate current I to said insulated gate bipolar transistorGObtained by integration, the threshold voltage V of the insulated gate bipolar transistorTHAccording to a data manual of the gate control device.
In particular, the machine learning algorithm comprises a clustering algorithm and a neural network algorithm,
the clustering algorithm is used for classifying the key parameters to obtain boundaries among working states in the switching process of the insulated gate bipolar transistor and dividing three working states of switching on, stable and switching off of the insulated gate bipolar transistor;
the neural network algorithm is used for learning and obtaining the relation between the key parameters of the insulated gate bipolar transistor and the collector current, and in different working states divided by the clustering algorithm, the neural network algorithm respectively fits the key parameters to the collector current curve to obtain the relation between the key parameters meeting the error requirement and the collector current.
The invention has the beneficial effects that: the soft measurement method capable of indirectly measuring the collector current of the grid control device is high in speed, low in cost and wide in application range; the influence of an application system on the characteristics of the grid control devices is considered, and the method is particularly suitable for measuring the collector current of each grid control device in the parallel current sharing control of the grid control devices
Drawings
Fig. 1 is a schematic diagram of a collector current soft measurement method for a gate control device according to the present invention.
Fig. 2 is a schematic diagram of boundary conditions in parallel application simulation of an insulated gate bipolar transistor IGBT in an embodiment, in which a collector current of one IGBT is divided by a clustering method, where (a) is a schematic diagram of boundary conditions of on and steady states, and (b) is a schematic diagram of boundary conditions of steady states and off states.
Fig. 3 is a diagram illustrating the prediction of collector current of one IGBT in an insulated gate bipolar transistor IGBT under a load condition of 18A by applying a simulated soft measurement method in parallel in the embodiment, where (a) is a diagram illustrating the prediction of on and steady state, and (b) is a diagram illustrating the prediction of steady state and off.
FIG. 4 is a diagram illustrating the prediction of collector current of one IGBT by using simulation soft measurement method in parallel connection of the insulated gate bipolar transistor IGBT under the load condition of 13.5A in the embodiment; the left graph is a prediction graph of switching-on and steady state, and the right graph is a prediction graph of steady state and switching-off.
FIG. 5 is a schematic diagram of prediction of IGBT collector current by soft measurement method in practical circuit of ZCS BUCK based on IGBT in the embodiment, wherein peak current load in graph (a) is 15A; the peak current load in graph (b) is 24A;
fig. 6 is a schematic diagram of IGBT parameter waveforms and boundary division of an IGBT-based ZCS BUCK practical application circuit in an embodiment, where (a) is an experimental waveform and (b) is a schematic diagram of boundary division.
Detailed Description
The invention is described in detail below with reference to the figures and with specific embodiments as collector current measurements for insulated gate bipolar transistors IGBT.
The collector current of the grid control device is measured by adopting a hybrid soft measurement method, as shown in figure 1, the method comprises analytic soft measurement modeling and empirical data soft measurement modeling, wherein the empirical data soft measurement modeling is modeling according to measured data under the guidance of a model established by the analytic soft measurement, key parameters are screened for an obtained IGBT collector global current model in the analytic soft measurement modeling, the screened parameters are variables of the empirical data soft measurement modeling, and the parameters can be selected to participate in modeling according to actual conditions in the empirical data soft measurement modeling.
The analytical soft measurement modeling comprises the establishment of an IGBT collector steady-state current analytical model, the establishment of an IGBT collector current transient current analytical model and an IGBT collector global current analytical model established according to the two analytical models, and key parameters which are convenient to measure in practical application are screened out from the IGBT collector global current analytical model. In this embodiment, the analytic soft measurement modeling is to derive a steady-state analytic model and a transient-state analytic model of the IGBT collector current according to the physical characteristics of the IGBT device, so as to obtain a global analytic model of the IGBT collector current, that is, including a steady state and a transient state, where the global analytic model is expressed as:
ICEglobal＝Fglobal(IG,QG,VCE,VGE,VTH,T) (1)
IGfor IGBT gate current, QGIs IGBT gate charge, VCEIs the voltage between the collector and emitter of the IGBT, VGEIs the voltage between the IGBT gate and emitter, VTHIs the IGBT threshold voltage, and T is the IGBT temperature. The above formula is a key parameter selected in parentheses, and these parameters are very easy to obtain in practical application based on the IGBT, so as to avoid using physical parameters such as electron hole mobility, doping concentration, and geometric dimension parameters related to device structure, process parameters, etc. to determine the IGBT collector current and difficult to obtain for IGBT users, but for the influence and determination effect of these parameters on the current, the equivalent is achieved by processing a large amount of measured data of the parameters in the above formula brackets.
