Elevator emergency stop people trapping fault detection method based on multi-model fusion

Abstract

The invention relates to an elevator emergency stop people trapping fault detection method based on multi-model fusion. The elevator emergency stop people trapping fault detection method comprises the steps that S1, acceleration data, audio data and image data in an elevator car in the elevator running process are collected; S2, data preprocessing is performed on the acceleration data and the audio data, and acceleration features in the acceleration data and audio features in the audio data are respectively extracted; S3, abnormal acceleration detection is carried out based on the acceleration features and an abnormal acceleration detection result is obtained, and abnormal sound detection is carried out based on the audio features and an abnormal sound detection result is obtained; and S4, the abnormal acceleration detection result and the abnormal sound detection result are fused, and whether an emergency stop fault alarm is generated for an elevator or not is judged based on the fusion result. The scheme can be realized only based on sensors and in cooperation with monitoring camera equipment, so that the use cost and the installation difficulty are greatly reduced.

Description

technical field [0001] The invention relates to the field of elevator operation monitoring, in particular to a multi-model fusion-based fault detection method for elevator emergency stop and trapped persons. Background technique [0002] With the development of society, the application of elevators has gradually become popular. However, while more and more elevators are being used, the number of elevator failure accidents is also increasing. The main reason is that relying on manual inspection of elevators and property or passenger repairs, there are missed reports, delayed reports, and false reports. situation, resulting in the accumulation of faults. At present, the methods used in elevator fault diagnosis mainly include rule reasoning, sensor detection, and model calculation. The rule reasoning method has low reasoning efficiency due to the difficulty of knowledge acquisition and self-learning; sensor detection, the mode is relatively single, and a large number of senso...

AI Technical Summary

Problems solved by technology

It can be seen that this method not only requires a large number of microcontrollers and complex circuits, but also requires the collection of a large amount of data to build expert knowledge

Furthermore, it is expensive to use and difficult to collect data

Images