A novel method for counting the number of particles of a capacitance inductance type abrasive grain sensor

Abstract

The application discloses a novel capacitive inductive abrasive particle sensor particle number counting method, relates to the technical field of abrasive particle detection, and comprises the following steps: constructing a signal base function library X, which is used for comparing a detected signal with standard templates in the library; initializing parameters, which are used for unifying signal processing benchmarks; judging residual signals, which can accurately determine whether other size abrasive particles exist and avoid misjudgment and missed detection; performing inner product operation, which can reflect the similarity between residual signals and each base function; determining optimal base functions, which accurately lock the abrasive particle characteristic templates most related to the current residual signals from the residual base function library; updating parameters, which are used for stripping the identified abrasive particle signals and reducing the base function library; and performing iterative operation and calculating the number of abrasive particles. The counting method can accurately distinguish the specific number of abrasive particles with different diameters by constructing a signal base function library, using a greedy algorithm to iteratively decompose and optimize the detected signal, and solves the technical defects that traditional sensors can only count the total number of abrasive particles.

Description

Technical Field

[0001] This invention relates to the field of abrasive particle detection technology, specifically a novel particle counting method using a capacitive-inductive abrasive particle sensor. Background Technology

[0002] In industrial production, transportation, aerospace, and other fields, the wear condition of mechanical equipment directly affects its operational safety and service life. Capacitive-inductive abrasive sensors, a commonly used wear condition monitoring device, work on the principle that abrasive particles cause changes in capacitance or inductance as they pass through the detection area. The sensor captures these signal fluctuations through a high-frequency oscillation circuit, thereby detecting the abrasive particles. These sensors can monitor the abrasive particles in oil or airflow in real time, providing data support for mechanical health assessment. They are widely used in industrial equipment maintenance, automotive engine monitoring, and aerospace component inspection, playing a crucial role in preventing equipment failures and optimizing maintenance cycles.

[0003] For example, patent CN120558798A discloses an abrasive particle detection system and method. The system and method include acquiring particle waveform signals output by an abrasive particle sensor; performing analog-to-digital conversion and filtering on the particle waveform signals; performing waveform interpretation and feature extraction on the processed particle waveform signals; and calculating the particle size based on the waveform interpretation and feature extraction results. This invention improves detection sensitivity and reduces the risk of false positives by first acquiring particle waveform data in real time, then performing analog-to-digital conversion and filtering to filter out interference signals; then, performing waveform interpretation and feature extraction on the filtered particle waveform signals; and finally, calculating the particle size and determining the particle material based on the waveform interpretation and feature extraction results.

[0004] For example, patent CN104697905B discloses a design method and device for an oil abrasive particle detection sensor. The method involves calculating the relationship between magnetic field strength B and the number of coil turns n using a formula after the sensor frame is determined. The n corresponding to the maximum value of B is taken as the number of turns of the induction coil and the two excitation coils. The device determines the number of turns of each coil in the sensor according to this invention. The output terminal of the induction coil is connected to a preamplifier circuit, a low-pass filter, a high-pass filter, and a modulation and analysis circuit of an oscilloscope, all connected in sequence. The device provides a modulation and analysis circuit including a preamplifier circuit, low-pass and high-pass filters for the signal output components. This method designs a sensor coil with the maximum number of turns to maximize the signal output amplitude. Furthermore, within the size of a mass-produced sensor frame, it is easy to install on various oil-lubricated engines. This device accurately acquires information related to abrasive particles and can detect metal abrasive particles with a diameter of only 500μm.

[0005] However, traditional capacitive and inductive abrasive sensors have significant drawbacks in practical applications: they can only count the total number of abrasive particles passing through the detection area, and cannot distinguish the specific number of abrasive particles of different diameters. Since abrasive particles of different diameters correspond to different degrees of wear on mechanical equipment, knowing only the total number of abrasive particles makes it difficult to accurately analyze the wear degree and potential faults of the equipment, resulting in limited accuracy in wear analysis and failing to meet the needs of high-precision mechanical health monitoring.