The key parameters of the IGBT in the embodiment comprise IGBT gate current IGVoltage V between collector and emitter of IGBTCEVoltage V between IGBT gate and emitterGERelated sampling circuits can be configured to directly measure, and the IGBT temperature T can be configured with a temperature sensor and a sampling circuit to measure, which are process parameters capable of being directly measured; the on-line application parameters need to be derived through process parameters, and IGBT gate charge QGThen the parameters can be extracted on-line by applying the parameter extraction method to the IGBT gate current IGObtained after integration, and the IGBT threshold voltage VTHCan directly utilize the IGBValues given in the T data manual or measured by the method of online application parameter extraction, as obtained from the data manual of IGBT, for example, V when the collector current reaches 15mAGEValue of as VTH。
The empirical data soft measurement modeling comprises 4 key components aiming at the process parameter measurement, online application parameter extraction, machine learning algorithm and statistical data modeling of the IGBT application circuit, so that the final IGBT collector current soft measurement model is obtained and used for measuring the IGBT collector current.
Firstly, an IGBT-based application circuit or system is established, and IGBT collector current is collected by using related instruments or current sensors. According to the working index of a specific IGBT application circuit or system, a large amount of test data can be collected in a fixed load or a certain load range, and each set of test data comprises key parameters in brackets in the formula (1) and an IGBT collector current value collected by an instrument or a current sensor.
And then establishing a relation between the parameters in brackets in the vertical type (1) and the IGBT collector current collected by an instrument or a current sensor by utilizing a machine learning algorithm. In this embodiment, the collected or extracted key parameters are classified by using a clustering algorithm, the operating state of the IGBT is divided into 3 or more processes of an on process, a steady state (on state), and an off state according to actual conditions, so as to obtain boundaries between the operating states of the IGBT switching process, and the number of the divided boundaries is determined according to experiments and experiences. After different working states of the IGBT are divided by the clustering algorithm, the key parameters are fitted to an IGBT collector current curve measured by an instrument or a current sensor by a BP neural network algorithm (or other learning algorithms) in each process divided by the clustering algorithm respectively, so that the defined error requirement is met, the relation between the key parameters and the IGBT collector current measured by the instrument or the current sensor is found out, and a final IGBT collector current soft measurement model is obtained. As shown in fig. 2, the boundary of the test data of one IGBT tube is divided by the clustering method when the IGBT parallel application simulation is performed, fig. 2(a) shows the on and steady states of the IGBT tube, and fig. 2(b) shows the steady state and the off states of the IGBT tube, it can be seen that the on and off states are divided into 4 stages, which is consistent with the process of physical division by the IGBT device, and this verifies the effectiveness of the clustering method. In this embodiment, in 4 processes divided by turning on and off, respectively, a verified soft measurement model is obtained by using a BP neural network algorithm or other machine learning algorithms, and it is found through verification that a predicted value and an actual measurement value of the model are consistent, as shown in fig. 3 and 4.
And entering a measurement stage after modeling, acquiring key parameters in real time in normal work based on the IGBT application system, substituting the key parameters into the finally obtained statistical data model, and calculating to obtain the indirectly measured IGBT collector current value.
