Self-adaptive filtering processing method and device for monitoring signals of hydraulic pump

Abstract

The invention provides a self-adaptive filtering processing method and device for monitoring signals of a hydraulic pump. The method provided by the invention comprises the following steps: acquiring a pure monitoring signal data set of the hydraulic pump in different health states and a noise-containing monitoring signal data set of the hydraulic pump in different airborne working scenes; performing feature extraction on the data sets, and determining a first feature corresponding to the pure monitoring signal data set and a second feature corresponding to the noise-containing monitoring signal data set; training a filtering model based on the distribution difference of the first feature and the second feature; and inputting the hydraulic pump monitoring signal to be filtered into the filtering model, and performing filtering processing on the hydraulic pump monitoring signal based on the filtering processing mode output by the filtering model to obtain a filtered hydraulic pump monitoring signal. The method and device provided by the invention are used for adaptively adjusting the filtering method according to the working state of the hydraulic pump and the monitoring signal, and filtering parameters do not need to be manually adjusted.

Description

Technical Field

[0001] The present application relates to the technical field of signal processing, and in particular to a method and device for adaptive filtering of hydraulic pump monitoring signals. Background Art

[0002] In the monitoring and fault diagnosis of aircraft onboard electromechanical systems, filtering is a key link in signal processing. Its main purpose is to extract target signal features from complex backgrounds to achieve accurate health status assessment and early detection of potential faults. As an important component of airborne electromechanical systems, hydraulic pumps operate in a complex environment. Vibration signals are often affected by coupling interference from neighboring equipment such as engines and casings, and are also accompanied by high-intensity noise pollution. These interferences and noises not only have a wide spectrum distribution and diverse statistical characteristics, but also have significant nonlinear and dynamic change characteristics, making the accurate extraction of target signals extremely challenging.

[0003] At present, the commonly used signal filtering methods include median filtering, mean filtering, low-pass filtering, high-pass filtering based on frequency bands, adaptive filtering, and filtering methods based on neural networks. Median filtering performs well in removing impulse noise, but has limited effect when facing Gaussian noise or other complex types of noise; mean filtering has a certain inhibitory effect on Gaussian noise, but it will introduce signal blurring problems and lead to the loss of important features; frequency band filtering is based on prior knowledge of interference frequency components and can effectively filter out-of-band noise, but it is powerless against complex noise within the band. These methods are highly dependent on the statistical characteristics of signals and noise. In practical applications, they can often only cope with specific scenarios and lack the flexibility to cope with changing environments. Especially under the complex and changeable operating conditions of airborne hydraulic pumps, these methods are difficult to meet the filtering requirements of time-varying signals, and the filtering performance is easily severely limited by environmental noise and nonlinear interference.

[0004] Therefore, there is an urgent need for an adaptive filtering processing method for hydraulic pump monitoring signals, which can adaptively adjust the filtering method according to the working state of the hydraulic pump and the monitoring signal, and there is no need to manually adjust the filtering parameters when the working state and noise characteristics of the hydraulic pump change. It has stronger adaptability to complex nonlinear and time-varying noise environments. Summary of the invention

[0005] In view of this, the present application provides a method and device for adaptive filtering processing of hydraulic pump monitoring signals, which is used to adaptively adjust the filtering method according to the working state of the hydraulic pump and the monitoring signal, and when the working state and noise characteristics of the hydraulic pump change, there is no need to manually adjust the filtering parameters, and it has stronger adaptability to complex nonlinear and time-varying noise environments.

[0006] Specifically, the present application is implemented through the following technical solutions: The first aspect of the present application provides a method for adaptive filtering of a hydraulic pump monitoring signal, the method comprising: Obtaining pure monitoring signals of the hydraulic pump in different health states and noisy monitoring signals of the hydraulic pump in different airborne working scenarios, and obtaining pure monitoring signal data sets and noisy monitoring signal data sets; Performing feature extraction on the pure monitoring signal data set and the noisy monitoring signal data set respectively, screening a first feature corresponding to the pure monitoring signal data set and a second feature corresponding to the noisy monitoring signal data set; the feature values ​​of the first feature and the second feature in a fault state and a healthy state are greater than feature values ​​corresponding to other unselected features; A filtering model is trained based on the distribution difference between the first feature and the second feature, and the filtering model outputs a filtering processing mode corresponding to the distribution difference, wherein the distribution difference is determined based on the overlap degree and frequency distribution consistency of the first feature and the second feature; The hydraulic pump monitoring signal to be filtered is input into the filtering model, and the hydraulic pump monitoring signal is filtered based on the filtering processing method output by the filtering model to obtain a filtered hydraulic pump monitoring signal.

[0007] A second aspect of the present application provides a hydraulic pump monitoring signal adaptive filtering processing device, the device comprising an acquisition module, a determination module, a training module and a filtering module; The acquisition module is used to acquire the pure monitoring signals of the hydraulic pump in different health states and the noisy monitoring signals of the hydraulic pump in different airborne working scenarios, and obtain the pure monitoring signal data set and the noisy monitoring signal data set; The determination module is used to perform feature extraction on the pure monitoring signal data set and the noisy monitoring signal data set respectively, and screen the first feature corresponding to the pure monitoring signal data set and the second feature corresponding to the noisy monitoring signal data set; the feature values ​​of the first feature and the second feature in the fault state and the healthy state are greater than the feature values ​​corresponding to other unselected features; The training module is used to train a filtering model based on the distribution difference between the first feature and the second feature, and the filtering model outputs a filtering processing mode corresponding to the distribution difference, wherein the distribution difference is determined based on the overlap degree and frequency distribution consistency of the first feature and the second feature; The filtering module is used to input the hydraulic pump monitoring signal to be filtered into the filtering model, and filter the hydraulic pump monitoring signal based on the filtering processing method output by the filtering model to obtain a filtered hydraulic pump monitoring signal.

