Door area outside elevator stopping detection method based on multisource sensor fault tag sample data

You are here: Homepage > Search

Door area outside elevator stopping detection method based on multisource sensor fault tag sample data

Free access

ZHANG XINGFENG / ZHANG LEI / WAN MIN / CAI WEIWEI / JIN XUZHE

The invention relates to a door area outside elevator stopping detection method based on multisource sensor fault tag sample data. The door area outside elevator stopping detection method includes thesteps that S1, a data collection device is used for conducting data information collection; S2, a fault tag sample library is established, calibrated fault tag samples are put into the fault tag sample library, and the data volume of the fault tag samples is judged; S3, if the data volume of the fault tag samples is smaller than the set threshold value, data information obtained in the step S1 isdetected through an anomaly detection model, if it is judged that the elevator stops outside a door area according to a detection result, the data information is adopted as new fault tag samples to be put into the fault tag sample library; and S4, if the number of the fault tag samples is larger than the set threshold value, data information obtained in the step S1 is detected through a supervised dichotomy model, and then a detection result is output under the condition that the detection precise meets the requirement. The method is low in cost and high in detection precision.

本发明涉及一种基于多源传感器故障标签样本数据门区外停梯检测方法，包括：S1、通过数据采集装置进行数据信息采集；S2、建立故障标签样本库，将标定好的故障标签样本放入所述故障标签样本库中，并判断所述故障标签样本的数据量；S3、若所述故障标签样本的数据量小于设定阈值，则将所述步骤S1获得的数据信息通过异常检测模型进行检测，若检测结果判定为电梯门区外停梯，则将所述数据信息作为新故障标签样本放入到所述故障标签样本库中；S4、若所述故障标签样本的数量大于设定阈值，则将所述步骤S1获得的数据信息通过有监督的二分类模型进行检测，再检测精度符合要求的情况下输出检测结果。本发明的方法成本低，检测精度高。