Intelligent power module based on SiC semiconductor and fault diagnosis system thereof

Abstract

The invention belongs to the field of information management systems, particularly relates to an intelligent power module based on a SiC semiconductor and a fault diagnosis system thereof, and aims to solve the problem of power module fault diagnosis in the intelligent development process of a power system. A SiC semiconductor device, a driving circuit, a protection circuit and a fault diagnosis circuit are integrated, and rapid and accurate fault detection and positioning are realized by using an advanced algorithm. The system improves the reliability and safety of a power system, reduces the maintenance cost, and has an important application value.

Description

Technical Field

[0001] The present invention relates to the field of information management systems, and in particular to an intelligent power module based on SiC semiconductors and a fault diagnosis system thereof. Background Art

[0002] In the context of the continuous development of power electronic systems, the performance requirements for power modules are increasing. With the intelligentization of power systems, traditional power modules have exposed many deficiencies in fault diagnosis and are unable to meet the needs of complex and changing operating environments.

[0003] SiC semiconductors have brought new opportunities for the development of power modules due to their excellent physical properties, such as high electron mobility, high breakdown electric field strength and good thermal conductivity. However, how to give full play to the advantages of SiC semiconductors and achieve efficient and accurate fault diagnosis and protection functions has become a key research issue.

[0004] At the same time, as the scale of power systems continues to expand, higher requirements are placed on the reliability, stability and safety of power modules. Traditional fault diagnosis technologies and methods can no longer meet these requirements, so it is necessary to develop more advanced and intelligent fault diagnosis systems. Summary of the invention

[0005] The present invention provides an intelligent power module integrating a SiC semiconductor device, a driving circuit, a protection circuit and a fault diagnosis circuit. The intelligent power module can monitor the electrical parameters and temperature information of the module in real time, and realize rapid and accurate detection and positioning of faults through a fault diagnosis system based on a multi-scale hyperbolic partial differential algorithm, a fuzzy random geometry algorithm, a quantum information processing algorithm, a multi-scale geometry fusion algorithm and an adaptive fault diagnosis algorithm.

[0006] The fault diagnosis system of the intelligent power module includes sensors for real-time monitoring of electrical parameters and temperature information, and fault diagnosis software based on the five algorithms, wherein the five algorithms are multi-scale hyperbolic partial differential algorithm, fuzzy random geometry algorithm, quantum information processing algorithm, multi-scale geometry fusion algorithm and adaptive fault diagnosis algorithm.

[0007] On the other hand, the present invention also provides a fault diagnosis method based on a multi-scale hyperbolic partial differential algorithm: it is used to perform multi-scale analysis on the electrical parameters of the power module according to the hyperbolic partial differential equation, solve the hyperbolic partial differential equation by the finite difference method, obtain the change law of the electrical parameters at different scales, and thus determine the fault type and location.

[0008] On the other hand, the present invention also provides a fault diagnosis method based on fuzzy random geometry algorithm: used to analyze the probability and location of fault occurrence by using fuzzy random process and non-Euclidean geometry, calculate the probability of fault occurrence by fuzzy membership function and probability distribution of random variables, and determine the fault location by using the distance and angle characteristics of non-Euclidean geometry.

[0009] On the other hand, the present invention also provides a fault diagnosis method based on quantum information processing algorithm: it is used to process the electrical information of the power module using quantum information theory and quantum computing technology, process the electrical information through the characteristics of the quantum state and quantum algorithm, and diagnose the fault type and location according to the changes in quantum information.

[0010] On the other hand, the present invention also provides a fault diagnosis method based on a multi-scale geometric fusion algorithm: it is used to combine multi-scale geometric analysis and non-Euclidean geometry, perform multi-scale fusion analysis on the electrical parameters of the power module, extract the characteristics of the electrical parameters at different scales through multi-scale geometric analysis, and use non-Euclidean geometry to analyze the relationship between these characteristics, so as to determine the fault type and location.

