Abrasive particle monitoring method and device and electronic equipment

Abstract

The invention discloses an abrasive particle monitoring method and device and electronic equipment, and belongs to the field of monitoring methods, and the method comprises the steps: collecting a to-be-detected signal based on each probe in an electrostatic sensor; performing variational mode decomposition processing on the to-be-detected signal to obtain a target eigenmode function corresponding to the to-be-detected signal; performing sparse representation based on the target eigenmode function to obtain a target signal corresponding to the to-be-detected signal; performing dynamic time warping on each group of target delay signals by using a sliding window to obtain similarity results respectively corresponding to each group of target delay signals; and if a wave crest meeting a preset condition exists in the sequence diagram presented by the similarity result, determining the number of the abrasive particles according to the wave crest, and obtaining an abrasive particle monitoring result. Through the method, the abrasive particle trigger signal is effectively extracted, the interference of abnormal pulses on abrasive particle identification is reduced, and the accuracy of abrasive particle identification is improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of monitoring methods, and particularly relates to a wear particle monitoring method and device and electronic equipment. BACKGROUND

[0002] The normal operation of gears and bearings and other components of heavy-duty equipment such as aircraft engines, conveyors, coal cutters and gearboxes is the most basic condition to ensure their normal operation. During the operation of the equipment, the lubrication system and its lubrication objects play a very key role in the safe, economic and reliable operation of large and complex rotating machinery systems. When these heavy-duty equipment is operating normally, the components are in an environment of heavy load, high temperature, high pressure and high speed, and failure phenomena such as cracks, pitting and spalling occur on the surface of the components, thereby causing oil line component failure. The most common cause of failure of bearings and gearboxes and other components is the fatigue cracks caused by high contact stress in the contact area, which further causes local damage such as cracks, pitting, spalling and failure on the rolling surface, thereby leaving a huge safety hazard. Under the lubricating oil medium, when a certain type of wear occurs, small particles of material and wear particles, i.e. wear particles, will fall off from the surface of the material. Therefore, the wear condition of the equipment can be understood in a timely manner by monitoring the wear particles in the equipment.

[0003] It can be understood that a double electric layer is spontaneously formed at any solid-liquid interface. In theory, when the fluid motion disturbs the double electric layer between the solid surface and the liquid phase, electric charges are generated. Therefore, a small part of the electric charges from the double layer is carried by the liquid, thereby causing the generation of free static charges. This phenomenon is called tribocharging. According to this mechanism, the wear particles present in the lubricating oil can absorb a certain amount of free static charges. When these charged particles pass through the internal space of the sensor probe, there is a static electric field between the particles and the probe. Due to the free electron flow caused by this field, an induced current is generated inside the probe. The induced current can be filtered and amplified by the acquisition circuit to provide a static voltage signal.

[0004] Therefore, in the related art, a corresponding relationship between the static electric signal and the friction and wear of the components under different working conditions can be established, so as to realize the diagnosis of the faulty components by being sensitive to the early signs of the fault through the static electric signal. However, in the above method, there are often strong noise interferences such as power frequency interference, random white noise and spikes in the collected static electric signal, which causes the extracted wear particle information to be often shallow, resulting in a large error in the number of detected wear particles and the like. SUMMARY

[0005] The purpose of the embodiments of the present application is to provide a wear particle monitoring method and device, electronic equipment, readable storage medium, chip and computer program product, which can effectively extract a wear particle trigger signal, reduce the interference of abnormal pulses on wear particle identification, and improve the accuracy of wear particle identification.

[0006] In a first aspect, embodiments of the present application provide a method for monitoring abrasive particles, the method comprising: collecting, by each probe in an electrostatic sensor, a signal to be detected; performing variational mode decomposition on the signal to be detected to obtain a target eigenmode function corresponding to the signal to be detected; and performing sparse representation based on the target eigenmode function to obtain a target signal corresponding to the signal to be detected; wherein target signals corresponding to two adjacent probes in the electrostatic sensor form a group of target delay signals; performing dynamic time warping on each group of target delay signals by using a sliding window to obtain a similarity result corresponding to each group of target delay signals; if a wave peak meeting a preset condition exists in a sequence diagram presented by the similarity result, determining the number of abrasive particles according to the wave peak to obtain a monitoring result of the abrasive particles.

[0007] In a second aspect, embodiments of the present application provide a device for monitoring abrasive particles, the device comprising: a signal collection module configured to collect, by each probe in an electrostatic sensor, a signal to be detected; a signal decomposition module configured to perform variational mode decomposition on the signal to be detected to obtain a target eigenmode function corresponding to the signal to be detected; a sparse representation module configured to perform sparse representation based on the target eigenmode function to obtain a target signal corresponding to the signal to be detected; wherein target signals corresponding to two adjacent probes in the electrostatic sensor form a group of target delay signals; and a dynamic time warping module configured to perform dynamic time warping on each group of target delay signals by using a sliding window to obtain a similarity result corresponding to each group of target delay signals; a result generation module configured to, if a wave peak meeting a preset condition exists in a sequence diagram presented by the similarity result, determine the number of abrasive particles according to the wave peak to obtain a monitoring result of the abrasive particles.

[0008] In a third aspect, embodiments of the present application provide an electronic device, which comprises a processor and a memory, the memory storing programs or instructions executable on the processor, and the programs or instructions, when executed by the processor, implement the steps of the method according to the first aspect.

[0009] In a fourth aspect, embodiments of the present application provide a readable storage medium, the readable storage medium storing programs or instructions, and the programs or instructions, when executed by a processor, implement the steps of the method according to the first aspect.

[0010] In a fifth aspect, an embodiment of the present application provides a chip, which comprises a processor and a communication interface, the communication interface is coupled with the processor, and the processor is configured to run programs or instructions to implement the method in the first aspect.

[0011] In a sixth aspect, an embodiment of the present application provides a computer program product stored in a storage medium, which is executed by at least one processor to implement the method in the first aspect.

