State detection method and related device, equipment, medium and program product

Abstract

The invention discloses a state detection method, a related device, equipment, a medium and a program product. The state detection method comprises the following steps: decomposing a first electric signal of an electric drive system into a plurality of first modal components; selecting at least one noise mode component from the plurality of first mode components; performing filtering processing on the at least one noise mode component to obtain at least one second mode component; performing signal reconstruction by using the at least one second modal component and each unfiltered modal component in the plurality of first modal components to obtain a second electric signal; and determining a state detection result of the electric drive system at least by using the second electric signal. In this way, the accuracy of state detection of the electric drive system can be improved.

Description

Technical Field

[0001] This application relates to the field of equipment monitoring technology, and in particular to a condition detection method and related devices, equipment, media and program products. Background Technology

[0002] In industrial settings, it is often necessary to monitor the status of electric drive systems to determine if any faults exist. This serves two purposes: firstly, it ensures the safe operation of the electric drive system and improves its reliability; secondly, it allows relevant personnel to promptly analyze and repair any malfunctioning systems.

[0003] Therefore, how to accurately detect the status of the electric drive system has become an urgent technical problem to be solved. Summary of the Invention

[0004] The main technical problem addressed by this application is to provide a condition detection method and related devices, equipment, media, and program products that can improve the accuracy of condition detection in electric drive systems.

[0005] To solve the above-mentioned technical problems, one technical solution adopted in this application is: to provide a state detection method, the method comprising: decomposing a first electrical signal of an electric drive system into several first mode components; selecting at least one noise mode component from the several first mode components; filtering the at least one noise mode component to obtain at least one second mode component; reconstructing the signal using the at least one second mode component and each unfiltered mode component from the several first mode components to obtain a second electrical signal; and determining the state detection result of the electric drive system using at least the second electrical signal.

[0006] In one embodiment, selecting at least one noise mode component from a plurality of first mode components includes: determining the pulse characterization value of each first mode component, wherein the pulse characterization value is used to characterize the intensity of the pulse signal in the mode component; and selecting at least one mode component whose pulse characterization value meets a preset requirement from the plurality of first mode components as at least one noise mode component.

[0007] In one embodiment, the pulse characterization value is spectral kurtosis, and the preset requirement includes: the spectral kurtosis of the modal component is greater than the spectral kurtosis threshold.

[0008] In one embodiment, filtering is performed on at least one noise mode component to obtain at least one second mode component, including: performing wavelet transform on each noise mode component to obtain a plurality of first wavelet coefficients corresponding to each noise mode component; performing wavelet threshold filtering on the plurality of first wavelet coefficients of each noise mode component to obtain a plurality of second wavelet coefficients corresponding to the noise mode component; and performing inverse wavelet transform on the plurality of second wavelet coefficients of each noise mode component to obtain the second mode component corresponding to the noise mode component.

[0009] In one embodiment, wavelet threshold filtering is performed on several first wavelet coefficients of the noise mode component to obtain several second wavelet coefficients corresponding to the noise mode component. This includes: for each first wavelet coefficient of the noise mode component, determining whether the first wavelet coefficient satisfies a filtering condition, which includes: the absolute value of the first wavelet coefficient being greater than or equal to a filtering threshold; if the first wavelet coefficient satisfies the filtering condition, then the difference between the first wavelet coefficient and the coefficient adjustment value is determined as the second wavelet coefficient corresponding to the first wavelet coefficient; if the first wavelet coefficient does not satisfy the filtering condition, then the product of the first wavelet coefficient and the attenuation coefficient is determined as the second wavelet coefficient corresponding to the first wavelet coefficient; wherein the coefficient adjustment value and the attenuation coefficient are both determined using the first wavelet coefficient and the filtering threshold.

[0010] In one embodiment, the step of determining the coefficient adjustment value includes: determining the sign value of the first wavelet coefficient using a sign function; determining a first product of the filter threshold and the sign value; determining a second product of a first preset value and the first wavelet coefficient, wherein the first preset value is negative; performing a power operation with the second preset value as the base and the second product as the exponent to obtain a first power operation value; summing the third preset value and the first power operation value to obtain a first sum value; and determining a first ratio of the first product to the first sum value as the coefficient adjustment value.

[0011] In one embodiment, the step of determining the attenuation coefficient includes: summing the square of the fourth preset value and the square of the first wavelet coefficient to obtain a second sum; determining the second ratio of the square of the filtering threshold to the second sum; determining the third product of the second ratio and the fifth preset value; performing a power operation with the sixth preset value as the base and the third product as the exponent to obtain a second power operation value; and using the second power operation value as the attenuation coefficient.

[0012] In one embodiment, the first electrical signal is the three-phase current signal of the motor.

[0013] And / or, the first electrical signal is one of several first signal segments, and the steps of obtaining several first signal segments include: dividing the original electrical signal of the electric drive system into several second signal segments; multiplying each second signal segment by a preset window function to obtain several first signal segments.

[0014] And / or, decompose the first electrical signal of the electric drive system into several first modal components, including: performing an empirical wavelet transform on the first electrical signal to obtain several first modal components; wherein the bandwidth parameter of each first modal component is proportional to the fundamental frequency of the first electrical signal.

[0015] In one embodiment, determining the state detection result of the electric drive system using at least the second electrical signal includes: performing a fast Fourier transform on the second electrical signal to obtain the frequency domain characteristics of the second electrical signal; extracting the harmonic distortion rate of the second electrical signal using the frequency domain characteristics; and predicting the state detection result of the electric drive system using at least the harmonic distortion rate of the second electrical signal.

[0016] In one embodiment, predicting the state detection result of the electric drive system using at least the harmonic distortion rate of the second electrical signal includes: acquiring dynamic temporal features and static features; wherein the dynamic temporal features include time-series sampling data of at least one first influencing factor, the at least one first influencing factor includes the harmonic distortion rate of the second electrical signal, and the static features include sampling data of several second influencing factors; extracting features from the dynamic temporal features using a time-series coding network to obtain a first feature; and extracting features from the static features using a deep residual network to obtain a second feature; fusing the first feature and the second feature to obtain a fused feature; and using a prediction network to make a prediction based on the fused feature to obtain the state detection result of the electric drive system.

[0017] In one embodiment, fusing a first feature and a second feature to obtain a fused feature includes: determining the attention weights corresponding to the first feature and the second feature respectively; weighting the first feature and the second feature using the attention weights corresponding to the first feature and the second feature respectively to obtain an attention feature; and using the attention feature as the fused feature.

[0018] And / or, before using a temporal coding network to extract features from dynamic temporal features and obtain the first feature, the method further includes: performing temporal alignment processing on the dynamic temporal features.

[0019] In one embodiment, at least one first influencing factor further includes the temperature of a plurality of temperature sampling points in the battery, and the method further includes: acquiring temperature sampling values ​​of each temperature sampling point at a first sampling frequency; and acquiring temperature values ​​of each temperature sampling point at a second sampling frequency in response to satisfying a sampling frequency switching condition; wherein the first sampling frequency is less than the second sampling frequency, and the sampling frequency switching condition includes at least one of the following conditions: there exists at least one set of temperature sampling points whose temperature difference at the same sampling time is greater than or equal to a first temperature threshold, and there exists at least one temperature sampling point whose temperature difference at two adjacent sampling times is greater than or equal to a second temperature threshold; or, the first sampling frequency is greater than the second sampling frequency, and the sampling frequency switching condition includes at least one of the following conditions: the temperature difference between any two temperature sampling points at the same sampling time is less than a third temperature threshold, and the temperature difference between each temperature sampling point at two adjacent sampling times is less than a fourth temperature threshold.

[0020] In one embodiment, the state detection result of the electric drive system includes the failure probability of the electric drive system. After obtaining the state detection result of the electric drive system by prediction using at least a second electrical signal, the method further includes: determining a failure probability threshold using current vehicle operating data; wherein the current vehicle operating data includes at least one of the current vehicle load rate and the current battery cycle count; and performing fault handling in response to the failure probability being greater than or equal to the failure probability threshold.

