P-0139

Trial Project P-0139
Matched I&T Wish Research and Design of a Machine Learning Based Adaptive Fault Prediction Data Analytic System for Lift Monitoring
(REF: W-0213)
Matched I&T Solution A deep learning based non-intrusive health monitoring and fault prognosis system for lifts
(REF: S-0220)
Solution Feature
  • Field measurement of lift system will be conducted to collect current and other important signals from lift system considering various normal and abnormal scenarios.
  • The lift electric signals can be simulated using our lab equipment. Typical lift faults will be considered. Then, an efficient data mining algorithm will be developed to extract the mostly important features that contribute to the specific faults.
  • Considering high volume and high sampling frequency of the collected data, we propose to employ deep learning framework to learn the complex relationship between the extracted features and different fault events. The model will be built and implemented efficiently on our powerful computing facilities.
  • The developed intelligent algorithms will be integrated to construct a software package with a friendly graphical user interface. The functions should include real-time lift condition display, abnormal events diagnosis and reporting, maintenance needs prediction.
  • Besides, an effective demonstration showcase to illustrate the proof-of-concept of the entire NICM system based on e.g. multimedia and animation technologies will be developed too.
Trial Information
Trial Site : EMSD HQ and Government Quarters
Trial Scale : 5 lifts
Trial Duration : January 2020 to June 2020
Info on I&T Solution Provider
Solution Provider:The Hong Kong Polytechnic University
Address:Room CF611, Department of Electrical Engineering, Hong Kong Polytechnic University, Hung Hom, Kowloon, HONG KONG
Contact Person:Prof Zhao XU
Position:Professor
Tel: +852 27666160
Email: eezhaoxu@polyu.edu.hk