[0012] The embodiment of the present application provides a kind of abrasive grain monitoring method. After each probe in based on electrostatic sensor collects the signal to be detected, the signal to be detected is carried out variational mode decomposition processing, obtains the target eigenfunction corresponding to the signal to be detected, carries out sparse representation based on the target eigenfunction, obtains the target signal corresponding to the signal to be detected, utilizes sliding window to each group target delay signal Dynamic Time Warping is carried out, obtains the similarity result corresponding to each group target delay signal respectively;If there is wave crest that meets preset condition in the sequence chart presented by the similarity result, then abrasive grain quantity is determined according to the wave crest, and abrasive grain monitoring result is obtained. By the above method, the abrasive grain triggered delay signal, i.e. target signal is effectively and cleanly extracted by VMD-SR processing to the signal to be detected, the interference of noise to subsequent signal identification is reduced, and DTW based on sliding window is carried out to two target signals, further reduce the interference of abnormal pulse to abrasive grain identification, improve the accuracy of abrasive grain identification. BRIEF DESCRIPTION OF DRAWINGS

[0013] Figure 1 The flow chart of the abrasive grain monitoring method provided by the embodiment of the present application is shown; Figure 2 The structural connection schematic diagram of the oil pipe and electrostatic sensor provided by the embodiment of the present application is shown; Figure 3 The structural schematic diagram of the electrostatic sensor provided by the embodiment of the present application is shown; Figure 4 The signal decomposition schematic diagram based on variational mode decomposition processing flow provided by the embodiment of the present application is shown; Figure 5 The signal power spectrum contrast diagram provided by the embodiment of the present application is shown; Figure 6 The result change diagram of the DTW based on sliding window provided by the embodiment of the present application is shown; Figure 7 The similarity result schematic diagram provided by the embodiment of the present application is shown; Figure 8 The group of analog signal schematic diagram provided by the embodiment of the present application is shown; Figure 9 This paper shows a signal diagram obtained after processing a set of analog signals provided in an embodiment of this application using VMD-SR. Figure 10 This illustration shows a DTW result diagram based on a processed analog signal provided in an embodiment of this application. Figure 11 A flowchart of another abrasive particle monitoring method provided in an embodiment of this application is shown; Figure 12 This paper illustrates a set of sampling signals provided in an embodiment of this application; Figure 13 This paper shows a signal diagram obtained after a set of sampled signals provided in an embodiment of this application have been processed by VMD-SR. Figure 14 This illustration shows a DTW result diagram based on a processed sampled signal, provided by an embodiment of this application. Figure 15 This paper presents a comparison chart of the number of abrasive particles detected based on different sensors, according to an embodiment of this application. Figure 16 This application provides another comparison chart of the number of abrasive particles detected based on different sensors. Figure 17 This application provides a comparison chart of similarity results based on different methods, as illustrated in an embodiment of the present application. Figure 18 A structural diagram of an abrasive monitoring device provided in an embodiment of this application is shown; Figure 19 A structural diagram of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0014] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0015] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0016] The following description, in conjunction with the accompanying drawings, details a method, apparatus, electronic device, readable storage medium, chip, and computer program product for abrasive particle monitoring provided in this application, through specific embodiments and application scenarios. Method Implementation Examples

[0017] The proper functioning of components such as gears and bearings in heavy-duty equipment like aircraft engines, conveyors, coal mining machines, and gearboxes is the most fundamental condition for their normal operation. During equipment operation, the lubrication system and its lubricated components play a crucial role in the safe, economical, and reliable operation of large, complex rotating machinery systems. Under normal operating conditions, these heavy-duty equipment components operate in environments of heavy load, high temperature, high pressure, and high speed, leading to failures such as cracking, pitting, and spalling on the component surfaces, resulting in oil circuit component malfunctions. The most common failure cause for components such as bearings and gearboxes is fatigue cracking due to high contact stress in the contact area, further causing localized damage to the rolling surfaces, such as cracking, pitting, spalling, and failure, leaving significant safety hazards. In lubricating oil environments, when certain types of wear occur, fine particles of material, known as abrasive grains, detach from the material surface. Therefore, monitoring abrasive grains in the equipment allows for timely understanding of the equipment's wear condition.

[0018] Studies have shown that abrasive particles generated by friction wear in lubrication systems typically carry a certain amount of charge. When these charged particles pass through a sensor, according to the principle of electrostatic induction, some electric field lines caused by the charge terminate on the sensor surface. Electrons in the sensor redistribute to balance the extra charge near the sensor, thus generating a current. This current can be converted into a voltage signal by an appropriate conditioning circuit, which is then collected by the acquisition circuit. Based on this principle, electrostatic sensors do not have special requirements regarding the material type of the abrasive particles and have a high monitoring capability for even minute abrasive particles.

[0019] The specific mechanisms of abrasive nuclear electromechanical interaction in oil circuits and the principles of abrasive electrostatic monitoring are as follows: The mechanism of abrasive particle charging in oil-lubricated media: In nature, electrification mainly arises from the contact and separation of two objects, accompanied by charge transfer. The magnitude of the charge is related to factors such as the Fermi level of the material, environmental factors, and the contact and separation conditions. Friction in a friction pair is essentially a continuous process of contact and separation between two friction surfaces. Under normal conditions, due to the good conductivity of metals, the charge is quickly conducted away after contact and separation, thus metal contacts often exhibit electrical neutrality. However, in lubricating oil media, when a certain type of wear occurs, abrasive particles detach from the material surface, and at the moment of detachment, their surface carries a portion of charge. After detachment, because the abrasive particles are suspended in the insulating lubricating oil, the charge on the abrasive particles does not dissipate and they remain charged. The charging situation of abrasive particles in oil media is quite complex. Studies have shown that when heavy-duty equipment such as engines and gearboxes experience cracks, pitting, spalling, or other failures, the generation of abrasive charge is related to tribo-charging, surface charge, triboemission, and the breaking of metallic bonds during abrasive grain formation. Abrasive grains generated during scuffing wear carry a positive charge, the magnitude of which is directly related to the total volume loss. These charged abrasive grains then enter the lubrication system along with the lubricating oil; this is the fundamental principle behind abrasive grain charging.

[0020] Principle of abrasive electrostatic monitoring: It is known that an electric double layer spontaneously forms at any solid-liquid interface. Theoretically, when fluid movement disturbs the electric double layer between the solid surface and the liquid phase, charge is generated. Therefore, a small portion of the charge from the double layer is carried by the liquid, resulting in the generation of free static charge. This phenomenon is called triboelectric charging. Based on this mechanism, wear particles present in the lubricating oil can absorb a certain amount of free static charge. When these charged particles pass through the internal space of the sensor probe, an electrostatic field exists between the particles and the probe. Due to the free electron flow caused by this field, an induced current is generated inside the probe. This induced current can be filtered and amplified by the acquisition circuit to provide an electrostatic voltage signal. The signal processing module receives the electrostatic signal and then performs information fusion, feature extraction, and noise reduction.