[0021] In one embodiment, the current vehicle operating data includes the current vehicle load rate and the current battery cycle count. Determining a fault probability threshold using the current vehicle operating data includes: determining a first adjustment coefficient using the current vehicle load rate, and determining a second adjustment coefficient using the current battery cycle count; wherein the first adjustment coefficient is positively correlated with the current vehicle load rate, and the second adjustment coefficient is positively correlated with the current battery cycle count; adjusting a base adjustment value using the first and second adjustment coefficients to obtain the fault probability threshold.

[0022] In one embodiment, determining a first adjustment coefficient using the current vehicle load rate includes: determining a fourth product of a first adjustment factor and the current vehicle load rate; and summing a preset constant and the fourth product to obtain the first adjustment coefficient.

[0023] And / or, determine the second adjustment factor using the current battery cycle count, including: performing a power operation with the second adjustment factor as the base and the current battery cycle count as the exponent to obtain the second adjustment factor.

[0024] And / or, adjust the base adjustment value using the first adjustment coefficient and the second adjustment coefficient to obtain the fault probability threshold, including: using the product of the first adjustment coefficient, the second adjustment coefficient and the base adjustment value as the fault probability threshold.

[0025] In one embodiment, in response to a fault probability being greater than or equal to a fault probability threshold, fault handling is performed, including: determining the target probability range to which the fault probability belongs; selecting a fault handling operation corresponding to the target probability range from a preset mapping relationship as a target fault handling operation; wherein the preset mapping relationship includes several preset probability ranges and fault handling operations corresponding to each preset probability range, and the target probability range is one of the several preset probability ranges; and executing the target fault handling operation.

[0026] To solve the above-mentioned technical problems, another technical solution adopted in this application is: providing a state detection device, the device comprising: a modal component decomposition module for decomposing a first electrical signal of an electric drive system into a plurality of first modal components; a selection module for selecting at least one noise modal component from the plurality of first modal components; a filtering module for filtering the at least one noise modal component to obtain at least one second modal component; a signal reconstruction module for reconstructing a signal using the at least one second modal component and each unfiltered modal component among the plurality of first modal components to obtain a second electrical signal; and a detection module for determining the state detection result of the electric drive system using at least the second electrical signal.

[0027] To solve the above-mentioned technical problems, another technical solution adopted in this application is: to provide an electronic device, including a memory and a processor coupled to each other, wherein the memory stores program instructions; and the processor is used to execute the program instructions stored in the memory to implement the above-mentioned state detection method.

[0028] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a computer-readable storage medium for storing program instructions, which, when executed by a processor, implement the above-mentioned state detection method.

[0029] To solve the above-mentioned technical problems, another technical solution adopted in this application is to provide a computer program product, which includes a computer program that implements the above-mentioned state detection method when executed by a processor.

[0030] The above scheme, in the process of electric drive system status detection, first decomposes the first electrical signal of the electric drive system into several first mode components, and filters at least one noise mode component among the several first mode components to obtain at least one second mode component; then, it uses at least one second mode component and each of the unfiltered mode components among the several first mode components to reconstruct the signal to obtain a second electrical signal; finally, it uses at least the second electrical signal to determine the status detection result of the electric drive system. When filtering the first electrical signal, it selectively filters the noise mode components of the first electrical signal, while not filtering the non-noise mode components. This reduces the false deletion of useful information in the non-noise mode components, reduces the distortion of the second electrical signal, and makes the second electrical signal obtained after signal reconstruction more effective. Therefore, the status detection result of the electric drive system determined using the second electrical signal is more accurate, thus improving the accuracy of electric drive status detection. Attached Figure Description

[0031] Figure 1 This is a flowchart illustrating an embodiment of the state detection method provided in this application; Figure 2 This is a flowchart illustrating another embodiment of the state detection method provided in this application; Figure 3 This is a schematic diagram of the network structure of the prediction model provided in this application; Figure 4 This is a flowchart illustrating one implementation of the method for determining the fault probability threshold provided in this application. Figure 5 This is a flowchart illustrating one embodiment of the fault handling method provided in this application; Figure 6 This is a schematic diagram of the framework of an embodiment of the state detection device provided in this application; Figure 7 This is a schematic diagram of the framework of an embodiment of the electronic device provided in this application; Figure 8 This is a schematic diagram of a framework of an embodiment of the computer-readable storage medium provided in this application. Detailed Implementation

[0032] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0033] It should be noted that the term "and / or" in this application is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The terms "first" and "second" in this application 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. Therefore, features defined as "first" or "second" may explicitly or implicitly include at least one of those features. The term "multiple" in this application means at least two, such as two, three, etc. The term "multiple" in this application means at least two. The term "several" in this application means at least two. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0034] Electric drive systems are widely used in many fields such as industry, transportation, home appliances, and new energy. For example, electric drive systems are used as the core power system for transportation tools (such as vehicles, ships, and aircraft), CNC machine tools, and robots.

[0035] In one embodiment, the electric drive system is a vehicle-mounted electric drive system. The electric drive system includes key components such as a battery, bus capacitor, inverter, motor, and motor controller. The battery is connected to the bus capacitor, which provides the DC bus voltage. The inverter's input terminal is connected to the bus capacitor, and its output terminal is connected to the motor. The control terminal of the switching transistor in the inverter is connected to the drive signal input terminal of the motor controller. The inverter converts the DC bus voltage into AC voltage to drive the motor. It is understood that abnormalities in the key components of the electric drive system or abnormal external environmental conditions may cause the electric drive system to malfunction.

[0036] The state detection method in this application can be executed by the motor controller in the electric drive system, or by other controllers in the vehicle (such as the vehicle controller), or by a cloud server. The cloud server can interact with the motor controller to obtain data related to the state of the electric drive system, thereby performing state detection on the electric drive system.

[0037] Please see Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the state detection method provided in this application, which is used to perform state detection on an electric drive system. Figure 1 As shown, the method includes the following steps: S11: Decompose the first electrical signal of the electric drive system into several first mode components.

[0038] In some implementations, the first electrical signal can be one of the following: a three-phase current signal of the motor, a three-phase voltage signal of the motor, an output current signal of the inverter, an output voltage signal of the inverter, a battery charging current signal, a battery discharging current signal, etc. For example, the first electrical signal is a three-phase current signal of the motor.

[0039] In some implementations, the first electrical signal can be decomposed into several first mode components using methods such as Empirical Wavelet Transform (EWT) and Variational Mode Decomposition (VMD). For details on Empirical Wavelet Transform and Variational Mode Decomposition, please refer to relevant technical documents; they will not be elaborated upon here.

[0040] S12: Select at least one noise mode component from several first mode components.

[0041] The plurality of first mode components may include at least one noise mode component and a non-noise mode component. At least one noise mode component is selected from the plurality of first mode components for filtering, while the non-noise mode components in the plurality of first mode components are not filtered.

[0042] S13: Filter at least one noise mode component to obtain at least one second mode component.

[0043] At least one noise mode component is filtered to remove impulse noise from each noise mode component, thereby obtaining at least one second mode component.

[0044] S14: Reconstruct the signal using at least one second mode component and each unfiltered mode component from several first mode components to obtain a second electrical signal.

[0045] The second electrical signal is obtained by superimposing and reconstructing the signals of at least one second mode component obtained after filtering and the unfiltered mode components retained in several first mode components.

[0046] S15: At least using the second electrical signal, determine the status detection result of the electric drive system.

[0047] In one embodiment, the status detection result of the electric drive system may include at least one of the following: electric drive system status type, electric drive system failure probability, electric drive system failure status level, electric drive system health status score, and electric drive system health status level.

[0048] The electric drive system is classified into two states: fault state and normal state.

[0049] The failure probability of an electric drive system reflects the likelihood of a failure, and can be used to further determine the state type of the electric drive system. For example, when the failure probability of the electric drive system is greater than or equal to a failure probability threshold, the electric drive system is determined to be in a fault state; when the failure probability of the electric drive system is less than the failure probability threshold, the electric drive system is determined to be in a normal state.

[0050] The fault status level of an electric drive system is used to reflect the severity of a fault in the system. For example, the fault status level of an electric drive system may be one of three fault levels: Level 1, Level 2, and Level 3, with the severity of the fault increasing sequentially from Level 1 to Level 3.

