Method and system for non-contact life evaluation of power devices based on near-field electromagnetic signals

Abstract

This invention provides a non-contact lifetime assessment method and system for power devices based on near-field electromagnetic signals, belonging to the field of semiconductor device condition monitoring technology. The method includes: acquiring near-field electromagnetic signals from the surrounding space of the power device under test during its operation; performing time-frequency transformation, frequency domain filtering, and normalization preprocessing on the near-field electromagnetic signals to obtain a standardized frequency domain signal; extracting feature parameters, including center frequency, maximum amplitude, spectral sequence, and spectral energy, from the standardized frequency domain signal to form a feature input sequence; and inputting the feature input sequence into an artificial intelligence model consisting of a high-dimensional information extraction module, a dynamic characteristic extraction module, and a health status assessment module for processing to obtain the lifetime assessment result of the power device. This invention achieves online, non-contact lifetime monitoring and assessment of in-service power devices without modifying the devices or circuits, and has the advantages of low deployment cost and good applicability.

Description

Technical Field

[0001] This invention relates to the field of semiconductor device condition monitoring technology, and in particular to a non-contact lifetime assessment method and system for power devices based on near-field electromagnetic signals. Background Technology

[0002] Power devices, as core components of power electronic systems, are widely used in radio frequency power amplification, power management, and high-reliability electronic systems. In practical applications, these devices must withstand harsh operating conditions such as high voltage, high current, and high power density for extended periods. Under the coupled effects of electrical and thermal stress, their internal structures gradually undergo irreversible performance degradation, ultimately leading to device failure. Therefore, accurate and timely assessment of the lifespan or remaining lifespan of power devices is crucial for ensuring the safety and reliability of system operation and for developing scientific maintenance strategies. Currently, power device lifespan assessments largely rely on accelerated aging tests conducted in laboratory environments or reliability models based on mathematical statistics. These methods are time-consuming and costly, and their assessment results are difficult to accurately map to the complex and variable actual operating conditions faced by devices in operation, lacking online monitoring capabilities.

[0003] To assess the health of devices in operation, some existing solutions attempt to infer their health status by monitoring operating electrical parameters such as on-state voltage drop and leakage current. However, such methods typically require establishing an electrical connection with the system under test, or adding sensors and data interfaces within the device and circuitry, resulting in high engineering deployment complexity, significant modifications to the existing system, and difficulty in achieving convenient non-intrusive applications. Furthermore, during normal operation, the evolution of internal carrier behavior, parasitic parameters, and electrothermal characteristics of power devices directly affects their switching transient processes, thereby generating near-field electromagnetic signals closely related to their internal states in the surrounding space. Currently, these near-field electromagnetic signals are mostly used for interference source analysis and localization in the field of electromagnetic compatibility, and there is a lack of a technical solution to effectively assess the operating stress and lifespan status of power devices using these signals. Therefore, there is an urgent need for a method and system that can perform non-contact, online, and accurate lifespan assessment of in-service power devices without altering the existing system hardware structure. Summary of the Invention

[0004] To address the aforementioned technical problems in related technologies, this invention proposes a non-contact lifetime assessment method and system for power devices based on near-field electromagnetic signals, which can overcome the above-mentioned shortcomings of the prior art.

[0005] To achieve the above-mentioned technical objectives, the technical solution of the present invention is implemented as follows: A non-contact lifetime assessment method for power devices based on near-field electromagnetic signals; This non-contact lifetime assessment method for power devices based on near-field electromagnetic signals includes the following steps: Step S1: Signal acquisition, acquiring the near-field electromagnetic signals of the surrounding space of the power device under test in operation; Step S2: Signal preprocessing, the acquired near-field electromagnetic signals are preprocessed to obtain standardized frequency domain signals; Step S3: Feature extraction, extracting at least one feature parameter from the standardized frequency domain signal to form a feature input sequence; Step S4: Artificial intelligence model processing, the feature input sequence is input into a pre-trained artificial intelligence model and processed by the artificial intelligence model to evaluate the lifetime status of the power device under test; The near-field electromagnetic signals are acquired in a non-contact manner.

