Search

CN-115541463-B - Weak output characteristic signal enhancement processing method for oil metal particle sensor

CN115541463BCN 115541463 BCN115541463 BCN 115541463BCN-115541463-B

Abstract

The invention discloses a method for enhancing and processing weak output characteristic signals of an oil metal particle sensor, which comprises the steps of setting sampling frequency and sampling point number, and collecting original output signals when tiny metal particles pass through the oil metal particle sensor; the method comprises the steps of carrying out mean value removal and amplitude scaling on an original output signal to obtain a weak characteristic signal, utilizing a nonlinear system to enhance the weak characteristic signal to obtain an initial enhanced output weak characteristic signal, carrying out point-by-point smooth filtering and mean value removal on the initial enhanced output weak characteristic signal to obtain a final enhanced output weak characteristic signal. The invention is applied to the field of weak digital signals, can effectively enhance the weak output characteristic signals when 100 mu m tiny metal particles pass through the sensor, and improves the detection sensitivity of the sensor. And the method has small calculated amount, better robustness and adaptability, can achieve standardization, and is convenient for the embedded microprocessor to realize the operation and the processing of signals.

Inventors

  • YANG DINGXIN
  • LIU XIAORONG

Assignees

  • 中国人民解放军国防科技大学

Dates

Publication Date
20260505
Application Date
20220919

Claims (5)

  1. 1. The method for enhancing the weak output characteristic signal of the oil metal particle sensor is characterized by comprising the following steps of: Step 1, setting sampling frequency and sampling point number, and collecting an original output signal s 0 (n) when tiny metal particles pass through an oil metal particle sensor based on the sampling frequency and the sampling point number; Step 2, preprocessing an original output signal s 0 (n) to obtain a weak characteristic signal s (n), wherein the preprocessing comprises mean value removing processing and amplitude scaling processing; Step 3, enhancing the weak characteristic signal s (n) by using a nonlinear system to obtain an initial enhancement output weak characteristic signal x (n), wherein the enhancement of the weak characteristic signal s (n) by using the nonlinear system, specifically, taking the weak characteristic signal s (n) as an excitation input nonlinear discrete equation is as follows: Wherein x (N) is a weak characteristic signal of the N-th initial enhancement output, x (n+1) is a weak characteristic signal of the n+1th initial enhancement output, h is a discrete step length, and N is a sampling point number; The initial value of the nonlinear discrete equation is set to x (0) =1 or x (0) = -1; h=1/M, where M is the number of samples corresponding to the pulse waveform width of the single output signal, for controlling the sampling frequency; And 4, performing point-by-point smoothing filtering and mean value removing processing on the weak characteristic signal x (n) of the initial enhancement output to obtain a weak characteristic signal s 2 (n) of the final enhancement output, wherein the method specifically comprises the following steps: Carrying out point-by-point smooth filtering on the weak characteristic signal x (n) of the initial enhancement output to obtain a weak characteristic signal s 1 (n) of the initial enhancement output after filtering, wherein the weak characteristic signal is; wherein K is the number of smooth points; And (3) carrying out mean value removal processing on the filtered weak characteristic signal s 1 (n) of the initial enhancement output to obtain a weak characteristic signal s 2 (n) of the final enhancement output.
  2. 2. The method for enhancing the weak output characteristic signal of the oil metal particle sensor according to claim 1, wherein in step 1, the setting of the sampling frequency and the sampling point number is specifically as follows: fs =M/T N≥ 3M wherein fs is the sampling frequency, T is the width of the pulse waveform of the characteristic signal output by the oil metal particle sensor, M is the sampling number corresponding to the width of the pulse waveform of the single output signal, and is used for controlling the sampling frequency, and N is the sampling number.
  3. 3. The method for enhancing the weak output characteristic signals of the oil metal particle sensor according to claim 1, wherein M is more than or equal to 100.
  4. 4. The method for enhancing the weak output characteristic signal of the oil metal particle sensor according to claim 1,2 or 3, wherein in the step 2, the preprocessing process specifically comprises the following steps: Firstly, carrying out mean value removal processing on an original output signal s 0 (n), and normalizing the signal amplitude of the original output signal s 0 (n) to be within the range of [ -1,1] through amplitude scaling processing to obtain a weak characteristic signal s (n).
  5. 5. The method for enhancing the weak output characteristic signal of the oil metal particle sensor according to claim 1,2 or 3, wherein k=m/10, where M is the number of samples corresponding to the pulse waveform width of the single output signal, and is used for controlling the sampling frequency.

