Real-time Lift Data Acquisition Device Based on Non-intrusive Sensors (REF:C-0029)
E&M InnoPortal Trial Project Ref. No.:

Overview

A Non-Intrusive Data Analytics based System for Adaptive Intelligent Condition Monitoring (NICM) of lifts is developed to diagnose the healthiness of thousands of multi-brand lift installations and send alerts about shabby lift conditions to maintenance agencies. The system utilises the current signals of key components and cost-effective sensors.

Problem Addressed

Condition monitoring is important to detect potential failures and reduce lift breakdowns. However, it requires intensive time to monitor the quality of lift works by regulatory bodies and also difficult to ensure satisfactory performance with a mix of ages, brands and types of lift equipment using different proprietary software programmes. Building owners and the management personnel often have little knowledge about lifts without recourse to lift manufacturer.

To address this issue, a novel non-intrusive AI-based data analytic condition monitoring system is developed. It requires limited sensor installations for monitoring electrical current without interfering existing circuitry. A unified monitoring platform is implemented across multi-brand lifts to enable effective communications among the different proprietary software programmes used by aged and modern lifts.
Innovation

The non-intrusive data acquisition system is composed of three modules, namely Clamp-type Current Transformer, Microcontroller with Data Transmission Module and Cloud-based Server with well-trained Deep Learning Model. Current sensors are installed non-intrusively on lift brake coil, traction motor, safety link and door circuit.

All data collected by the sensors is converted into digital form with high-resolution ADC modules and transferred to the embedded system for further processing and analysis. After preprocessing the raw dataset to address tasks such as loss data imputation, data cleaning, maintenance data removal and other related challenges, two monitoring systems, namely the expert system and the AI-based system, will operate simultaneously to ensure precise detection of faulty events.

With the advanced communication technologies of RS485 and 4G, the collected data is efficiently uploaded to a cloud platform periodically. A Graphical User Interface (GUI) system is developed for data monitoring and visualisation.
Key Benefits
  • Leveraging AI and big operation datasets, it delivers a new cost-effective approach that enables the timely scheduling and quality monitoring of maintenance jobs, therefore shortening or reducing lift breakdowns
  • The deep learning algorithm automatically learns important feature patterns in data samples according to given targets, viably exploiting complicated features and performing preventive and predictive fault detection
Patent and Award
  • HK Patent (Granted)
  • International Exhibition of Inventions of Geneva 2021 (Silver Medal)     
  • Build for Asia 2020 Outstanding Award (Building)
  • City I&T Grand Challenge Innovation Award
Project Reference

Government buildings, Hing Wah Estate and Housing Authority properties