[0006] To address the aforementioned issues, there is an urgent need for innovative designs based on existing methods. Summary of the Invention

[0007] The purpose of this invention is to provide a novel capacitive-inductive abrasive particle counting method to solve the problem in the prior art that can only count the total number of abrasive particles passing through the detection area, but cannot distinguish the specific number of abrasive particles of different diameters. Since abrasive particles of different diameters correspond to different degrees of wear on mechanical equipment, knowing only the total number of abrasive particles makes it difficult to accurately analyze the wear degree and potential faults of the equipment, resulting in limited accuracy of wear analysis and failing to meet the needs of high-precision mechanical health monitoring.

[0008] To achieve the above objectives, the present invention provides the following technical solution: a novel method for counting particles using a capacitive-inductive abrasive sensor, comprising the following steps:

[0009] Step S1: Construct a signal basis function library X, which is used to compare the detected signal with the standard templates in the library to achieve accurate mapping of the signal to the diameter of the abrasive grain and determine which diameter of abrasive grain caused the fluctuation.

[0010] Step S2: Initialize parameters to unify the signal processing benchmark, ensure the effectiveness of the comparison, and provide an initial state for subsequent iterative calculations;

[0011] Step S3, residual signal judgment, can accurately determine whether there are abrasive particles of other sizes, ensuring the integrity of the count and improving the counting accuracy;

[0012] Step S4, inner product operation, can reflect the similarity between the residual signal and each basis function, and provide a basis for selecting the abrasive feature signal that best matches the current residual signal;

[0013] Step S5, determining the optimal basis function, accurately locking the abrasive feature template most relevant to the current residual signal from the remaining basis function library, provides a clear target for signal stripping, content quantification and iterative advancement, and is the core link to realize diameter counting;

[0014] Step S6: Update parameters by stripping away the identified abrasive grain signals and reducing the basis function library to clear interference and clarify the target for the next iteration, ensuring that the algorithm can gradually decompose the mixed signals and accurately identify abrasive grains of all diameters.

[0015] Step S7, iterative calculation, through repeated filtering-stripping-updating cycle, gradually separates the signal components of all diameter abrasive grains from the mixed signal, solving the problem that single-round calculation cannot identify multiple abrasive grains;

[0016] Step S8: Calculate the number of abrasive particles and convert the relative content coefficient obtained from the iterative calculation into the actual number of abrasive particles of different diameters. This provides direct and applicable data support for the assessment of wear status of mechanical equipment and solves the problem from signal analysis to practical application.

[0017] Preferably, step S1 is performed as follows: first, a fixed diameter is added to the pure oil. Metal abrasive particles were uniformly stirred to obtain a standard solution containing H particles per milliliter. The standard solution was then passed through a lubricating oil pipe, and the voltage change signal was collected using the induction coil of a capacitive-inductive abrasive particle sensor. After amplification, filtering, and sampling, a base signal with the same length as the signal to be detected was obtained. The basic signals corresponding to several different diameters Composition of signal basis function library Where j is the abrasive grain diameter type number, representing the diameter of the metal abrasive grain. Several different specifications are selected based on actual testing needs.

[0018] Preferably, step S2 is specifically implemented by recording the signal collected by the sensor, amplified, filtered, and sampled as the oil to be detected passes through the lubricating oil pipeline. Then set the residual signal Set up the remaining base function library The processing method of the signal to be detected y is the same as that of the basic signal. To maintain consistency and ensure signal comparability, the initialized residual signal and remaining basis function library provide initial data for subsequent iterative operations.