[0008] The present application provides a method and device for adaptive filtering of hydraulic pump monitoring signals. In the first aspect, a comprehensive data set is established by acquiring a pure monitoring signal data set of the hydraulic pump in different health states and a noisy monitoring signal data set in different airborne scenarios. This multi-state and multi-scenario data coverage lays the foundation for subsequent feature extraction, enabling the filtering model to adapt to diverse signal inputs. Further, through feature extraction and feature screening, the first feature (pure monitoring signal feature) and the second feature (noise-containing monitoring signal feature) extracted are more significant in health and fault states, providing highly distinguishable input variables. Based on the distribution differences of the two types of features (including overlap and frequency distribution consistency), the trained filtering model can dynamically adjust the filtering strategy to achieve adaptive filtering processing for different noise environments and health states. In the second aspect, after the first feature and the second feature are screened, the present application uses the distribution differences of the first feature and the second feature to train the filtering model, ensuring that the filtering strategy is highly correlated with the feature distribution differences. Since the distribution of the first and second features is significantly different in the noise environment and the pure monitoring signal, the filtering model can focus on a specific frequency band or a specific signal mode, thereby effectively avoiding interference from noise or non-target signals. At the same time, the filtering strategy based on distribution difference can be adjusted dynamically to adaptively optimize the filtering parameters for different input signals. This fast response mechanism based on feature distribution difference significantly improves the filtering speed and accuracy, and realizes fast, accurate and undisturbed filtering processing. On the third aspect, when performing feature screening, the present application ensures that the first feature and the second feature screened have strong distinguishing and representativeness through the principle that the feature values ​​of the first feature and the second feature are higher than the feature values ​​of the unselected feature. When training the filtering model, the overlap degree of the features and the consistency of the frequency distribution are used as the basis for judgment, so that the filtering model can more accurately identify and process the noise signal. At the same time, the filtering processing mode output by the filtering model is directly associated with the feature distribution difference, so that the filtering strategy always matches the current signal state. Therefore, the present application can dynamically adapt to different noise environments and health states, and maintain the stability and high precision of the filtering effect. In addition, when the working state of the hydraulic pump and the interference signal change dynamically, the need for manual intervention is reduced through intelligent filtering model training and adaptive adjustment, and the complexity of manually adjusting the filter parameters is reduced. This can make the processing of hydraulic pump monitoring signals more automated, efficient, and adaptable, and can process and optimize signals under different working conditions in real time. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A flowchart of a method for adaptively filtering a hydraulic pump monitoring signal provided in Embodiment 1 of the present application; Figure 2 This is a schematic diagram of the structure of the hydraulic pump monitoring signal adaptive filtering processing device provided in Example 2 of the present application. DETAILED DESCRIPTION

[0010] Here, exemplary embodiments will be described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application.

[0011] The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in this article refers to and includes any or all possible combinations of one or more associated listed items.

[0012] It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0013] Specific embodiments are given below to introduce the technical solution of the present application in detail.

[0014] Figure 1 This is a flow chart of the method for adaptive filtering of hydraulic pump monitoring signals provided in Example 1 of the present application. Figure 1 , the method provided in this embodiment may include: S101. Acquire pure monitoring signals of a hydraulic pump in different health states and noisy monitoring signals of a hydraulic pump in different airborne working scenarios to obtain a pure monitoring signal data set and a noisy monitoring signal data set.

[0015] Specifically, a pure monitoring signal refers to a signal collected when the hydraulic pump is running under an ideal environment, which is not interfered by external noise and only contains the vibration characteristics and operating status information of the hydraulic pump itself. Different health states refer to the working status of the hydraulic pump at different stages of its operating life or at different levels of failure, including healthy state (normal state), slightly degraded state, severely degraded state, fault state, etc. Noise-containing monitoring signals refer to signals collected in a real airborne environment during the operation of the hydraulic pump. In addition to the vibration characteristics of the hydraulic pump itself, they also contain various external noises and interferences, such as coupled vibrations of adjacent equipment (the impact of vibrations of the engine, casing, etc. on the signal), environmental noise, and system noise. Different airborne working scenarios refer to a variety of typical operating conditions and environmental conditions that a hydraulic pump may encounter during aircraft operation, including startup and shutdown stages, stable flight stages, takeoff and landing stages, and sudden operating conditions.

[0016] It should be noted that the pure monitoring signal and the noisy monitoring signal can be obtained through experimental means or through simulation means.

[0017] In specific implementation, the acquisition of pure monitoring signals of the hydraulic pump in different health states and noisy monitoring signals of the hydraulic pump in different airborne working scenarios includes: acquiring pure monitoring signals of the hydraulic pump in different health states on a standard test bench; the pure monitoring signals cover all working conditions of the hydraulic pump within the actual working envelope; acquiring environmental interference noise signals of the hydraulic pump when not working in an airborne environment on a vibration test bench; and acquiring noisy monitoring signals of the hydraulic pump when working under different working conditions in an airborne environment on the vibration test bench.

[0018] Specifically, a standard hydraulic pump test bench is used to check the test environment to ensure that there is no additional vibration interference or electromagnetic noise in the test environment. High-precision sensors (such as accelerometers, pressure sensors, flow sensors, etc.) are installed in key parts of the hydraulic pump, such as the pump housing and bearing seat. Furthermore, according to the working envelope range of the hydraulic pump, a variety of typical working conditions are set, among which the speed range includes low speed, medium speed, and high speed operation, the load range includes light load, medium load, and full load operation, and the pressure range includes low pressure, medium pressure, and high pressure. Under each working condition, different health states are simulated by artificially adjusting the internal components of the hydraulic pump (such as creating local wear and adding leakage channels). For example, when simulating the normal state, the hydraulic pump has no faults or performance degradation; when simulating the slightly degraded state, the hydraulic pump slipper and bearing are slightly worn or there is a small amount of leakage in the pump chamber; when simulating the severely degraded state, the performance of the hydraulic pump is significantly reduced, such as severe leakage or significantly insufficient flow. When simulating the fault state, the key components of the hydraulic pump are damaged, such as seal rupture or bearing failure. In each health state, pure monitoring signals are collected to obtain pure monitoring signals in different health states under all working conditions.

[0019] Furthermore, a vibration test bench is used to simulate the vibration environment of the aircraft casing and external noise sources (such as engine operation vibration). The hydraulic pump and sensor are fixed on the vibration test bench, but the hydraulic pump is not running. A variety of vibration frequencies and amplitudes are set. For example, the frequency range is from 10Hz to 10kHz, covering the typical frequency band of airborne equipment vibration; the amplitude range includes low, medium and high vibration intensities, simulating different flight states (such as cruising, take-off, landing), and recording the environmental interference noise signals of the hydraulic pump under different vibration conditions.

[0020] Furthermore, the hydraulic pump is operated on the vibration test bench and loaded with different health states and working conditions, where the health state includes normal state, slightly degraded state, severely degraded state and fault state as in the above embodiment. The working conditions include various combinations such as low speed and low load, high speed and high load, etc., covering the actual working envelope. The vibration test bench applies vibration conditions consistent with the airborne environment, and the sensor synchronously records the noise-containing monitoring signal of the hydraulic pump (including the operating characteristic signal of the hydraulic pump and the external vibration noise signal).