[0051] Both the health status score and the health status level of an electric drive system can be used to reflect the health level of the system. For example, the health status level of an electric drive system can be one of three levels: Level 1, Level 2, and Level 3, with the health level increasing sequentially from Level 1 to Level 3.

[0052] In one embodiment, the state detection result of the electric drive system can be predicted by a pre-trained prediction model. For example, the second electrical signal or the harmonic distortion rate of the extracted second electrical signal can be input into the prediction model to obtain the state detection result output by the prediction model. Alternatively, other data related to the state detection of the electric drive system can also be input into the prediction model to obtain the state detection result output by the prediction model.

[0053] In this embodiment, during the electric drive system status detection process, the first electrical signal of the electric drive system is first decomposed into several first mode components, and at least one noise mode component among the several first mode components is filtered to obtain at least one second mode component. Then, the signal is reconstructed using the at least one second mode component and each of the unfiltered mode components among the several first mode components to obtain a second electrical signal. Afterwards, the status detection result of the electric drive system is determined using at least the second electrical signal. When filtering the first electrical signal, the noise mode components of the first electrical signal are filtered selectively, while the non-noise mode components of the first electrical signal are not filtered. This reduces the false deletion of useful information in the non-noise mode components, reduces the distortion of the second electrical signal, and makes the second electrical signal obtained after signal reconstruction more effective. As a result, the status detection result of the electric drive system determined using the second electrical signal is more accurate, thus improving the accuracy of electric drive status detection.

[0054] Please see Figure 2 , Figure 2This is a flowchart illustrating another embodiment of the state detection method provided in this application, which is used to perform state detection on an electric drive system. It should be noted that if substantially the same result is obtained, the method of this application does not necessarily require further clarification. Figure 2 The illustrated process sequence is limited. For example... Figure 2 As shown, the method includes the following steps: S21: Decompose the first electrical signal of the electric drive system into several first mode components.

[0055] In one embodiment, the first electrical signal is one of several first signal segments, which can be obtained by performing frame-segmentation and windowing processing on the original electrical signal of the electric drive system.

[0056] In this embodiment, the step of obtaining several first signal segments includes: dividing the original electrical signal of the electric drive system into several second signal segments; multiplying each second signal segment by a preset window function to obtain several first signal segments.

[0057] In this context, a second signal segment can be considered as a frame of the original electrical signal, and there may be partial signal overlap between adjacent second signal segments.

[0058] For example, a continuous original electrical signal can be divided into several second signal segments by using a fixed frame length and frame shift. The frame length can be an integer multiple of the period of the original electrical signal. It is understood that when the frame shift is less than the frame length, there is partial signal overlap between adjacent second signal segments.

[0059] For example, the preset window function can be a Hanning window, a Hamming window, or other window functions.

[0060] In this embodiment, by performing frame-by-frame windowing processing on the original electrical signal of the electric drive system, the unstable original electrical signal of the electric drive system can be converted into locally stable signal segments. Furthermore, by using a preset window function to window each second signal segment, the spectral leakage caused by signal truncation can be effectively reduced, thereby providing a first electrical signal with higher effectiveness for subsequent data processing.

[0061] In one embodiment, the first electrical signal is decomposed into several first mode components by performing an empirical wavelet transform on the first electrical signal. Each first mode component represents a component of a specific frequency band in the first electrical signal. In this embodiment, the first electrical signal is specifically an electrical signal of a motor, such as the three-phase current signal of a motor. Furthermore, the bandwidth parameter of each first mode component is proportional to the fundamental frequency of the first electrical signal. Here, the bandwidth parameter refers to the frequency band width.

[0062] For example, the bandwidth parameter of each first mode component can be expressed by the following formula:

[0063] in, The bandwidth parameter represents each first mode component; This represents the fundamental frequency of the first electrical signal.

[0064] The fundamental frequency of the first electrical signal It can be expressed by the following formula:

[0065] in, Indicates the number of pole pairs of the motor. This indicates the motor speed.

[0066] In this embodiment, since the fundamental frequency of the first electrical signal is related to the motor speed and changes with the motor speed, the frequency band parameters of each first mode component can be adaptively adjusted by the fundamental frequency of the first electrical signal, thereby improving the adaptability of the mode component decomposition of the complex first electrical signal.

[0067] S22: Determine the pulse characterization value of each first mode component.

[0068] The pulse characterization value of the first mode component is used to characterize the intensity of the pulse signal in the first mode component.

[0069] S23: Select at least one modal component whose pulse characterization value meets the preset requirements from a plurality of first modal components, and use it as at least one noise modal component.

[0070] The default requirement is that the pulse characterization value is greater than a characterization value threshold. That is, the first mode component with a pulse characterization value greater than the characterization value threshold is identified as a noise mode component, and the first mode component with a pulse characterization value less than or equal to the characterization value threshold is identified as a non-noise mode component. The characterization value threshold can be set according to actual needs.

[0071] In one embodiment, the pulse characterization value in steps S22 and S23 is the spectral kurtosis. The spectral kurtosis of the first mode component can be determined by the following formula:

[0072] in, For the k-th first mode component among the several first mode components obtained by decomposition, This represents the spectral kurtosis of the k-th first mode component. This represents the L2 norm.

[0073] The preset requirement in this implementation includes that the spectral kurtosis of the modal components is greater than the spectral kurtosis threshold. The spectral kurtosis threshold corresponds to the aforementioned characterization value threshold. For example, the spectral kurtosis threshold ranges from [3.5, 5.0]. For instance, the spectral kurtosis threshold is 4.2.

[0074] It should be noted that in other embodiments, other parameters can also be used as pulse characterization values ​​to measure the pulse signal strength, and this embodiment does not specifically limit this.

[0075] S24: Perform wavelet transform on each noise mode component to obtain several first wavelet coefficients corresponding to each noise mode component.

[0076] For details on wavelet transform, please refer to the relevant technical documentation; further explanation will not be provided here.

[0077] S25: For each noise mode component, perform wavelet threshold filtering on several first wavelet coefficients of the noise mode component to obtain several second wavelet coefficients corresponding to the noise mode component.

[0078] In step S25, wavelet threshold filtering is performed on each of the first wavelet coefficients of the noise mode components, including the following steps: Step 1: Determine whether the first wavelet coefficients meet the filtering conditions.

[0079] The filtering conditions include: the absolute value of the first wavelet coefficient is greater than or equal to the filtering threshold.

[0080] In one embodiment, to reduce the false deletion of effective components and lower the signal distortion rate, the filtering threshold can be set to a dynamically changing value. In this embodiment, the filtering threshold can be determined by the following formula:

[0081] in, Indicates the filter threshold. This represents the number of the first wavelet coefficients of the noise modal components. Indicates the standard deviation of noise. With N and It changes dynamically with the changes.

[0082] In step one, if the first wavelet coefficients meet the filtering condition (i.e., the absolute value of the first wavelet coefficients is greater than or equal to the filtering threshold), then proceed to step two. If the first wavelet coefficients do not meet the filtering condition (i.e., the absolute value of the first wavelet coefficients is less than the filtering threshold), then proceed to step three.

[0083] Step 2: Determine the difference between the first wavelet coefficient and the coefficient adjustment value, and use it as the second wavelet coefficient corresponding to the first wavelet coefficient.

[0084] The coefficient adjustment value is determined using the first wavelet coefficient and the filtering threshold.

[0085] In one embodiment, determining the coefficient adjustment value includes the following sub-steps: Sub-step one: Use the sign function to determine the sign value of the first wavelet coefficient.

[0086] When the sign of the first wavelet coefficient is positive, the sign value of the first wavelet coefficient is 1. When the sign of the first wavelet coefficient is negative, the sign value of the first wavelet coefficient is -1.

[0087] Sub-step two: Determine the first product of the filtering threshold and the sign value.

[0088] Sub-step three: Determine the second product of the first preset value and the first wavelet coefficient.

[0089] The first preset value is negative. For example, the first preset value is -1.

[0090] Sub-step four involves exponentiation using the second preset value as the base and the second product as the exponent to obtain the first exponentiation value.

[0091] For example, the second preset value is e.

[0092] Sub-step five: Summing the third preset value and the first power operation value to obtain the first sum value.