[0019] Preferably, the residual signal determination method in step S3 is as follows: first calculate the residual signal. If the 2-norm of the residual signal is not greater than the preset threshold h, it is determined that there are no abrasive particles in the oil to be detected, and the counting ends; if the 2-norm is greater than the threshold h, then step S4 is executed. The setting of the threshold h needs to take into account the detection accuracy of the sensor and the noise level. When the 2-norm of the residual signal is less than or equal to h, it indicates that the residual signal is mainly composed of noise, and it can be considered that there is no effective abrasive particle signal, thus avoiding false counting. The threshold h is preset according to the detection accuracy requirements of the sensor and the noise level.

[0020] Preferably, the inner product operation in step S4 is to process the residual signal. With the remaining basis function library Perform inner product operations on all basis functions.

[0021] Preferably, in step S5, the basis function with the largest inner product value is selected, and the corresponding number of the basis function is denoted as k. The coefficients of the basis function are then calculated. The largest inner product value indicates that the abrasive grain diameter corresponding to the basis function best matches the diameter of the main abrasive grains in the current residual signal. The coefficients... This reflects the relative content of abrasive grains of this diameter in the oil being tested.

[0022] Preferably, the update parameter in step S6 refers to the update of the residual signal. And from the remaining basis function library Remove the basis function numbered k from the library to obtain a new library of remaining basis functions. By subtracting the signal components corresponding to the identified abrasive particles, the updated residual signal contains only the signals and noise of the unidentified abrasive particles, while removing the matched basis functions to avoid double counting.

[0023] Preferably, step S7 is performed by returning to step S3 and repeating steps S3-S6 until a residual signal is obtained. If the 2-norm is not greater than the threshold h, output all non-zero coefficients. , where K is the number of abrasive grain diameter types involved in the counting.

[0024] Preferably, in step S8, the calculation formula is used. The diameter of the oil to be tested is obtained Number of abrasive grains Where L is the volume of the oil to be tested in the lubricating oil pipeline, the relative content coefficient is calculated using this formula based on the concentration of the standard solution and the volume of the oil in the lubricating oil pipeline. This is converted into the actual number of abrasive grains, enabling accurate counting of abrasive grains of different diameters.

[0025] Preferably, the capacitive-inductive abrasive sensor includes an induction coil, an excitation coil, a lubricating oil pipe, a post-processing circuit, a power supply, and a display module. The excitation coil is used to generate a magnetic field, the induction coil is used to detect changes in the magnetic field strength of the oil in the lubricating oil pipe and convert them into a voltage signal, and the post-processing circuit is used to amplify, filter, and sample the voltage signal.

[0026] The amplification process uses a conventional amplification circuit, the filtering process uses a low-pass filter, the sampling frequency is set to 120Hz, and the sampling frequency is greater than twice the cutoff frequency of the low-pass filter.

[0027] The induction coil measures the magnetic field strength of the oil in the lubricating oil pipeline for 1 second.

[0028] Compared with the prior art, the beneficial effects of the present invention are:

[0029] 1. By constructing a signal basis function library and using a greedy algorithm to iteratively decompose and optimize the detection signal, the specific number of abrasive particles of different diameters can be accurately distinguished. This solves the technical defect of traditional sensors that can only count the total number of abrasive particles. Moreover, this method is simple to operate and does not require major modifications to the hardware structure of existing capacitive and inductive abrasive particle sensors. It can be achieved simply by optimizing the signal processing algorithm, thus reducing application costs. At the same time, this method has high detection accuracy and strong practicality, and can provide more comprehensive and accurate data support for the wear condition assessment of mechanical equipment. It helps to predict equipment failures in advance, optimize maintenance strategies, and extend the service life of equipment, and has broad application prospects.

[0030] 2. This method can effectively cope with complex interferences such as oil flow disturbances and circuit noise. It can still work stably in harsh environments such as industrial workshops and automobile engine compartments. Moreover, the threshold h can be flexibly set according to the scenario, and the detection requirements of different devices can be met without modifying the algorithm logic. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the counting method of the present invention. Detailed Implementation

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

[0033] This application provides a novel method for counting particles using a capacitive-inductive abrasive sensor. To better understand the above technical solution, it will be described in detail below with reference to the accompanying drawings and specific embodiments. Figure 1 As shown in this embodiment of the present application, a novel capacitive-inductive abrasive particle counting method includes the following steps:

[0034] Step S1: Construct a signal basis function library X, which is used to compare the detected signal with the standard templates in the library to achieve accurate mapping of the signal to the diameter of the abrasive grain and determine which diameter of abrasive grain caused the fluctuation.