[0006] Further, step S2 specifically includes: The acquired near-field electromagnetic signal in the time domain is subjected to time-frequency transformation to obtain the frequency domain amplitude spectrum; The frequency domain amplitude spectrum is filtered in the frequency domain to retain the spectral components in the key frequency bands that reflect the device state; The filtered frequency domain amplitude spectrum is normalized to obtain the standardized frequency domain signal.

[0007] Furthermore, the frequency domain filtering presets a characteristic passband frequency range based on the current switching frequency of the power device under test, and filters out spectral components located outside the characteristic passband frequency range.

[0008] Further, step S3 specifically includes: Extract the center frequency of the standardized frequency domain signal. f 0; Extract the maximum amplitude of the standardized frequency domain signal. S max ; Extract at the center frequency f A spectral signal sequence centered at 0 and within a preset bandwidth B. S f0 ; Calculate the spectral energy within the preset bandwidth B range. E f0 ; The center frequency f 0. Maximum amplitude S max Spectral signal sequence S f0 and spectral energy E f0Combined together or in part, to form the feature input sequence. X input .

[0009] Furthermore, the spectral energy E f0 Calculated using the following formula: ; in, The frequency amplitude corresponding to the nth sampling point within the preset bandwidth B frequency band, where N is the total number of sampling points within that frequency band. This represents the frequency resolution of the signal.

[0010] Furthermore, the artificial intelligence model includes: The high-dimensional information extraction module is used to extract high-dimensional information from the feature input sequence; A dynamic feature extraction module is used to perceive and process the dynamic changes of the high-dimensional information in the reduced dimension; The health status assessment module is used to assess the lifetime status of the power device under test by providing the high-dimensional information and dynamic change information.

[0011] Furthermore, the high-dimensional information extraction module is implemented using a convolutional neural network, the dynamic feature extraction module is implemented using a recurrent neural network or a long short-term memory neural network, and the health status assessment module is implemented using a fully connected neural network or an extreme gradient boosting algorithm.

[0012] Furthermore, the method periodically performs signal acquisition, signal preprocessing, feature extraction, and artificial intelligence model processing steps at fixed time intervals to dynamically monitor and evaluate the lifetime status of the power device under test.

[0013] According to another aspect of the present invention, a non-contact lifetime assessment system for power devices based on near-field electromagnetic signals is provided; This non-contact lifetime assessment system for power devices based on near-field electromagnetic signals includes: Near-field electromagnetic probes are used to collect near-field electromagnetic signals from the surrounding space of a power device under test in a non-contact manner during operation. A data processing module is communicatively connected to the near-field electromagnetic probe, and the data processing module is configured to perform the steps of the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals as described in any one of claims 1 to 8, so as to output the lifetime assessment result of the power device.

[0014] Furthermore, the near-field electromagnetic probe is positioned at a fixed height above the power device under test.

[0015] The beneficial effects of this invention are as follows: By collecting the near-field electromagnetic signals spontaneously emitted by power devices under normal operating conditions and performing standardized preprocessing and multi-dimensional feature extraction from the time domain to the frequency domain, the subsequent artificial intelligence model can accurately perceive the device degradation information hidden in the signal spectrum, thereby achieving the purpose of online life assessment of power devices under non-contact conditions and without modifying the system under test, effectively improving the safety of in-service system operation and the economy of maintenance. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart of the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals according to the present invention. Figure 2 This is a diagram of the electromagnetic radiation signal acquisition structure of the non-contact lifetime assessment system for power devices based on near-field electromagnetic signals described in this invention. Figure 3 This is a flowchart of the signal processing and artificial intelligence model processing of the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals described in this invention. Detailed Implementation

[0018] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention.