[0093] For example, the third preset value is 1.

[0094] Sub-step six: Determine the first ratio of the first product to the first sum, as the coefficient adjustment value.

[0095] For example, the second wavelet coefficients obtained after adjusting the first wavelet coefficients can be expressed by the following formula:

[0096] in, represents the first wavelet coefficient, j represents the scale, and k represents the position; This represents the second wavelet coefficient corresponding to the first wavelet coefficient; This indicates the aforementioned coefficient adjustment value.

[0097] The above formula ( In the case of (the situation), In the form of the Sigmoid function, when Much larger than the dynamic threshold hour, Approximately 1, coefficient adjustment value Approximate dynamic threshold At this point, the contraction is close to the soft threshold, which can better preserve effective high-frequency components (such as harmonic distortion characteristics). When Approaching dynamic threshold At that time, the coefficient adjustment value Less than via the Sigmoid function form accomplish Smooth adjustment reduces signal step distortion.

[0098] Step 3: If the first wavelet coefficient does not meet the filtering conditions, determine the product of the first wavelet coefficient and the attenuation coefficient as the second wavelet coefficient corresponding to the first wavelet coefficient.

[0099] The attenuation coefficient is determined using the first wavelet coefficient and the filtering threshold.

[0100] In one embodiment, determining the attenuation coefficient includes the following sub-steps: Sub-step one: sum the fourth preset value and the square of the first wavelet coefficient to obtain the second sum value.

[0101] The fourth preset value is a small value. For example, the fourth preset value is... .

[0102] Sub-step two: Determine the second ratio of the square of the filter threshold to the second sum.

[0103] Sub-step three: Determine the third product of the second ratio and the fifth preset value.

[0104] For example, the fifth preset value is -1.

[0105] Sub-step four involves exponentiation using the sixth preset value as the base and the third product as the exponent to obtain the second exponentiation value.

[0106] For example, the sixth preset value is e.

[0107] Sub-step five: Use the value of the second power operation as the attenuation coefficient.

[0108] For example, the second wavelet coefficients obtained after adjusting the first wavelet coefficients can be expressed by the following formula:

[0109] in, Indicates the attenuation coefficient; The aforementioned fourth preset value is used to prevent numerical overflow caused by a denominator of zero.

[0110] The above formula ( In the case of (the situation where), the first wavelet coefficient is affected by the exponential decay coefficient. Rapid decay can effectively suppress the first wavelet coefficients. Random noise.

[0111] Steps one through three are divided into and Two scenarios, each using different methods to remove the first wavelet coefficients. It can effectively suppress noise and improve the signal-to-noise ratio.

[0112] S26: For each noise mode component, perform inverse wavelet transform on several second wavelet coefficients of the noise mode component to obtain the second mode component corresponding to the noise mode component.

[0113] For details on inverse wavelet transform, please refer to the relevant technical documentation; further explanation will not be provided here.

[0114] S27: Reconstruct the signal using at least one second mode component and each unfiltered mode component from several first mode components to obtain a second electrical signal.

[0115] S28: At least the second electrical signal is used to determine the status detection result of the electric drive system.

[0116] In step S28, the second electrical signal is first processed to obtain its harmonic distortion rate. Then, using at least the harmonic distortion rate of the second electrical signal, the state detection result of the electric drive system is predicted. The harmonic distortion rate of the second electrical signal is used to measure the degree to which the waveform of the second electrical signal deviates from a sine wave.

[0117] When critical components in an electric drive system malfunction, the waveform of the system's electrical signals often changes, introducing additional harmonic components and thus increasing the harmonic distortion rate of the electrical signals. For example, when a motor malfunctions (such as a stator inter-turn short circuit), it causes spikes in the motor's three-phase current signals, introducing additional harmonic components. Therefore, the harmonic distortion rate of the electric drive system's electrical signals can effectively reflect whether a fault exists in the system, and thus, the state of the electric drive system can be predicted by obtaining its harmonic distortion rate.

[0118] Since the second electrical signal is obtained by filtering the first electrical signal of the electric drive system (i.e. filtering noise), obtaining the harmonic distortion rate of the second electrical signal can reduce the interference of noise on the harmonic distortion rate, so that the harmonic distortion rate of the second electrical signal can more realistically reflect the state of the electric drive system, thereby improving the accuracy of the electric drive system state detection.

[0119] In one embodiment, the step of obtaining the harmonic distortion rate of the second electrical signal includes: performing a fast Fourier transform on the second electrical signal to obtain its frequency domain characteristics; and extracting the harmonic distortion rate of the second electrical signal using its frequency domain characteristics. Specifically, the fundamental component and all harmonic components are extracted from the frequency domain characteristics of the second electrical signal, the effective value of the fundamental component and the effective value of each harmonic component are determined, and the ratio of the sum of the effective values ​​of all harmonic components to the effective value of the fundamental component is taken as the harmonic distortion rate of the second electrical signal.

[0120] In one embodiment, the time-series sampling data of harmonic distortion rate is input into a pre-trained prediction model. The prediction model uses the time-series sampling data of harmonic distortion rate to make predictions and outputs the state detection results of the electric drive system.

[0121] In another embodiment, the time-series sampling data of harmonic distortion rate and other data related to the state detection of the electric drive system can be input into a pre-trained prediction model to obtain the state detection results of the electric drive system. In this embodiment, the time-series sampling data of harmonic distortion rate and other data related to the state detection of the electric drive system can be further divided into dynamic time-series features and static features. Under normal operation of the electric drive system, the dynamic time-series features change with time, while the static features remain essentially unchanged. Then, the pre-trained prediction model is used to predict the state detection results of the electric drive system based on the dynamic time-series features and static features. Obtaining the state detection results of the electric drive system includes the following steps: Step 1: Obtain dynamic time-series features and static features.

[0122] In step one, the dynamic time series characteristics include time series sampling data of at least one first influencing factor, and the at least one first influencing factor includes the harmonic distortion rate of the aforementioned second electrical signal.

[0123] In one example, in addition to the harmonic distortion rate of the second electrical signal, at least one first influencing factor may also include the temperature of several temperature sampling points in the battery, the average temperature of the battery, the voltage of each cell in the battery, the vehicle load rate of the motor, the ambient dust concentration, the status indicators of several monitoring parameters, the three-phase current signal of the motor, the three-phase voltage signal of the motor, the output current signal of the inverter, the output voltage signal of the inverter, etc.

[0124] The temperature sampling points can be any temperature-sensitive area within the battery, or they can correspond one-to-one with any individual cell within the battery, meaning one temperature sampling point is located on each individual cell. The temperature of these sampling points can be obtained through a distributed temperature sensor network. The average battery temperature is the average of the temperatures of all individual cells.

[0125] Vehicle load factor reflects the load status of a vehicle. The vehicle load factor can be determined using the following formula:

[0126] The concentration of ambient dust can be obtained using a laser scattering particulate matter sensor, which is installed near the air filter inlet.

[0127] Several monitoring parameters may include the operating status of the code embedded in the motor controller and parameters detected by the motor controller through sensors (such as motor speed, motor torque, etc.). The status identifier of the monitoring parameter is used to indicate whether the monitoring parameter is normal or abnormal. For example, when the monitoring parameter is normal, the corresponding status identifier can be 01; when the monitoring parameter is abnormal, the corresponding status identifier can be 02. The motor controller's storage log stores the status identifiers corresponding to each of the several monitoring parameters. For example, when the status detection method is executed by the motor controller, the status identifiers of several monitoring parameters can be directly read from the motor controller's storage log; when the status detection method is executed by other controllers in the vehicle, data interaction with the motor controller can be performed via the vehicle's CAN bus to obtain the status identifiers of several monitoring parameters.

[0128] In one example, each of the first influencing factors can be sampled using a suitable sampling frequency. For instance, for slowly changing first influencing factors, such as ambient dust concentration and the average battery temperature during normal operation, low-frequency sampling can be used. For moderately changing first influencing factors, such as the three-phase current signal of a motor, medium-frequency sampling can be used; and for drastically changing first influencing factors, such as the harmonic distortion rate of a second electrical signal, high-frequency sampling can be used. For example, the low-frequency, medium-frequency, and high-frequency sampling frequencies can be set to 1Hz, 100Hz, and 10kHz, respectively.