[0035] In practice, firstly, based on the actual testing requirements (such as the diameter range of common wear abrasive grains in the equipment), select j different fixed diameter metal abrasive grains, each diameter denoted as . (j=1, 2, ..., M, where M is the total number of abrasive grain diameter types). The diameter accuracy of each abrasive grain meets the testing standards, and the material is consistent with the wear abrasive grains that may occur in the testing scenario (such as steel, iron, and other metal materials) to avoid signal characteristic deviations due to material differences. Then, pure oil (of the same type as the oil to be tested, without any impurity abrasive grains) is used as the substrate, and each diameter is... A standard solution of abrasive particles was prepared by adding a measured amount of abrasive particles to pure oil. Abrasive particles were uniformly stirred until completely dispersed, ultimately preparing a standard solution containing H abrasive particles per milliliter. Then, a capacitive-inductive abrasive sensor was activated, ensuring its hardware parameters were consistent with those used in subsequent testing, including the magnetic field strength of the excitation coil, the installation position of the induction coil, and the flow rate in the lubricating oil pipeline. The post-processing circuit was then activated and adjusted according to preset parameters. A standard amplifier circuit was used for signal amplification, with the low-pass filter enabled and the sampling frequency set to 120Hz to ensure standardization of the signal processing chain. Finally, the prepared diameter... A standard solution with a concentration of H particles / mL is introduced into the lubricating oil pipe of the sensor at a stable flow rate, and the detection time of the induction coil is controlled to be 1 second (consistent with the measurement time during subsequent detection). The original signal of voltage change generated by the passage of abrasive particles in the induction coil is collected. (The last sentence appears to be incomplete and possibly refers to a different measurement method.) The standard solution needs to be sampled 3-5 times to ensure signal stability and repeatability. Abnormal fluctuations in the original signal are discarded. The acquired original voltage signal is processed according to a fixed procedure of amplification → filtering → sampling. The processed discrete signal sequence is denoted as... ={ For i=1...N, this signal is the diameter. The standard characteristic signal corresponding to the abrasive grains is finally used to classify all different diameters. Corresponding standard characteristic signal (j=1, 2, ..., M) are integrated to form a signal basis function library. For each basis function in the library Mark the corresponding abrasive grain diameter. Set the standard solution concentration H to ensure accurate association between the basis function, abrasive diameter, and standard concentration during subsequent calls. Save the basis function library X for direct use during subsequent testing. If the sensor hardware parameters or oil type change, the basis function library needs to be recalibrated according to the above steps.

[0036] In this embodiment, by preparing standard solutions of abrasive grains with different diameters, the characteristic signals corresponding to each diameter abrasive grain can be obtained, laying the foundation for subsequent differentiation of abrasive grains with different diameters. H is a fixed value that can be set according to the detection accuracy requirements to ensure the concentration stability of the standard solution.

[0037] Step S2: Initialize parameters to unify the signal processing benchmark, ensure the effectiveness of the comparison, and provide an initial state for subsequent iterative calculations;