[0020] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0021] like Figure 1 and Figure 3 As shown in the embodiment of the present invention, the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals includes the following steps: Step S1: Signal acquisition, acquiring the near-field electromagnetic signals of the surrounding space of the power device under test in operation; Step S2: Signal preprocessing, the acquired near-field electromagnetic signals are preprocessed to obtain standardized frequency domain signals; Step S3: Feature extraction, extracting at least one feature parameter from the standardized frequency domain signal to form a feature input sequence; Step S4: Artificial intelligence model processing, the feature input sequence is input into a pre-trained artificial intelligence model and processed by the artificial intelligence model to evaluate the lifetime status of the power device under test; The near-field electromagnetic signals are acquired in a non-contact manner.

[0022] According to an embodiment of the present invention, the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals, in a specific embodiment, step S2 specifically includes: The acquired near-field electromagnetic signal in the time domain is subjected to time-frequency transformation to obtain the frequency domain amplitude spectrum; The frequency domain amplitude spectrum is filtered in the frequency domain to retain the spectral components in the key frequency bands that reflect the device state; The filtered frequency domain amplitude spectrum is normalized to obtain the standardized frequency domain signal.

[0023] According to an embodiment of the present invention, the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals, in a specific implementation, the frequency domain filtering presets a characteristic passband frequency range based on the current switching frequency of the power device under test, and filters out spectral components located outside the characteristic passband frequency range.

[0024] According to an embodiment of the power device non-contact lifetime assessment method based on near-field electromagnetic signals described in this invention, in a specific embodiment, step S3 specifically includes: Extract the center frequency of the standardized frequency domain signal. f 0; Extract the maximum amplitude of the standardized frequency domain signal. S max ; Extract at the center frequency f A spectral signal sequence centered at 0 and within a preset bandwidth B. S f0 ; Calculate the spectral energy within the preset bandwidth B range. E f0 ; The center frequency f 0. Maximum amplitude S max Spectral signal sequence S f0 and spectral energy E f0 Combined together or in part, to form the feature input sequence. X input .

[0025] According to an embodiment of the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals described in this invention, in a specific implementation, the spectral energy... E f0 Calculated using the following formula: ; in, The frequency amplitude corresponding to the nth sampling point within the preset bandwidth B frequency band, where N is the total number of sampling points within that frequency band. This represents the frequency resolution of the signal.

[0026] According to an embodiment of the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals described in this invention, in a specific implementation, the artificial intelligence model includes: The high-dimensional information extraction module is used to extract high-dimensional information from the feature input sequence; The dynamic feature extraction module is used to sense and process the dynamic changes of the high-dimensional information in the time dimension. The health status assessment module is used to assess the lifetime status of the power device based on the high-dimensional information and dynamic change information.

[0027] According to an embodiment of the present invention, the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals, in a specific implementation, the high-dimensional information extraction module is implemented using a convolutional neural network, the dynamic characteristic extraction module is implemented using a recurrent neural network or a long short-term memory neural network, and the health status assessment module is implemented using a fully connected neural network or a limit gradient boosting algorithm.

[0028] According to an embodiment of the present invention, the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals, in a specific implementation, the method periodically performs signal acquisition, signal preprocessing, feature extraction, and artificial intelligence model processing steps at fixed time intervals to dynamically monitor and assess the lifetime status of the power devices.

[0029] Secondly, such as Figure 2 As shown, the non-contact lifetime assessment system for power devices based on near-field electromagnetic signals according to an embodiment of the present invention includes: Near-field electromagnetic probes are used to collect near-field electromagnetic signals from the surrounding space of a power device under test in a non-contact manner during operation. A data processing module is communicatively connected to the near-field electromagnetic probe, and the data processing module is configured to perform the steps of the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals as described in any one of claims 1 to 8, so as to output the lifetime assessment result of the power device.