[0129] In one example, at least one primary influencing factor includes the temperature at several temperature sampling points within the battery. The data sampling process may also include the following sub-steps: Sub-step one: Obtain the temperature sampling value of each temperature sampling point at the first sampling frequency.

[0130] Sub-step two: Determine whether the sampling frequency switching conditions are met.

[0131] If the sampling frequency switching condition is met, proceed to sub-step three. If the sampling frequency switching condition is not met, proceed to sub-step one.

[0132] Sub-step three: Obtain the temperature values ​​of each temperature sampling point at the second sampling frequency.

[0133] In one scenario, the first sampling frequency is less than the second sampling frequency. For example, the first sampling frequency is the aforementioned intermediate frequency sampling frequency, and the second sampling frequency is the aforementioned high frequency sampling frequency. The sampling frequency switching condition includes at least one of the following conditions: there exists at least one set of temperature sampling points whose temperature difference at the same sampling time is greater than or equal to a first temperature threshold; or there exists at least one temperature sampling point whose temperature difference between two adjacent sampling times is greater than or equal to a second temperature threshold. Each set of temperature sampling points includes two temperature sampling points.

[0134] When the sampling frequency switching condition is met under this circumstance, it indicates that the battery has a local overheating problem. This local overheating may be caused by a micro-short circuit in the battery. By switching the first sampling frequency to the second sampling frequency, i.e. increasing the sampling frequency, it is easier to capture more subtle micro-short circuit faults in the battery and reduce the risk of thermal runaway in the battery.

[0135] In another scenario, the first sampling frequency is greater than the second sampling frequency. For example, the first sampling frequency is the aforementioned high-frequency sampling frequency, and the second sampling frequency is the aforementioned intermediate-frequency sampling frequency. The sampling frequency switching condition includes at least one of the following: the temperature difference between any two temperature sampling points at the same sampling time is less than a third temperature threshold; and the temperature difference between any two adjacent sampling times is less than a fourth temperature threshold. The first temperature threshold and the second temperature threshold may be the same or different. For example, both the first temperature threshold and the second temperature threshold are 5°C.

[0136] When the sampling frequency switching condition is met, it means that the battery does not have a local overheating problem. By switching the first sampling frequency to the second sampling frequency, i.e. reducing the sampling frequency, power consumption can be reduced.

[0137] The first temperature threshold, the second temperature threshold, the third temperature threshold, and the fourth temperature threshold may be the same or different. For example, the first temperature threshold, the second temperature threshold, the third temperature threshold, and the fourth temperature threshold are all 5°C.

[0138] In one example, considering that the sampling frequencies of each first influence factor may be different, the dynamic time series features are time-aligned before proceeding to step two. Time-aligning the dynamic time series features can be understood as aligning the sampling data of all first influence factors acquired at different sampling frequencies on the time axis. The time-aligned dynamic time series features include the sampling data of each first influence factor at several times; that is, each time step includes the sampling data of all first influence factors at that time.

[0139] By performing temporal alignment processing on dynamic temporal features before inputting them into a temporal coding network, the temporal coding network can take into account the sampling data of all first influencing factors at the same time during feature extraction, thereby improving the feature extraction effect.

[0140] In step one, the static features include sampling data of several second influencing factors. For example, the several second influencing factors may include battery DC internal resistance, battery self-discharge rate, insulation resistance value, coolant conductivity in the electric drive cooling system, and bus capacitor capacitance, etc.

[0141] Step 2: Use the temporal coding network of the prediction model to extract features from the dynamic temporal features to obtain the first feature.

[0142] In some examples, the temporal coding network can be an LSTM (Long Short-Term Memory) network, a GRU (Gated Recurrent Unit), a Transformer, or other coding networks that can handle temporal features.

[0143] Step 3: Use the deep residual network of the prediction model to extract features from the static features to obtain the second feature.

[0144] In one example, the deep residual network consists of 3 residual blocks, each of which contains 2 convolutional layers.

[0145] Step four: Fuse the first feature and the second feature to obtain the fused feature.

[0146] In one example, the first feature and the second feature are concatenated, and the concatenated feature is used as the fused feature.

[0147] In another example, fusion features are obtained through an attention mechanism. This example involves the following sub-steps: Sub-step one: Determine the attention weights corresponding to the first feature and the second feature, respectively.

[0148] The attention weights of the first feature and the second feature represent the importance of the first feature and the second feature, respectively.

[0149] The concatenated first and second features are input into the attention network of the prediction model to obtain the attention weights corresponding to the first and second features, respectively.

[0150] Sub-step two involves weighting the first and second features using the attention weights corresponding to the first and second features respectively, to obtain the attention features.

[0151] Sub-step three involves using attention features as fusion features.

[0152] In sub-steps one through three, attention weights corresponding to the first and second features are dynamically allocated through an attention mechanism, and the first and second features are weighted to obtain a deeper feature representation, which is beneficial to improving the accuracy of subsequent motor system state detection.

[0153] Step 5: Use the prediction network to make predictions based on the fused features to obtain the state detection results.

[0154] In steps one through five, because time-series coding networks excel at extracting long-term dependencies and temporal evolution patterns from time-series data, and deep residual networks effectively extract deep spatial features from static features, the features input to the prediction model are divided into dynamic temporal features and static features. The first feature is extracted from the dynamic temporal features using the time-series coding network, and the second feature is extracted from the static features using the deep residual network. This makes the extracted first and second features more accurate, which is beneficial for improving the accuracy of subsequent motor system status detection. Furthermore, compared to using single data (such as motor current or battery temperature) to detect motor system status, fusing multiple data (such as harmonic distortion rate, average battery temperature, vehicle load rate, and environmental dust concentration) allows for the analysis of coupling effects between various data, improving the accuracy of early fault detection in the electric drive system.

[0155] Figure 3 This is a schematic diagram of the network structure of the prediction model provided in this application. For example... Figure 3 As shown, the prediction model includes a temporal coding network, a deep residual network, an attention network, and a prediction network. The temporal coding network extracts features from dynamic temporal features to obtain the first feature; the deep residual network extracts features from static features to obtain the second feature; the attention network performs attention processing on the first and second features to obtain a fused feature; and the prediction network makes predictions based on the fused feature and outputs the state detection results of the electric drive system.

[0156] In some examples, the prediction network can be implemented using long short-term memory networks, deep residual networks, fully connected layers, etc.

[0157] In this embodiment, the status detection result of the electric drive system includes the failure probability of the electric drive system. After obtaining the status detection result of the electric drive system, the following steps can be performed: using the current vehicle operating data, a failure probability threshold is determined; when the failure probability of the electric drive system is greater than or equal to the failure probability threshold, it is determined that the electric drive system has a failure, and failure handling is performed. The current vehicle operating data includes at least one of the current vehicle load rate and the current battery cycle count. The current vehicle load rate reflects the current load condition of the vehicle, and the current battery cycle count reflects the current degree of battery aging.

[0158] Figure 4 This is a flowchart illustrating one embodiment of the method for determining the fault probability threshold provided in this application. In this embodiment, the current vehicle operating data includes the current vehicle load rate and the current battery cycle count. Figure 4 As shown, the method includes the following steps: S401: Determine the first adjustment factor using the current vehicle load rate.

[0159] The first adjustment factor is positively correlated with the current vehicle load rate.

[0160] In one example, the fourth product of the first adjustment factor and the current vehicle load rate is determined; the first adjustment coefficient is obtained by summing the preset constant and the fourth product.

[0161] For example, the default constant is 1.

[0162] S402: Determine the second adjustment factor using the current number of battery cycles.

[0163] The second adjustment factor is positively correlated with the current number of battery cycles.

[0164] In one example, the second adjustment coefficient is obtained by exponentiation of the second adjustment factor as the base and the current battery cycle count as the exponent.

[0165] S403: Adjust the base adjustment value using the first adjustment coefficient and the second adjustment coefficient to obtain the fault probability threshold.

[0166] In one example, the product of the first adjustment factor, the second adjustment factor, and the base adjustment value is used as the failure probability threshold.