[0021] Therefore, based on the above principles, in related technologies, a correspondence between electrostatic signals and the friction and wear of components under different operating conditions can be established. This aims to diagnose faulty components by using electrostatic signals to detect early signs of failure. Furthermore, electrostatic monitoring technology is a cost-effective and time-saving passive monitoring method because it does not require voltage excitation. This characteristic offers a significant advantage over inductive sensors in monitoring wear particles in coal mine equipment. It is not limited by the maximum equipment voltage and safety standards in mines.

[0022] Based on the aforementioned advantages of wear particle electrostatic monitoring technology, the University of Southampton conducted several investigations to better understand the charging and electrostatic monitoring of wear particles. Harvey studied the influence of lubricating oil quality on triboelectric charging and highlighted the presence of charged wear particles within the lubricating oil. He used charge level measurements to analyze the relationship between wear particle charging and wear rate. Using a pin-disc triboelectric testing apparatus, Morris et al. conducted a series of studies to detect the presence of abnormal electrostatic signals before wear failure. Wang et al. studied electrostatic signals under oil-lubricated contact sliding friction and found a significant correlation between the signal and the coefficient of friction. Mao and Liu explored the application potential of electrostatic sensors in wind turbine gearboxes, and their results showed a close relationship between wear state and electrostatic signal amplitude. Previous studies have repeatedly demonstrated the sensitivity of electrostatic sensors to wear particles in the early stages of wear, proving their practicality in engineering applications.

[0023] In terms of sensing models, measurement applications, and data processing, Yan et al. first established a mathematical model of point charges in the ring probe space, revealing the characteristics of the output signal. They also conducted modeling and simulation studies on sensors of various shapes, successfully verifying the aerodynamic characteristics of the sensor prototype. This research achieved satisfactory application results. Fort et al. studied an electrostatic sensor modeling method for gas turbine monitoring. Addabbo et al. further developed an intelligent measurement system with enhanced low-frequency response to detect moving charged particles. Tang et al. used a compressed sensing method to process signals from array-type electrostatic sensors in gas turbines for signal or data processing. This method effectively achieved in-situ particle inversion based on the processing results; however, the computational cost of this method is too high, resulting in low efficiency.

[0024] Furthermore, the aforementioned methods often suffer from strong noise interference, such as power frequency interference, random white noise, and spikes, in the electrostatic signals triggered by abrasive particles. This results in superficial abrasive particle information, leading to significant errors in the detected number of abrasive particles and hindering accurate identification of wear particle array electrostatic probes. Moreover, there are few publicly available and reliable identification techniques or tested experiments. This makes accurate identification and statistical analysis of wear particles through electrostatic signals difficult and challenging in engineering applications.

[0025] To address the aforementioned issues, the equipment used in the simulation experiments and implementation of the scheme described in this application includes, but is not limited to, dampers, gearboxes, motors, oil pumps, oil pipes, signal lines, sampling pumps, electrostatic sensors, inductive sensors, controllers, data acquisition units, and physicochemical parameter sensors. Of course, in the specific implementation process, the above equipment can be adjusted according to actual needs, and this application does not limit this.

[0026] ]Reference Figure 1 The flowchart illustrates a wear particle monitoring method provided in an embodiment of this application. Specific steps may include: Step 101: Collect the signal to be detected based on each probe in the electrostatic sensor.

[0027] Understandably, because the charge on abrasive particles is very small, the detection device needs high sensitivity and good anti-interference ability; and the lubrication system itself has high requirements for the flow characteristics of lubricating oil, and there should be no obstacles to the flow; at the same time, the diameter of the lubrication pipeline is usually very small (less than 50mm). If the probe of the electrostatic sensor adopts a ring structure, it can not only better distinguish and compare the electrostatic signal and noise in the abrasive particles, but also facilitate installation and disassembly; the ring probe makes full use of the area of ​​the ceramic pipe, making the monitoring signal more accurate.

[0028] Therefore, refer to Figure 2 The diagram shown is a structural connection diagram of an oil pipe and an electrostatic sensor provided in this application embodiment. In this embodiment, the probes of the oil circuit electrostatic sensor can adopt a non-contact ring structure and be arranged coaxially. This application does not limit the number of probes. When abrasive particles pass through the oil circuit, they trigger charge movement in the probes, thereby forming an induced current within the probes. This induced current can be filtered and amplified by the acquisition circuit to provide an electrostatic voltage signal, which can also be called a particle-triggered delay signal or a particle-triggered pulse signal. Simultaneously, considering different forms of noise interference, the signal to be detected is essentially composed of various noise interference signals and the particle-triggered delay signal acquired based on the electrostatic induction principle of abrasive particles.

[0029] Specifically, refer to Figure 3 The schematic diagram shown in this application illustrates the structure of an electrostatic sensor. When charged abrasive particles pass through each coaxial probe, a delay signal is triggered. Simultaneously, each channel corresponds to one probe and one delay signal, allowing those skilled in the art to obtain the corresponding delay signal through each channel. The time delay between the two delay signals is determined by the oil speed and the distance L between the two probes.

[0030] After acquiring the signal to be detected, based on the correspondence between the electrostatic signal and the friction and wear of the parts under different working conditions, the wear particles can be correctly identified and the wear condition of the corresponding device can be determined by denoising, feature extraction, and recognition of the signal to be detected.

[0031] In addition, based on the principle of interaction between charged particles in the oil circuit, electrostatic sensors can actually be divided into two types: direct contact and non-contact. In specific experiments and data processing, technicians can select electrostatic sensors according to specific circumstances and actual effects. This application does not limit the use of only non-contact probes.

[0032] Step 102: Perform variational mode decomposition on the signal to be detected to obtain the target intrinsic mode function corresponding to the signal to be detected.

[0033] Variational Mode Decomposition (VMD) is an adaptive, fully non-recursive variational mode and signal processing method. It adaptively decomposes the signal to be detected into a set of ideal Intrinsic Mode Functions (IMFs). These IMFs differ in that they possess the minimum combined spectral bandwidth while satisfying the requirement that their temporal superposition equals the unprocessed signal. Using VMD to decompose the signal to be detected effectively filters out noise components. Furthermore, by identifying a target IMF from a set of IMFs to replace the signal to be detected in subsequent processing, it effectively reduces signal sparsity, thereby reducing the computational load in signal recognition and improving computational efficiency.

