Predictive Maintenance and fault diagnosis of lift key components based on a hybrid deep learning framework

I&T Solution Predictive Maintenance and fault diagnosis of lift key components based on a hybrid deep learning framework
(REF: S-1136)
Trial Project
Solution Feature
  • This solution adopts a hybrid deep learning framework, comprised by a supervised learning model, a self-supervised/unsupervised learning model and an expert system, to predict the abnormal/fault events. The architecture and of underlying framework and the functionality of each model can be referred in the uploaded graph.
  • In order to improve the learning efficiency and accuracy, transfer learning technology will be applied if multiple cross-brand/model lifts are considered.
  • In order to identify the correlation across components and other potential impact factors, both linear correlation analysis and machine learning based feature significance analysis will be performed.
  • If the dataset is collected with high sampling frequency, the features of both frequency domain and frequency domain need to be studied to give a full portrait of each sample.
  • The AI model will be deployed on the designated server (either cloud or local) to process the online data stream and render the results in realtime. A dedicated web-based dashboard will be also developed to showcase the AI results, including the component status and timely warnings of the detected faults.
Trial Application and Expected Outcome
  • Literation review, existing relevant technology summarize and preliminary validation of technology feasibility.
  • Preprocess the collected dataset, offline training of the AI candidate models and identify an “optimal” generalized model dedicated for our dataset through continuous performance analysis and comparison.
  • A web-based dashboard will be provided to demonstrate the effectiveness of proposed solution, the functions include but not limited to the abnormal/fault event logging and alert, statistics of KPI and trend analysis.
  • A Interim Study Report (at the middle stage of project) and a Final Study Report (at project end) will be provided, including but not limited to the following aspects: a) Project Summary and Plan; b) Method statement; c) Performance Verification; 4) Conclusion.
  • Source code and OM manual for the developed AI models will be provided at the end of project.
Additional Solution Information W-0394_Schematic of Our Solution.pdf
Info on I&T Solution Provider
Solution Provider:Ergatian Limited
Address:Office Room 18, Smart-Space IT Street, Level 3, IT Street, Cyberport 3, 100 Cyberport Road, Hong Kong
Contact Person:Cynthia Yu
Position:Project Manager
Tel:6231 7146
Email: chaisongjian@gmail.com

For details of the above I&T solution, please contact the I&T solution provider.