[0167] For example, the failure probability threshold can be determined by the following formula:

[0168] in, Indicates the failure probability threshold. Indicates the base adjustment value. Indicates the first adjustment factor. Indicates the second adjustment factor. This represents the first adjustment factor. This represents the second adjustment factor, where L represents the current vehicle load rate. This indicates the current battery cycle count.

[0169] In the above formula, the base adjustment value α is the fault probability threshold. The baseline value determines the failure probability threshold. The starting point. Because the current vehicle load rate L is 0 and the current battery cycle count... When it is 0, the failure probability threshold The base adjustment value is α.

[0170] First adjustment factor Used to adjust the current vehicle load rate L against the fault probability threshold. The degree of impact. When the current vehicle load rate L increases, although the risk of electric drive system failure increases, the fluctuations in the electric drive system's own data (such as temperature and current) also increase. These fluctuations may be normal data fluctuations. If the failure probability threshold based on a low load rate is still used, it is easy to misjudge normal data fluctuations as failures. Therefore, the first adjustment factor... Used to set the failure probability threshold An appropriate increase should be made to balance the sensitivity and stability of fault diagnosis.

[0171] The second adjustment factor r is greater than 1; for example, the second adjustment factor r is a number greater than 1 and close to 1. The second adjustment factor is used to make the failure probability threshold... The probability increases exponentially with the current battery cycle count. On the one hand, this reduces the likelihood of misdiagnosing normal degradation caused by battery aging as a fault in the electric drive system. On the other hand, it increases the fault probability threshold. It can detect serious malfunctions caused by battery aging.

[0172] Base adjustment value α, first adjustment factor The specific value of the second adjustment factor r can be set according to actual needs. For example, the base adjustment value α is 0.15; the first adjustment factor... The first adjustment factor is 0.7; the second adjustment factor r is 1.02.

[0173] In steps S401 to S403, the fault probability threshold is dynamically adjusted in real time, taking into account the vehicle load and battery aging conditions. This can improve the accuracy and reliability of fault determination in the electric drive system and reduce the false alarm rate of the electric drive system.

[0174] Figure 5 This is a flowchart illustrating one embodiment of the fault handling method provided in this application. This method is performed when the fault probability of the electric drive system is determined to be greater than or equal to a fault probability threshold, i.e., when a fault is determined to exist in the electric drive system. Figure 5 As shown, the troubleshooting method includes the following steps: S501: Determine the target probability range to which the failure probability of the electric drive system belongs.

[0175] The target probability range is one of several preset probability ranges.

[0176] S502: Select the fault handling operation corresponding to the target probability range from the preset mapping relationship and use it as the target fault handling operation.

[0177] The preset mapping relationship includes several preset probability ranges and the corresponding fault handling operations for each preset probability range. Different preset probability ranges correspond to different fault state levels, and the corresponding fault handling operations are different.

[0178] In one example, several preset probability ranges include a first probability range, a second probability range, and a third probability range. The first probability range is [0.6, 0.8), the second probability range is [0.8, 0.95), and the third probability range is [0.95, 1]. The first probability range, the second probability range, and the third probability range correspond to the first fault level, the second fault level, and the third fault level, respectively.

[0179] The fault handling operations corresponding to the first fault level may include at least one of sending a fault warning message to a first terminal and reducing the vehicle speed to a preset speed. The first terminal may be an in-vehicle terminal, a user's mobile phone, computer, tablet, or smart wearable device, etc. The fault warning message is used to indicate that there is a first-level fault in the electric drive system, and the fault warning message may include the fault code of the electric drive system. The preset speed can be set to 20km / h.

[0180] The fault handling operations corresponding to the second fault level may include at least one of switching to a safety mode and sending vehicle location information to a second terminal. The safety mode may include switching to a backup electric drive system or limiting vehicle output power. The second terminal may be a remote monitoring platform.

[0181] The fault handling operations corresponding to the third fault level may include at least one of emergency shutdown and sending a rescue message.

[0182] S503: Perform the target fault handling operation.

[0183] In this embodiment, when a fault is determined in the electric drive system, the time-series sampling data of at least one first influencing factor and the sampling data of several second influencing factors can be analyzed to determine which influencing factor data anomalies caused the electric drive system fault. For example, if a fault is detected in the electric drive system and only the ambient dust concentration is within an abnormal concentration range, the electric drive system fault can be considered to be caused by the ambient dust concentration. As another example, if a fault is detected in the electric drive system and only the average battery temperature is within an abnormal temperature range, the electric drive system fault can be considered to be caused by the average battery temperature.

[0184] Please see Figure 6 , Figure 6 This is a schematic diagram of a framework of an embodiment of the state detection device provided in this application. In this embodiment, the state detection device 600 includes a modal component decomposition module 601, a selection module 602, a filtering module 603, a signal reconstruction module 604, and a detection module 605.

[0185] The modal component decomposition module 601 is used to decompose the first electrical signal of the electric drive system into several first modal components; the selection module 602 is used to select at least one noise modal component from the several first modal components; the filtering module 603 is used to filter the at least one noise modal component to obtain at least one second modal component; the signal reconstruction module 604 is used to reconstruct the signal using at least one second modal component and each unfiltered modal component from the several first modal components to obtain a second electrical signal; and the detection module 605 is used to determine the state detection result of the electric drive system using at least the second electrical signal.

[0186] In one embodiment, the selection module 602 is used to determine the pulse characterization value of each first mode component, the pulse characterization value being used to characterize the intensity of the pulse signal in the mode component; and to select at least one mode component whose pulse characterization value meets the preset requirements from a plurality of first mode components as at least one noise mode component.

[0187] In one embodiment, the pulse characterization value is spectral kurtosis, and the preset requirement includes: the spectral kurtosis of the modal component is greater than the spectral kurtosis threshold.

[0188] In one embodiment, the filtering module 603 is used to perform wavelet transform on each noise mode component to obtain a plurality of first wavelet coefficients corresponding to each noise mode component; for each noise mode component, wavelet threshold filtering is performed on the plurality of first wavelet coefficients of the noise mode component to obtain a plurality of second wavelet coefficients corresponding to the noise mode component; for each noise mode component, inverse wavelet transform is performed on the plurality of second wavelet coefficients of the noise mode component to obtain the second mode component corresponding to the noise mode component.

[0189] In one embodiment, for each first wavelet coefficient of the noise mode component, the filtering module 603 is used to determine whether the first wavelet coefficient meets the filtering conditions. The filtering conditions include: the absolute value of the first wavelet coefficient is greater than or equal to the filtering threshold. If the first wavelet coefficient meets the filtering conditions, the difference between the first wavelet coefficient and the coefficient adjustment value is determined as the second wavelet coefficient corresponding to the first wavelet coefficient. If the first wavelet coefficient does not meet the filtering conditions, the product of the first wavelet coefficient and the attenuation coefficient is determined as the second wavelet coefficient corresponding to the first wavelet coefficient. The coefficient adjustment value and the attenuation coefficient are both determined using the first wavelet coefficient and the filtering threshold.

[0190] In one embodiment, the filtering module 603 is used to determine the sign value of the first wavelet coefficient using a sign function; determine the first product of the filtering threshold and the sign value; determine the second product of the first preset value and the first wavelet coefficient, wherein the first preset value is negative; perform a power operation with the second preset value as the base and the second product as the exponent to obtain a first power operation value; sum the third preset value and the first power operation value to obtain a first sum value; and determine a first ratio of the first product to the first sum value as a coefficient adjustment value.

[0191] In one embodiment, the filtering module 603 is used to sum the fourth preset value and the square of the first wavelet coefficient to obtain a second sum; determine the second ratio of the square of the filtering threshold to the second sum; determine the third product of the second ratio and the fifth preset value; perform a power operation with the sixth preset value as the base and the third product as the exponent to obtain a second power operation value; and use the second power operation value as the attenuation coefficient.

[0192] In one embodiment, the first electrical signal is the three-phase current signal of the motor.

[0193] And / or, the first electrical signal is one of several first signal segments, and the state detection device 600 further includes a signal acquisition module 606. The signal acquisition module 606 is used to divide the original electrical signal of the electric drive system into several second signal segments; and to multiply each second signal segment by a preset window function to obtain several first signal segments.

