Action state detection method and system for elevator brake

An action state detection method for an elevator brake having brake linings, comprising the following steps:

• acquiring mechanical waveform signals of two of the brake linings in the elevator brake respectively, and determining frequency domain features based on a difference between the two mechanical waveform signals;
• at the same time, acquiring angular acceleration data of braking wheels in the brake, the elevator braking duration and the elevator running speed of two braking states at the moment of starting braking by the braking wheels in the brake and the middle moment of a braking process, respectively, and obtaining a braking power descriptor of the elevator based on the elevator braking duration and the elevator running speed of the two braking states;
• and constructing an angular acceleration sequence of braking according to the angular acceleration data of the two braking states;
• acquiring degrees of similarity of different braking states of the brake linings;
• selecting a minimum value in the degrees of similarity based on the degrees of similarity;
• obtaining a pairing decision factor based on the minimum value;
• and pairing any two of the brake linings based on the pairing decision factor to obtain a matching pair of the brake linings in the same braking state;
• obtaining the degrees of similarity of the brake linings in the matching pair based on the braking power descriptor, the angular acceleration sequence and the frequency domain features of each of the brake linings in the matching pair;
• acquiring a domain set of each of the brake linings in the matching pair, and determining the braking performance distance of any of the brake linings for sorting all the corresponding braking performance distances of the two brake linings from small to large based on the degrees of similarity, the weight of any sample in the domain set of one brake lining and the other brake lining and mechanical waveform signals;
• acquiring sample data corresponding to the first K′ braking performance distances;
• and constructing a normal record data set, with K′ greater than or equal to 1;
• grouping the normal record data sets based on braking power corresponding to samples in the normal record data sets to determine a median of elevator running speed corresponding to the samples in each group;
• dividing the operating state levels of the elevator brake according to the size of the median;
• and forming a data training set by the frequency domain features and the operating state levels;
• training a constructed network model by the data training set to obtain a trained network model;
• acquiring the braking power descriptor and the elevator running speed of an elevator to be detected in real time;
• inputting the braking power descriptor and the frequency domain features of the elevator into the trained network model;
• outputting the operating state levels of the elevator brake;
• evaluating the elevator brake according to the operating state levels;
• and controlling the elevator.