Online fault diagnosis method for solenoid valve based on analysis of time-frequency domain characteristics

An online fault diagnosis method for a solenoid valve based on analysis of time-frequency domain characteristics, comprising the following steps:

• acquiring present timing current data, historical normal valve core timing current data, historical spring breakage timing current data, historical stuck valve core timing current data, and historical slightly stuck valve core timing current data for a solenoid valve drive end, and preprocessing the data;
• performing time domain analysis and frequency domain analysis on the preprocessed present timing current data, historical normal valve core timing current data, historical spring breakage timing current data, historical stuck valve core timing current data, and historical slightly stuck valve core timing current data for the solenoid valve drive end to obtain present time domain and frequency domain feature vectors, time domain and frequency domain feature vectors for a normal valve core, time domain and frequency domain feature vectors for a broken spring, time domain and frequency domain feature vectors for a stuck valve core, and time domain and frequency domain feature vectors for a slightly stuck valve core, respectively;
• performing fusion processing on the present time domain and frequency domain feature vectors, the time domain and frequency domain feature vectors for the normal valve core, the time domain and frequency domain feature vectors for the broken spring, the time domain and frequency domain feature vectors for the stuck valve core, and the time domain and frequency domain feature vectors for the slightly stuck valve core to obtain a present complex feature vector, a complex feature vector for the normal valve core, a complex feature vector for the broken spring, a complex feature vector for the stuck valve core, and a complex feature vector for the slightly stuck valve core, respectively;
• establishing a vector dataset based on the complex feature vector for the normal valve core, the complex feature vector for the broken spring, the complex feature vector for the stuck valve core, and the complex feature vector for the slightly stuck valve core;
• and
• training a pre-established random forest model with the vector dataset, and diagnosing and identifying the present complex feature vector after the completion of training the random forest model;
• determining a fault type of the solenoid valve based on a diagnostic identification result;
• and
• providing fault warning and maintenance recommendations based on the fault type of the solenoid valve, and performing maintenance on the solenoid valve based on the maintenance recommendations;
• wherein the present timing current data for the solenoid valve drive end is specifically current values at a plurality of time points when the solenoid valve drive end is in a present state;
• the historical normal valve core timing current data for the solenoid valve drive end is specifically historical current values at a plurality of historical time points when the solenoid valve drive end is in a normal valve core state;
• the historical spring breakage timing current data for the solenoid valve drive end is specifically historical current values at a plurality of historical time points when the solenoid valve drive end is in a spring breakage state;
• the historical stuck valve core timing current data for the solenoid valve drive end is specifically historical current values at a plurality of historical time points when the solenoid valve drive end is in a stuck valve core state;
• and the historical slightly stuck valve core timing current data for the solenoid valve drive end is specifically historical current values at a plurality of historical time points when the solenoid valve drive end is in a slightly stuck valve core state;
• obtaining the present time domain feature vector, the time domain feature vector for the normal valve core, the time domain feature vector for the broken spring, the time domain feature vector for the stuck valve core, and the time domain feature vector for the slightly stuck valve core specifically comprises:
• separately reading preprocessed current values at the plurality of time points when the solenoid valve drive end is in the present state and preprocessed historical current values at the plurality of historical time points when the solenoid valve is in the normal valve core state, in the spring breakage state, in the stuck valve core state, and in the slightly stuck valve core state, and performing mean value analysis and arrangement in a descending order on the current values to obtain a current mean value, a current effective value, a maximum current value, and a minimum current value in the present state, and current mean values, current effective values, maximum current values, and minimum current values in the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state, respectively;
• performing difference analysis on the maximum current value and the minimum current value in the present state, and the maximum current values and the minimum current values in the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state to obtain a current extreme difference value in the present state, and current extreme difference values in the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state, respectively;
• and
• establishing the present time domain feature vector, the time domain feature vector for the normal valve core, the time domain feature vector for the broken spring, the time domain feature vector for the stuck valve core, and the time domain feature vector for the slightly stuck valve core based on the current mean value, the current effective value, and the current extreme difference value in the present state, and the current mean values, the current effective values, and the current extreme difference values in the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state, respectively;
• and
• obtaining the present frequency domain feature vector, the frequency domain feature vector for the normal valve core, the frequency domain feature vector for the broken spring, the frequency domain feature vector for the stuck valve core, and the frequency domain feature vector for the slightly stuck valve core specifically comprises:
• separately reading preprocessed current values at the plurality of time points when the solenoid valve drive end is in the present state and preprocessed historical current values at the plurality of historical time points when the solenoid valve is in the normal valve core state, in the spring breakage state, in the stuck valve core state, and in the slightly stuck valve core state, and performing second derivative analysis on the current values to obtain second-order derivatives in the present state, the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state, respectively;
• performing binary discrete wavelet packet decomposition on the second-order derivatives in the present state, the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state to obtain wavelet functions in the present state, the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state, respectively;
• performing energy analysis on the wavelet functions in the present state, the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state to obtain energy values in the present state, the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state, respectively;
• and
• performing normalization processing on the energy values in the present state, the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state to obtain normalized energy features in the present state, the normal valve core state, the spring breakage state, the stuck valve core state, and the slightly stuck valve core state, respectively, which are denoted as the present frequency domain feature vector, the frequency domain feature vector for the normal valve core, the frequency domain feature vector for the broken spring, the frequency domain feature vector for the stuck valve core, and the frequency domain feature vector for the slightly stuck valve core, respectively.