[0034] Reference Figure 4 The diagram shown is a signal decomposition schematic based on variational mode decomposition processing flow provided in this application embodiment. Figure 4 -(a) represents the original signal, which includes various noise interference signals and particle triggering signals carrying abrasive grain information. After performing VMD processing on the original signal, the following can be obtained: Figure 4 -(b) shows the six IMFs. Among them, IMF1-IMF5 are noise signals. And as... Figure 4 As shown in (c), IMF6 has sparser features than the original signal to be detected, and Figure 4 The signal segment features shown in the elliptical region of -(c) are more pronounced, and the signal-to-noise ratio is relatively lower. Therefore, it can be determined that IMF6 is the target modulus function to replace the signal to be detected for subsequent processing. However, there are still some issues in the IMF. Figure 4 The signal drift and residual noise shown in the box area of ​​-(c) still require a subsequent noise reduction process to reduce interference in the recognition process and improve the accuracy of abrasive particle recognition.

[0035] Optionally, to improve the accuracy of abrasive particle identification, step 102 involves performing variational mode decomposition on the signal to be detected to obtain the target intrinsic mode function corresponding to the signal to be detected, including: Step S11: Perform DC removal processing on the signal to be detected to obtain the signal to be decomposed; Step S12: Perform variational mode decomposition on the signal to be decomposed to obtain the intrinsic mode function corresponding to the signal to be detected; Step S13: Determine the target intrinsic modulus function from the intrinsic modulus functions; The penalty factor for the variational mode decomposition process is 2000, and the number of decompositions is 6.

[0036] With a VMD penalty factor of 2000 and a decomposition number of 6, it exhibits good signal decomposition and noise reduction effects, effectively extracting particle trigger signals carrying abrasive grain information. Of course, the specific parameter settings depend on the actual results, and this application does not impose any limitations on them.

[0037] Regarding step S13, in the embodiments of this application, a suitable intrinsic mode function can be determined by setting thresholds such as signal-to-noise ratio and power, or it can be determined manually. This application does not impose any limitations on this method.

[0038] Step 103: Perform sparse representation based on the target intrinsic mode function to obtain the target signal corresponding to the signal to be detected; wherein, the target signals corresponding to two adjacent probes in the electrostatic sensor constitute a set of target delay signals.

[0039] Sparse representations (SR) are an important signal processing method that aims to represent a signal with as few atoms as possible from a given overcomplete dictionary. This results in a more concise representation of the signal, making it easier for technicians to extract the information contained in the signal and to further process it.

[0040] In the embodiments of this application, sparse representation can be performed using analytical dictionaries such as wavelet dictionaries, overcomplete DCT dictionaries, and curvelet dictionaries. Alternatively, dictionary learning can be performed using methods such as the Method of Optimal Directions (MOD) algorithm and the FOCUSS dictionary learning algorithm; this application does not impose any limitations on this approach. Furthermore, in the embodiments of this application, the most relevant basis functions can be selected from the determined dictionary to represent the target intrinsic mode function using methods such as the Orthogonal Matching Pursuit (OMP) algorithm and the Subspace Pursuit (SP) algorithm; this application also does not impose any limitations on this approach.

[0041] Reference Figure 5 The above-described embodiment of this application provides a signal power spectrum comparison diagram, wherein... Figure 5 -(a) represents the original signal to be detected and the signal to be detected after VMD-SR processing. Figure 5 -(b) is the power spectrum of the original signal to be detected. Figure 5 -(c) is the power spectrum of the signal to be detected after VMD-SR processing. Figure 5 -(b) and Figure 5 -(c) Comparison shows that the power of the processed signal is mainly concentrated in the lower frequency band, while the power of the noisy original signal is not concentrated, indicating that it has a better denoising effect.

[0042] Therefore, firstly, following the method described in step 102, VMD decomposition is used to reduce noise components, giving it a sparser characteristic than the original signal. Then, the VMD decomposition result, i.e., the target intrinsic mode function, is used as the input to the SR according to the method described in step 103 to obtain the target signal corresponding to the signal to be detected. Compared with directly inputting the original signal, the computational burden of the process is effectively reduced. Therefore, a higher signal-to-noise ratio is achieved along a cleaner useful pulse. This method reduces interference in abrasive particle identification and improves the accuracy of abrasive particle identification.

[0043] Optionally, step 103, which involves performing sparse representation based on the target intrinsic modulus function to obtain the target signal corresponding to the signal to be detected, includes: Step S21: Use orthogonal matching pursuit to perform sparse representation of the target intrinsic modulus function to obtain the target signal corresponding to the signal to be detected.

[0044] Specifically, to handle electrostatic signals with high sampling rates and large amounts of data, i.e., the target eigenfunction, an approximation algorithm can be used. A wavelet dictionary is selected as the complete dictionary. By employing the OMP algorithm, the most relevant basis functions are selected from the wavelet dictionary to represent the target eigenfunction.

[0045] Step 104: Use a sliding window to perform dynamic time warping on each group of target delay signals to obtain the similarity results corresponding to each group of target delay signals.

[0046] Dynamic Time Warping (DTW) can calculate the similarity between two time series, and is particularly suitable for time series of different lengths and rhythms (such as audio sequences of different people reading the same word). DTW automatically warps the time series (i.e., performs local scaling on the time axis) to make the two sequences as consistent in shape as possible, thus obtaining the maximum possible similarity. Unlike traditional distance metrics based on Euclidean or Mahala-Nobis distance, DTW has a stronger ability to identify waveform similarity.

[0047] Specifically, refer to Figure 6 The embodiment of this application provides a result variation diagram of DTW based on a sliding window. The process of dynamically time-normalizing two delayed signals using a sliding window is as follows: (1) As Figure 6 As shown in (a), when the levels of the two signals have not yet changed, that is, when the sliding window of the two dual-channel signals has not yet reached the particle signal, the amplitude of the window slice signal remains close to zero. Performing DTW on these two signal segments should result in a result close to zero. (2) Figure 6 -(b) As shown in the upper part, when window 1 reaches the particle signal in channel 1, the amplitude of the window slice signal in channel 2 remains zero. The DTW results of the two slice signals will start to generate non-zero values. Since window 2 has not yet reached the pulse in channel 2, the DTW results will gradually increase. (3) Figure 6 As shown in the upper part of (b), when window 2 reaches the delayed particle signal while window 1 still contains a portion of the pulse, the DTW result will not be zero. As the window continues to slide, window 2 gradually covers the particle signal in channel 2, resulting in greater similarity in the shape of the window slice signals. The DTW calculation result decreases, and eventually, there will be two positions where two windows completely contain both pulses. At this point, the DTW calculation should reach a local minimum, i.e., as shown in (b). Figure 6 The segment signal shown in -(c); (4) Figure 6 As shown in the upper part of (d), as the two windows continue to slide, the similarity between the two slice signals gradually decreases, resulting in an increase in the DTW calculation value; (5) Figure 6 As shown in the lower half of (d), as the two sliding windows move away from the particle signal in turn, the DTW result gradually decreases to 0, a process that is symmetrical to the process shown in (1)-(2).

