Electrical Drives, Power and Control Research Group

1. Power System Planning, Operation and Control with Impact of New Technology

Professor Q H Wu and Mr D R Turner

The rapid expansion of electrical power systems in China has received worldwide attention.  Energy efficiency has been listed with a high priority in research programmes in Britain.  The University of Liverpool, UK, has a traditional strength in electrical engineering research and carries out active research in the area of emerging knowledge and new technology.  Electric Power Research Institute (EPRI), China, is the top organisation engaged in industrial and academic research responsible for the Chinese electrical power industry.  Based on these merits collaboration between the University of Liverpool and EPRI, in the area of applications of emerging knowledge and new technology in electrical power systems, has been identified as being significant and funded by the British Council under its Academic Links Scheme.

The proposed links will establish an environment for Chinese and British academic researchers and industrial engineers to jointly investigate the impact of new technology on power system planning, operation and control.  The links involve not only specific research projects, which focus on the frontier of power system research and development, but technical exchanges and collaboration between Chinese and British industries.  The research projects are concerned with the studies of power system planning, operation and control with the impact of new technology, which includes FACTS, CUSTOM POWER and emerging knowledge of Intelligence Engineering.

2. Power System Load Modelling and Network Reduction

Professor Q H Wu

Various mathematical models are required by the electricity supply industry to study power system performance and control, and to undertake daily industrial operation and dispatch.  The accuracy of the models determines the quality of the research studies and the efficiency of industrial electric power production.  The conventional methodology of power system modelling cannot guarantee the quality of models. In fact, the process of power system modelling, like that of any other industrial systems, consists of physical analysis, mathematical deduction and modelling induction.  Based on the existing knowledge of mechanical motion, electromagnetic phenomena and network dynamics of the power system components, a trial and error method is commonly (inevitably in practice) employed to develop power system models as addressed in the above four modelling steps.  Due to the complexity of power systems, most efforts made to construct accurate models are very dependent on the trial and error involved in the above modelling process.

Machine learning provides a new way to achieve system modelling, in particular for large-scale and complex industrial systems.  Intelligent learning offers a powerful methodology of systematic deduction and induction to construct power system models based on real-world data obtained on-site at industrial power systems.  This project is concerned with an investigation of a methodology using evolutionary computation techniques based on system measurements to construct power system area load models alongside with distribution network reduction.

3. Economical Power System Operation and Voltage Control

Professor Q H Wu

The power system economical operation has been considered conventionally as two separate problems: P- and Q-problems.  The P-problem is to regulate active power outputs of generators to minimise fuel costs.  The Q-problem is to control voltages of PV-buses and tap settings of the under load tap changing transformers to minimise network power loss.  Various methods including conventional optimization techniques and evolutionary computation algorithms are investigated to solve the economical operation problem, which are complicated by the regulation rules introduced recently in the British electricity market.  The research project attempts to tackle the integrated problem of power dispatch, voltage control and electricity market.

4. Learning Coordinated Control of Multi-machine Power Systems

Professor Q H Wu and Mr K H Chan

With the advances in power electronics technology, the Flexible AC Transmission Systems (FACTS) have been regarded as a powerful means to improve the power transmission and system stability.  This project is concerned with an investigation into fuzzy logic control of FACTS and an application of reinforcement learning techniques for optimisation of the parameters of the fuzzy logic controller employed for FACTS control to enhance the stability of electric power systems.  At first a fuzzy logic control strategy is proposed for FACTS control, based on the analysis of the transient power.  Then a reinforcement learning method is introduced to optimise the parameters of the fuzzy logic controller without requiring the system model.  In the learning process, the power system is assumed to be an unknown dynamic system.  A team of interconnected learning automata is used to search for the optimal parameters of the fuzzy logic controller according to a given performance index. 

The research work is generally concerned with investigation into applications of learning automata based and dynamic programming based reinforcement learning techniques for coordinated control of power system devices.  Design of learning controllers does not require models of power systems.  The learning controllers are installed locally and are able to control the individual devices in a co-ordinated manner.  The aim of learning coordinated control is to explore the potential of system stability, security and economic operation.

