| Engineering home | Future students | Current students | Alumni | Partnerships | Staff (Intranet) | Contact Engineering |
| Research | Departments and schools | Staff directory | A-Z index | Site map |
| Engineering home | Future students | Current students | Alumni | Partnerships | Staff (Intranet) | Contact Engineering |
| Research | Departments and schools | Staff directory | A-Z index | Site map |
Research Projects
Projects A, B and C below are major projects. I also have some interest in Projects D - G which are developing. Project A: Fault Detection and Fault-Tolerant ControlPut simply this project looks at how to deal with failures in a given control system. Normally when all is as it should be, a controlled system that has been properly designed should perform to specifications. As an example, consider aircraft flight control. Aircraft have various control systems in place for attitude control and trajectory control. When all is functioning well there are no problems. However, if there is a failure in one of control surfaces (actuators) e.g., the elevator, rudder, or ailerons, how will the controlled aircraft perform? There'll certainly be some performance degradation the consequences of which could range from mild to drastic. Similarly consequences exist in the case where a sensor fails, eg the pitch angle or pitch rate sensors. The purpose of fault tolerant control is to first detect when such a fault occurs, and then to do whatever is possible to minimise the performance degradation that follows. It turns out that there is quite an interesting range of problems involved here and the first problem of reliably detecting and isolating the fault is in itself not trivial. Project B: Robust ControlRobust control is about designing a controller for a plant in the case where the plant model is not precisely known. It's always a good idea to ensure that any controller designed contains some robustness to unmodelled dynamics. In a mathematical sense this is achieved by assuming that the plant may be modelled in terms of a nominal model embedded within a set of models, and designing a controller that will provide acceptable performance over the full set of plant models. In special cases where the plant may be modelled as a linear state-space model within the km/mu-sunthesis framework, then the relatively recent field of semidefinite programming is particularly applicable though one soon finds that the more interesting/realistic problems are generally not linear and at best are bilinear, and therefore very hard to solve. Project C: Modelling and Control of Short- and Mid-term Dynamics of Power SystemsThis is a major project. We have developed models suitable for each component within the power system tailored to the short- to mid-term dynamics. Inter- and intra-area oscillations may be modelled. We have chosen to implement these using a Simulink(TM) framework and written our own software components from the ground up because we wanted to have the flexibility to do certain tasks that canned package simulators do not offer (particularly for fault detection). Eight-order models of synchronous machines with coordinate change to Park/Blondel system, various blocks for modelling excitation systems, turbines and their associated speed-governing systems, power system stabilisers, transformers, transmission lines, loads, etc are all included in the simulation package. It can be used for oscillatory stability studies, transient stability studies and fault-tolerant control type studies. Nonlinear, high-order mathematical models are produced and there are some quite interesting challenges in controller design that present themselves. Project D: Neuropsychology From An Engineering PerspectivePart of the difficulty with this area is that it is relatively difficult to measure and influence signals. Where are the sensors and actuators to monitor consciousness, or personality disorders, for example? How much do we really know about the human mind? From my perspective, as one moves from engineering, to medicine/physiology, to psychology the field becomes more complex by several orders of magnitude. Systems that engineers deal with are for the most part simple, causal, and predictable. Engineers have sensors which they use to measure various critical signals within a system, and they have actuators that can be used to manipulate certain critical variables within a system. For example, we can precisely control frequency and voltage in an electrical power system, and we can use these same technologies to guide the most sophisticated high performance aircraft through highly complex manouevers. To a certain extent there has been success in carrying these ideas to the field of biomedical engineering. For example, we can automate drug infusion for patients, and there are sensors (eg X-ray machines, nuclear magnetic resonance imaging ) that can be successfully used to "see" within the human body. But there is not really much in the way of sensors/actuators for dealing with the mind. Sometimes certain quantities can be measured but cannot be isolated from other quantities. So it's a challenging new field to do research in. Project E: Financial Modelling and ControlMost of the modelling in finance/econometrics thus far has been done by mathematicians/economists. Examples include the highly renowned Black-Scholes model for option pricing, various approaches for computing statistics relating to evolving time signals, and portfolio optimisations. There might well be some added insights that we as engineers can bring to the field - particularly in the area of modelling. Project F: Synchrotron ControlMajor Projects Victoria (Victorian Government) are building the Australian Synchrotron . There are a number of control loops in the project that could be the subject of serious research projects. As well as more conventional plant type control in cooling the delicate equipment, water flow sensing, ultra high vacuum control and the protection of both personnel and valuable research equipment in a complex environment there is an entire research field devoted to the feed-forward and feed-back control of electrons in a large and high energy orbit. Orbit stability is of paramount concern for high brightness synchrotron sources, and this stability is critical to the end user scientific community. Special techniques are used to instrument the "beam", to manipulate it and to ensure stability under tiny vibrational movements, temperature changes and magnetic effects. This is one place where high energy physics meets real time process control!
Project G: Towards Photonic Atomic Clocks Using Ultra-fast Ultra-stability Phenomena in Fibre Lasers: Control Models and ImplementationThis project is lead by a colleague, Assoc. Prof Le Binh who is looking for some input on the nonlinear control related issues. Reference standard time is important and attracts much attention since the introduction of the atomic clock in 1948. Since then and more recently a number of laboratories in Europe and America have tried to achieved the most accurate and stable time reference systems using the interaction between microwaves, lightwaves and the atomic transitions in gaseous atoms such as Cs and Sr. On the other hand the transition of electrons in an optical amplification solid state medium would also offer an alternative atomic transition medium that can be incorporated in a photonic resonance structure. This project is proposed to tackle this new attempt to generate the most accurate and stable photonic atomic clock in competition to those bulk types reported by l’Observatoire de Paris of France and the National Measurement Laboratory of the UK. Objectives are:
Other important applications of the fibre lasers can be for optical sampling, photonic packet switching, soliton sources for solitonic logics and soliton transmission over long haul optically amplified fibre communications etc.
Other projectsThere are also some minor projects that I do with undergraduate and postgraduate students. Back to home page |