[0011] On the other hand, the present invention also provides a fault diagnosis method based on an adaptive fault diagnosis algorithm: used to automatically adjust the fault diagnosis algorithm and system parameters according to the fault diagnosis results, and automatically adjust the parameters of the diagnosis algorithm by real-time monitoring of the fault type and degree to improve the accuracy and efficiency of fault diagnosis.

[0012] Beneficial effects:

[0013] (I) Multi-scale hyperbolic partial differentiation algorithm

[0014] 1 Algorithm principle: Combining multi-scale geometric analysis and hyperbolic partial differential equations, multi-scale analysis of power module electrical parameters is achieved. By solving the hyperbolic partial differential equations, the variation rules of electrical parameters at different scales are obtained.

[0015] 2 Advantages: It can more accurately capture the characteristics of electrical parameters at different scales and improve the accuracy and efficiency of fault diagnosis. For example, when processing complex signals, the multi-scale hyperbolic partial differential algorithm can better analyze the local and global characteristics of the signal.

[0016] 3 Application: Applied to fault diagnosis and protection of power systems. For example, during the operation of the power system, the multi-scale hyperbolic partial differential algorithm is used to monitor and analyze electrical parameters in real time, detect faults in time and take corresponding protection measures.

[0017] (II) Fuzzy random geometric algorithm

[0018] 1. Algorithm core: Optimize fault diagnosis by combining fuzzy random process and non-Euclidean geometry. Use fuzzy random process to describe the probability and uncertainty of fault occurrence, and use non-Euclidean geometry to analyze the location and cause of fault occurrence.

[0019] 2 Innovation: In the process of fault diagnosis, the fuzziness and randomness of electrical parameters are fully considered to improve the reliability and accuracy of fault diagnosis. For example, the location and probability of fault occurrence can be more accurately determined through fuzzy random geometric algorithms.

[0020] 3 Application: Applied to fault diagnosis and protection of power systems. For example, during the operation of the power system, fuzzy random geometric algorithms are used to analyze and predict faults, providing guarantees for the safe operation of the system.

[0021] 3. Quantum Information Processing Algorithms

[0022] 1 Algorithm features: Use quantum information theory and quantum computing technology to process the electrical information of the power module. Improve the accuracy and efficiency of fault diagnosis through the characteristics of quantum states and the application of quantum algorithms.

[0023] 2 Advantages: Ability to process complex electrical information and improve the system's fault tolerance and anti-interference ability. For example, quantum information processing algorithms can detect faults more accurately and have strong anti-interference capabilities.

[0024] 3 Application: Applied to fault diagnosis and protection of power systems. For example, during the operation of the power system, the electrical parameters are analyzed and processed through quantum information processing algorithms to improve the reliability and safety of the system.

[0025] (IV) Multi-scale geometric fusion algorithm

[0026] 1 Algorithm principle: Combine multi-scale geometric analysis and non-Euclidean geometry to achieve multi-scale fusion of power module electrical parameters. Through the multi-scale geometric analysis method, the characteristics of electrical parameters at different scales are extracted, and the relationship between these characteristics is analyzed using non-Euclidean geometry.

[0027] 2 Advantages: It can provide a more comprehensive understanding of the operating status of the power module and improve the accuracy and efficiency of fault diagnosis. For example, the multi-scale geometric fusion algorithm can better analyze the local and global characteristics of electrical parameters, thereby more accurately determining the fault type and location.

[0028] 3 Application: Applied to fault diagnosis and protection of power systems. For example, during the operation of the power system, the electrical parameters are analyzed and processed through multi-scale geometric fusion algorithms to improve the reliability and safety of the system.

[0029] (V) Adaptive fault diagnosis algorithm

[0030] 1 Algorithm implementation: Automatically adjust the fault diagnosis algorithm and system parameters according to the fault diagnosis results. For example, when a fault occurs, the parameters of the diagnosis algorithm are automatically adjusted according to the fault type and degree to improve the accuracy and efficiency of diagnosis.

[0031] 2 Innovation: The adaptive adjustment of the fault diagnosis system is realized, which improves the reliability and stability of the system. For example, the adaptive fault diagnosis algorithm can automatically adjust the operating parameters of the system according to the fault situation to ensure the safe operation of the system.