[0194] And / or, the modal component decomposition module 601 is used to perform empirical wavelet transform on the first electrical signal to obtain several first modal components; wherein, the bandwidth parameter of each first modal component is proportional to the fundamental frequency of the first electrical signal.

[0195] In one embodiment, the detection module 605 is used to perform a fast Fourier transform on the second electrical signal to obtain the frequency domain characteristics of the second electrical signal; extract the harmonic distortion rate of the second electrical signal using the frequency domain characteristics; and predict the state detection result of the electric drive system using at least the harmonic distortion rate of the second electrical signal.

[0196] In one embodiment, the detection module 605 is used to acquire dynamic temporal features and static features; wherein, the dynamic temporal features include time-series sampling data of at least one first influencing factor, the at least one first influencing factor including the harmonic distortion rate of a second electrical signal, and the static features include sampling data of several second influencing factors; the dynamic temporal features are extracted using a time-series coding network to obtain a first feature; and the static features are extracted using a deep residual network to obtain a second feature; the first feature and the second feature are fused to obtain a fused feature; and a prediction network is used to make a prediction based on the fused feature to obtain the state detection result of the electric drive system.

[0197] In one embodiment, the detection module 605 is used to determine the attention weights corresponding to the first feature and the second feature respectively; to weight the first feature and the second feature using the attention weights corresponding to the first feature and the second feature respectively, to obtain attention features; and to use the attention features as fusion features.

[0198] And / or, before using the temporal coding network to extract features from the dynamic temporal features and obtain the first feature, the detection module 605 also performs temporal alignment processing on the dynamic temporal features.

[0199] In one embodiment, at least one first influencing factor further includes the temperature of a plurality of temperature sampling points in the battery, and the state detection device further includes a temperature sampling module 607. The temperature sampling module 607 is used to acquire the temperature sampling value of each temperature sampling point at a first sampling frequency; and in response to satisfying a sampling frequency switching condition, to acquire the temperature value of each temperature sampling point at a second sampling frequency. Wherein, the first sampling frequency is less than the second sampling frequency, and the sampling frequency switching condition includes at least one of the following conditions: there exists at least one set of temperature sampling points whose temperature difference at the same sampling time is greater than or equal to a first temperature threshold; or there exists at least one temperature sampling point whose temperature difference at two adjacent sampling times is greater than or equal to a second temperature threshold; or, the first sampling frequency is greater than the second sampling frequency, and the sampling frequency switching condition includes at least one of the following conditions: the temperature difference between any two temperature sampling points at the same sampling time is less than a third temperature threshold; or the temperature difference between each temperature sampling point at two adjacent sampling times is less than a fourth temperature threshold.

[0200] In one embodiment, the condition detection device 600 further includes a probability threshold determination module 608. The condition detection result of the electric drive system includes the failure probability of the electric drive system. After the detection module 605 predicts the condition detection result of the electric drive system using at least a second electrical signal, the probability threshold determination module 608 is used to determine a failure probability threshold using current vehicle operating data; wherein, the current vehicle operating data includes at least one of the current vehicle load rate and the current battery cycle count. The detection module 605 is also used to perform fault handling in response to a failure probability greater than or equal to the failure probability threshold.

[0201] In one embodiment, the current vehicle operating data includes the current vehicle load rate and the current battery cycle count. The probability threshold determination module 608 is used to determine a first adjustment coefficient using the current vehicle load rate and a second adjustment coefficient using the current battery cycle count; wherein the first adjustment coefficient is positively correlated with the current vehicle load rate, and the second adjustment coefficient is positively correlated with the current battery cycle count; the base adjustment value is adjusted using the first and second adjustment coefficients to obtain the fault probability threshold.

[0202] In one embodiment, the probability threshold determination module 608 is used to determine the fourth product of the first adjustment factor and the current vehicle load rate; and to sum the preset constant and the fourth product to obtain the first adjustment coefficient.

[0203] And / or, the probability threshold determination module 608 is used to perform a power operation with the second adjustment factor as the base and the current battery cycle number as the exponent to obtain the second adjustment coefficient.

[0204] And / or, the probability threshold determination module 608 is used to take the product of the first adjustment coefficient, the second adjustment coefficient and the base adjustment value as the fault probability threshold.

[0205] In one embodiment, the detection module 605 is used to determine the target probability range to which the fault probability belongs; select the fault handling operation corresponding to the target probability range from the preset mapping relationship as the target fault handling operation; wherein, the preset mapping relationship includes several preset probability ranges and fault handling operations corresponding to each preset probability range, and the target probability range is one of the several preset probability ranges; and execute the target fault handling operation.

[0206] It should be noted that the apparatus of this embodiment can perform the steps in the above method. For detailed descriptions of the relevant content, please refer to the method section above, which will not be repeated here.

[0207] Please see Figure 7 , Figure 7 This is a schematic diagram of a framework of an embodiment of the electronic device provided in this application. In this embodiment, the electronic device 70 includes a memory 71 and a processor 72.

[0208] Processor 72 can also be referred to as a CPU (Central Processing Unit). Processor 72 may be an integrated circuit chip with signal processing capabilities. Processor 72 can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor can be a microprocessor, or processor 72 can be any conventional processor 72, etc.

[0209] The memory 71 in the electronic device 70 is used to store the program instructions required for the processor 72 to run.

[0210] The processor 72 is used to execute program instructions to implement the state detection method in this application.

[0211] Please see Figure 8 , Figure 8This is a schematic diagram of a framework of an embodiment of the computer-readable storage medium provided in this application. The computer-readable storage medium 80 of this application embodiment stores program instructions 81, which, when executed, implement the state detection method provided in this application. The program instructions 81 can form a program file and be stored in the aforementioned computer-readable storage medium 80 in the form of a software product, so that a computer device (which may be a personal computer, server, or network device, etc.) can execute all or part of the steps of the methods of various embodiments of this application. The aforementioned computer-readable storage medium 80 includes various media capable of storing program code, such as a USB flash drive, mobile hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, or terminal devices such as computers, servers, mobile phones, and tablets.

[0212] This embodiment also provides a computer program product, which includes a computer program that, when executed by a processor, implements the state detection method in this application.

[0213] Understandably, a computer program product can be a computer program product contained on a tangible computer-readable medium, which includes program code for performing the state detection method in this application. In some embodiments, the computer program product can be downloaded and installed from a network, and can also be copied, transferred, and installed between different computer hardware. Its wireless transmission method can include the Internet, Bluetooth, WIFI, etc., and its wired transmission method can include USB, Lightning, Type-C, etc.

[0214] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0215] The description of the various embodiments above tends to emphasize the differences between the various embodiments. The similarities or similarities between them can be referred to, and for the sake of brevity, they will not be repeated here.

[0216] In the several embodiments provided in this application, it should be understood that the disclosed methods, apparatuses, and systems can be implemented in other ways. For example, the apparatus implementations described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of apparatuses or units may be electrical, mechanical, or other forms.

[0217] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0218] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0219] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0220] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A state detection method characterized by, The method comprises: decomposing a first electrical signal of an electric drive system into a plurality of first modal components; selecting at least one noise modal component from the plurality of first modal components; respectively filtering the at least one noise modal component to obtain at least one second modal component; reconstructing a second electrical signal by using the at least one second modal component and each modal component of the plurality of first modal components which has not been filtered; determining a state detection result of the electric drive system by using at least the second electrical signal.

2. The method of claim 1, wherein, The selecting at least one noise modal component from the plurality of first modal components comprises: respectively determining a pulse characteristic value of each first modal component, the pulse characteristic value being used to represent an intensity of a pulse signal in the modal component; selecting at least one modal component whose pulse characteristic value meets a preset requirement from the plurality of first modal components as the at least one noise modal component.

3. The method of claim 2, wherein: the pulse characteristic value is spectral kurtosis, and the preset requirement comprises that the spectral kurtosis of the modal component is greater than a spectral kurtosis threshold.

4. The method of claim 1, wherein, The filtering the at least one noise modal component to obtain at least one second modal component comprises: respectively performing wavelet transform on each noise modal component to obtain a plurality of first wavelet coefficients corresponding to the noise modal component; performing wavelet threshold filtering on the plurality of first wavelet coefficients of the noise modal component to obtain a plurality of second wavelet coefficients corresponding to the noise modal component; performing inverse wavelet transform on the plurality of second wavelet coefficients of the noise modal component to obtain the second modal component corresponding to the noise modal component.