[0021] Optionally, the method of obtaining a pure monitoring signal of the hydraulic pump in different health states and a noisy monitoring signal of the hydraulic pump in different airborne working scenarios includes: modeling the noise environment according to the working scenario of the hydraulic pump, generating a noise signal based on a vibration load spectrum, and obtaining an environmental interference noise signal; and fusing the environmental interference noise signal with the pure monitoring signal to obtain the noisy monitoring signal of the hydraulic pump when working under different working conditions.

[0022] In the specific implementation, by analyzing the working environment of the hydraulic pump, possible noise sources are identified, such as engine vibration, casing conduction vibration, structural noise caused by flight operations, and vibration of other onboard equipment. According to the actual airborne environment, a mathematical model of the noise environment is established, and the vibration data in the external environment is used as input. According to the working conditions of the hydraulic pump, appropriate modeling is performed, and the frequency components and power spectra of the vibration signal are extracted using the spectrum analysis method. According to different working conditions, the noise characteristics of different frequency bands are considered to simulate the interference to the hydraulic pump in the airborne environment. Furthermore, according to the specific working conditions (such as speed, load, etc.) of the hydraulic pump, combined with the noise sources of the airborne environment, a load spectrum suitable for the hydraulic pump is designed. The load spectrum can contain vibration components of multiple frequency bands, and it is usually necessary to comprehensively consider low-frequency (such as low-frequency vibration of the engine) and high-frequency (such as sensor noise and mechanical resonance) signals. The vibration simulation tool is used to generate a signal that meets the set load spectrum and simulate the time domain and frequency domain characteristics of the noise. The frequency components, amplitude, phase and other information of the environmental noise are combined with the vibration load spectrum, and an environmental interference noise signal consistent with the airborne environment is generated through a suitable noise model (such as white noise, color noise, etc.). Furthermore, a simulation tool is used to weightedly fuse the generated environmental interference noise signal with the pure monitoring signal collected from the healthy state of the hydraulic pump to obtain a noisy monitoring signal.

[0023] It should be noted that the relationship between the intensity and frequency components of the environmental interference noise signal and the pure monitoring signal needs to be considered during the fusion process to ensure that the noisy monitoring signal is close to the actual working condition. In addition, the noise intensity can be adjusted under different working conditions to simulate the noisy monitoring signal of the hydraulic pump under different airborne environments. For example, under high load or high speed, the signal generated by the hydraulic pump is more susceptible to external noise, so the noise component can be appropriately enhanced.

[0024] S102, perform feature extraction on the pure monitoring signal data set and the noisy monitoring signal data set respectively, and select the first feature corresponding to the pure monitoring signal data set and the second feature corresponding to the noisy monitoring signal data set; the feature values ​​of the first feature and the second feature in the fault state and the healthy state are greater than the feature values ​​corresponding to other unselected features.

[0025] Specifically, the first feature corresponding to the pure monitoring signal data set and the second feature corresponding to the noisy monitoring signal data set are different, but the process and implementation principle of obtaining the first feature and obtaining the second feature are similar. In this embodiment, only the acquisition of the first feature is introduced. The second feature is similar and will not be repeated in this embodiment.

[0026] In a specific implementation, the feature extraction is performed on the pure monitoring signal data set, and the first feature corresponding to the pure monitoring signal data set is screened, including: denoising and standardizing the pure monitoring signal data set; extracting features from the pure monitoring signal data set based on Fourier transform to obtain time domain features and frequency domain features of pure monitoring signals under different health states; analyzing the time domain features and frequency domain features to determine the first feature of a sudden change from a healthy state to a faulty state.

[0027] Specifically, filtering methods (such as wavelet denoising, median filtering or low-pass filtering) are used to remove high-frequency noise or abnormal spikes in the pure monitoring signal data set. A denoising method suitable for the hydraulic pump monitoring signal is selected to ensure that the main components of the signal are retained without excessive smoothing. The signal amplitude of the pure monitoring signal data set is normalized to a uniform range (such as [0, 1] or a standard normal distribution with a mean of 0 and a variance of 1) to eliminate the impact of dimensional differences. Furthermore, statistical features such as mean, variance and standard deviation, skewness and kurtosis, and peak factor are directly extracted from the time domain signal, and the time domain signal is converted into a frequency domain signal through Fourier transform to extract key frequency domain features such as main frequency component, bandwidth, harmonic component, and energy distribution. The distribution differences of time domain features and frequency domain features under different health states are compared to identify the significant change points of the signal from the healthy state to the fault state, and the features with obvious feature changes are determined as the first features using statistical analysis methods (such as variance analysis and eigenvalue curve drawing).

[0028] It should be noted that the first feature is determined based on the change trend of the pure monitoring signal between different health states (fault state and healthy state), that is, the characteristic value of the first feature in the fault state and the healthy state is greater than the characteristic value corresponding to other unselected features. However, the first feature may change with the working condition of the hydraulic pump, or the first feature may not change with the working condition of the hydraulic pump. In this step, the situation where the first feature changes with the working condition of the hydraulic pump is first introduced, that is, the change amplitude of the first feature with the change of the working condition of the hydraulic pump is higher than the change amplitude corresponding to other unselected features.

[0029] Optionally, after screening the first feature corresponding to the pure monitoring signal data set and the second feature corresponding to the noisy monitoring signal data set, it also includes: when the changing trend of the first feature and the second feature with the hydraulic pump operating condition is higher than a first preset value, constructing a mapping relationship between the first feature, the second feature and the operating condition based on a neural network as a mark of the first feature and the second feature.

[0030] Specifically, in combination with the above description, the first preset value is set according to actual needs. In this embodiment, this is not limited.

[0031] In specific implementation, by fitting the relationship curve between the first feature and the hydraulic pump working condition, the gradient or rate of change of the first feature as the hydraulic pump working condition changes is calculated, and the change trend is measured by slope, change amplitude or standard deviation. Further, the change trend of each first feature as the hydraulic pump working condition changes is traversed, and when it is determined that the change trend is greater than the first preset value, it is determined that the first feature is more sensitive to the change in working condition, and a neural network (such as a multi-layer perceptron, a convolutional neural network, etc.) is used to construct a mapping relationship between the first feature and the hydraulic pump working condition. Among them, the input is the working condition data of the hydraulic pump (such as speed, pressure), and the output is the first characteristic value. A large amount of data is used to train and verify the neural network to ensure the fitting accuracy of the first feature and the working condition mapping.