[0032] 3 Application: Applied to fault diagnosis and protection of power systems. For example, during the operation of the power system, the faults can be analyzed and processed through adaptive fault diagnosis algorithms to improve the reliability and safety of the system.

[0034] Example 1

[0035] 1. Fault diagnosis model based on hyperbolic partial differential equations

[0036] 1. Model construction: Use hyperbolic partial differential equations to describe the changing patterns of electrical parameters of power modules during operation. For example, the relationship between current, voltage and other parameters changing with time and space can be accurately modeled using hyperbolic partial differential equations.

[0037] 2 Solution method: The finite difference method is used to solve the hyperbolic partial differential equation. The time and space of the power module are discretized, and the electrical parameter values ​​at different times and locations are obtained through iterative calculation.

[0038] 3 Application: By monitoring the changes in electrical parameters, the fault type and location can be determined. For example, when the current or voltage fluctuates abnormally, the solution of the hyperbolic partial differential equation can be used to determine the location and extent of the fault.

[0039] Example 2

[0040] 2. Fault diagnosis model based on multi-scale geometric analysis

[0041] 1 Model principle: The electrical signal of the power module is analyzed using a multi-scale geometric analysis method. This method can capture the characteristics of the signal at different scales, thereby providing a more comprehensive understanding of the operating status of the power module.

[0042] 2 Feature extraction: Through tools such as wavelet transform, the electrical signal is decomposed into components at different scales. For example, in wavelet transform, wavelet coefficients of different scales reflect the different frequency characteristics of the signal.

[0043] 3. Fault diagnosis: Identify the fault type based on the extracted features. For example, when some frequency features of the signal are abnormal, determine the possible fault type.

[0044] Example 3

[0045] 3. Fault diagnosis model based on fuzzy random process

[0046] 1 Model establishment: Considering the uncertainty factors in the operation of the power module, a fuzzy random process is used to describe the probability of fault occurrence. For example, factors such as the time and degree of fault occurrence have certain randomness and fuzziness.

[0047] 2 Probability calculation: Calculate the probability of a fault through the theory of fuzzy mathematics and random processes. For example, use fuzzy membership functions and the probability distribution of random variables to determine the probability of a fault.

[0048] 3. Fault prediction: Based on the fault probability and related parameters, predict the time and extent of the fault. For example, when the fault probability reaches a certain threshold, early warning is given and corresponding protective measures are taken.

[0049] Example 4

[0050] (IV) Fault diagnosis model based on non-Euclidean geometry

[0051] 1. Geometric model: Introduce non-Euclidean geometry concepts to model the electrical parameter space of the power module. For example, in non-Euclidean space, concepts such as distance and angle are different from those in traditional Euclidean space.

[0052] 2. Feature analysis: Use the properties of non-Euclidean geometry to analyze the relationship between electrical parameters. For example, describe the change trend of electrical parameters by using distance and angle in non-Euclidean space.

[0053] 3. Fault diagnosis: Determine the fault type and location based on the analysis results. For example, when the distribution of electrical parameters in non-Euclidean space is abnormal, determine the location and cause of the fault.

[0054] Example 5

[0055] (V) Fault diagnosis model based on quantum information theory

[0056] 1. Quantum information representation: Using the characteristics of quantum states to represent the electrical information of the power module. For example, quantum bits can carry more information and have characteristics such as quantum entanglement.

[0057] 2. Information processing: Processing quantum information through quantum algorithms. For example, using the parallelism of quantum computing to quickly process electrical signals.

[0058] 3. Fault diagnosis: Diagnose the fault type and location based on the changes in quantum information. For example, when the quantum state changes, analyze the cause of the fault and possible impact.

[0059] Example 6

[0060] (I) Fault diagnosis and protection optimization solution

[0061] 1. Optimization algorithm: Combine the fault diagnosis model based on hyperbolic partial differential equations and multi-scale geometric analysis to improve the accuracy and efficiency of fault diagnosis. For example, in the process of solving hyperbolic partial differential equations, the multi-scale geometric analysis method is used to pre-process the signal to improve the accuracy and speed of the solution.