5. Intelligent Condition Monitoring and Fault Diagnosis of Electrical Power Apparatus

Professor Q H Wu, Dr J Z Lu and Mr W H Tang

The power transformer is a major apparatus in a power system, and its correct functioning is vital to system operation.  It is therefore very necessary to closely monitor their in-service behaviour, in order to avoid catastrophic failures and costly outages and improve the management of maintenance and servicing.  The project focuses on the development of an intelligent condition monitoring and fault diagnosis system for power transformers.  The system design is based on the analysis of power transformer thermal dynamics, dissolved gas and partial discharges.  Intelligent detection of transformer winding distortion is also investigated.

The integration of machine learning, information extraction, intelligent classifier, case-based reasoning and advanced software development, undertaken in the project, provides an intelligent platform for condition monitoring and fault diagnosis of power transformers.

6. The Study of Chaos in Stochastic Aspects of Dynamics

Professor Q H Wu, Professor R J Bhansali and Mr K W Lau

The study of chaotic dynamics has been of interest to many physicists, chemists and mathematicians and it has received attention from researchers working in other fields of science and engineering.  Chaotic dynamics are always found in many engineering systems, where there exist complex, hierarchical and distributed structures, incomplete system information, uncertainties and immeasurable states.  However, there is an obvious lack of methodologies to analyse and control the chaos in such chaotic systems.

This project investigates an equivalence relation between chaotic and stochastic systems, and develops an equivalent stochastic system model and a learning control strategy for control of chaotic dynamics.  The reconstruction of the equivalent stochastic system model using learning techniques based on measurable time series of a chaotic system will be then investigated.  This will lead to a realistic approach to the control of high-dimensional chaotic systems using low-dimensional equivalent models and provides a smooth path for development of a learning control methodology for the control of unknown complex chaotic systems.  University of Liverpool Graduates Association (Hong Kong) supports the study.

7. Coordinated Control of Large-scale Nonlinear Dynamical Systems

Professor Q H Wu and Dr L Jiang

This is an EPSRC ROPA project entitled ‘The coordinated control of large-scale nonlinear time-varying dynamical systems’.  The methodology is developed for design of the distributed controllers that employ a minimum number of local measurements to coordinate global system performance, by using of estimation of system nonlinearities, perturbations and interconnecting dynamics.  The problem has been formulated from a wide range of applications in complex inter-connected industrial systems, such as power systems, joint-action robots, and telecommunication systems, where there exists complex system structure which is difficult to model, time-varying characteristic and unpredictable disturbances, strong nonlinearities which degrade the control performance, local dynamics (subsystems) interacting with each other and a limited number of local states (in particular avoiding using of remote states) for control implementation.  For example, Thyristor Controlled Series Compensators (TCSCs) are installed on transmission lines in power systems in a distributed manner but their control must be implemented locally. 

The project aims to:  (1) investigating system decoupling by devising a fictitious states and designing an observer in local systems for estimation of system nonlinearities, perturbation and interacted dynamics between subsystems; (2) study coordination of local controllers contributing to global system performance.  The application work will be concerned with the control of FACTS (Flexible AC Transmission System) devices installed in power systems.

8. Development of Intelligent Power Electronic Drive for Permanent Magnet DC Motors

Dr K I Nuttall and Professor Q H Wu

This research study concerns the development of an electronic drive for permanent magnet DC motors in which the drive electronics is to be integrated within the motor to give a high efficiency, flexible and compact motor with low electromagnetic interference.  A self-commutation scheme has been developed to further improve the reliability of the design and embedded intelligence is to be included to optimise the performance with regard to efficiency, torque/speed and ripple characteristics.  A special feature of the design is the arrangement of the power drive system that consists of distributed integrated power drive units, allowing the power to be efficiently dissipated and facilitating the use of the latest state-of-the-art electronic devices.

An embedded intelligence system will allow the performance of the motor to be effectively monitored and appropriate commutation control techniques to be applied.  Application of these methods is expected to provide the basis of a novel new machine with higher power/size and power/weight ratio that conventional separated drive and motor systems, as well as substantially lower electromagnetic radiated interference.  The project will appeal to applicants with interests in the development of novel intelligence algorithms and their application to a modern electronic electric motor drive system.