[0032] The method provided in this embodiment uses the mutation of different signals from healthy state to fault state as the standard to accurately screen out the features that are sensitive to healthy state and fault state, highlighting the distinguishability and representativeness of the features. The screened first and second features not only significantly reduce the interference of redundant features, but also reduce the data dimension, improve the feature extraction efficiency and the processing speed and accuracy of the filtering model. Subsequently, for the features with a high trend of change with the working condition of the hydraulic pump, a mapping relationship between the features and the working condition of the hydraulic pump is established through a neural network, and the features are dynamically associated with the working condition of the hydraulic pump, so that the feature screening has higher adaptability and real-time adjustment ability. This feature mark based on the change of working condition gives the feature stronger stability and generalization ability, and can maintain consistency in different scenarios, thereby enhancing the robustness of the filtering model to cope with complex working conditions of the hydraulic pump. In addition, by combining the correlation between the changing trend of the feature and the working condition, the screened features not only optimize the filtering performance, but also better meet the actual engineering needs and have good application value. Therefore, the feature screening method greatly improves the accuracy, adaptability and practical application effect of the filtering model by highlighting the working condition correlation and mark attributes of the feature.

[0033] Further, in combination with the above description, the first feature may change with the working condition of the hydraulic pump, or the first feature may not change with the working condition of the hydraulic pump. In the above embodiment, the case where the first feature changes with the working condition of the hydraulic pump is first introduced. In this step, the case where the first feature does not change with the working condition of the hydraulic pump is introduced, that is, the variation range of the first feature with the working condition of the hydraulic pump is lower than the variation range corresponding to other unselected features.

[0034] In specific implementation, by fitting the relationship curve between the first feature and the hydraulic pump working condition, the gradient or rate of change of the first feature with the hydraulic pump working condition is calculated, and the change trend is measured by slope, change amplitude or standard deviation. Further, the change trend of each first feature with the hydraulic pump working condition is traversed, and when it is determined that the change trend is less than the second preset value, it is determined that the first feature has good working condition invariance, is suitable for independent evaluation of the health state, can stably characterize the health state, and is directly retained as the first feature.

[0035] S103: training a filtering model based on the distribution difference between the first feature and the second feature, wherein the filtering model outputs a filtering processing method corresponding to the distribution difference.

[0036] Specifically, the input of the filtering model is the first feature and the second feature, and the output is a filtering processing method. The filtering processing method includes a frequency band filtering method, a signal adaptive decomposition method, an adaptive filtering method, a method based on a neural network filtering, and the like.

[0037] The training filtering model based on the distribution difference of the first feature and the second feature includes: comparing the frequency band overlap of the first feature and the second feature; if the frequency band overlap is less than a threshold, filtering the noise of the noise-containing monitoring signal data set based on the non-overlapping frequency band; if the frequency band overlap is greater than or equal to the threshold, judging the difference between the fault feature in the second feature and the noise feature in the second feature and the similarity between the feature distribution and the known distribution to determine the target filtering method, and different target filtering methods have different feature recognition capabilities. That is to say, if the overlap is low, the frequency band can be used to distinguish them directly, and the filtering based on the frequency band is the most direct, fast and accurate. However, if the overlap is high, it is necessary to further see whether there is a significant difference between the noise and the effective signal features. If there is a significant difference, the noise can still be directly filtered based on the filtering method, that is, the signal adaptive decomposition method is used. However, if the difference is not obvious, the signal decomposition cannot distinguish the two signals. At this time, the noise and effective signal feature distribution features are compared to see if they are the same as the known signal type. If they are the same, the adaptive filtering method can be used again based on our known information to achieve rapid differentiation filtering. If they are not the same, the unknown signal is faced, the two are mixed and the difference is not obvious. At this time, the neural network recognition method is used. The method provided by the present invention adopts filtering methods with different precisions and efficiencies according to the differences and known characteristics of the mixed contents in the mixed signals. For signals that can be clearly distinguished, a fast and simple filtering method is directly adopted, which ensures the accuracy of filtering while saving filtering time.

[0038] In a specific implementation, the distribution difference training filtering model based on the first feature and the second feature includes: performing spectral analysis on the first feature and the second feature to determine the spectral components affected by noise; when the spectral components are not within the operating frequency range of the pure monitoring signal, determining that the distribution difference is out-of-band noise; when the spectral components are within the operating frequency range of the pure monitoring signal, determining that the distribution difference is in-band noise.

[0039] Specifically, the first feature and the second feature are Fourier transformed to obtain a spectrum, and the first feature and the second feature are converted into spectrum expressions to obtain amplitude and frequency distribution, and the energy concentration distribution and noise characteristics in the spectrum are analyzed to determine the spectrum components of the first feature and the second feature and their frequency distribution relationship with the pure monitoring signal. Further, the spectrum components of the first feature and the second feature are compared with the spectrum of the pure monitoring signal to determine the newly added or abnormally enhanced frequency components in the characteristic spectrum, such as abnormal increase in amplitude, sudden increase in high-frequency energy, etc. When the abnormal frequency component is not within the operating frequency range of the pure monitoring signal, it indicates that the noise feature and the fault feature have no cross-influence within the frequency band, and it is determined to be out-of-band noise (such as a hydraulic pump working at a specific speed, its main frequency range is known, and any frequency component within this range is out-of-band noise). Out-of-band noise may come from external equipment vibration, environmental vibration, etc. When the abnormal frequency component is within the operating frequency range of the pure monitoring signal, it indicates that the noise characteristics and the fault characteristics have cross-influence within the frequency band, which is determined to be in-band noise. In-band noise may be caused by internal faults or coupling interference (such as harmonic distortion or bearing failure) of the hydraulic pump, which manifests as abnormal increase in amplitude, harmonic distortion, etc.

[0040] Optionally, the filtering model outputs a filtering processing mode corresponding to the distribution difference, including: (1) When it is determined that the distribution difference is out-of-band noise, the filtering model determines that the corresponding filtering processing method is a frequency band filtering method.

[0041] Specifically, according to the different characteristics of out-of-band noise and in-band noise, an adaptive filtering method needs to be designed. When out-of-band noise is determined, a frequency band filtering method is usually used to selectively retain the signal within the operating frequency range of the hydraulic pump and suppress noise in other frequency bands.

[0042] (2) When it is determined that the distribution difference is in-band noise and the second feature is smaller than the first feature, the filtering model determines that the corresponding filtering processing method is a signal adaptive decomposition method.

[0043] Specifically, when it is determined to be in-band noise and the second feature is smaller than the first feature, it indicates that the noise feature and the fault feature partially overlap within the frequency band, and the amplitude or energy of the fault feature is significantly higher than that of the noise feature. Therefore, the characteristic frequency related to the fault can be extracted by the signal adaptive decomposition method, while the frequency components related to the noise can be removed.