[0062] 2 Real-time monitoring: By real-time monitoring of the electrical parameters and temperature of the power module, faults can be discovered in time and corresponding protection measures can be taken. For example, sensors can be used to collect data in real time and transmit the data to the fault diagnosis system for analysis.

[0063] 3 Adaptive adjustment: Automatically adjust protection measures based on fault diagnosis results. For example, when a fault occurs, the parameters of the protection circuit are automatically adjusted according to the type and severity of the fault to ensure safe operation of the system.

[0064] (II) System performance improvement plan

[0065] 1. Performance optimization: Optimize system performance based on the fault diagnosis model of fuzzy random process and non-Euclidean geometry. For example, through the analysis of fuzzy random process, we can better understand the probability and influencing factors of fault occurrence, so as to optimize the operating parameters of the system.

[0066] 2. Reliability enhancement: Using quantum information theory and multi-scale geometric analysis methods to improve system reliability. For example, the processing and analysis of quantum information can detect faults more accurately and improve the fault tolerance of the system.

[0067] 3. Efficiency improvement: Improve the operating efficiency of the system by optimizing the fault diagnosis algorithm and system structure. For example, adopt parallel computing and intelligent algorithms to speed up the fault diagnosis and processing.

[0068] Example 7

[0069] (I) Hardware Design

[0070] 1. Power module: Use SiC semiconductor devices as core components to design power modules with high power density and low loss. For example, use SiC MOSFET as a power switch to improve the switching speed and efficiency of the module.

[0071] 2. Driving circuit: Design an efficient driving circuit to ensure the normal operation of SiC devices. For example, use an isolated driving circuit to improve the stability and anti-interference ability of the driving signal.

[0072] 3 Protection circuit: Set up a protection circuit to protect the power module from overcurrent, overvoltage, etc. For example, use an overcurrent protection circuit and an overvoltage protection circuit to prevent the power module from being damaged under abnormal conditions.

[0073] 4. Fault diagnosis circuit: Install a fault diagnosis circuit to monitor the electrical parameters and temperature of the power module in real time. For example, use sensors and signal processing circuits to transmit the monitored information to the fault diagnosis system for analysis.

[0074] (2. Software Implementation

[0075] 1. Fault diagnosis algorithm: Develop fault diagnosis software based on multi-scale hyperbolic partial differential algorithm, fuzzy random geometry algorithm, quantum information processing algorithm, multi-scale geometry fusion algorithm and adaptive fault diagnosis algorithm. For example, write a program to realize real-time monitoring and fault diagnosis functions of electrical parameters.

[0076] 2. Data processing: Process and analyze the collected data. For example, use data processing algorithms to filter and reduce noise on electrical parameters to improve data quality.

[0077] 3. Fault diagnosis result display: Display the fault diagnosis result on the display screen. For example, display the fault type, location and treatment measures through a graphical interface.

[0078] (III) Testing and Verification

[0079] 1. Test method: Test the power module to verify its performance and fault diagnosis capability. For example, by simulating fault scenarios, test the accuracy and reliability of the fault diagnosis system.

[0080] 2 Test results: According to the test results, the fault diagnosis system is optimized and improved. For example, the parameters of the fault diagnosis algorithm are adjusted according to the test results to improve the accuracy and efficiency of diagnosis.

[0081] 3. Verification: Verify the fault diagnosis system to ensure that it meets the actual application requirements. For example, verify the effectiveness and reliability of the fault diagnosis system through actual operation tests.

[0082] Example 8

[0083] Modeling and solving based on multi-scale hyperbolic partial differential algorithm

[0084] 1. Model establishment: The electrical characteristics and physical structure of the power module determine the variation of its electrical parameters. We construct a multi-scale hyperbolic partial differential equation based on the variation of current, voltage and other parameters inside the power module. This equation can describe the variation of electrical parameters at different scales. For example, at the macro scale, it can reflect the overall electrical performance of the power module; at the micro scale, it can reflect the local characteristics inside the chip.

[0085] 2. Discretization: To facilitate the solution, we use the finite difference method. The solution area of ​​the power module is divided into multiple grid points, and time and space are discretized. In this way, each grid point represents the electrical parameters of the power module at a specific time and location. By approximating the time and space derivatives, we convert the continuous equation into a discrete set of equations.