In order to verify the effectiveness of the method in practical application, a zero current turn-off (ZCS) voltage reduction circuit (BUCK) based on an IGBT is taken as an experimental object, the collector current of the IGBT is subjected to soft measurement by the method, the result is shown in figure 5, the predicted value of the soft measurement is consistent with the measured value of an instrument or a current sensor through figure 5, in order to reduce the operation amount in the example, the operation of the IGBT is divided into 3 processes, namely turn-on, steady-state and turn-off through a clustering algorithm, as shown in figure 6.
The method provided by the invention can be realized by one or more of embedded system software, desktop software, board-level circuit system, FPGA soft core and application-specific integrated circuit hard core modes. Besides the IGBT, the IGBT can also be applied to power devices such as MOSFET, MCT, GTO and the like.
Compared with the traditional method, the soft measurement method for the collector current of the grid control device can indirectly measure the collector current of the grid control device through other parameters of the grid control device on the premise of not needing a current sensor; the method is suitable for the grid control devices of any manufacturer and any model, does not need to obtain the structural parameters and the process parameters of the grid control devices, is suitable for measuring steady-state and transient currents, and supports global or local current measurement.
Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.
Claims (4)
1. A collector current soft measurement method for a grid control device is characterized by comprising the stages of analytic soft measurement modeling, empirical data soft measurement modeling and measurement,
the analytical soft measurement modeling comprises the following steps:
1.1 establishing a steady-state current analytic model of the grid control device;
1.2 establishing a transient current analytic model of the grid control device;
1.3, establishing a global current analytic model of the grid control device according to a steady-state current analytic model and a transient current analytic model of the grid control device, and screening key parameters of the grid control device;
the empirical data soft measurement modeling comprises the following steps:
2.1, establishing an application circuit of the grid control device, and setting a measurement interface of the key parameter screened in the step 1.3 on the application circuit;
2.2 collecting multiple groups of test data in the application circuit established in the step 2.1, wherein the test data comprises the key parameters of the grid control device screened out by the analytic soft measurement modeling and the collector current of the corresponding grid control device;
2.3, establishing a statistical data model according to the plurality of groups of test data acquired in the step 2.2, namely establishing a relation between the key parameters and the collector current of the corresponding grid control device by using a machine learning algorithm;
the measuring stage is as follows: and in an application circuit of the grid control device, acquiring the key parameters in real time, and substituting the acquired key parameters into a statistical data model established by empirical data soft measurement modeling for calculation to obtain the collector current of the grid control device.
2. The method of claim 1, wherein the key parameters comprise directly measurable process parameters and on-line application parameters derived from the process parameters.
3. The method of claim 2, wherein when the gated device is an insulated gate bipolar transistor, the process parameter comprises a gate current I of the insulated gate bipolar transistorGVoltage V between collector and emitter of insulated gate bipolar transistorCEVoltage V between gate and emitter of insulated gate bipolar transistorGEAnd the temperature T of the insulated gate bipolar transistor, the on-line application parameter comprising the gate charge Q of the insulated gate bipolar transistorGAnd threshold voltage V of the insulated gate bipolar transistorTHGate charge Q of said insulated gate bipolar transistorGBy means of a gate current I to said insulated gate bipolar transistorGObtained by integration, the threshold voltage V of the insulated gate bipolar transistorTHAnd determining according to a data manual of the grid control device.