[0044] (3) When it is determined that the distribution difference is in-band noise and the distribution of the first feature is determined or the distribution of the second feature is determined, the filtering model determines that the corresponding filtering processing method is an adaptive filtering method or a method based on neural network filtering.

[0045] Specifically, when it is determined to be in-band noise and the distribution of the first feature or the distribution of the second feature is determined, it indicates that the noise feature and the fault feature partially overlap within the frequency band, and the statistical characteristics of the noise feature or the fault feature (such as distribution shape, energy distribution or spectrum characteristics) are known, a more accurate filtering method can be selected based on its distribution information, that is, an adaptive filtering method or a method based on a neural network filter. Among them, the adaptive filtering method can adjust the filtering parameters in real time according to the signal characteristics, and the method based on the neural network filter can accurately separate the noise and fault features when the distribution characteristics are known.

[0046] Optionally, after the filtering model is trained based on the distribution difference between the first feature and the second feature, and the filtering model outputs a filtering processing method corresponding to the distribution difference, the method further includes: acquiring actual onboard data of the hydraulic pump under an actual working environment based on a sensor; comparing the actual onboard data with the hydraulic pump monitoring signal after filtering by the filtering model, and adjusting the filtering model based on the comparison result.

[0047] Specifically, the actual onboard data of the hydraulic pump under the actual working environment is obtained. These data are usually collected in real time by sensors, reflecting the actual operating status of the hydraulic pump under different working conditions. The actual onboard data contains the operating characteristics of the hydraulic pump under complex working conditions, including the health status, vibration, pressure, temperature and other signals of the equipment. These signals are usually affected by noise and may include interference from the onboard environment, adjacent equipment, etc.

[0048] Furthermore, these actual airborne data are compared with the hydraulic pump monitoring signals processed by the filter model, that is, the signals optimized by the filter model are compared with the real airborne signals directly collected by the sensor, with the focus on detecting the differences between the two, especially the noise removal effect, the accuracy of feature extraction, and the clarity of the signal. Through this comparison, possible problems in the filtering process can be identified, such as the filter's poor processing effect on certain noise signals, or the failure to fully restore the health status characteristics of the hydraulic pump. According to the comparison results, if it is found that the processing effect of the filter model is not ideal, it may be necessary to adjust the parameters or training process of the filter model. The adjustment process may include optimizing the weights of the filter, reselecting features, or modifying the structure of the filter model to improve its adaptability and accuracy. In this way, the filter model can be continuously adjusted and optimized based on the feedback of the actual data to achieve better signal processing effects, especially in noise processing and fault feature extraction in actual working environments.

[0049] In the specific implementation, the actual airborne data of the hydraulic pump in the actual working environment is collected in real time through the sensors installed on the hydraulic pump (such as pressure sensors, vibration sensors or acoustic sensors). The collected data is usually uploaded to the central control unit or storage device through the data acquisition module. Further, the collected actual airborne data is input into the pre-trained filtering model. The filtering model processes the actual airborne data according to the difference in feature distribution in the actual airborne data, and outputs the monitoring signal after filtering and denoising. The actual airborne data is compared with the monitoring signal after filtering and denoising, and the same features are extracted from the actual airborne data and the monitoring signal after filtering and denoising, such as spectrum distribution, time domain waveform, noise power, etc. The difference between the monitoring signal after filtering and denoising and the actual airborne data is quantified through error measurement methods (such as mean square error, signal correlation coefficient, signal-to-noise ratio improvement, etc.). It is judged whether the monitoring signal after filtering and denoising has reduced the noise level while retaining the key hydraulic pump operation characteristics. If the monitoring signal after filtering and denoising has lost important features, it means that the filtering model needs to be adjusted. Based on the comparison results, the filter model is dynamically optimized, including: (1) Analyzing the source of the difference. If the denoising is insufficient (too much noise remains), analyze the noise recognition ability of the filter model for specific frequencies or features, and adjust the parameters of the filter model; if the feature is lost (the useful signal is filtered): check whether the weight distribution of the filter model to the feature is reasonable, and retrain the filter model. (2) Fine-tune the filter model. Use the features of the actual airborne data as a supplementary data set to optimize the filter model through incremental training or online learning to make it more suitable for the current working conditions. (3) Based on the difference analysis results, dynamically adjust the processing logic of the filter model, such as changing the weight and frequency range of the filter algorithm, or adding a processing module for specific interference.

[0050] The method provided in this embodiment further enhances the adaptability and accuracy of the filtering process by dynamically adjusting the filtering model. First, in the training stage, the filtering model is trained based on the distribution difference between the first feature and the second feature, and the filtering processing mode that matches the distribution difference is outputted, which provides a strong basis for the initial effect of the filtering. In practical applications, the airborne data of the hydraulic pump in the actual working environment is collected by the sensor, and compared with the monitoring signal after filtering processing to evaluate the deviation of the filtering effect. This comparison method can dynamically identify the deficiencies of the filtering model in different environments or working conditions, thereby providing an accurate basis for adjusting the model parameters. Through this feedback mechanism, the filtering model can be continuously optimized, gradually adapted to the characteristics of the actual airborne data, and reduce the errors caused by changes in working conditions or different signal characteristics. In addition, this adjustment of the filtering model based on actual data can effectively compensate for the distribution deviation between the training data and the actual working environment, ensuring that the filtering process not only performs well under ideal conditions, but also operates stably in complex and changeable actual working conditions. This application realizes the dynamic adaptive optimization of the filtering model by introducing a closed-loop mechanism of actual data comparison and filtering model adjustment, and significantly improves the robustness, generalization ability and practical application effect of the model.

[0051] S104, inputting the hydraulic pump monitoring signal to be filtered into the filtering model, filtering the hydraulic pump monitoring signal based on the filtering processing method output by the filtering model, and obtaining a filtered hydraulic pump monitoring signal.

[0052] Specifically, in combination with the above description, the filtering model is based on a filtering processing method corresponding to the characteristics of the input hydraulic pump monitoring signal and the characteristic output of the pure monitoring signal. When the filtering processing method is a frequency band filtering method, in this step, a frequency band filter is used to filter the hydraulic pump monitoring signal, wherein the types of frequency band filters include bandpass filters and low-pass / high-pass filters. The bandpass filter is suitable for effectively isolating the out-of-band noise in the hydraulic pump monitoring signal within a known operating frequency range (such as the main frequency and harmonic range). The low-pass / high-pass filter is suitable for specific situations where only high-frequency or low-frequency noise in the hydraulic pump monitoring signal needs to be removed.

[0053] For example, for the following hydraulic pump monitoring signal: target frequency: hydraulic pump main frequency is 100 Hz, harmonic range is 100-500 Hz, interference frequency: low frequency noise is 0-50 Hz, high frequency noise is above 600 Hz. The frequency band filters that can be used are as follows: low pass filter: retain signals with frequencies below 500 Hz and suppress high frequency noise, band pass filter: retain signals within the range of 100-500 Hz and suppress low and high frequency noise.

[0054] Optionally, when the filtering processing method is the signal adaptive decomposition method, in this step, when the variational modal decomposition method in the signal adaptive decomposition method is selected to filter the hydraulic pump monitoring signal, firstly, the spectrum analysis of the hydraulic pump monitoring signal is performed to identify the frequency characteristics related to the fault state: the hydraulic pump monitoring signal is subjected to a fast Fourier transform to obtain its frequency distribution. The fault characteristic frequency range (such as certain frequency multiples or modulation frequencies) is identified, and the center frequency of these frequency bands is determined. When determining the fault characteristic frequency range, the frequency components with significant mutations (i.e., the frequencies with the largest energy change from the healthy state to the fault state) are selected, and specific mechanical characteristic frequencies (such as the fundamental frequency, frequency multiples or modulation frequencies of the hydraulic pump, etc.) are considered. Furthermore, the variational modal decomposition method is used to decompose the hydraulic pump monitoring signal, and the narrowband signal components corresponding to each center frequency are extracted: the variational modal decomposition method decomposes the hydraulic pump monitoring signal into several modes, each of which is a narrowband signal in the frequency domain. Reconstruct the components in each modal signal that match the fault characteristic frequency range to form a new signal that only contains the fault characteristic distribution frequency band, that is, the filtered hydraulic pump monitoring signal: Analyze the spectrum distribution of each mode and match it with the fault center frequency set. Exclude the modes related to noise (such as modes whose frequencies do not overlap with the fault characteristic frequencies, or modes with lower energy). Superimpose the filtered modal signals to reconstruct the new signal.

[0055] Optionally, when the filtering processing method is an adaptive filtering method, in this step, the hydraulic pump monitoring signal is filtered based on the adaptive filtering method. The adaptive filtering method is that the hydraulic pump monitoring signal passes through a parameter-adjustable digital filter to generate a filtered output signal, which is compared with a pure monitoring signal to form an error signal, and the filter parameters are adjusted through an adaptive algorithm to ultimately minimize the mean square value of the error signal. Adaptive filtering can use the results of the filter parameters obtained at the previous moment to automatically adjust the filter parameters at the current moment to adapt to the unknown or time-varying statistical characteristics of the signal and noise, thereby achieving optimal filtering.

[0056] In specific implementation, the hydraulic pump monitoring signal is processed to make it close to the pure monitoring signal, thereby removing noise. The pure monitoring signal is usually obtained through experiments or simulations. When designing the filter, a digital filter with adjustable parameters is used, such as a finite impulse response (FIR) or infinite impulse response (IIR) filter. During initialization, the parameters of the filter are usually set to zero or random values, and the output signal of the filter is the result of the hydraulic pump monitoring signal after the filter weighted processing.

[0057] Further, the error between the filter output and the pure monitoring signal is calculated to form an error signal. This error represents the difference between the filter output signal and the ideal target signal. The error signal is used to guide the update of the filter parameters. By minimizing the error signal, the filter output is close to the pure monitoring signal. Use an adaptive algorithm to update the filter parameters. Common adaptive algorithms include the least mean square (LMS) algorithm and the recursive least squares (RLS) algorithm. The filter adjusts its weight value through error feedback at each moment, so that the output signal (hydraulic pump monitoring signal) gradually approaches the target signal (pure monitoring signal). This process is iterative, and the filter parameters are adjusted and updated multiple times until the error reaches a stable minimum value or meets the convergence condition. At each moment, the filter adjusts the parameters of the current moment according to the parameter value of the previous moment, thereby continuously optimizing the filtering effect. Finally, after adaptive adjustment, the filter output signal will be closest to the target signal, the mean square value of the error signal will be the smallest, and the optimal filtering effect will be achieved.

[0058] Optionally, when the filtering processing method is a method based on neural network filtering, in this step, the hydraulic pump monitoring signal is filtered based on the method based on neural network filtering. The neural network filtering method is usually implemented through a generative adversarial network (GAN) framework, involving two main components: a generator and a discriminator. First, in the training phase, the neural network is trained by providing a large number of noisy monitoring signals and corresponding pure monitoring signals. The task of the generator is to receive the noisy monitoring signal as input and generate a filtered signal, which should be as close to the target pure monitoring signal as possible. The task of the discriminator is to determine whether the input signal comes from a real pure monitoring signal data set or a filtered signal generated by the generator.

[0059] The generator and discriminator are optimized through competitive training. The generator tries to generate a signal that is as "real" as possible, while the discriminator tries to accurately distinguish between real signals and generated signals. This process is evaluated through a loss function. A common loss function is the residual loss function, which calculates the difference between the generated signal and the pure monitoring signal. Through the alternating optimization of the generator and the discriminator, the generator gradually learns to generate higher quality signals in different noise environments, and is able to remove noise and retain the main features of the signal. After multiple rounds of iterative optimization, a trained filtering network is obtained, which has strong adaptability, can adapt to different types of noise and effectively denoise.

[0060] After the training is completed, for the hydraulic pump monitoring signal, the trained adaptive filtering network will directly process the signal and generate a filtered output signal to achieve the purpose of feature extraction and noise reduction. Compared with traditional filtering methods, this process has stronger nonlinear modeling capabilities, especially capable of handling complex and unknown noise environments, thereby providing more accurate and stable noise suppression effects.

[0061] In the specific implementation, based on the noise characteristics and fault characteristics in the hydraulic pump monitoring signal to be filtered, a suitable filtering method (such as frequency band filtering method, signal adaptive decomposition method, adaptive filtering method or neural network filtering method, etc.) is selected. This selection determines the basic type of the filter, but the parameters of the filter need to be further adjusted. Furthermore, the actual airborne data of the hydraulic pump under different working conditions are collected, and the signal is feature extracted to analyze the noise characteristics and fault characteristics. Based on the selected filtering method, the filtering model is initialized and the basic parameters (such as frequency range, filter order, gain, etc.) are set. These parameters can be set by some general rules or previous experience. The filter model is trained using the labeled training data (including noise and fault signals), and the parameters of the filter are adjusted so that it can adaptively optimize the noise suppression effect according to the different working conditions and fault states of the hydraulic pump. During the training process, the parameters of the filter are continuously optimized according to the performance of the filtering model on the verification data set. Through error analysis (such as the mean square error of the signal, signal-to-noise ratio, etc.), the effect of the filtering model in denoising and retaining key signals is judged. If the filtering effect is not ideal, the parameters of the filter can be fine-tuned, or multiple filtering methods can be combined for optimization. The filter model can further adjust the filter response based on real-time data feedback. For example, the filter can dynamically adjust its parameters according to the working state of the hydraulic pump (such as load changes, speed changes, etc.) to ensure that the filtering effect is always maintained in the best state. After obtaining the trained filter model, a suitable filtering method is reselected based on the trained filter model, and the hydraulic pump monitoring signal to be filtered is filtered based on the selected filtering method to obtain the filtered hydraulic pump monitoring signal.