[0038] In practice, the capacitive-inductive abrasive sensor is first activated with hardware parameters identical to those used when building the basis function library. The excitation coil magnetic field strength, lubricating oil flow rate, and post-processing circuit parameters (amplification factor, low-pass filter cutoff frequency) are kept constant. The oil to be detected is introduced into the lubricating oil pipeline at a stable flow rate. The detection time of the induction coil is controlled to be 1 second (consistent with the basis function acquisition time). The original signal of voltage change output by the induction coil is acquired. Then, the original signal is standardized. Finally, the discrete signal sequence processed in step 1 is formally defined as the signal to be detected. The format is (i is the sampling point index, N is the total number of sampling points, and is consistent with the base function library) (The lengths are completely identical). This variable serves as the original reference signal throughout the process, and is not modified after assignment. It is only used as a benchmark for subsequent iterations before defining the residual signal variable. Its format and dimensions must match the signal to be detected. Completely identical, the signal to be detected Complete data assignment to ,Right now ,at this time Representing the original mixed signal before any wear grain signals have been stripped, it is the core dynamic variable of the iterative operation, defining the variables of the remaining basis function library. Its data structure needs to be compatible with a pre-built complete signal basis function library. Consistent, a complete signal basis function library Complete assignment to ,Right now ,at this time The standard feature signal containing all diameter abrasive grains represents the set of all unmatched basis functions. In each subsequent iteration, matched basis functions are removed.

[0039] In this embodiment, the algorithm's operational boundaries can be clearly defined by initialization parameters, avoiding logical confusion. This defines clear operational boundaries for the algorithm and provides an initial basis for determining termination conditions. It establishes a unified and standardized starting state for the iterative algorithm, ensuring the effectiveness of the comparison between the signal to be detected and the basis function library. It clarifies the roles and boundaries of the original signal, residual signal, and basis function library during the iteration process, avoiding logical confusion. This provides an initial basis for determining the start and termination of the algorithm, allowing the entire counting process to start in an orderly manner or terminate quickly.

[0040] Step S3, residual signal judgment, can accurately determine whether there are abrasive particles of other sizes, ensuring the integrity of the count and improving the counting accuracy;

[0041] In practice, the residual signal is calculated first. The 2-norm of the residual signal is used to determine whether there are any abrasive particles in the oil being tested. If the 2-norm is not greater than the preset threshold h, then the counting ends. If the 2-norm is greater than the threshold h, then step S4 is executed. The threshold h needs to be set by comprehensively considering the sensor's detection accuracy and noise level. When the 2-norm of the residual signal is less than or equal to h, it indicates that the residual signal is mainly composed of noise, and it can be considered that there is no effective abrasive particle signal, thus avoiding false counting. The threshold h is preset according to the sensor's detection accuracy requirements and noise level. The formula for calculating the norm is: The threshold h needs to be calibrated in advance through experiments. The preset threshold h is retrieved from the parameter library to ensure that the threshold used in each round of judgment is consistent.

[0042] In this embodiment, the presence or absence of abrasive particles can be accurately determined by residual signal judgment, avoiding misjudgment and missed detection. It can also control the termination time of the iterative process to ensure the integrity of the count. It can separate effective signals from noise, improve counting accuracy, adapt to different detection scenarios, and ensure the universality of the method. That is, the threshold h can be flexibly adjusted according to the actual scenario, and the residual signal judgment is made by quantitative comparison, so that the algorithm can be adapted to different scenarios.

[0043] Step S4, inner product operation, can reflect the similarity between the residual signal and each basis function, providing a basis for selecting the abrasive feature signal that best matches the current residual signal.

[0044] In practice, the first step is to define the computational object, namely the residual signal of the current round. and the remaining basis function library , With all The lengths of all base functions must be exactly the same (N), otherwise the inner product operation cannot be performed. Then, an empty array list is created to store the inner product value of each basis function and the residual signal. The array length is the same as the remaining basis function library. The number of basis functions M is consistent, each position corresponds to the inner product result of a basis function, and the corresponding basis function number is labeled (e.g., To avoid confusion, then traverse in order. Each basis function in (k=1, 2, ..., m), perform calculations for each basis function, i.e., the residual signal The i-th sampling point i, and basis functions The i-th sampling point Multiply to get the product value Adding the product values ​​of all sampling points, we obtain the inner product of the basis function and the residual signal, as shown in the formula: Then check the number of inner product results to avoid omissions or duplicate calculations, and then output the inner product operation results.