5. The method of claim 4, wherein, The performing wavelet threshold filtering on the plurality of first wavelet coefficients of the noise modal component to obtain a plurality of second wavelet coefficients corresponding to the noise modal component comprises: for each first wavelet coefficient of the noise modal component, determining whether the first wavelet coefficient meets a filtering condition, the filtering condition comprising that an absolute value of the first wavelet coefficient is greater than or equal to a filtering threshold; if the first wavelet coefficient meets the filtering condition, determining a difference between the first wavelet coefficient and a coefficient adjustment value as the second wavelet coefficient corresponding to the first wavelet coefficient; if the first wavelet coefficient does not meet the filtering condition, determining a product of the first wavelet coefficient and an attenuation coefficient as the second wavelet coefficient corresponding to the first wavelet coefficient; wherein the coefficient adjustment value and the attenuation coefficient are determined by using the first wavelet coefficient and the filtering threshold.

6. The method of claim 5, wherein, The determining the coefficient adjustment value comprises: determining a sign value of the first wavelet coefficient by using a sign function; determining a first product of the filtering threshold and the sign value; determining a second product of a first preset value and the first wavelet coefficient, wherein the first preset value is a negative value; performing power operation on the second product with a second preset value as a base number to obtain a first power operation value; sum the third preset value and the first power operation value to obtain a first sum value; determine a first ratio of the first product and the first sum value as the coefficient adjustment value.

7. The method of claim 5, wherein, The step of determining the attenuation coefficient comprises: sum a fourth preset value and a square of the first wavelet coefficient to obtain a second sum value; determine a second ratio of a square of the filter threshold value and the second sum value; determine a third product of the second ratio and a fifth preset value; perform power operation on the third product with a sixth preset value as the base number to obtain a second power operation value; use the second power operation value as the attenuation coefficient.

8. The method of claim 1, wherein, the first electrical signal is a three-phase current signal of the motor; and / or, the first electrical signal is one signal segment in a plurality of first signal segments, and the step of obtaining the plurality of first signal segments comprises: segmenting an original electrical signal of the electric drive system into a plurality of second signal segments; multiplying each of the second signal segments by a preset window function to obtain the plurality of first signal segments; and / or, the step of decomposing the first electrical signal of the electric drive system into a plurality of first modal components comprises: performing empirical wavelet transform on the first electrical signal to obtain the plurality of first modal components; wherein a bandwidth parameter of each of the first modal components is proportional to a fundamental frequency of the first electrical signal.

9. The method of claim 1, wherein, The step of determining the state detection result of the electric drive system at least by using the second electrical signal comprises: performing fast Fourier transform on the second electrical signal to obtain a frequency domain feature of the second electrical signal; extracting a harmonic distortion rate of the second electrical signal by using the frequency domain feature; predicting the state detection result at least by using the harmonic distortion rate of the second electrical signal.

10. The method of claim 9, wherein, The step of predicting the state detection result at least by using the harmonic distortion rate of the second electrical signal comprises: obtaining dynamic time sequence features and static features; wherein the dynamic time sequence features comprise time sequence sampling data of at least one first influence factor, the at least one first influence factor comprises the harmonic distortion rate of the second electrical signal, and the static features comprise sampling data of a plurality of second influence factors; extracting first features by using a time sequence coding network on the dynamic time sequence features, and extracting second features by using a deep residual network on the static features; fusing the first features and the second features to obtain fused features; predicting the state detection result based on the fused features by using a prediction network.

11. The method of claim 10, wherein, the step of fusing the first features and the second features to obtain fused features comprises: determining attention weights corresponding to the first features and the second features, respectively; weighting the first features and the second features by using the attention weights corresponding to the first features and the second features, respectively, to obtain attention features; using the attention features as the fused features. And / or, before the feature extraction of the dynamic time sequence feature by using the time sequence coding network, the method further comprises: performing time sequence alignment processing on the dynamic time sequence feature.

12. The method of claim 10, wherein, The at least one first influencing factor further comprises temperatures of a plurality of temperature sampling points in the battery, and the method further comprises: obtaining temperature sampling values of each of the temperature sampling points at a first sampling frequency; in response to a sampling frequency switching condition being met, obtaining temperature values of each of the temperature sampling points at a second sampling frequency; wherein the first sampling frequency is less than the second sampling frequency, and the sampling frequency switching condition comprises at least one of the following conditions: there are at least one group of temperature sampling points whose temperature difference at the same sampling time is greater than or equal to a first temperature threshold, or there is at least one temperature sampling point whose temperature difference at adjacent two sampling times is greater than or equal to a second temperature threshold; or, the first sampling frequency is greater than the second sampling frequency, and the sampling frequency switching condition comprises at least one of the following conditions: the temperature difference at the same sampling time of any two temperature sampling points is less than a third temperature threshold, or the temperature difference at adjacent two sampling times of each temperature sampling point is less than a fourth temperature threshold.

13. The method of claim 1, wherein, The state detection result comprises a failure probability of the electric drive system, and after the state detection result of the electric drive system is determined by using at least the second electric signal, the method further comprises: determining a failure probability threshold by using current vehicle operating data; wherein the current vehicle operating data comprises at least one of a current vehicle load rate and a current battery cycle number; in response to the failure probability being greater than or equal to the failure probability threshold, performing failure processing.

14. The method of claim 13, wherein, The current vehicle operating data comprises the current vehicle load rate and the current battery cycle number, and the determination of the failure probability threshold by using the current vehicle operating data comprises: determining a first adjustment coefficient by using the current vehicle load rate, and determining a second adjustment coefficient by using the current battery cycle number; wherein the first adjustment coefficient is positively correlated with the current vehicle load rate, and the second adjustment coefficient is positively correlated with the current battery cycle number; adjusting a basic adjustment value by using the first adjustment coefficient and the second adjustment coefficient to obtain the failure probability threshold.

15. The method of claim 14, wherein the determination of the first adjustment coefficient by using the current vehicle load rate comprises: determining a fourth product of a first adjustment factor and the current vehicle load rate; summing a preset constant and the fourth product to obtain the first adjustment coefficient; and / or, the determination of the second adjustment coefficient by using the current battery cycle number comprises: performing power operation with the second adjustment factor as the base number and the current battery cycle number as the exponent to obtain the second adjustment coefficient; and / or, the adjustment of the basic adjustment value by using the first adjustment coefficient and the second adjustment coefficient to obtain the failure probability threshold comprises: The product of the first adjustment coefficient, the second adjustment coefficient and the base adjustment value is taken as the failure probability threshold.

16. The method of claim 13, wherein, The failure processing includes: determining a target probability range to which the failure probability belongs; selecting a failure processing operation corresponding to the target probability range from a preset mapping relationship as a target failure processing operation, wherein the preset mapping relationship includes a plurality of preset probability ranges and failure processing operations corresponding to each of the preset probability ranges, and the target probability range is one of the plurality of preset probability ranges; performing the target failure processing operation.

17. A state detection device characterized by comprising: The device includes: a modal component decomposition module configured to decompose a first electrical signal of an electric drive system into a plurality of first modal components; a selection module configured to select at least one noise modal component from the plurality of first modal components; a filtering module configured to perform filtering processing on the at least one noise modal component to obtain at least one second modal component; a signal reconstruction module configured to perform signal reconstruction using the at least one second modal component and each modal component of the plurality of first modal components that has not been subjected to the filtering processing to obtain a second electrical signal; a detection module configured to determine a state detection result of the electric drive system using at least the second electrical signal.

18. An electronic device, comprising: The device includes a memory and a processor coupled to each other, The memory stores program instructions; The processor is configured to execute the program instructions stored in the memory to implement the method of any one of claims 1-16.

19. A computer-readable storage medium, characterized in that, The computer readable storage medium is configured to store program instructions, which are executed by the processor to implement the method of any one of claims 1-16.

20. A computer program product, characterised in that, The computer program product includes a computer program, which is executed by the processor to implement the method of any one of claims 1-16.

Images