[0048] Optionally, the electrostatic sensor is used to perform electrostatic monitoring of abrasive particles in the target oil passage; step 104, which involves using a sliding window to dynamically time-normalize each group of target delay signals to obtain similarity results for each group of target delay signals, includes: Step S31: Obtain the oil flow rate in the target oil circuit and the structural parameters of the probe; Step S32: Determine the width and step size of the sliding window based on the oil flow rate and the classical parameters; Step S33: Perform dynamic time warping on each group of target delay signals using a sliding window constructed based on the width and the step size to obtain the similarity results corresponding to each group of target delay signals.

[0049] The structural parameters of the probe can be obtained from the probe manufacturer's specifications or from publicly available datasets, and the oil flow rate of the target oil circuit can be obtained from the monitoring devices in the equipment or from calculations. This application does not limit either of these.

[0050] It is understood that a sliding window is a set of elements defined by dynamic boundaries. Therefore, the width of the sliding window refers to the size defined by the boundaries of the sliding window, and the magnitude of the boundary change is called the step size, i.e., the length of each slide. In the embodiments of this application, the width and step size of the sliding window can be determined according to the structural parameters of the probe and the oil velocity in the target oil path. This avoids the DTW results being too coarse due to inappropriate width and step size of the sliding window, refines the DTW results of the target delay signal, and thus improves the accuracy of abrasive particle identification.

[0051] Step 105: If there is a peak in the sequence graph presented by the similarity result that meets the preset conditions, then determine the number of abrasive particles based on the peak and obtain the abrasive particle monitoring result.

[0052] Reference Figure 7 The illustration shown is a similarity result diagram, i.e., a DTW result, provided by an embodiment of this application. The bimodal shape corresponds to a charged particle passing through two probes. Based on this, in this embodiment, three or more probes can be introduced for signal acquisition. Then, through steps 102-103, the target signal corresponding to each probe is obtained, and DTW is performed between each pair of target signals. By extending the time period, a bimodal structure is searched in the similarity results corresponding to adjacent probes to determine the presence of abrasive particles and improve the accuracy of abrasive particle identification.

[0053] As an example, when an additional sensor probe is introduced, generating three channels of output signal, performing DTW calculations between the middle probe signal and the signals from the two side probes will produce two DTW sequences, resulting in two bimodal structures. Conversely, if at similar locations, one DTW result shows a bimodal structure and the other a single peak, it indicates that at the corresponding time point, one channel's signal has an unexpected pulse, interfering with the identification of abrasive particles. Therefore, by using multiple probes in a coaxial array sensor, the presence of particles can be determined more reliably, thereby improving the accuracy of identification.

[0054] Based on the above, in this embodiment, to avoid the influence of abnormal pulse signals, the detection of a single abrasive grain passing through two probes is taken as the standard. That is, when a double-peak shape is displayed in the DTW result, the number of abrasive grains is determined based on the number of double-peak shapes, and the corresponding device damage condition, i.e., the abrasive grain monitoring result, is determined based on the specific parameters of the double-peak shapes. Therefore, the preset condition described in this application can be that the peak shape is a double-peak shape. Of course, depending on the different abrasive grain monitoring scenarios, different preset conditions can be determined according to the actual situation in the specific implementation process, such as two independent but relatively close single peaks. This application does not limit this.

[0055] Furthermore, in this embodiment, the criteria for determining the bimodal shape can be: the two peaks are relatively close and there is a local minimum between them. Both the proximity of the peaks and the local minimum can be determined by setting thresholds. Of course, the bimodal shape can also be determined by manual comparison; this application does not limit this approach.

[0056] In summary, this application provides a method for abrasive particle monitoring. Based on the signals collected by various probes in an electrostatic sensor, variational mode decomposition (VMD) is performed on the signals to be detected to obtain the target intrinsic mode functions (IEMs) corresponding to the signals. Sparse representation is then performed on the IEMs to obtain the target signals corresponding to the signals to be detected. Dynamic time warping (DTW) is applied to each group of target delay signals using a sliding window to obtain similarity results for each group of target delay signals. If a peak meeting preset conditions exists in the sequence graph presented by the similarity results, the number of abrasive particles is determined based on the peak, thus obtaining the abrasive particle monitoring result. Through the above method, VMD-SR processing of the signals to be detected effectively and cleanly extracts the delayed signals triggered by abrasive particles, i.e., the target signals, reducing noise interference for subsequent signal recognition. Furthermore, DTW based on a sliding window is applied to the two target signals to further reduce the interference of abnormal pulses on abrasive particle recognition and improve the accuracy of abrasive particle recognition.

[0057] Optionally, step 105, if the sequence graph presented by the similarity result contains a peak that meets preset conditions, then the number of abrasive particles is determined based on the peak to obtain the abrasive particle monitoring result, including: Step S41: When the number of probes is 2, if the sequence graph presented by the similarity result has a bimodal shape, the number of abrasive particles is determined according to the number of bimodal shapes.

[0058] When the number of probes is 2, meaning there are only two signals to be detected, there is only one DTW result, i.e., a similarity result. Therefore, according to the aforementioned... Figure 7 In this embodiment of the application, the number of bimodal shapes is equal to the number of abrasive grains passing through the electrostatic sensor.

[0059] Specifically, with a probe count of 2, the embodiments of this application demonstrate the superior performance of VMD-SR-DTW for abrasive particle identification through the following processed data: Reference Figure 8 The illustration shows a set of analog signal diagrams provided in the embodiments of this application. The analog signals of channel 1 and channel 2 are generated by combining sinusoidal pulses with random signals, which can simulate noisy particle signals. At the same time, additional abnormal amplitude spikes are introduced in channel 2 without adding them to the corresponding time points in channel 1. Figure 8 The signal segment shown within the elliptical region is the segment with more obvious characteristics in the channel 1 signal, while the remaining region mainly represents the channel 2 signal.