[0045] Step S5, determining the optimal basis function, accurately locking the abrasive feature template most relevant to the current residual signal from the remaining basis function library, provides a clear target for signal stripping, content quantification and iterative advancement, and is the core link to realize diameter counting;

[0046] In practice, the process first obtains the association information between the inner product operation result and the basis function, retrieves the complete result array of the inner product operation output from the previous step, and then filters for the maximum inner product value. If multiple inner product values ​​are equal and all are the maximum, the decision can be aided by the priority of the abrasive grain diameter corresponding to the basis function (e.g., sorted by diameter from largest to smallest, prioritizing the basis function corresponding to the largest diameter abrasive grain) or the stability of the signal energy (e.g., the magnitude of the 2-norm of the basis function), ensuring that only one optimal basis function is selected. Then, the basis function corresponding to the maximum inner product value is located, and the remaining basis function library is used. Extracting the basis function corresponding to this number is the optimal basis function for the current round, denoted as . ( = (i=1...N), and simultaneously record the abrasive grain diameter corresponding to the optimal basis function. (Retrieved from the annotation information of the basis function library), establish the correlation between the optimal basis function and the abrasive diameter, and then output the optimal basis function and related parameters.

[0047] Step S6: Update parameters by stripping away the identified abrasive grain signals and reducing the basis function library to clear interference and clarify the target for the next iteration, ensuring that the algorithm can gradually decompose the mixed signals and accurately identify abrasive grains of all diameters.

[0048] In practice, the object to be updated and the input parameters must first be clearly defined. The core object to be updated is the residual signal. and the remaining basis function library That is, update the residual signal And from the remaining basis function library Remove the basis function numbered k from the library to obtain a new library of remaining basis functions. By subtracting the signal components corresponding to the identified abrasive particles, the updated residual signal only contains the signals and noise of the unidentified abrasive particles. At the same time, the matched basis functions are removed to avoid double counting. Then, the updated parameters are saved to prepare for the next iteration.

[0049] Step S7, iterative calculation, through repeated filtering-stripping-updating cycle, gradually separates the signal components of all diameter abrasive grains from the mixed signal, solving the problem that single-round calculation cannot identify multiple abrasive grains;

[0050] In practice, it is necessary to return to step S3 and repeat steps S3-S6 until the residual signal is obtained. If the 2-norm is not greater than the threshold h, output all non-zero coefficients. Where K is the number of abrasive grain diameter types participating in the counting, through a multi-round closed-loop process, the limitation of a single round of calculation being able to identify only one type of abrasive grain is overcome, and full coverage identification of abrasive grains with multiple diameters is achieved. Each iteration uses residual signal judgment as the entry point and parameter update as the exit point to ensure that the iteration proceeds in an orderly manner, without missing any valid abrasive grains or performing invalid calculations. The final output of abrasive grain diameter and content coefficient results is the core data support for achieving accurate counting by diameter, and directly connects to the subsequent actual quantity calculation.

[0051] Step S8: Calculate the number of abrasive particles and convert the relative content coefficient obtained from the iterative calculation into the actual number of abrasive particles of different diameters. This provides direct and applicable data support for the assessment of wear status of mechanical equipment and solves the problem from signal analysis to practical application.

[0052] In practical implementation, the input parameters and core formulas must first be clearly defined, that is, based on the calculation formulas... The diameter of the oil to be tested is obtained Number of abrasive grains Where L is the volume of the oil to be tested in the lubricating oil pipeline, the relative content coefficient is calculated using this formula based on the concentration of the standard solution and the volume of the oil in the lubricating oil pipeline. The data is converted into the actual number of abrasive grains to achieve accurate counting of abrasive grains of different diameters. Then, the completeness and validity of the input parameters are verified, and a final result table is constructed to specify the diameter and corresponding quantity of each type of abrasive grain. The tabular results are then displayed directly through the sensor's display module or exported to a data terminal to support subsequent trend analysis.