[0030] In a specific embodiment of the non-contact lifetime assessment system for power devices based on near-field electromagnetic signals according to an embodiment of the present invention, the near-field electromagnetic probe is positioned at a fixed height above the power device under test.

[0031] To facilitate understanding of the above technical solutions of the present invention, the following detailed description of the above technical solutions of the present invention is provided through specific embodiments and working principles.

[0032] Example 1 System Architecture and Signal Acquisition The system architecture proposed in this invention mainly consists of the power device under test, a near-field electromagnetic probe, and a data processing module that is communicatively connected to the near-field electromagnetic probe. The data processing module is responsible for performing subsequent signal preprocessing, feature extraction, and artificial intelligence model analysis. Its specific implementation can be an industrial computer, embedded system, or cloud server integrated with a processor and memory; no limitation is imposed here.

[0033] During the signal acquisition phase, the power device under test (DUT) is placed in its normal high-frequency operating environment. Its internal periodic switching action generates near-field electromagnetic radiation signals in the surrounding space. A near-field electromagnetic probe is positioned non-contactly at a fixed height on the surface of the DUT to capture these near-field time-domain electromagnetic radiation signals. This non-contact acquisition method is a key premise of this invention, ensuring that the measurement process does not introduce additional electrical loads, does not alter the original operating state of the power device, and requires no invasive hardware modifications to the system under test.

[0034] Signal preprocessing and feature extraction After receiving the raw time-domain signal from the near-field electromagnetic probe, the data processing module first performs a signal preprocessing procedure. The purpose of the preprocessing procedure is to eliminate environmental noise interference, stabilize the signal amplitude, and provide high-quality, standardized data input for subsequent feature extraction.

[0035] Specifically, the preprocessing process includes the following three core operations: The first step is time-frequency transformation. A Fast Fourier Transform (FFT) is applied to the acquired raw time-domain signal to transform it from the time domain to the frequency domain, obtaining a frequency-domain amplitude spectrum that characterizes the distribution of signal energy with frequency. This step converts the time-varying electromagnetic fluctuations in the physical world into a frequency function that is mathematically easier to analyze and extract features from.

[0036] The second step is frequency domain filtering. This step serves the dual functions of noise suppression and feature focusing. Based on the current switching frequency of the power device under test, this invention presets a characteristic passband frequency range. In the obtained frequency domain amplitude spectrum, only the spectral components falling within this passband range are retained, while components outside this passband range, mainly composed of noise and interference, are removed. This effectively suppresses irrelevant background electromagnetic interference, allowing subsequent analysis to focus on the sensitive frequency bands that best reflect the device's switching behavior and internal state.

[0037] The third step is signal normalization. The filtered frequency domain amplitude spectrum is normalized to eliminate overall signal amplitude variations that may be caused by minor fluctuations in the device's operating point or probe position deviations, thereby enhancing the generalization ability and evaluation stability of subsequent artificial intelligence models. After these three steps, a standardized frequency domain signal with stable amplitude, high signal-to-noise ratio, and dedicated to subsequent analysis is finally obtained.

[0038] Next, feature extraction is performed on the standardized frequency domain signal. To comprehensively and accurately capture signal changes related to device lifetime degradation, this invention extracts feature parameters from multiple dimensions to form the input sequence of the artificial intelligence model. X input The extracted feature parameters specifically include: Center frequency f 0, which is the frequency value corresponding to the highest amplitude point in the standardized frequency domain signal, is directly related to the core switching frequency of the power device.

[0039] Maximum amplitude S max That is, the peak amplitude of the standardized frequency domain signal spectrum reflects the transient energy intensity of the switching process.

[0040] Spectral signal sequences within a bandwidth B near the center frequency: ( S f0 = [ S 1, S 2, S 3,…, S N The sequence is specifically in | f – f Within the frequency band of 0|≤0.5B, the frequency amplitude corresponding to each sampling point S N It is obtained through sequential sampling. It depicts the spectral envelope details near the center of the switching frequency, containing rich device parasitic parameters and state information.

