Expertise within dynamics and control systems
MEMARC has been working with dynamics and control systems since the 1980s.
In that time, MEMARC has attracted funding from the Engineering and Physical Sciences Research Council, Royal Society, Royal Society of Engineering, North West Water and British Nuclear Fuels Limited within this specialist area.
Within the area of dynamics and control systems, MEMARC’s researchers are specifically interested in the following:
- Modelling and control of automotive engines and vehicle engineering
- Process modelling, control and fault detection
- Neural networks and applications to control systems
- Fault detection and identification for nonlinear systems
Modelling and control of automotive engines and vehicle engineering
MEMARC began research into modelling and control of automotive engines and vehicle engineering in 2000. As part of this research, MEMARC collaborated with BMW AG E30 in Munich, conducting investigations on engine control and engine on-board monitoring. The real engine data was used to model and design air/fuel ratio control systems for VVT SI engines. As part of this work, MEMARC and BMW trained two PhD students.
Later, researchers worked with Cummins Inc. in Daventry on fault detection, isolation, as well as diesel engine modelling and control.
The research from these projects have been published in more than 40 journal and international conference papers, including more than 10 SCI papers and seven papers in IMechE Journal of Automobile Engineering. Another key area of expertise for researchers involved in this work is vehicle dynamics, including:
- Control monitoring
- Optimal design and management for suspension systems
- Powertrain systems
- Chassis systems
- Automatic controlled brake systems
By conducting research into vehicle dynamics, the researchers hope to obtain solutions for practical automotive engineering problems with modern AI techniques. To achieve this, MEMARC researchers have formed numerous industrial partnerships.
In terms of facilities, within this specialist area researchers can use two engine test beds, which are equipped with various sensors and interfaced to computer systems for data collection, adaptive control and engine monitoring. Researchers can also make sue of an engineering work bed for vehicle dynamics test and design.
Process modelling, control and fault detection
Within this specialism, researchers conduct work into process modelling, control and fault detection. The control techniques developed include:
- Adaptive robust control
- Adaptive decoupling control
- Nonlinear adaptive model-based predictive control
- Adaptive neural network model-based control
Researchers working in these areas make use of MEMARC’s multivariable biochemical reactor rig, which is sponsored by the Engineering and Physical Sciences Research Council, as well as an in-line waste water treatment process with pH neutralization, which is sponsored by British Nuclear Fuels Limited and several nonlinear multi-tank processes.
Neural networks and applications to control systems
Researchers within this specialism have worked on intelligent control for many years and are committed to bridging the gap between theoretical development and practical applications. New methods and techniques worked on by researchers within this area include:
- Structure adaptive RBF networks (published in IEEE Trans. On Neural Networks, 2000, in excess of 150 citations) and MLP networks
- Neural-fuzzy inference systems
- Nonlinear control system design using genetic algorithm (IEE Proceedings: Control Theory and Applications, 1998)
- Recursive Orthogonal Least Squares (ROLS) training for RBF networks (Neural Processing Letters, 2000)
- Neural network-based fault tolerant control (IEEE Trans. on Systems, Man Cybernetics, 2005)
- Neural network model based predictive control for internal combustion engines (Engineering Applications of Artificial Intelligence, 2014, this paper received Elsevier 2006-2010 top 10 citation paper award)
- Higher-order sliding mode control with neural network compensation (Neural Networks, 2008)
- Neural model-based robust and stable control (Neural Networks, 2008). Robust Control (ASME Journal: Dynamic Systems, Measurement and Control, 2007). Neural model-based model predictive control (IEEE Trans. on Control Systems Technology, 2006)
- Adaptive Control (IEEE Trans. on Systems, Man and Cybernetics, 2005)
- Neural network structure adaptation (IEEE Trans. on Neural Networks, 2000)
- Bilinear Parity space method (IEEE Trans. on Control Systems Technology, 2000)
- Bilinear fault detection filter (BFDF) (I.J. Control, 1997), Bilinear fault detection observer (BFDO) (Automatica, 1996)
The methods developed have been applied to hydraulic drive systems, gas-fired industrial furnaces, biochemical reactors, in-line pH neutralization process, internal combustion engines, PEM fuel cell stacks, glass furnace and wind turbine generation systems.
Loading staff profiles…
Contact the dynamics and control systems specialists
If you’d like to ask a question or find out more about information about this specialist area, please contact the team using the details below.
Director: Professor Ian Jones
Call: 0151 231 2506
Co-Director: Professor James Ren
Call: 0151 231 2525