Development of AI Models for Predictive Maintenance of Lifts

I&T Wish Development of AI Models for Predictive Maintenance of Lifts
(REF: W-0394)
Matched I&T Solution
Trial Project
Summary and Challenges
  • To achieve predictive maintenance of the components, we target to develop generalized AI models to detect the potential faults using the collected data.
  • The models will be based on the historical performance of a single component and the relationship among multiple components.
  • The collected data is in the format of text files.
  • The solution shall include Extract-Transform-Load of the mentioned data.
  • The AI model shall be applicable on an existing server in EMSD.
  • Additional hardware can be proposed if deemed necessary.
  • A dashboard shall be created to visualize the performance of each critical lift component, including the detected faults from the historical data, potential performance degradation and estimated end of life of the component.
Expected Outcome
  1. Lifespan prediction model and/or trend analysis of lift component performance. The components include but not limited to motor, ropes, lift door, lift car, brake and levelling;
  2. Findings of correlation across components / datasets and other potential impact factors;
  3. A dashboard :
    • showing the performance of each critical component and the trend analysis of the components;
    • prompting alerts when abnormalities are detected;
  4. Source code for the AI models;
  5. Operation and maintenance manual for the AI models;
  6. Provide recommendations on implementing predictive maintenance system including but not limited to the required data quality, choice of machine learning algorithms and conclude the project with a suggestion report.
Expected Trial Duration 15-month
Contact Information
I&T Wish Proposer:Electrical and Mechanical Services Department (EMSD)
Contact Person:Leung Kin Fung
Position:Building Services Engineer/General Engineering Services/F3
Tel:3757 6069
Email: leungkf@emsd.gov.hk
Upload Date 2022-01-11
Closing Date 2022-01-25