[0041] Spectral energy within bandwidth B near the center frequency E f0 This energy value is obtained through the formula: The calculation shows that, among which This represents the frequency resolution of the signal. E f0 It represents the overall radiated energy level within the sensitive frequency band and is a key indicator for evaluating the operating stress of devices.

[0042] Finally, the above feature parameters are integrated to form a multi-dimensional feature input sequence: X input =[ f 0, S max , S f0 , E f0 ] -1 This data will serve as the foundation for subsequent analysis of artificial intelligence models.

[0043] Artificial intelligence models and life assessment The artificial intelligence model is the core algorithm module of this invention for mapping electromagnetic signal features to lifetime states. It receives the processed feature input sequence for each frame. X input After internal calculations, the final lifespan assessment results of the power devices are output.

[0044] In a preferred embodiment, the artificial intelligence model employs a cascaded neural network architecture, consisting of three functionally interconnected core modules: The first module is the high-dimensional information extraction module. This module can be implemented using a convolutional neural network (CNN). Through its internal convolutional and pooling layers, CNN can abstract and compress the input feature sequence layer by layer, effectively extracting the high-dimensional spatial structure information hidden in the original features. While reducing the data dimensionality, it retains the most critical state representation.

[0045] Secondly, there is the dynamic characteristic extraction module. Since the degradation of power devices is a continuous process evolving over time, the current state is closely related to historical states. Therefore, this embodiment introduces a dynamic characteristic extraction module to capture this temporal dependency. This module can be implemented using a recurrent neural network (RNN) or its variant, a long short-term memory neural network (LSTM). The RNN / LSTM network structure can "remember" historical information from multiple past frames and couple it with the information of the current frame for analysis, thereby perceiving the dynamic evolution trend of the device's internal characteristics over a long time dimension, providing the necessary historical background reference for the final lifetime assessment.

[0046] Finally, there is the health status assessment module. This module receives high-dimensional features and dynamic information processed by the first two modules, and performs the final regression or classification calculation using a fully connected neural network (FCNN) or extreme gradient boosting (XGBoost) algorithm. The fully connected layer outputs a specific numerical value or status level as an assessment result of the remaining lifespan or health status of the current power device through weighted summation and nonlinear mapping of the input information.

[0047] During system operation, to balance continuous monitoring with efficient use of computing resources, the data processing module periodically reads a frame of time-domain signal from the near-field electromagnetic probe at fixed time intervals T, and repeatedly executes the aforementioned preprocessing, feature extraction, and model analysis processes. The dynamic characteristic extraction module in the artificial intelligence model continuously iterates and updates its internal state memory, ensuring that each lifetime assessment result is not only based on a static judgment of the current single frame signal, but also on a dynamic comprehensive assessment based on the cumulative coupling of all historical electromagnetic signal information since the start of monitoring, thereby guaranteeing the accuracy and stability of the assessment results.

[0048] In summary, the non-contact lifetime assessment method and system for power devices based on near-field electromagnetic signals provided by this invention non-contactly senses the near-field electromagnetic signals inevitably generated by the power device during operation, and uses artificial intelligence algorithms to deeply mine the implicit degradation information in these signals, achieving online, real-time, and accurate assessment of the remaining lifetime of in-service power devices. This method is flexible in deployment, requires no hardware modifications to existing circuit systems, greatly reducing the application threshold and implementation cost of lifetime monitoring, and has significant practical value for improving the operational safety and maintenance economy of various power electronic systems.

