Inverter IGBT state monitoring method and system based on quintuple harmonic content characteristics

Abstract

The invention provides an inverter IGBT (Insulated Gate Bipolar Translator) state monitoring method and system based on quintuple harmonic content characteristics. The method comprises the following steps of 1, collecting output current data of a three-phase inverter; 2, Fourier transform is carried out on the output current data, and the fifth harmonic content is extracted to serve as a characteristic quantity; step 3, establishing an artificial intelligence data model, and performing association training on the fifth harmonic content and the IGBT on-resistance to obtain an IGBT on-resistance prediction model; 4, collecting output current data of the three-phase inverter in real time, repeating the step 2 to extract real-time fifth harmonic content, and inputting the real-time fifth harmonic content into the IGBT on-resistance prediction model to obtain real-time IGBT on-resistance; and step 5, judging the health state of the IGBT according to the real-time IGBT on-resistance, and realizing on-line state monitoring of the IGBT. According to the technical scheme, a sensor does not need to be additionally arranged in the converter, IGBT state monitoring can be achieved only by analyzing the output current, and the application range is wide.

Description

Technical Field

[0001] This invention relates to the fields of inverter reliability analysis and power device condition monitoring technology, and in particular to an inverter IGBT condition monitoring method and system based on the fifth harmonic content characteristics. Background Technology

[0002] Extensive research has been conducted by scholars both domestically and internationally on IGBT module condition monitoring solutions. These can be broadly categorized into three methods: traditional thermoelectric model coupling, data-driven methods, and intelligent sensor monitoring. The traditional thermoelectric model coupling method primarily establishes an electrothermal model of the IGBT module and monitors temperature-sensitive parameters to analyze junction temperature changes, thereby monitoring the IGBT module's condition. While this method is low-cost and highly reliable, its applicability is limited, and its effectiveness is lower under complex operating conditions. The data-driven method relies heavily on artificial intelligence, introducing human intelligence to analyze various parameters of the IGBT module and extract patterns from the data to accurately predict the IGBT's operating status. This method boasts powerful processing capabilities, can solve nonlinear fault problems, and is applicable to many complex operating conditions; however, it is highly dependent on the quantity and quality of data, requiring extensive and comprehensive historical data for training, resulting in high costs. The intelligent sensor monitoring method relies on current sensing technology to directly acquire key state parameters for IGBT condition monitoring. This method directly acquires key parameters, has strong anti-interference capabilities, and extremely high accuracy; however, its high cost hinders widespread adoption. Summary of the Invention

[0003] In view of this, the purpose of this invention is to provide an inverter IGBT state monitoring method and system based on the fifth harmonic content characteristics. This method eliminates the need to install sensors inside the inverter and only requires analysis of the output current to monitor the IGBT state. It has a wide range of applications. At the same time, based on the representativeness of the selected characteristics, it greatly reduces the computational load of the data model and improves the calculation speed.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a method for monitoring the state of an inverter IGBT based on the characteristics of the fifth harmonic content, comprising the following steps:

[0005] Step 1: Collect the output current data of the three-phase inverter;

[0006] Step 2: Perform Fourier transform on the output current data and extract the fifth harmonic content as a feature quantity;

[0007] Step 3: Establish an artificial intelligence data model, and perform correlation training between the fifth harmonic content and the IGBT on-resistance to obtain an IGBT on-resistance prediction model;

[0008] Step 4: Collect the output current data of the three-phase inverter in real time, repeat step 2 to extract the real-time fifth harmonic content, input it into the IGBT on-resistance prediction model, and obtain the real-time IGBT on-resistance.

[0009] Step 5: Determine the health status of the IGBT based on the real-time IGBT on-resistance to achieve online IGBT status monitoring.

[0010] In a preferred embodiment, in step 2, the Fourier transform is used to analyze the distortion of the output current waveform and extract the fifth harmonic content from the low-order harmonics. The fifth harmonic content is monotonically correlated with the IGBT on-resistance.

[0011] In a preferred embodiment, the Fourier expansion of the output current data is based on the three-phase load symmetry condition, where the sum of the three-phase voltages is 0. By analyzing the voltage relationship between the DC side midpoint and the output side midpoint, the expression for the inverter output side phase voltage is obtained, and the harmonic components of the output current are derived from this expression.

[0012] In a preferred embodiment, the voltage between the DC side midpoint N1 and the output side midpoint N2 of the three-phase inverter system is u. 12 ,

[0013] Among them, u x1 U represents the midpoint voltage of each phase upper arm of the bridge. 12 This represents the voltage difference between the midpoint of the DC side and the output side.

[0014] For the output phase voltage u of the inverter system al Expanding this into a Fourier series yields:

[0015]

[0016] Where u dc The DC side voltage is given by ω, where w is the fundamental angular frequency, and u is the DC side voltage. ao This is the Fourier transform expression for the phase voltage of phase a.