[0060] by Figure 8 Taking the analog signal shown as an example, refer to Figure 9 It shows a signal diagram obtained after processing a set of analog signals provided in an embodiment of this application using VMD-SR, wherein, Figure 9 -(a) is for Figure 5 The image shown is a signal obtained after VMD processing of the analog signal. Figure 9 -(b) is for Figure 9 -The signal diagram shown in (a) is obtained after SR processing. Figure 9 -(b) and Figure 9 -(a) It can be clearly seen that although some random noise remains after VMD processing, the random noise has been successfully removed after SR processing, leaving only the particle trigger delay signal and abnormal amplitude spikes.

[0061] This highlights the powerful capabilities of VMD-SR in identifying particle-triggered delay signals and unexpected pulses (such as anomalous amplitude spikes). Therefore, signal denoising using VMD-SR not only enables accurate identification of particle-triggered delay signals but also helps filter out anomalous amplitude pulses triggered by non-wear particles, reducing the error in abrasive particle identification and thus further improving the accuracy of abrasive particle identification.

[0062] Finally, refer to Figure 10 The illustration shows a DTW result diagram based on a processed analog signal provided in this application embodiment. The presence of the added spike is clearly visible, and it is this spike that causes the singular peak in the DTW sequence. Furthermore, the two delayed signals result in two bimodal shapes in the DTW result, meaning that two abrasive particles pass through within the time interval of the two delayed signals. This observation highlights the powerful ability of the proposed method to identify particle-triggered delayed signals and unexpected pulses (such as anomalous amplitude spikes).

[0063] Optionally, step 105, if the sequence graph presented by the similarity result contains a peak that meets preset conditions, then the number of abrasive particles is determined based on the peak to obtain the abrasive particle monitoring result, including: Step S51: When the number of probes is 3, if both sets of target delay signals have a bimodal shape in the sequence diagrams of the similarity results, the number of abrasive particles is determined according to the position of the bimodal shape in the similarity results.

[0064] Specifically, refer to Figure 11 The embodiment of this application provides a DTW result diagram based on a processed analog signal, and the specific steps include: (1) Obtain the signal to be detected y1 from the first detection channel according to the method described in step 101, and remove the DC component; (2) Perform VMD on y1 according to the method described in step S12, with a penalty factor of 2000, and decompose it into 6 modes; (3) Extract the 6th IMF as the initial denoising signal, that is, determine the 6th IMF as the target intrinsic mode function according to the method described in step S13; (4) Construct a complete dictionary according to the method described in step 103, and perform sparse representation on the initial denoised signal with a sparsity of 5; reconstruct the particle signal after sparse representation to obtain the secondary denoised signal sig1. (5) Input the signals of the second and third probes corresponding to the same time period, and repeat steps (1) to (4) to obtain sig2 and sig3; (6) Set the width and step size of the sliding window according to the probe structure parameters and oil flow rate as described in step S31; (7) Following the method described in step 104, simultaneously advance two sliding windows in sig1 and sig2, generating a pair of amplitude-limiting signals and a DTW calculation value at each step. A DTW result sequence can be obtained, marked by a distance of 1. (8) Repeat process (7) between sig2 and sig3 to obtain distance 2; (9) Search for a bimodal shape at distance 1. The following criteria can be applied to identify wear particle signals: find two peaks that are very close to each other, and the two peaks have a local minimum. If these conditions are met, it is likely caused by wear particles. Search for a bimodal shape at distance 2. Specifically, the method described in step 105 can be used to determine whether the peaks are close and whether a local minimum exists; (10) Compare the results of distances 1 and 2. If a bimodal shape is found in both sequences and their positions are close, it can be further confirmed that the abrasive has passed through. Of course, in order to further improve the accuracy of abrasive identification, the presence of abrasive can also be determined based on whether the bimodal shapes are the same or similar. This application does not limit the comparison. (11) Increment the abrasive count by 1 and input the multi-channel signal for the next time period.

[0065] As an example, with a probe count of 3, the embodiments of this application demonstrate the excellent performance of VMD-SR-DTW for abrasive particle identification through the following experimental data.

[0066] Reference Figure 12 The schematic diagram of a set of sampling signals provided in the embodiment of this application shows the monitoring data of the original signals of three channels for about 40 seconds. Due to excessive interference caused by noise signals, abnormal pulses, etc., the signal diagram will not be described in detail.

[0067] Reference Figure 13 The image shown is a signal diagram obtained after processing a set of sampled signals using VMD-SR, as provided in this embodiment of the application. Figure 13 -(a) shows the signal graph after VMD-SR processing of three signals: signal 1 (sig1), signal 2 (sig2), and signal 3 (sig3). The vertical axis is in units of 10— 4 V-shaped image, Figure 13 -(b) shows Figure 13 -A magnified detail of the signal segment in (a). Figure 13 In (b), the curve describing signal 2 is the lightest, the curve describing signal 3 is the darkest, and the curve describing signal 1 is in the middle. In the embodiments of this application, this can be specifically utilized... Figure 13 -(b) Auxiliary Viewing Figure 13 -(a).

[0068] likeFigure 13 As shown, VMD-SR processing successfully eliminated a large amount of noise while retaining the particle-triggered pulses, resulting in a higher signal-to-noise ratio and reducing the error in subsequent abrasive particle identification.

[0069] After denoising, DTW calculation based on a sliding window is performed between sig1, sig2, and sig3, referring to... Figure 14 This illustrates a DTW result diagram based on a processed sampled signal, provided in an embodiment of this application. Figure 14 -(a) is the DTW result graph with a detection duration of 40s. Figure 14 -(b) is Figure 14 - (a) Enlarged images of regions 1, 2, and 3. Within a local area of ​​the two images, the DTW result curve between sig2 and sig3 lags behind the DTW result curve between sig1 and sig2 in position.

[0070] according to Figure 14 As shown, the DTW sequence obtained from sig1 and sig2 contains 13 bimodal shapes, while the DTW sequence obtained from sig2 and sig3 contains 12 bimodal shapes and one unimodal shape.

[0071] By comparison Figure 13 The signal images show that sig1 and sig2 both have pulse signals around 32s, while sig3 does not. Therefore, sig2 and sig3 do not have a double-peak shape near the 800th order. Thus, it can be inferred that the pulse signal around 32s is not a wear-particle-triggered signal, meaning that a total of 12 wear particles passed through the electrostatic sensor during this period. This indicates that... Figure 11 The illustrated abrasive monitoring method demonstrates its effectiveness in capturing time-delay signals and filtering out unexpected signals.

