Search

CN-121993391-A - Online testing system for full life cycle performance degradation of reciprocating pump

CN121993391ACN 121993391 ACN121993391 ACN 121993391ACN-121993391-A

Abstract

The invention discloses an online test system for full life cycle performance degradation of a reciprocating pump, and particularly relates to the technical field of state monitoring of the reciprocating pump, wherein a dynamic characteristic matching module identifies working conditions and outputs a fault characteristic set; the invention relates to a method for testing the performance degradation of a reciprocating pump full life cycle, which comprises the steps of enhancing and extracting nonlinear weak fault characteristics from an original signal by a weak signal enhancement module to output an enhanced characteristic signal, performing multi-source decoupling on the enhanced signal by a coupling decoupling analysis module to output an independent fault characteristic value sequence and a fault propagation path, generating a health state quantization index by a full life cycle mapping module according to the sequence and the path, performing dynamic space-time reference alignment by a reference alignment module, and outputting an aligned track geometric data set.

Inventors

  • LI HAOHAI
  • LI JIE

Assignees

  • 潍坊亚油机械有限公司

Dates

Publication Date
20260508
Application Date
20260408

Claims (9)

  1. 1. An on-line testing system for full life cycle performance degradation of a reciprocating pump, comprising: The dynamic characteristic matching module is used for identifying the working condition change of the reciprocating pump in real time and dynamically outputting a corresponding fault characteristic set; the weak signal enhancement module is used for receiving the fault feature set, enhancing and extracting nonlinear weak fault features from the original sensing signals and outputting enhanced feature signals; The coupling and decoupling analysis module is used for carrying out multi-source information coupling and decoupling analysis on the enhanced characteristic signals and outputting a decoupled independent fault characteristic value sequence and a fault propagation path; and the full life cycle mapping module is used for generating the current health state quantization index of the reciprocating pump according to the independent fault characteristic value sequence and the fault propagation path mapping.
  2. 2. The system for online testing full life cycle performance degradation of a reciprocating pump according to claim 1, wherein the working conditions of the reciprocating pump obtained by the dynamic characteristic matching module at least comprise rotation speed, output pressure, instantaneous flow and characteristics of a conveying medium, wherein the characteristics of the conveying medium comprise medium density, viscosity and sand content.
  3. 3. The online test system for full life cycle performance degradation of a reciprocating pump according to claim 2, wherein the specific steps of the dynamic feature matching module identifying the working condition change of the reciprocating pump in real time and dynamically outputting the corresponding fault feature set are as follows: continuously acquiring real-time rotating speed, output pressure, instantaneous flow, medium density, viscosity and sand content of the reciprocating pump through a rotating speed sensor, a pressure sensor, a flow sensor and a medium characteristic sensor to form a working condition parameter sequence; Performing steady state judgment on the working condition parameter sequence by adopting a sliding time window, calculating the variation coefficient of each parameter in the time window, judging a steady state working condition when the variation coefficient of all the parameters is lower than a corresponding preset variation threshold value, and judging a dynamic working condition otherwise; invoking a pre-trained working condition classification model according to a steady state or dynamic discrimination result, and outputting a working condition classification label; and according to the working condition type labels, searching and dynamically outputting a corresponding fault feature set from a preset fault feature knowledge base, wherein the fault feature set comprises a time domain feature set and a wavelet packet energy feature set.
  4. 4. An on-line testing system for full life cycle performance degradation of a reciprocating pump according to claim 3, wherein the condition classification model maps current condition parameter vectors to condition class clusters based on a self-organizing map neural network.
  5. 5. The on-line testing system for full life cycle performance degradation of a reciprocating pump of claim 1, wherein the weak signal enhancement module receives the set of fault characteristics in the following specific manner: Analyzing a fault characteristic set, and extracting sensitive frequency bands and sensitive time scale parameters related to weak faults from the fault characteristic set, wherein an original sensing signal is a vibration acceleration signal of a plurality of measuring points of each cylinder body of the reciprocating pump, which is acquired by a vibration sensor; the method for extracting the sensitive frequency band comprises the steps of taking a vibration acceleration signal of the reciprocating pump when the reciprocating pump operates in a healthy state as a reference, carrying out three-layer wavelet packet decomposition on the original sensing signal to obtain eight frequency band subspaces, calculating the change rate of wavelet packet energy of each subspace relative to the healthy reference, and determining a frequency range corresponding to the first three subspaces with kurtosis indexes larger than three as the sensitive frequency band; The extraction mode of the sensitive time scale parameter is that the number of the wavelet packet decomposition layers is determined according to the ratio of the lowest frequency to the sampling frequency in the sensitive frequency band, and the time resolution corresponding to the number of the wavelet packet decomposition layers is used as the sensitive time scale parameter.
  6. 6. The on-line testing system for full life cycle performance degradation of a reciprocating pump according to claim 5, wherein the specific steps of the weak signal enhancement module enhancing and extracting nonlinear weak fault features from the original sensing signal and outputting enhanced feature signals are as follows: According to the sensitive time scale parameter, intercepting an original sensing signal according to the time scale section to obtain a plurality of signal fragments; Carrying out maximum and minimum normalization processing on each signal segment to ensure that all signal amplitudes are between negative ones and positive ones, thereby obtaining normalized signal segments; The method comprises the steps of respectively inputting each normalized signal segment into a bistable random resonance system, wherein system parameters of the bistable random resonance system comprise potential well parameters and noise intensity parameters, adopting a signal-to-noise ratio gain maximization criterion as an objective function of parameter optimization, setting the searching range of the potential well parameters to be the lowest value zero point to the highest value ten and the searching range of the noise intensity parameters to be the lowest value zero point to the highest value one hundred, and determining the optimal potential well parameters and the noise intensity parameters through iterative searching to enable the system to be in a random resonance state; splicing the output signals of each normalized signal segment after stochastic resonance treatment according to the original time sequence to obtain stochastic resonance enhancement signals; and carrying out three-layer wavelet soft threshold noise reduction processing on the stochastic resonance enhancement signal to obtain a final enhancement characteristic signal.
  7. 7. The on-line testing system for full life cycle performance degradation of a reciprocating pump of claim 6, wherein the enhanced characteristic signals received by the coupling and decoupling analysis module are enhanced characteristic signals corresponding to a plurality of measuring points of each cylinder of the reciprocating pump; the multi-source information coupling and decoupling analysis is to separate coupling components in the enhanced characteristic signals of the measuring points of each cylinder body, eliminate mutual interference between vibration signals of adjacent cylinder bodies and output independent fault characteristic value sequences after decoupling, which are in one-to-one correspondence with each cylinder body.
  8. 8. The online test system for full life cycle performance degradation of a reciprocating pump according to claim 7, wherein the specific steps of the coupling and decoupling analysis module outputting the decoupled independent fault characteristic value sequence and the fault propagation path are as follows: Constructing an enhanced characteristic signal of each cylinder measuring point into an observed signal matrix, adopting an independent component analysis algorithm, taking kurtosis as a comparison function, and carrying out iterative computation on a unmixed matrix to decompose the observed signal matrix into mutually statistically independent source signal components, wherein each source signal component corresponds to the vibration contribution of one cylinder, so as to obtain an independent fault characteristic value sequence of each cylinder; calculating a cross correlation function of the independent fault characteristic value sequences between adjacent cylinders according to time sequence by taking the obtained independent fault characteristic value sequences of each cylinder as input, and determining a propagation direction according to the positive and negative of the time delay corresponding to the cross correlation function peak value when the cross correlation function peak value exceeds a preset cross correlation threshold value and the time delay exists; and outputting the fault propagation path in the form of a cylinder number sequence to mark an initial fault source cylinder and a subsequent cylinder sequence affected by propagation.
  9. 9. The online testing system for full life cycle performance degradation of a reciprocating pump according to claim 8, wherein the specific steps of the full life cycle mapping module generating the current health status quantization index of the reciprocating pump according to the independent fault characteristic value sequence and the fault propagation path map are as follows: Comparing the current independent fault characteristic value sequence with a health reference characteristic value sequence for each cylinder body, and calculating the health degree value of each cylinder body, wherein the health reference characteristic value sequence is the independent fault characteristic value sequence of each cylinder body of the reciprocating pump in a health state, the health degree value is equal to one minus the ratio of the Euclidean distance between the current independent fault characteristic value sequence and the health reference characteristic value sequence to the sum of the modular length of the health reference characteristic value sequence and the modular length of the current independent fault characteristic value sequence, and the value is zero when the calculation result is smaller than zero; Determining the sequence position of each cylinder body in the propagation path according to the fault propagation path, and correcting the health degree score multiplied by the propagation attenuation coefficient of the cylinder body with the propagation sequence, wherein the propagation attenuation coefficient takes the value range from zero to eight to one; and carrying out weighted average on the health degree scores of all the cylinders after correction, and presetting and distributing weights of all the cylinders according to the propagation sequence of the cylinders in the fault propagation path to obtain a health state quantification index between zero and one.

Description

Online testing system for full life cycle performance degradation of reciprocating pump Technical Field The invention relates to the technical field of monitoring of the state of a reciprocating pump, in particular to an online test system for the performance degradation of the full life cycle of the reciprocating pump. Background The reciprocating pump is a positive displacement pump for realizing fluid delivery through the reciprocating motion of a piston or a plunger, and is widely applied to the industrial fields of petroleum drilling, chemical processes, mine drainage, power delivery and the like. In the long-term operation process, key components such as a pump valve, a plunger, a sealing element, a bearing and the like of the reciprocating pump gradually wear, fatigue, corrosion and other performance degradation phenomena, and timely detection of the degradation states has important significance for guaranteeing safe operation of equipment and reducing maintenance cost. At present, the performance monitoring of the reciprocating pump mainly adopts a mode of combining periodic inspection with on-line parameter threshold alarming. Common online monitoring means comprise vibration monitoring, temperature monitoring, pressure pulsation analysis, flow monitoring and the like, and whether equipment is abnormal or not is judged by setting a fixed threshold value or simple statistical indexes (such as vibration amplitude and kurtosis factor). Part of advanced test systems introduce time domain statistical features, frequency domain features (such as side band amplitude) and machine learning classification methods (such as support vector machines and neural networks) to perform fault identification, and the methods can obtain good diagnosis effects under steady-state working conditions. The running condition of the reciprocating pump in the actual industrial field is often not constant, factors such as rotation speed adjustment, load change, conveying medium characteristic change and the like can cause the condition to present strong time variability, and most of the existing monitoring methods depend on steady state assumption or a diagnosis model trained for specific conditions, so that the detection accuracy can fluctuate when the conditions change. In addition, the amplitude of a fault signal generated by early performance degradation is low, the fault signal is easy to be interfered by normal operation vibration of equipment and environmental noise, and the weak characteristics are difficult to stably extract by conventional signal processing means. For a multi-cylinder reciprocating pump, vibration signals among cylinders are mutually coupled through a mechanical structure, signals collected by a sensor are usually superposition of vibration of a plurality of cylinders, and extra difficulties are brought to accurate positioning of fault sources and tracing of propagation paths among the cylinders. Therefore, the invention provides an online test system for full life cycle performance degradation of a reciprocating pump, so as to solve the problems. Disclosure of Invention In order to achieve the above purpose, the present invention provides the following technical solutions: an on-line testing system for full life cycle performance decay of a reciprocating pump, comprising: The dynamic characteristic matching module is used for identifying the working condition change of the reciprocating pump in real time and dynamically outputting a corresponding fault characteristic set; the weak signal enhancement module is used for receiving the fault feature set, enhancing and extracting nonlinear weak fault features from the original sensing signals and outputting enhanced feature signals; The coupling and decoupling analysis module is used for carrying out multi-source information coupling and decoupling analysis on the enhanced characteristic signals and outputting a decoupled independent fault characteristic value sequence and a fault propagation path; and the full life cycle mapping module is used for generating the current health state quantization index of the reciprocating pump according to the independent fault characteristic value sequence and the fault propagation path mapping. In a preferred embodiment, the working conditions of the reciprocating pump obtained by the dynamic characteristic matching module at least comprise rotating speed, output pressure, instantaneous flow and conveying medium characteristics, wherein the conveying medium characteristics comprise medium density, viscosity and sand content. The dynamic characteristic matching module is used for identifying the working condition change of the reciprocating pump in real time and dynamically outputting a corresponding fault characteristic set, and comprises the following specific steps of: continuously acquiring real-time rotating speed, output pressure, instantaneous flow, medium density, viscosity and sand content of the reciprocating p