[0017] Based on this, the Fourier expansion of the output current is:

[0018] .

[0019] Among them, u dc Where is the DC side voltage, r is the equivalent resistance of the load side, w is the fundamental angular frequency, and I... ao Let be the Fourier transform expression of the output current of phase a.

[0020] In a preferred embodiment, in step 3, the artificial intelligence data model is a convolutional neural network model. During the training process, the IGBT on-resistance is changed, the corresponding output current data is collected, and the fifth harmonic content is extracted to form a training dataset. The training dataset contains a label group of the fifth harmonic content and the corresponding IGBT on-resistance.

[0021] In a preferred embodiment, in step 3, the change in the IGBT on-resistance is caused by the failure mechanism of the IGBT, which includes bond wire failure, solder layer cracking, etc. By analyzing the change in on-resistance, the health status of the IGBT can be determined.

[0022] In a preferred embodiment, in step 5, the health status judgment is based on the following: when the real-time IGBT on-resistance exceeds a preset threshold, the IGBT is determined to be in a degraded or faulty state, and a warning signal is issued.

[0023] In a preferred embodiment, the three-phase inverter is a 6-pulse photovoltaic inverter used in photovoltaic systems and wind power systems.

[0024] The present invention also provides an inverter IGBT state monitoring system based on the fifth harmonic content characteristics, including a processor, a memory and a bus, wherein the memory stores machine-readable instructions executed by the processor;

[0025] When the system is running, the processor and the memory communicate via a bus, and the machine-readable instructions are executed by the processor as described in the inverter IGBT state monitoring method based on the fifth harmonic content characteristics.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] 1. This application treats the converter as a black box and judges the working status of the IGBT by analyzing the output.

[0028] 2. This application proposes a condition monitoring method based on harmonic content analysis. The innovation lies in selecting harmonic content as a characteristic quantity to form corresponding tags with key IGBT health parameters (on-resistance), and using an artificial intelligence data model for analysis to achieve online monitoring of the IGBT's operating status.

[0029] 3. This application differs from existing data-driven methods. This scheme requires less computation, only harmonic content data needs to be analyzed, and the calculation is simple.

[0030] 4. This application differs from existing thermoelectric models or intelligent sensor monitoring methods. This solution does not require additional sensors to be installed in the converter. It only analyzes the output and the data acquisition is simple. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of a common three-phase inverter topology; Figure 2 The diagram shows the waveforms of the three-phase output circuit, where (a) is the three-phase output current when R=0.0001Ω and (b) is the three-phase output current when R=0.01Ω.

[0032] Figure 3 This is a histogram of harmonic content according to a preferred embodiment of the present invention;

[0033] Figure 4 This is a graph showing the variation of the fifth harmonic content in a preferred embodiment of the present invention;

[0034] Figure 5 This is a schematic diagram of the implementation process of a preferred embodiment of the present invention;

[0035] Figure 6 This is a schematic diagram of a detailed simulation verification scheme for a preferred embodiment of the present invention;

[0036] Figure 7 This is a simulation circuit diagram of a three-phase inverter topology according to a preferred embodiment of the present invention.

[0037] Figure 8 This is a schematic diagram of the simulation waveform of a preferred embodiment of the present invention. Detailed Implementation

[0038] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0039] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0040] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations according to this application; as used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise; furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0041] refer to Figure 1-8 This application provides an inverter IGBT condition monitoring method based on the fifth harmonic content characteristics, and an online condition monitoring method based on the relationship between harmonic content and IGBT on-resistance.

[0042] Three-phase inverters are widely used in many important fields such as photovoltaic systems and wind power systems due to the expansion of new energy power generation. Common topologies for them include: Figure 1 As shown in the diagram, the IGBT (Inverter Box) power semiconductor device plays a crucial role, but it is also one of the weakest links in the entire inverter system. Statistics show that the failure rate of IGBTs during system operation can be as high as 34%. Under the influence of cyclical electrothermal and mechanical stresses caused by load fluctuations and changes in operating conditions, the performance of IGBTs gradually deteriorates. This change poses a significant potential risk to the system's operation, easily leading to system failure and substantial losses. Therefore, online status monitoring of IGBT devices, assessing their health status, and timely early warning and maintenance measures before degradation failures occur are of great significance to the healthy operation of the entire system.

[0043] Studies have shown that most IGBT failure mechanisms (such as bond wire failure and solder layer cracking) lead to changes in their on-resistance. Online monitoring of the on-resistance of IGBTs in inverters allows for real-time health assessment, thereby ensuring the safe operation of the system.