[0053] In summary, by establishing a unique mapping between diameter and feature signals using a signal basis function library and combining iterative decomposition with a greedy algorithm, this method achieves, for the first time, the separate counting of abrasive particles of different diameters, rather than simply counting the total number. By quantifying similarity through inner product operations and filtering effective signals using the 2-norm of the residual signal, noise and abrasive particle signals are effectively separated, avoiding false detections and missed detections. The counting error is significantly lower than that of traditional methods. This provides precise data support for equipment health management; different diameter abrasive particles correspond to different wear stages. Precise data from diameter to quantity helps maintenance personnel determine the wear type and locate the fault location, rather than simply knowing that wear exists. By tracking different diameters... The changing trend of the number of abrasive particles can predict faults in advance, avoid over-maintenance or under-maintenance, and reduce equipment downtime losses. The abrasive particle diameter types in the basis function library can be flexibly added or removed according to the wear abrasive particle diameter range of different equipment without modifying the core algorithm. Its adaptability is far superior to traditional fixed threshold counting methods. Through residual signal judgment and iterative signal stripping design, it can effectively deal with complex interferences such as oil flow disturbances and circuit noise. It can still work stably in harsh environments such as industrial workshops and automobile engine compartments. Moreover, the threshold h can be flexibly set according to the scenario, and can meet the detection requirements of different equipment without modifying the algorithm logic.

[0054] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A novel method for counting particles using a capacitive-inductive abrasive sensor, characterized in that, The counting method includes the following steps: Step S1: Construct a signal basis function library X, which is used to compare the detected signal with the standard templates in the library to achieve accurate mapping of the signal to the diameter of the abrasive grain and determine which diameter of abrasive grain caused the fluctuation. Step S2: Initialize parameters to unify the signal processing benchmark, ensure the effectiveness of the comparison, and provide an initial state for subsequent iterative calculations; Step S3, residual signal judgment, can accurately determine whether there are abrasive particles of other sizes, ensuring the integrity of the count and improving the counting accuracy; Step S4, inner product operation, can reflect the similarity between the residual signal and each basis function, and provide a basis for selecting the abrasive feature signal that best matches the current residual signal; Step S5, determining the optimal basis function, accurately locking the abrasive feature template most relevant to the current residual signal from the remaining basis function library, provides a clear target for signal stripping, content quantification and iterative advancement, and is the core link to realize diameter counting; Step S6: Update parameters by stripping away the identified abrasive grain signals and reducing the basis function library to clear interference and clarify the target for the next iteration, ensuring that the algorithm can gradually decompose the mixed signals and accurately identify abrasive grains of all diameters. Step S7, iterative calculation, through repeated filtering-stripping-updating cycle, gradually separates the signal components of all diameter abrasive grains from the mixed signal, solving the problem that single-round calculation cannot identify multiple abrasive grains; Step S8: Calculate the number of abrasive particles and convert the relative content coefficient obtained from the iterative calculation into the actual number of abrasive particles of different diameters. This provides direct and applicable data support for the assessment of wear status of mechanical equipment and solves the problem from signal analysis to practical application.

2. The novel capacitive-inductive abrasive particle counting method according to claim 1, characterized in that: The specific method of step S1 is as follows: First, add a fixed diameter to the pure oil. Metal abrasive particles were uniformly stirred to obtain a standard solution containing H particles per milliliter. The standard solution was then passed through a lubricating oil pipe, and the voltage change signal was collected using the induction coil of a capacitive-inductive abrasive particle sensor. After amplification, filtering, and sampling, a base signal with the same length as the signal to be detected was obtained. The basic signals corresponding to several different diameters Composition of signal basis function library Where j is the abrasive grain diameter type number, representing the diameter of the metal abrasive grain. Several different specifications are selected based on actual testing needs.