[0086] 3. Solution process: Use the iteration method to solve the discretized equations. During the iteration process, the electrical parameter values ​​at the grid points are gradually updated according to the initial conditions and boundary conditions. This process is repeated until the convergence condition is met. The convergence condition can be set as the error between two adjacent iteration results is less than a given threshold.

[0087] Modeling and solving based on fuzzy random geometry algorithm

[0088] 1 Fuzzy random process modeling: The process of power module failure is regarded as a fuzzy random process. Fuzzy random variables are defined to describe the possibility and location of failure. For example, we can set a fuzzy membership function to represent the degree of possibility of failure at different locations. The form of this function can be adjusted according to the actual situation, such as the distance to the fault location, the change of electrical parameters and other factors.

[0089] 2 Non-Euclidean geometry analysis: Map the geometric structure of the power module into non-Euclidean space. In this way, we can use the characteristics of non-Euclidean geometry to analyze the relationship between the fault location and electrical parameters. For example, in non-Euclidean space, the definitions of distance and angle are different from those in traditional Euclidean space, which helps us describe the geometric characteristics of the fault location more accurately.

[0090] 3 Probability calculation and diagnosis: Calculate the probability of fault occurrence based on the probability distribution of fuzzy random variables and the results of non-Euclidean geometry analysis. For example, by integrating the fuzzy membership function and combining it with non-Euclidean geometry, the probability value of fault occurrence is obtained. According to the size and distribution of the probability value, determine the most likely location and type of fault.

[0091] Modeling and solving based on quantum information processing algorithms

[0092] 1Quantum state representation: Encode the electrical parameters and state information of the power module into quantum states. Quantum states can represent different states, such as the spin state and energy state of electrons. The electrical parameters are represented by the superposition state of quantum bits, making the expression of information richer and more accurate.

[0093] 2 Quantum gate operation: quantum gates are used to operate and transform quantum states. Quantum gates can realize operations such as rotation and entanglement of quantum states, and simulate physical processes and fault conditions in power modules. For example, through a specific combination of quantum gates, the state of the quantum state can be changed, thereby simulating the changes in electrical parameters and the impact of faults.

[0094] 3. Measurement and diagnosis: Measure the quantum state after operation to obtain the measurement result. According to the measurement result and the evolution law of the quantum state, analyze the fault of the power module. For example, by measuring the state of the quantum bit, the value range and probability distribution of the electrical parameter can be obtained, so as to determine whether there is a fault and the type and location of the fault.

[0095] Modeling and solving based on multi-scale geometric fusion algorithm

[0096] 1. Multi-scale geometric analysis: Multi-scale analysis of the geometric structure of the power module is performed, from macro to micro, to extract geometric features at different scales. For example, at the macro scale, we can focus on the overall shape, size and layout of the power module; at the micro scale, we can deeply analyze the geometric details of the chip, circuit, etc. Through various mathematical tools, such as wavelet transform and fractal geometry, we can obtain geometric feature parameters at different scales.

[0097] 2 Non-Euclidean geometry fusion: The characteristic parameters obtained from multi-scale geometric analysis are combined with non-Euclidean geometry to establish a fusion model. Non-Euclidean geometry provides a more flexible geometric description method that can better handle complex geometric shapes and structures. By constructing geometric manifolds and mapping functions, microscopic geometric features are linked to macroscopic geometric features, thereby more comprehensively describing the geometric characteristics of the power module.

[0098] 3. Fault diagnosis and analysis: Fault diagnosis and analysis are performed based on the fusion model and the changes in electrical parameters. The fault type and location are determined by comparing the differences in geometric features and electrical parameters under normal and fault conditions. For example, when a short circuit occurs, the geometry of the local area may change, and the electrical parameters may also become abnormal. The fusion model can accurately detect and locate this fault.