4. The collector current soft measurement method for a gated device according to claim 3, wherein the machine learning algorithm comprises a clustering algorithm and a neural network algorithm,
the clustering algorithm is used for classifying the key parameters to obtain boundaries among working states in the switching process of the insulated gate bipolar transistor and dividing three working states of switching on, stable and switching off of the insulated gate bipolar transistor;
the neural network algorithm is used for learning and obtaining the relation between the key parameters of the insulated gate bipolar transistor and the collector current, and in different working states divided by the clustering algorithm, the neural network algorithm respectively fits the key parameters to the collector current curve to obtain the relation between the key parameters meeting the error requirement and the collector current.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710879802.8A CN107622167B (en) | 2017-09-26 | 2017-09-26 | Collector current soft measurement method for grid control device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710879802.8A CN107622167B (en) | 2017-09-26 | 2017-09-26 | Collector current soft measurement method for grid control device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107622167A CN107622167A (en) | 2018-01-23 |
| CN107622167B true CN107622167B (en) | 2020-04-17 |
Family
ID=61090667
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710879802.8A Active CN107622167B (en) | 2017-09-26 | 2017-09-26 | Collector current soft measurement method for grid control device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107622167B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110502777A (en) * | 2019-07-03 | 2019-11-26 | 国网江苏省电力有限公司南京供电分公司 | IGBT module condition detecting system and method based on neural network prediction |
| CN111487520B (en) * | 2020-04-27 | 2022-04-01 | 全球能源互联网研究院有限公司 | IGBT module testing method and device and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103208984A (en) * | 2012-01-13 | 2013-07-17 | Abb研究有限公司 | Active gate drive circuit |
| CN104242890A (en) * | 2013-06-20 | 2014-12-24 | Abb研究有限公司 | Active gate drive circuit |
| CN106991221A (en) * | 2017-03-24 | 2017-07-28 | 清华大学 | A kind of sectional broken line model based on IGBT device transient physical process |
-
2017
- 2017-09-26 CN CN201710879802.8A patent/CN107622167B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103208984A (en) * | 2012-01-13 | 2013-07-17 | Abb研究有限公司 | Active gate drive circuit |
| CN104242890A (en) * | 2013-06-20 | 2014-12-24 | Abb研究有限公司 | Active gate drive circuit |
| CN106991221A (en) * | 2017-03-24 | 2017-07-28 | 清华大学 | A kind of sectional broken line model based on IGBT device transient physical process |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107622167A (en) | 2018-01-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111368454B (en) | SiC MOSFET SPICE model establishment method based on bare chip packaging structure | |
| JP6251310B2 (en) | Measuring method of current-voltage characteristics | |
| CN103105571B (en) | Simulated measurement method of current characteristics of insulated gate bipolar transistor | |
| CN110502805A (en) | IGBT physical model statistic property extracting method | |
| CN107622167B (en) | Collector current soft measurement method for grid control device | |
| Jing et al. | An improved behavior model for IGBT modules driven by datasheet and measurement | |
| Yang et al. | Hybrid data-driven modeling methodology for fast and accurate transient simulation of SiC MOSFETs | |
| CN112098798A (en) | Neural network-based junction temperature on-line measurement method for silicon carbide MOS (Metal oxide semiconductor) device | |
| CN110807516A (en) | Junction temperature prediction method of IGBT module for driver | |
| CN114004067A (en) | MOSFET switching tube characteristic model establishing method, device, equipment and medium | |
| CN112630544B (en) | High-voltage SiC MOSFET drain-source interelectrode nonlinear capacitance measurement and modeling method | |
| Zhang et al. | Characterization of SiC MOSFET switching performance | |
| CN106202590B (en) | IGBT module switching transients model parameter acquisition methods and method for establishing model | |
| Xu et al. | A datasheet-based behavioral model of SiC MOSFET for power loss prediction in electromagnetic transient simulation | |
| WO2019146460A1 (en) | Method for measuring current-voltage characteristic | |
| Navarro et al. | A sequential model parameter extraction technique for physics-based IGBT compact models | |
| CN103792478B (en) | A kind of emulation mode of insulated gate bipolar transistor DC characteristic | |
| CN111898281B (en) | SiC IGBT device behavior model building method | |
| Zhuolin et al. | A simple SiC power MOSFET model | |
| CN112213562A (en) | Method for measuring and calculating internal resistance of grid electrode of power semiconductor device | |
| CN113092863B (en) | Method and device for measuring and calculating internal resistance of grid electrode of power device and storage medium | |
| Song et al. | An improved transient model of high voltage IGBT based on vector fitting | |
| Hirose et al. | Development of the HiSIM-IGBT model for EV/HV electric circuit simulation | |
| Wu et al. | A Novel Compact Model of Gate Capacitance for SiC MOSFET with Easy Parameter Extraction Method | |
| Zhang et al. | Characteristics and Drive Design Analysis of SiC |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |