15. Nonlinear Control Theory and Applications to Drive Systems

Dr J Wang and Mr L Yang

In general, dynamical systems cannot always be precisely defined because some approximation, imprecision or uncertainty may have been introduced during the modelling procedure.  Mathematical models of actual systems always contain uncertain elements, which model the designer's lack of knowledge about parameter values and disturbances.  A class of nonlinear uncertain dynamical systems, which is affine in the control input, is considered.  To encompass all possible realisation of uncertainty in the systems, differential inclusion is adopted to model this class of uncertain dynamic systems.  Deterministic feedback control proposes the use of linear or non-linear feedback control functions, which operate effectively over a specified magnitude range of system parameter variations and disturbances.  Applying these deterministic methods, the synthesis of stabilising feedback controls may give rise to controls with 'high' gain in order to address the problem of uncertainty in the system.  For all physically realisable realistic systems, however, there is some implicit bound on the gain that is allowable for such stabilising controllers.  Therefore, the problem of constructing a stabilising controller, whose gain satisfies a constraint defined in terms of a practical acceptable threshold level, needs to be considered.  The objective of the project is to develop a constrained generalised adaptive feedback control strategy for stabilisation of nonlinear uncertain dynamic systems.  Applications include power system, electrical drives, and pneumatic actuator control through real-time implementation of the control strategy for a back-up generation system.

16. Real-time Modelling and Control of HCCI Engines

Dr J Wang, Dr K I Nuttall and Mr N Jia

Homogeneous Charge Compression Ignition (HCCI) engines are being actively researched as an alternative to spark ignition and conventional compression ignition engines due to their low emissions and high efficiency.  Significant advances have been made in expanding the operating window HCCI operation, which has been limited due to violent combustion at the higher range and misfire at lower engine loads.  However, combustion control is a challenge for HCCI engines to become a commercial success.  The project objective is therefore to evaluate potential HCCI engine combustion control strategies, in particular with reference to varying load and speed conditions.  The project is organised in two phases: 1) HCCI engine modelling for control - Development of an HCCI engine model for the purpose of real-time control evaluation.  The model will place emphasis on describing the relationship between the engine output performance and the controllable input parameters.  2) Development of control strategies and systems - Evaluation of potential closed-loop control strategies for HCCI engines in order to maintain proper ignition timing as load and speed are varied and to keep the combustion optimised over an appropriate operating range.

17. Robust Control of Time-delay Systems

Dr Q C Zhong

Systems with delays frequently appear in engineering.  Typical examples of time-delay systems are communication networks, chemical processes, teleoperation systems, biosystems, underwater vehicles and so on.  The robust control of such systems is quite difficult and involves complex mathematical tools, such as chain-scattering representation, J-spectral factorizations, differential and algebraic Riccati equations, L2[0,h]-induction norm, frequency-domain techniques etc.  As part of EPSRC-funded research, the project focuses on some challenging problems like the H-infinity control of systems with multiple delays, where the different channels in the plant may have different delays and these delays cannot be separated into input and/or output delays; the application of unified Smith predictor in H-infinity control an delay-type Nehari problem; the comparison of unified Smith predictor and modified Smith predictor, etc.

18. Implementation of Distributed Delay in Control Laws

Dr Q C Zhong

Distributed delays, i.e. finite integrals over time, also called finite-impulse response (FIR) blocks, often appear as part of dead-time compensators for processes with dead-time.  They also appear in H∞ control of (even, stable) dead-time systems and continuous-time deadbeat control.  Due to the requirement of internal stability, such an FIR block has to be, approximately, implemented as a stable block without hidden unstable poles.  This problem is not trivial.  The project investigates various ways to implement distributed delay in control laws so that the stability of the system is guaranteed.  These include the implementation using discrete delays in the s-domain and in the z-domain; rational implementations using the delta-operator and the gamma-operator etc.

19. Stability Analysis of Time-delay Systems

Dr Q C Zhong

The robust stability analysis of time-delay systems is a very important topic, e.g, in communication networks and network-controlled systems etc.  Both delay-dependent and delay-independent stability criteria for a congestion control algorithm and a mass-spring-damper system controlled via the internet have been obtained using a dual-locus diagram method.  We have used the concept of J-lossless bistable matrices in solving some H-infinity control problems and the underlying idea will be further investigated to develop a new method for stability analysis.  The results will be applied to the stability analysis of combustion systems.

20. Advanced Process Control

Dr Q C Zhong

This project focuses on the investigation of advanced control theories for chemical processes.  These include dynamic matrix control, PID control and auto tuning, control of inverse-response processes, control of processes with dead-time etc.

21. Control of Grid-connected DC/AC Power Converters

Dr Q C Zhong

This project concerns DC/AC power converters, which take electricity from DC energy sources such as batteries, fuel cells, solar panels, and, more possibly, rectified DC supplies from variable frequency AC sources.  The converters feed power into local grid and into the public grid.  Controlling such a converter is a delicate task, since the local grid voltage should be clean sinusoidal and synchronized with the public grid, and the active power is maintained at a desirable level and the reactive power is small.  The control design involves advanced robust control, repetitive control, infinite-dimensional systems, power electronics, DSP and PLL.  The main control problems include H-infinity repetitive control of the output voltage, active power and reactive power control, H-infinity control of a balanced neutral point, which is crucial for 3-phase 4-wire systems.

22. Control Using Delay Elements (Time-delay Control)

Dr Q C Zhong

It has been well known that delay is bad for control.  However, it has also been shown that using delay elements, intentionally and reasonably, can improve some system performance, even stabilizing a system.  Repetitive control (or Time Delay Learning Control, TDLC) is one of the possibilities.  Delay elements can be used to learn periodic signals so that periodic signals can be perfectly tracked or rejected.  This has been widely used in power electronics.  Another way to use delay elements is to construct time-delay filters to reduce the residual vibration of flexible structures.  This has been widely applied to spacecraft, robots, hard disk drives, servo systems, etc.

23. Control of Integral Processes with Time Delay

Dr Q C Zhong

Integral processes with time-delay are simple, but difficult to control because of the instability and the inherent time-delay.  Various schemes were proposed to control the system and detailed analysis was made.  These include a disturbance observer-based control scheme, quantitative analysis of achievable performance and stability regions.