Technology Lecture: Model-based Health Diagnosis of Hybrid Systems: A Bond Graph Approach

Date: 03 Feb 2009 - 03 Feb 2009

Venue: SIMTech Training Room, Tower Block, Level 3

The objective of this technology lecture is to create awareness of the current trends and technologies in the area of model-based equipment health diagnosis for manufacturing industry.  This lecture is sponsored by IEEE Singapore Robotics and Automation Chapter.

This presentation focuses on the model–based diagnosis of systems with hybrid dynamics, using bond graph method. Hybrid systems consist of continuous behaviour and discrete states represented by modes. In each mode, the system is governed by continuous dynamics and different modes correspond to different continuous models. System health diagnosis is a key feature for early detection of faults, failure prevention, reliability and condition based maintenance. A health diagnosis system integrates three main skills, namely fault detection, fault isolation and fault parameter estimation.

Bond Graph (BG) is an effective tool for modelling complex systems and it has been proven useful for Fault Detection and Isolation (FDI) in continuous systems. BG provides the casual relations between the system's variables which allows health diagnosis algorithms to be developed systematically from the graph. In the same spirit, Hybrid Bond Graph (HBG) is a BG-based modelling approach which provides an avenue to modelling of complex hybrid systems.

A novel health diagnosis method for systems with hybrid dynamics will be presented. This novel method can apply not only to hybrid systems such as vehicles and manufacturing machines but also to the applications of intelligent maintenance systems for industry. The method is based on the hybrid bond graph modelling approach and a new concept of Global Analytical Redundancy Relations (GARR).  The GARRs are unified relations which describe the system behaviour over all its operating modes, and form a set of residuals for FDI. GARRs are derived systematically from the HBG. One of the advantages of this approach is that a separate analysis of individual modes is not required. In addition, properties such as monitoring-ability and cause-effect relations between faults, mode-changes and residuals, are directly and systematically analysed from the GARRs.

About the Speaker
Born in Israel, Dr Shai Arogeti received his BSc degree in Mechanical Engineering, MSc and PhD in 1997, 2000 and 2006 respectively from Ben-Gurion University of the Negev (Israel). During his academic studies, he won several awards, including the 1994 and 1995 mechanical engineering department Chairman award for outstanding academic achievements, the Jacob's award of 1995 for outstanding achievements in the field of thermo-sciences, the Wolf Foundation stipend 2000/01 for outstanding MSc students, VATAT stipend for outstanding PhD students who specialised in high technologies.

He began his career with the Ben-Gurio University of the Negev (Israel) as an instructor to teach mechatronics, micro-computers in mechanical systems, digital control systems and assumed the post of principal developer of MATLAB interactive course in 2005. 

In 2006, he moved on to join the School of Electrical and Electronic Engineering, Nanyang Technological University as Research Fellow. His latest project on the "performance monitoring, diagnosis and prognosis (PMDP)" module is a collaborative A*STAR research project encompassing  "integrative serviceability management and supervisory control of networking manufacturing systems and virtual enterprises". He has been proactively involved in developing health monitoring methods for mechanical and electrical equipment. His area of research interests include nonlinear control, robotics and automotive control, health monitoring diagnosis and prognosis. He has authored and co-authored extensively over 30 international professional journal and conference papers on model-based fault diagnosis, robotics and automotive control. 


1.45pm     Registration

2.00pm     Presentation by Dr Shai Arogeti

3.00pm    Q&A 

3.30pm    Refreshment & Networking

4.00pm    End

Who Should Attend
R&D managers, researchers, scientists, engineers, academic staff and engineering students.

Pre-registration for the lecture is free of charge. Seats are available on a first-come, first-served basis.
For research and technical enquiries, please contact: 
Dr Luo Ming, Research Scientist, Email:

For general enquiries, please contact: 
Alice Koh, Email: