The invention discloses a hidden danger pre-diagnosis method of an
elevator traction system, relates to the technical field of
elevator safety monitoring, and is used for solving the technical problems of poor real-time performance, low accuracy and difficulty in defect positioning of hidden danger diagnosis of the
elevator traction system in the prior art. The hidden danger pre-diagnosis method for the elevator
traction system comprises the steps that multi-
source data are synchronously collected, specifically, a
current transformer is deployed in a power supply loop at the input end of an elevator
traction motor to collect a three-phase current instantaneous value, a
speed measurement encoder is used for measuring the speed, the elevator running state, the speed, the floor position and the time-
variable load weight are synchronously obtained, and a time-space correlation
data set is constructed;
signal self-adaptive preprocessing: performing layered
noise reduction, working condition alignment and dynamic normalization
processing on the acquired current signals; multi-scale
feature fusion extraction: current
signal features are extracted from a basic feature layer, a
time domain feature layer, a
frequency domain and a time-
frequency domain feature layer; and feature optimization and
decision making: after
feature screening and dimension reduction fusion, inputting a pre-training classification model to output hidden danger types, positions and risk levels.