Description

Weak output characteristic signal enhancement processing method for oil metal particle sensor Technical Field The invention relates to the technical field of weak digital signals, in particular to a weak output characteristic signal enhancement processing method for tiny metal particles passing through an oil liquid metal particle sensor, which can improve the detection sensitivity of the sensor to the tiny metal particles. Background The oil liquid metal particle sensor is widely applied to the on-line monitoring and fault prediction of the health state of complex mechanical systems such as engines, wind power gearboxes and the like. The sensor belongs to an online oil detection sensor, is generally arranged in a lubricating oil way of a mechanical system, and monitors whether abnormal wear and faults occur on mechanical parts or not by detecting the quantity and the size of metal particles contained in lubricating oil flowing through the sensor. According to experimental statistics, the equivalent diameter size of metal particles generated during normal wear of mechanical parts is mostly below 100 mu m, and the size of the generated metal particles is in the range of 100 mu m-150 mu m during early abnormal wear. Effective monitoring of metal wear particles with equivalent diameter dimensions of around 100 μm is therefore crucial for finding early wear failures. However, it is difficult to detect the metal abrasion particles with the size of about 100 μm, on the one hand, the output characteristic signals of the particles passing through the sensor are very weak due to the small particle size, and on the other hand, the sensor is affected by site noise such as vibration, environmental interference and the like, so that the original weak signals are submerged in the noise, and the particle characteristic signals are difficult to detect. At present, methods such as wavelet analysis and time domain filtering are often adopted to eliminate noise, but the methods are required to adjust more parameters, have large calculation amount and poor instantaneity, and can not effectively extract the characteristic signals output by the tiny metal particles which are almost submerged in noise, so that the detection sensitivity of the sensor to the tiny metal particles is greatly influenced. Disclosure of Invention Aiming at the defects in the prior art, the invention provides a method and a system for enhancing the weak output characteristic signals of an oil liquid metal particle sensor, which are used for enhancing the output characteristic signals of tiny metal particles of the metal particle sensor so as to improve the detection sensitivity of the sensor to tiny metal particles with equivalent diameter size of about 100 mu m. In order to achieve the purpose, the invention provides a weak output characteristic signal enhancement processing method of an oil metal particle sensor, which comprises the following steps: Step 1, setting sampling frequency and sampling point number, and collecting an original output signal s 0 (n) when tiny metal particles pass through an oil metal particle sensor based on the sampling frequency and the sampling point number; Step 2, preprocessing an original output signal s 0 (n) to obtain a weak characteristic signal s (n), wherein the preprocessing comprises mean value removing processing and amplitude scaling processing; Step 3, enhancing the weak characteristic signal s (n) by using a nonlinear system to obtain an initial enhancement output weak characteristic signal x (n); And 4, carrying out point-by-point smooth filtering and mean value removal processing on the weak characteristic signal x (n) of the initial enhancement output to obtain a weak characteristic signal s 2 (n) of the final enhancement output. In one embodiment, in step 1, the setting the sampling frequency and the sampling point number is specifically: fs=M/T N≥3M wherein fs is sampling frequency, M is sampling number corresponding to the pulse waveform width of a single output signal, the sampling frequency is controlled, M is required to be more than or equal to 100, T is the pulse waveform width of the characteristic signal output by the oil metal particle sensor, and N is the sampling point number. In one embodiment, in step2, the preprocessing specifically includes: Firstly, carrying out mean value removal processing on an original output signal s 0 (n), and normalizing the signal amplitude of the original output signal s 0 (n) to be within the range of [ -1,1] through amplitude scaling processing to obtain a weak characteristic signal s (n). In one embodiment, in step 3, the enhancing the weak characteristic signal s (n) by using a nonlinear system, specifically, using the weak characteristic signal s (n) as an excitation input nonlinear discrete equation is: x(n+1)=(h+1)·x(n)+h[s(n)-x3(n)]n=0,1,...N-1 Wherein x (N) is a weak characteristic signal of the nth initial enhancement output, x (n+1) is a weak characteri