[0044] (I) Extraction of IGBT Output Feature Signals Based on Harmonic Content Analysis

[0045] When inverter operating conditions change, both its on-resistance and three-phase output current will change accordingly, and the IGBT on-resistance can serve as an important indicator of its health status. Based on this, an artificial intelligence data model can be introduced to achieve online monitoring of the IGBT status. This method does not require establishing a complex thermoelectric coupling model, nor does it require monitoring the specific electrical parameters of the IGBTs in the inverter. The inverter can be treated as a black box, and its operating status can be judged by analyzing its output, making it versatile.

[0046] Because the on-resistance is low under normal conditions and can be almost ignored, and because of the output filter circuit, its impact on the output current is negligible (e.g., Figure 2 The two are directly used as corresponding feature quantities for model training. Since the IGBT is a switching device, it performs high-frequency switching operations in the converter. The voltage difference caused by its on-resistance will cause frequency-related changes in the IGBT to a certain extent. Therefore, the output current is further analyzed in detail.

[0047] Considering that the operating conditions of IGBTs change under different operating conditions, and this change directly leads to distortion of the output current waveform, thus generating harmonics, this study analyzes the harmonic generation mechanism and, combined with actual test data, establishes an accurate model to explore the relationship between changes in IGBT on-resistance and changes in output current harmonics.

[0048] When an IGBT is turned on, a voltage drop will occur due to the influence of the on-resistance, which can be simply expressed as:

[0049] (1)

[0050] Among them, R on (T) is the IGBT on-resistance, I out V is the output current flowing through the IGBT when it is turned on. drop This is the IGBT on-state voltage drop.

[0051] The voltage drop changes nonlinearly with the load current. Especially when the load current is large or the temperature changes significantly, the voltage drop will significantly affect the output voltage, causing asymmetrical waveform distortion and thus introducing low-order harmonics.

[0052] In PWM control, the ideal duty cycle and output voltage relationship should be:

[0053] (2)

[0054] Among them, V out V is the output voltage. dc D is the DC side voltage. ideal For the ideal duty cycle.

[0055] Due to the influence of the on-state impedance, the actual duty cycle needs to compensate for the on-state impedance voltage drop, thus requiring the actual duty cycle to be corrected as follows:

[0056] (3)

[0057] Among them, D ideal For the ideal duty cycle, D act V represents the actual duty cycle. dc V is the DC side voltage. drop This is the IGBT on-state voltage drop.

[0058] This nonlinear change can also cause asymmetrical waveform distortion in the PWM waveform, introducing low-order harmonics.

[0059] Fourier transform calculations are introduced to analyze the output and extract harmonic variation parameters.

[0060] Let the voltage between the DC side midpoint N1 and the output side midpoint N2 of the three-phase inverter system be u. 12 Then the phase voltages of each phase of the load are as follows:

[0061] (4)

[0062] Where u x1 U represents the midpoint voltage of each phase upper arm of the bridge. x2u represents the phase voltage of each phase of the load. 12 This represents the voltage difference between the midpoint of the DC side and the output side.

[0063] Because the three-phase load is symmetrical, the sum of the three-phase voltages is 0.

[0064] By organizing, we can obtain:

[0065] (5)

[0066] Among them, u x1 U represents the midpoint voltage of each phase upper arm of the bridge. 12 This represents the voltage difference between the midpoint of the DC side and the output side.

[0067] The output phase voltage u of the inverter system al Expanding this into a Fourier series yields:

[0068] (6)

[0069] Where u dc The DC side voltage is given by ω, where w is the fundamental angular frequency, and u is the DC side voltage. ao This is the Fourier transform expression for the phase voltage of phase a.

[0070] Based on this, the Fourier expansion of the output current can be obtained as follows:

[0071] (7)

[0072] Among them, u dc Where is the DC side voltage, r is the equivalent resistance of the load side, w is the fundamental angular frequency, and I... ao Let be the Fourier transform expression of the output current of phase a.

[0073] The above analysis shows that the low-order harmonics output by a 6-pulse photovoltaic inverter are mainly the 5th and 7th harmonics.

[0074] Harmonic content analysis of the output was performed using Fourier analysis to obtain the following results: Figure 3 As a result, the fifth harmonic content was relatively prominent and could be used as a characteristic quantity for analysis.

[0075] Through simulation experiments, the on-resistance was varied, and output current data was collected under the corresponding resistance. The fifth harmonic content was extracted using Fourier transform and correlation analysis was performed. Taking phase A as an example, it was found that the fifth harmonic content of phase A gradually decreases as the IGBT on-resistance increases. Figure 4 As shown, it exhibits a certain degree of monotonicity.

