Elevator fault pre-diagnosis method

You are here: Homepage > Search

Elevator fault pre-diagnosis method

Free access

WANG SUYU / GUO JUN / ZHOU HAN / WANG LINGZHU

The invention discloses an elevator fault pre-diagnosis method which comprises the following steps: extracting a characteristic variable related to the working state of an elevator from structured data, taking the characteristic variable as a corresponding variable, taking whether a class C fault occurs as a result, and taking occurrence and non-occurrence as two levels to carry out variance analysis; according to a variance analysis result, judging whether the corresponding characteristic variable has an influence on the C-type fault so as to obtain a variable influencing the occurrence of the C-type fault, and in combination with a variable correlation analysis result, screening out a key variable influencing the occurrence of the C-type fault so as to establish a fault prediction model and carry out model training; and if the accuracy rate, the recall rate, the correct rate and the F1 value of the fault prediction model all meet the corresponding preset requirements, the fault prediction model is output to predict whether the elevator can have C-type faults or not in a period of time in the future. According to the elevator fault pre-diagnosis method, key factors generating elevator faults are quantified from different dimensions, and pre-diagnosis of the elevator faults is achieved.

本发明公开了一种电梯故障预诊断方法，包括：从结构化数据中提取出与电梯工作状态有关的特征变量，将特征变量作为相应变量，将C类故障是否发生作为结果，将发生与不发生作为两个水平，以进行方差分析，根据方差分析结果判断相应特征变量是否对C类故障有影响，进而得到影响C类故障发生的变量，并结合变量相关性分析结果，筛选出影响C类故障发生的关键变量，以此建立故障预测模型并进行模型训练；若故障预测模型的准确率、召回率、正确率和F1值均满足相应的预设要求，则输出故障预测模型以预测电梯在未来一段时间内是否会发生C类故障。本发明提供的电梯故障预诊断方法从不同维度来量化产生电梯故障的关键因素，实现电梯故障的预诊断。