[0072] Meanwhile, to verify the above observations, we retrieved monitoring data from the inductive sensor during this period, with a data refresh rate of 0.1 Hz. The comparison results are as follows... Figure 15 As shown, a statistical error of 1 exists between the two types of sensors during the first and fourth data refresh cycles. Specifically, within a 30–40 s time interval, the electrostatic sensor detects two particles, while the inductive sensor detects only one. This can be verified by examining... Figure 14 In the magnified details in (b), we observed two very close bimodal shapes, indicating that the two particles were likely very close to each other. However, due to the limited capability of continuous particle detection, the sensor could only identify one particle at a time.

[0073] To further evaluate the differences between electrostatic and inductive sensors in abrasive particle monitoring, this embodiment of the application performs random data extraction from multiple sampling stages to identify abrasive particles within a consistent time span of 40 seconds. (Refer to...) Figure 16 The illustrated embodiment of this application provides another comparison chart of the number of abrasive particles detected based on different sensors, which shows the difference between the identification results obtained by the method described in this application and the monitoring results from the sensing sensors.

[0074] Although some differences exist in the monitoring results based on the two sensors, they exhibit similar trends. Furthermore, the electrostatic sensor detects a slightly higher number of particles compared to the inductive sensor. Therefore, the method described in this application demonstrates higher sensitivity and accuracy compared to abrasive particle monitoring methods utilizing inductive sensors. It is understandable that achieving absolute accuracy in particle measurement is challenging with any type of sensor. However, multiple sampling surveys have shown that the method described in this application can still provide a relatively accurate assessment of the number of abrasive particles, thus offering a reliable approach for subsequent engineering practice.

[0075] In addition, to verify Figure 1 The advantages of the method in abrasive particle identification are illustrated in this application embodiment, which uses the original signals from two channels as an example and conducts four comparative experiments. (Refer to...) Figure 17 The embodiment of this application shows a comparison chart of similarity results based on different methods, wherein... Figure 17 -(a) is the similarity result sequence 1 obtained after directly performing DTW on the original signal. Figure 17 -(b) is the similarity result sequence 2 obtained by performing DTW on the original signal after VMD processing. Figure 17 -(c) is the similarity result sequence 3 obtained after performing DTW on the original signal after wavelet decomposition. Figure 17 -(d) is the similarity result sequence 4 obtained after the original signal has been processed by the VMD-SR-correlation coefficient method.

[0076] Depend on Figure 17 As can be seen, sequences 1-3 reveal that the computational results obtained from the original signals exhibit numerous interference peaks, and the representation of wear particles does not perfectly exhibit a bimodal shape. The DTW results obtained solely through VMD denoising also fail to identify all bimodal shapes, as some wear particle signals appear as single peaks in the DTW results. Similarly, the DTW results after wavelet decomposition denoising also exhibit the same multi-peak aggregation characteristic, resulting in non-uniform waveform shapes. By comparing sequence 4 with the results obtained from the VMD-SR-DTW model, it is observed that the correlation coefficient method cannot identify specific regular shapes in a pair of time-delayed signals. These results all demonstrate that... Figure 1The wear particle monitoring method shown is important and has advantages. It can effectively remove interference from factors such as noise signals and abnormal pulses, and has high accuracy.

[0077] Therefore, the abrasive particle detection method described in this application achieves better results in enhancing the wear particle signal compared to using wavelet or VMD decomposition alone. Similarly, when waveform shape indices are involved, the results obtained using the DTW algorithm are superior to those obtained using the correlation coefficient method. Device Examples

[0078] Reference Figure 18 The diagram shows a structural diagram of an abrasive monitoring device according to an embodiment of this application. The device 200 may include: a signal acquisition module 201, used to acquire the signal to be detected based on each probe in the electrostatic sensor; Signal decomposition module 202 is used to perform variational mode decomposition on the signal to be detected to obtain the target intrinsic mode function corresponding to the signal to be detected; The sparse representation module 203 is used to perform sparse representation based on the target intrinsic mode function to obtain the target signal corresponding to the signal to be detected; wherein, the target signals corresponding to two adjacent probes in the electrostatic sensor constitute a set of target delay signals; The dynamic time warping module 204 is used to perform dynamic time warping on each group of target delay signals using a sliding window, so as to obtain the similarity results corresponding to each group of target delay signals. The result generation module 205 is used to determine the number of abrasive particles based on the peaks in the sequence graph presented by the similarity results if there are peaks that meet preset conditions, and to obtain abrasive particle monitoring results.

[0079] Optionally, the signal decomposition module may include: The DC removal submodule is used to perform DC removal processing on the signal to be detected to obtain the signal to be decomposed. The decomposition submodule is used to perform variational mode decomposition on the signal to be decomposed to obtain the intrinsic mode function corresponding to the signal to be detected. The function determination submodule is used to determine the target intrinsic modulus function from the intrinsic modulus functions; The penalty factor for the variational mode decomposition process is 2000, and the number of decompositions is 6.

[0080] Optionally, the sparse representation module may include: The orthogonal matching pursuit submodule is used to perform sparse representation of the target intrinsic modulus function using the orthogonal matching pursuit method to obtain the target signal corresponding to the signal to be detected.

[0081] Optionally, the electrostatic sensor is used to perform electrostatic monitoring of abrasive particles in the target oil passage.

[0082] The dynamic time warping module may include: The parameter acquisition submodule is used to acquire the oil flow rate in the target oil circuit and the structural parameters of the probe. The parameter determination submodule is used to determine the width and step size of the sliding window based on the oil flow rate and the classical parameters. The dynamic time warping submodule is used to perform dynamic time warping on each group of target delay signals based on a sliding window constructed based on the width and the step size, so as to obtain the similarity results corresponding to each group of target delay signals.

[0083] Optionally, the preset condition includes the shape of the peak being a bimodal shape.

[0084] Optionally, the result generation module may include: The first quantity determination submodule is used to determine the number of abrasive particles based on the number of bimodal shapes in the sequence graph presented by the similarity results when the number of probes is 2.

[0085] Optionally, the result generation module may include: The second quantity determination submodule is used to determine the number of abrasive particles based on the position of the bimodal shape in the similarity results when the number of probes is 3, if both sets of target delay signals have a bimodal shape in the sequence graphs.