[0076] (II) IGBT Condition Monitoring Scheme Based on Harmonic Content Analysis

[0077] Implementation process: Based on the proposed harmonic content correspondence approach, the implementation scheme for IGBT state monitoring is as follows: Figure 4 As shown, by introducing an artificial intelligence data model for training and prediction of IGBT on-resistance, and by collecting inverter output current data, the on-resistance of the IGBT can be predicted, thereby enabling real-time monitoring of the IGBT's operating status.

[0078] Detailed simulation verification scheme as follows Figure 6 As shown, the simulation model uses a Matlab program to control the change in on-resistance and collect the output current data at the corresponding resistance, which is then sent to the data processing module. The data undergoes a Fourier transform to obtain the fifth harmonic content, forming a corresponding group with the on-resistance at that point. The data is then imported into the model for training.

[0079] Case Study: Taking a three-phase inverter as an example, Figure 7 The simulation circuit topology is given. Figure 8 Simulation waveforms are presented, and a comprehensive IGBT on-resistance prediction model is established using a convolutional neural network model for training. The results show that the predicted values ​​almost perfectly match the actual values, proving the feasibility of the proposed method.

Claims

1. A method for monitoring the condition of an inverter IGBT based on the characteristics of its fifth harmonic content, characterized in that, Includes the following steps: Step 1: Collect the output current data of the three-phase inverter; Step 2: Perform Fourier transform on the output current data and extract the fifth harmonic content as a feature quantity; Step 3: Establish an artificial intelligence data model, and perform correlation training between the fifth harmonic content and the IGBT on-resistance to obtain an IGBT on-resistance prediction model; Step 4: Collect the output current data of the three-phase inverter in real time, repeat step 2 to extract the real-time fifth harmonic content, input it into the IGBT on-resistance prediction model, and obtain the real-time IGBT on-resistance. Step 5: Determine the health status of the IGBT based on the real-time IGBT on-resistance to achieve online IGBT status monitoring.

2. The inverter IGBT condition monitoring method based on fifth harmonic content characteristics according to claim 1, characterized in that, In step 2, the Fourier transform is used to analyze the distortion of the output current waveform and extract the fifth harmonic content in the low-order harmonics. The fifth harmonic content is monotonically correlated with the IGBT on-resistance.

3. The inverter IGBT condition monitoring method based on fifth harmonic content characteristics according to claim 2, characterized in that, The Fourier expansion of the output current data is based on the three-phase load symmetry condition, where the sum of the three-phase voltages is 0. The harmonic components of the output current are derived by analyzing the voltage between the DC side midpoint and the output side midpoint.

4. The inverter IGBT condition monitoring method based on fifth harmonic content characteristics according to claim 3, characterized in that, The voltage between the DC side midpoint N1 and the output side midpoint N2 of the three-phase inverter system is u. 12 , Among them, u x1 U represents the midpoint voltage of each phase upper arm of the bridge. 12 This represents the voltage difference between the DC side and the output side midpoint. For the output phase voltage u of the inverter system al Expanding this into a Fourier series yields: Where u dc The DC side voltage is given by ω, where w is the fundamental angular frequency, and u is the DC side voltage. ao This is the Fourier transform expression for the phase voltage of phase a; Based on this, the Fourier expansion of the output current is: 。 Among them, u dc Where is the DC side voltage, r is the equivalent resistance of the load side, w is the fundamental angular frequency, and I... ao Let be the Fourier transform expression of the output current of phase a.

5. The inverter IGBT condition monitoring method based on fifth harmonic content characteristics according to claim 1, characterized in that, In step 3, the artificial intelligence data model is a convolutional neural network model. During the training process, the IGBT on-resistance is changed, the corresponding output current data is collected, and the fifth harmonic content is extracted to form a training dataset. The training dataset contains a label group of the fifth harmonic content and the corresponding IGBT on-resistance.

6. The inverter IGBT condition monitoring method based on fifth harmonic content characteristics according to claim 1, characterized in that, In step 3, the change in the IGBT on-resistance is caused by the failure mechanism of the IGBT, which includes bond wire failure and solder layer cracking.

7. The inverter IGBT condition monitoring method based on fifth harmonic content characteristics according to claim 1, characterized in that, In step 5, the health status is determined as follows: when the real-time IGBT on-resistance exceeds a preset threshold, the IGBT is determined to be in a degraded or faulty state, and a warning signal is issued.

8. The inverter IGBT condition monitoring method based on the fifth harmonic content characteristics according to claim 1, characterized in that, The three-phase inverter is a 6-pulse photovoltaic inverter used in photovoltaic systems and wind power systems.

9. An inverter IGBT state monitoring system based on fifth harmonic content characteristics, comprising a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executed by the processor; characterized in that, When the system is running, the processor and the memory communicate via a bus, and the machine-readable instructions are executed by the processor as described in any one of claims 1 to 8. This is an inverter IGBT state monitoring method based on the fifth harmonic content characteristics.

Images