[0062] The adaptive filtering processing method for the monitoring signal of a hydraulic pump provided in this embodiment, in the first aspect, establishes a comprehensive data set by acquiring a pure monitoring signal data set of the hydraulic pump in different health states and a noisy monitoring signal data set in different airborne scenarios. This multi-state and multi-scenario data coverage lays the foundation for subsequent feature extraction, enabling the filtering model to adapt to diverse signal inputs. Further, through feature extraction and feature screening, the performance of the extracted first feature (pure monitoring signal feature) and the second feature (noise-containing monitoring signal feature) in the health and fault states is more significant, providing a highly distinguishable input variable. Based on the distribution difference of the two types of features (including overlap and frequency distribution consistency), the trained filtering model can dynamically adjust the filtering strategy to achieve adaptive filtering processing for different noise environments and health states. In the second aspect, after the first feature and the second feature are screened, the present application uses the distribution difference of the first feature and the second feature to train the filtering model, ensuring that the filtering strategy is highly correlated with the feature distribution difference. Since the distribution of the first and second features is significantly different in the noise environment and the pure monitoring signal, the filtering model can focus on a specific frequency band or a specific signal mode, thereby effectively avoiding interference from noise or non-target signals. At the same time, the filtering strategy based on distribution difference can be adjusted dynamically to adaptively optimize the filtering parameters for different input signals. This fast response mechanism based on feature distribution difference significantly improves the filtering speed and accuracy, and realizes fast, accurate and undisturbed filtering processing. On the third aspect, when the present application performs feature screening, the principle that the feature value is higher than the feature value of the unselected feature is adopted to ensure that the screened features have strong distinguishing and representativeness. When training the filtering model, the coincidence degree of the features and the consistency of the frequency distribution are used as the basis for judgment, so that the filtering model can more accurately identify and process the noise signal. At the same time, the filtering processing mode output by the filtering model is directly associated with the feature distribution difference, so that the filtering strategy always matches the current signal state. Therefore, the present application can dynamically adapt to different noise environments and health states, and maintain the stability and high precision of the filtering effect. Moreover, when the working state of the hydraulic pump and the interference signal change dynamically, the need for manual intervention is reduced through intelligent filtering model training and adaptive adjustment, and the complexity of manually adjusting the filter parameters is reduced. In this way, the processing process of the monitoring signal of the hydraulic pump can be more automated and efficient, and the adaptability is also stronger, and the signals under different working conditions can be processed and optimized in real time. Fourthly, when determining the features, the selection of features not only depends on the change from healthy state to faulty state, but also takes into account the changing trend of the hydraulic pump working condition. The signal characteristics of the hydraulic pump will be different under different working conditions, so the features should be able to adapt to these changes. By analyzing the changing trend of these features with working conditions, features with good recognition ability under various working conditions can be selected, thereby ensuring the stability and accuracy of the filtering model in different practical application scenarios.Through this screening process, the features finally screened out are features with significant distinguishing ability, and these features have good performance under different working conditions and health states of the hydraulic pump. This can enhance the robustness and generalization ability of the subsequent fault diagnosis model, and ensure that the health state of the hydraulic pump can be accurately identified under different working environments. Fifthly, through the dynamic adjustment of the filtering model, the adaptability and accuracy of the filtering process are further enhanced. First, in the training stage, the filtering model is trained based on the distribution difference between the first feature and the second feature, and the filtering processing method that matches its output with the distribution difference provides a strong foundation for the initial effect of the filtering. In practical applications, the airborne data of the hydraulic pump in the actual working environment is collected by sensors, and compared with the monitoring signal after filtering processing to evaluate the deviation of the filtering effect. This comparison method can dynamically identify the deficiencies of the filtering model in different environments or working conditions, thereby providing an accurate basis for adjusting the model parameters. Through this feedback mechanism, the filtering model can be continuously optimized, gradually adapt to the characteristics of the actual airborne data, and reduce the errors caused by changes in working conditions or different signal characteristics. In addition, this filter model adjustment based on actual data can effectively compensate for the distribution deviation between the training data and the actual working environment, ensuring that the filter processing not only performs well under ideal conditions, but also operates stably under complex and changeable actual working conditions. This application realizes dynamic adaptive optimization of the filter model by introducing a closed-loop mechanism of actual data comparison and filter model adjustment, significantly improving the robustness, generalization ability and practical application effect of the model.

[0063] Corresponding to the aforementioned embodiment of a method for adaptively filtering and processing a hydraulic pump monitoring signal, the present application also provides an embodiment of a device for adaptively filtering and processing a hydraulic pump monitoring signal.

[0064] Figure 2 This is a schematic diagram of the structure of the hydraulic pump monitoring signal adaptive filtering processing device provided in Example 2 of the present application. Figure 2 , the device provided in this embodiment includes an acquisition module 210, a determination module 220, a training module 230 and a filtering module 240; The acquisition module is used to acquire the pure monitoring signals of the hydraulic pump in different health states and the noisy monitoring signals of the hydraulic pump in different airborne working scenarios, and obtain the pure monitoring signal data set and the noisy monitoring signal data set; The determination module is used to perform feature extraction on the pure monitoring signal data set and the noisy monitoring signal data set respectively, and screen the first feature corresponding to the pure monitoring signal data set and the second feature corresponding to the noisy monitoring signal data set; the feature values ​​of the first feature and the second feature in the fault state and the healthy state are greater than the feature values ​​corresponding to other unselected features; The training module is used to train a filtering model based on the distribution difference between the first feature and the second feature, and the filtering model outputs a filtering processing mode corresponding to the distribution difference, wherein the distribution difference is determined based on the overlap degree and frequency distribution consistency of the first feature and the second feature; The filtering module is used to input the hydraulic pump monitoring signal to be filtered into the filtering model, and filter the hydraulic pump monitoring signal based on the filtering processing method output by the filtering model to obtain a filtered hydraulic pump monitoring signal.