3. The novel capacitive-inductive abrasive particle counting method according to claim 1, characterized in that: The specific method of step S2 is as follows: when the oil to be detected passes through the lubricating oil pipeline, the signal collected by the sensor and after amplification, filtering, and sampling processing is recorded as follows: Then set the residual signal Set up the remaining base function library The processing method of the signal to be detected y is the same as that of the basic signal. To maintain consistency and ensure signal comparability, the initialized residual signal and remaining basis function library provide initial data for subsequent iterative operations.

4. The novel particle counting method for a capacitive-inductive abrasive sensor according to claim 1, characterized in that: The residual signal determination method in step S3 is as follows: first calculate the residual signal. If the 2-norm of the residual signal is not greater than the preset threshold h, it is determined that there are no abrasive particles in the oil to be detected, and the counting ends; if the 2-norm is greater than the threshold h, then step S4 is executed. The setting of the threshold h needs to take into account the detection accuracy of the sensor and the noise level. When the 2-norm of the residual signal is less than or equal to h, it indicates that the residual signal is mainly composed of noise, and it can be considered that there is no effective abrasive particle signal, thus avoiding false counting. The threshold h is preset according to the detection accuracy requirements of the sensor and the noise level.

5. A novel method for counting particles using a capacitive-inductive abrasive sensor according to claim 1, characterized in that: The inner product operation in step S4 is to process the residual signal. With the remaining basis function library Perform inner product operations on all basis functions.

6. The novel particle counting method for a capacitive-inductive abrasive sensor according to claim 1, characterized in that: In step S5, the basis function with the largest inner product value is selected, and its corresponding index is denoted as k. The coefficients of the basis function are then calculated. The largest inner product value indicates that the abrasive grain diameter corresponding to the basis function best matches the diameter of the main abrasive grains in the current residual signal. The coefficients... This reflects the relative content of abrasive grains of this diameter in the oil being tested.

7. A novel method for counting particles using a capacitive-inductive abrasive sensor according to claim 1, characterized in that: The update parameter in step S6 refers to updating the residual signal. And from the remaining basis function library Remove the basis function numbered k from the library to obtain a new library of remaining basis functions. By subtracting the signal components corresponding to the identified abrasive particles, the updated residual signal contains only the signals and noise of the unidentified abrasive particles, while removing the matched basis functions to avoid double counting.

8. A novel method for counting particles using a capacitive-inductive abrasive sensor according to claim 1, characterized in that: The method for step S7 is as follows: return to step S3, and repeat steps S3-S6 until a residual signal is obtained. If the 2-norm is not greater than the threshold h, output all non-zero coefficients. , where K is the number of abrasive grain diameter types involved in the counting.

9. A novel method for counting particles using a capacitive-inductive abrasive sensor according to claim 1, characterized in that: In step S8, the calculation formula is used. The diameter of the oil to be tested is obtained Number of abrasive grains Where L is the volume of the oil to be tested in the lubricating oil pipeline, the relative content coefficient is calculated using this formula based on the concentration of the standard solution and the volume of the oil in the lubricating oil pipeline. This is converted into the actual number of abrasive grains, enabling accurate counting of abrasive grains of different diameters.

10. A novel method for counting particles in a capacitive-inductive abrasive sensor according to claim 1, characterized in that: The capacitive-inductive abrasive sensor includes an induction coil, an excitation coil, a lubricating oil pipe, a post-processing circuit, a power supply, and a display module. The excitation coil is used to generate a magnetic field, the induction coil is used to detect changes in the magnetic field strength of the oil in the lubricating oil pipe and convert them into a voltage signal, and the post-processing circuit is used to amplify, filter, and sample the voltage signal. The amplification process uses a conventional amplification circuit, the filtering process uses a low-pass filter, the sampling frequency is set to 120Hz, and the sampling frequency is greater than twice the cutoff frequency of the low-pass filter. The induction coil measures the magnetic field strength of the oil in the lubricating oil pipeline for 1 second.

Images