[0049] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A non-contact lifetime assessment method for power devices based on near-field electromagnetic signals, characterized in that, Includes the following steps: Step S1: Signal acquisition, acquiring the near-field electromagnetic signals of the surrounding space of the power device under test in operation; Step S2: Signal preprocessing, the acquired near-field electromagnetic signals are preprocessed to obtain standardized frequency domain signals; Step S3: Feature extraction, extracting at least one feature parameter from the standardized frequency domain signal to form a feature input sequence; Step S4: Artificial intelligence model processing, the feature input sequence is input into a pre-trained artificial intelligence model and processed by the artificial intelligence model to evaluate the lifetime status of the power device under test; The near-field electromagnetic signals are acquired in a non-contact manner.

2. The non-contact lifetime assessment method for power devices based on near-field electromagnetic signals according to claim 1, characterized in that, Step S2 specifically includes: The acquired near-field electromagnetic signal in the time domain is subjected to time-frequency transformation to obtain the frequency domain amplitude spectrum; The frequency domain amplitude spectrum is filtered in the frequency domain to retain the spectral components in the key frequency bands that reflect the device state; The filtered frequency domain amplitude spectrum is normalized to obtain the standardized frequency domain signal.

3. The non-contact lifetime assessment method for power devices based on near-field electromagnetic signals according to claim 2, characterized in that, The frequency domain filtering presets a characteristic passband frequency range based on the current switching frequency of the power device under test, and filters out spectral components located outside the characteristic passband frequency range.

4. The non-contact lifetime assessment method for power devices based on near-field electromagnetic signals according to claim 1, characterized in that, Step S3 specifically includes: Extract the center frequency of the standardized frequency domain signal. f 0; Extract the maximum amplitude of the standardized frequency domain signal. S max ; Extract at the center frequency f A spectral signal sequence centered at 0 and within a preset bandwidth B. S f0 ; Calculate the spectral energy within the preset bandwidth B range. E f0 ; The center frequency f 0. Maximum amplitude S max Spectral signal sequence S f0 and spectral energy E f0 Combined together or in part, to form the feature input sequence. X input .

5. The non-contact lifetime assessment method for power devices based on near-field electromagnetic signals according to claim 4, characterized in that, The spectral energy E f0 Calculated using the following formula: ； in, The frequency amplitude corresponding to the nth sampling point within the preset bandwidth B frequency band, where N is the total number of sampling points within that frequency band. This represents the frequency resolution of the signal.

6. The non-contact lifetime assessment method for power devices based on near-field electromagnetic signals according to claim 1, characterized in that, The artificial intelligence model includes: The high-dimensional information extraction module is used to extract high-dimensional information from the feature input sequence; The dynamic feature extraction module is used to sense and process the dynamic changes of the high-dimensional information in the time dimension. The health status assessment module is used to assess the lifetime status of the power device based on the high-dimensional information and dynamic change information.

7. The non-contact lifetime assessment method for power devices based on near-field electromagnetic signals according to claim 6, characterized in that, The high-dimensional information extraction module is implemented using a convolutional neural network, the dynamic feature extraction module is implemented using a recurrent neural network or a long short-term memory neural network, and the health status assessment module is implemented using a fully connected neural network or an extreme gradient boosting algorithm.

8. The non-contact lifetime assessment method for power devices based on near-field electromagnetic signals according to claim 1, characterized in that, The method periodically performs signal acquisition, signal preprocessing, feature extraction, and artificial intelligence model processing steps at fixed time intervals to dynamically monitor and evaluate the lifetime status of the power device under test.

9. A non-contact lifetime assessment system for power devices based on near-field electromagnetic signals, characterized in that, include: Near-field electromagnetic probes are used to collect near-field electromagnetic signals from the surrounding space of a power device under test in a non-contact manner during operation. A data processing module is communicatively connected to the near-field electromagnetic probe, and the data processing module is configured to perform the steps of the non-contact lifetime assessment method for power devices based on near-field electromagnetic signals as described in any one of claims 1 to 8, so as to output the lifetime assessment result of the power device.

10. The non-contact lifetime assessment system for power devices based on near-field electromagnetic signals according to claim 9, characterized in that, The near-field electromagnetic probe is positioned at a fixed height above the power device under test.

Images