[0099] Modeling and solving based on adaptive fault diagnosis algorithm

[0100] 1. Fault feature extraction: Real-time monitoring of the electrical parameters and status information of the power module to extract features related to the fault. These features may include changes in current and voltage, frequency, phase, etc. For example, by analyzing the changes in electrical parameters, feature vectors that can reflect the occurrence of a fault can be extracted.

[0101] 2. Diagnostic model establishment and update: Based on the initial fault data and expert knowledge, a fault diagnosis model is established. This model can use machine learning algorithms such as neural networks, support vector machines, etc. During operation, the parameters of the diagnostic model are continuously updated and optimized according to new fault data. For example, for a neural network model, the connection weights between neurons are adjusted through the back propagation algorithm to improve the accuracy and adaptability of the diagnostic model.

[0102] 3 Adaptive adjustment and diagnosis: Adaptively adjust the diagnostic algorithm and model parameters according to the fault diagnosis results and the system's operating status. For example, after a certain type of fault is diagnosed, adjust the feature extraction method and the structure of the diagnostic model according to the characteristics and historical data of the fault to better adapt to the diagnosis of similar faults. At the same time, dynamically adjust the threshold and weight of the diagnostic model according to the system's operating environment and load changes to improve the reliability and real-time performance of fault diagnosis.

[0103] The present invention is based on the intelligent power module and fault diagnosis system of SiC semiconductors, and improves the accuracy and efficiency of fault diagnosis by introducing a variety of advanced mathematical models and algorithms. By optimizing the system structure and performance, the safe and stable operation of the power system is achieved. The innovation and technical solution of the present invention have important application value and provide new ideas and methods for the development of the power electronics field.

Claims

1. An intelligent power module: characterized in that The intelligent power module integrates SiC semiconductor devices, drive circuits, protection circuits and fault diagnosis circuits. It can monitor the electrical parameters and temperature information of the module in real time, and realize fast and accurate detection and positioning of faults through a fault diagnosis system based on multi-scale hyperbolic partial differential algorithm, fuzzy random geometry algorithm, quantum information processing algorithm, multi-scale geometry fusion algorithm and adaptive fault diagnosis algorithm.

2. The fault diagnosis system of the intelligent power module according to claim 1: comprising sensors for real-time monitoring of electrical parameters and temperature information, and fault diagnosis software based on the five algorithms. The five algorithms are respectively a multi-scale hyperbolic partial differential algorithm, a fuzzy random geometry algorithm, a quantum information processing algorithm, a multi-scale geometry fusion algorithm and an adaptive fault diagnosis algorithm.

3. A fault diagnosis method based on a multi-scale hyperbolic partial differential algorithm: used to perform multi-scale analysis of the electrical parameters of the power module according to the hyperbolic partial differential equation, characterized in that The hyperbolic partial differential equation is solved by the finite difference method to obtain the variation law of electrical parameters at different scales, so as to determine the fault type and location.

4. A fault diagnosis method based on fuzzy random geometry algorithm: used to analyze the probability and location of fault occurrence using fuzzy random process and non-Euclidean geometry, characterized in that The probability of fault occurrence is calculated through fuzzy membership functions and probability distribution of random variables, and the distance and angle characteristics of non-Euclidean geometry are used to determine the fault location.

5. A fault diagnosis method based on quantum information processing algorithm: used to process the electrical information of the power module using quantum information theory and quantum computing technology, characterized in that The electrical information is processed through the characteristics of the quantum state and quantum algorithms, and the fault type and location are diagnosed based on the changes in quantum information.

6. A fault diagnosis method based on a multi-scale geometric fusion algorithm: used to combine multi-scale geometric analysis and non-Euclidean geometry to perform multi-scale fusion analysis on power module electrical parameters, characterized in that The characteristics of electrical parameters at different scales are extracted through multi-scale geometric analysis, and the relationship between these characteristics is analyzed using non-Euclidean geometry to determine the fault type and location.

7. A fault diagnosis method based on an adaptive fault diagnosis algorithm: used to automatically adjust the fault diagnosis algorithm and system parameters according to the fault diagnosis results, characterized in that By real-time monitoring of fault types and severity, the parameters of the diagnostic algorithm are automatically adjusted to improve the accuracy and efficiency of fault diagnosis.