[0086] Optionally, the probe has a non-contact ring structure.

[0087] In summary, this application provides an abrasive particle monitoring device. After acquiring the signal to be detected based on each probe in the electrostatic sensor, the signal to be detected is subjected to variational mode decomposition (VMD) to obtain the target intrinsic mode function (IEM) corresponding to the signal to be detected. Based on the IEM, sparse representation is performed to obtain the target signal corresponding to the signal to be detected. Dynamic time warping (DTW) is performed on each group of target delay signals using a sliding window to obtain the similarity results corresponding to each group of target delay signals. If there is a peak in the sequence graph presented by the similarity results that meets preset conditions, the number of abrasive particles is determined based on the peak to obtain the abrasive particle monitoring result. Through the above method, VMD-SR processing of the signal to be detected effectively and cleanly extracts the delay signal triggered by the abrasive particles, i.e., the target signal, reducing the interference of noise on subsequent signal recognition. Furthermore, DTW based on a sliding window is performed on the two target signals to further reduce the interference of abnormal pulses on abrasive particle recognition and improve the accuracy of abrasive particle recognition.

[0088] The abrasive particle monitoring device in this application embodiment can be an electronic device or a component within an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices besides a terminal. As an example, the electronic device can be a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM, or self-service machine, etc. This application embodiment does not specifically limit the device.

[0089] The abrasive particle monitoring device in this application embodiment can be a device with an operating system. This operating system can be Android, iOS, or other possible operating systems; this application embodiment does not specifically limit it.

[0090] The abrasive monitoring device provided in this application embodiment can realize the various processes implemented in the above embodiments. To avoid repetition, it will not be described again here.

[0091] Optionally, such as Figure 19 As shown, this application embodiment also provides an electronic device 500, including a processor 501 and a memory 502. The memory 502 stores a program or instructions that can run on the processor 501. When the program or instructions are executed by the processor 501, they implement the various steps of the above-described wear particle monitoring method embodiment and can achieve the same technical effect. To avoid repetition, they will not be described again here.

[0092] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0093] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described wear particle monitoring method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0094] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0095] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described wear particle monitoring method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0096] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0097] This application provides a computer program product, which is stored in a storage medium and executed by at least one processor to implement the various processes of the above-described wear particle monitoring method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0098] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0099] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0100] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for monitoring abrasive particles, characterized in that, The method includes: The signal to be detected is acquired by each probe in the electrostatic sensor; The signal to be detected is subjected to variational mode decomposition to obtain the target intrinsic mode function corresponding to the signal to be detected; Based on the sparse representation of the target intrinsic mode function, the target signal corresponding to the signal to be detected is obtained; wherein, the target signals corresponding to two adjacent probes in the electrostatic sensor constitute a set of target delay signals; By using a sliding window to perform dynamic time warping on each group of target delay signals, similarity results corresponding to each group of target delay signals are obtained; If a peak that meets a preset condition exists in the sequence graph presented by the similarity results, the number of abrasive particles is determined based on the peak, and the abrasive particle monitoring result is obtained.

2. The method according to claim 1, characterized in that, The step of performing variational mode decomposition on the signal to be detected to obtain the target intrinsic mode function corresponding to the signal to be detected includes: The signal to be detected is subjected to DC removal processing to obtain the signal to be decomposed; The signal to be decomposed is subjected to variational mode decomposition to obtain the intrinsic mode function corresponding to the signal to be detected; The target intrinsic modulus function is determined from the intrinsic modulus functions; The penalty factor for the variational mode decomposition process is 2000, and the number of decompositions is 6.

3. The method according to claim 1, characterized in that, The step of obtaining the target signal corresponding to the signal to be detected by performing sparse representation based on the target intrinsic mode function includes: The target intrinsic modulus function is sparsely represented by orthogonal matching pursuit to obtain the target signal corresponding to the signal to be detected.

4. The method according to claim 1, characterized in that, The electrostatic sensor is used for electrostatic monitoring of abrasive particles in the target oil circuit; the dynamic time warping of each group of target delay signals using a sliding window to obtain the similarity results corresponding to each group of target delay signals includes: Obtain the oil flow rate in the target oil circuit and the structural parameters of the probe; The width and step size of the sliding window are determined based on the oil flow rate and the classical parameters. The sliding window constructed based on the width and the step size is used to perform dynamic time warping on each group of target delay signals to obtain the similarity results corresponding to each group of target delay signals.

5. The method according to claim 1, characterized in that, The preset condition includes that the shape of the wave peak is a bimodal shape.

6. The method according to claim 5, characterized in that, If the sequence graph presented by the similarity results contains a peak that meets preset conditions, then the number of abrasive particles is determined based on the peak to obtain the abrasive particle monitoring result, including: When the number of probes is 2, if the sequence graph presented by the similarity results has a bimodal shape, the number of abrasive particles is determined according to the number of bimodal shapes.

7. The method according to claim 5, characterized in that, If the sequence graph presented by the similarity results contains a peak that meets preset conditions, then the number of abrasive particles is determined based on the peak to obtain the abrasive particle monitoring result, including: When the number of probes is 3, if both sets of target delay signals present a bimodal shape in the sequence diagram of the similarity results, the number of abrasive particles is determined according to the position of the bimodal shape in the similarity results.

8. The method according to claim 1, characterized in that, The probe has a non-contact ring structure.

9. An abrasive particle monitoring device, characterized in that, The device includes: The signal acquisition module is used to acquire the signal to be detected based on the various probes in the electrostatic sensor. The signal decomposition module is used to perform variational mode decomposition on the signal to be detected to obtain the target intrinsic mode function corresponding to the signal to be detected. The sparse representation module is used to perform sparse representation based on the target intrinsic mode function to obtain the target signal corresponding to the signal to be detected; wherein, the target signals corresponding to two adjacent probes in the electrostatic sensor constitute a set of target delay signals; The dynamic time warping module is used to perform dynamic time warping on each group of target delay signals using a sliding window, and obtain the similarity results corresponding to each group of target delay signals. The result generation module is used to determine the number of abrasive particles based on the peaks that meet preset conditions in the sequence graph presented by the similarity results, and to obtain the abrasive particle monitoring results.

10. An electronic device, characterized in that, It includes a processor and a memory, the memory storing a program or instructions that can run on the processor, the program or instructions being executed by the processor to implement the steps of the abrasive monitoring method as described in any one of claims 1-8.

Images