[0065] The device of this embodiment can be used to perform Figure 1 The steps, specific implementation principles and implementation processes of the method embodiment shown are similar and will not be repeated here.

[0066] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0067] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present application scheme. A person of ordinary skill in the art can understand and implement it without paying any creative work.

[0068] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for adaptive filtering of hydraulic pump monitoring signals, characterized in that: The method comprises: Obtaining pure monitoring signals of the hydraulic pump in different health states and noisy monitoring signals of the hydraulic pump in different airborne working scenarios, and obtaining pure monitoring signal data sets and noisy monitoring signal data sets; Performing feature extraction on the pure monitoring signal data set and the noisy monitoring signal data set respectively, screening a first feature corresponding to the pure monitoring signal data set and a second feature corresponding to the noisy monitoring signal data set; the feature values ​​of the first feature and the second feature in a fault state and a healthy state are greater than feature values ​​corresponding to other unselected features; A filtering model is trained based on the distribution difference between the first feature and the second feature, and the filtering model outputs a filtering processing mode corresponding to the distribution difference, wherein the distribution difference is determined based on the overlap degree and frequency distribution consistency of the first feature and the second feature; The hydraulic pump monitoring signal to be filtered is input into the filtering model, and the hydraulic pump monitoring signal is filtered based on the filtering processing method output by the filtering model to obtain a filtered hydraulic pump monitoring signal.

2. The method according to claim 1, characterized in that The extracting features of the clean monitoring signal data set and screening the first feature corresponding to the clean monitoring signal data set includes: Performing denoising and standardization processing on the pure monitoring signal data set; Performing feature extraction on the pure monitoring signal data set based on Fourier transform to obtain time domain features and frequency domain features of the pure monitoring signals under different health states; The time domain features and frequency domain features are analyzed to determine a first feature of a sudden change from a healthy state to a fault state.

3. The method according to claim 1, characterized in that After screening the first feature corresponding to the pure monitoring signal data set and the second feature corresponding to the noisy monitoring signal data set, the method further includes: When the changing trend of the first feature and the second feature with the hydraulic pump working condition is higher than a first preset value, a mapping relationship between the first feature, the second feature and the working condition is constructed based on a neural network as a sign of the first feature and the second feature.

4. The method according to claim 1, characterized in that: The filtering model is trained based on the distribution difference of the first feature and the second feature, including: comparing the frequency band overlap between the first feature and the second feature; If the frequency band overlap is less than a threshold, filtering the noise of the noisy monitoring signal data set based on non-overlapping frequency bands; If the frequency band overlap is greater than or equal to a threshold, the difference between the fault feature in the second feature and the noise feature in the second feature and the similarity between the feature distribution and the known distribution are judged to determine the target filtering method, and different target filtering methods have different feature recognition capabilities.

5. The method according to claim 1, characterized in that The filtering model is trained based on the distribution difference of the first feature and the second feature, comprising: Performing spectrum analysis on the first feature and the second feature to determine spectrum components affected by noise; When the frequency spectrum component is not within the operating frequency range of the pure monitoring signal, determining that the distribution difference is out-of-band noise; When the frequency spectrum component is within the operating frequency range of the pure monitoring signal, the distribution difference is determined to be in-band noise.

6. The method according to claim 5, characterized in that The filtering model outputs a filtering processing method corresponding to the distribution difference, including: When it is determined that the distribution difference is out-of-band noise, the filtering model determines that the corresponding filtering processing method is a frequency band filtering method; When it is determined that the distribution difference is in-band noise and the second feature is smaller than the first feature, the filtering model determines that the corresponding filtering processing method is a signal adaptive decomposition method; When it is determined that the distribution difference is in-band noise and the distribution of the first feature is determined or the distribution of the second feature is determined, the filtering model determines that the corresponding filtering processing method is an adaptive filtering method or a method based on neural network filtering.

7. The method according to claim 1, characterized in that The method of obtaining the pure monitoring signals of the hydraulic pump in different health states and the noisy monitoring signals of the hydraulic pump in different airborne working scenarios includes: Acquire pure monitoring signals of the hydraulic pump in different health states on a standard test bench; the pure monitoring signals cover all working conditions of the hydraulic pump within the actual working envelope; Acquire the environmental interference noise signal of the hydraulic pump when it is not working in an airborne environment on a vibration test bench; The noise-containing monitoring signals of the hydraulic pump working under different working conditions in an airborne environment are obtained on the vibration test bench.

8. The method according to claim 1, characterized in that: The method of obtaining the pure monitoring signals of the hydraulic pump in different health states and the noisy monitoring signals of the hydraulic pump in different airborne working scenarios includes: Noise environment modeling is performed according to the working scenario of the hydraulic pump, noise signals are generated based on the vibration load spectrum, and environmental interference noise signals are obtained; The environmental interference noise signal is fused with the pure monitoring signal to obtain the noise-containing monitoring signal when the hydraulic pump works under different working conditions.

9. The method according to claim 1, characterized in that: After the filtering model is trained based on the distribution difference between the first feature and the second feature, and the filtering model outputs a filtering processing mode corresponding to the distribution difference, the method further includes: Acquire actual onboard data of the hydraulic pump in actual working environment based on sensors; The actual onboard data is compared with the hydraulic pump monitoring signal after filtering by the filtering model, and the filtering model is adjusted based on the comparison result.

10. A hydraulic pump monitoring signal adaptive filtering processing device, characterized in that: The device comprises an acquisition module, a determination module, a training module and a filtering module; The acquisition module is used to acquire the pure monitoring signals of the hydraulic pump in different health states and the noisy monitoring signals of the hydraulic pump in different airborne working scenarios, and obtain the pure monitoring signal data set and the noisy monitoring signal data set; The determination module is used to perform feature extraction on the pure monitoring signal data set and the noisy monitoring signal data set respectively, and screen the first feature corresponding to the pure monitoring signal data set and the second feature corresponding to the noisy monitoring signal data set; the feature values ​​of the first feature and the second feature in the fault state and the healthy state are greater than the feature values ​​corresponding to other unselected features; The training module is used to train a filtering model based on the distribution difference between the first feature and the second feature, and the filtering model outputs a filtering processing mode corresponding to the distribution difference, wherein the distribution difference is determined based on the overlap degree and frequency distribution consistency of the first feature and the second feature; The filtering module is used to input the hydraulic pump monitoring signal to be filtered into the filtering model, and filter the hydraulic pump monitoring signal based on the filtering processing method output by the filtering model to obtain a filtered hydraulic pump monitoring signal.

Images