9. Reduced Order Controller Design

Dr E Prempain

The Static Output Feedback control problem has been the subject of much research.  The literature on the design of linear static output feedback controllers is rich and contains many references.  It is known that a large number of systems can be controlled by low-order controller with good closed-loop performance.  But, the synthesis of a static output feedback gain or a controller of fixed order is still an open problem.

Recently, linear matrix inequalities (LMI's) have attained much attention in control engineering (see for example Boyd et. al. 1994) since many control problems can be formulated in terms of LMI's and thus solved via convex programming approaches.  In this project attention is focused on developing new sufficient LMI conditions for solving the SOF problem via LMI optimization.

10. Control of an Induction Machine using an H-infinity Approach

Dr E Prempain

Induction motors are widely used to convert electrical energy into mechanical energy.  Induction motors are extremely cheap, easy to produce and very robust.  They are good at producing mechanical torques/ forces where accuracy is not the main issue (transportation, drilling machines etc…).  But they are not still used in robotic if high precision tasks are required.  Actually, in robotic, most of the actuation is performed by DC motors.  DC motors have good precision capabilities despite some major drawbacks (high initial and maintenance costs, mechanically more fragile).

To make the induction motors more accurate one can use feedback control.  Recent, theoretical works show that high performance controllers can make induction motors competing with the DC ones.  However, such feedback ideas are not yet well understood or even known in the industrial world.

The aim of this project is to produce some generic software tools to simplify the task of the practical engineer to design such high performance controllers.  Also, our objective is to illustrate the usefulness of the proposed tools on an induction motor bench.

11. R&D of Power Quality Conditioner for Electric Railway Traction Systems Based on IGBT Inverter

Professor Q H Wu, Mr D R Turner, Mr H F Ding, Mr D J Zhang and Dr L Jiang

As a major load operating in electric railway traction systems, the electric locomotive has various detrimental characteristics such as single phase, nonlinearity, impact, fluctuation and heave-duty, etc.  All of them have been raising great attention from both customers and engineers over the past decades with respect to the issues of power pollution, stability of the power grid and cost effectiveness.  This project takes advantages of the modern power electronics conversion and switching technology techniques under the concept of Flexible AC Transmission System (FACTS) with the aim of improving the power quality.  Mathematical modelling and performance analysis are involved in the project and a prototype of the power quality conditioner is under development, which is capable of regulating the voltage at the load bus or compensating the harmonics current and the reactive power generated by the nonlinear load, with advanced DSP control technology.

12. Advanced Power System Protection Relays

Professor Q H Wu and Mr D J Zhang

Relaying technology has been playing a crucial role in the operation and control of power system.  It has advanced dramatically in recent years along with the progress of computer and microprocessor techniques.  However, most of the protection relays employed today just take the advantage of computer techniques to realize traditional principles.  How to improve performance substantially based on the emerging knowledge and measurement technology is a significant field of research.  This project is to focus on introducing forefront methodologies, for example Wavelet Transformation, Mathematical Morphology and Support vector, along with Internet IP technology, to develop a new generation of protective relays.

13. Coordinated Voltage Control Based on Short-term Load Forecasting

Professor Q H Wu and Dr J Y Wen

Automatic tap change control (ATCC) and automatic reactive switching (ARS) equipments are usually used in HV substations to keep the busbar voltage within their limits.  They only use the current information of the substation (LV and HV busbar voltages) to calculate the control strategy.  The current voltages are com-pared with the set limits (deadbands) at each sample time, if they are outside of the deadbands, the automatic equipments start to count up the outside times.  When the total time reaches the pre-set delay time, the operation of transformer tap and shunt capacitors are initiated.  As the load of the substation is always in changing, some of the operations are not unnecessary because the voltage would return to the set limits itself sometime.  To decrease the operation times will prolong the life of transform tap and capacitor switchgear.  The conventional methods use pre-set delay time to avoid unnecessary actions, as it used only the system current states, it can not obtain an optimal result for a period of times, such as one day

Coordinated voltage control (CVC) provides a new way to prevent unnecessary tap changer and capacitor switching actions based on short term load forecasting.  It forecasts the voltage and reactive power of the substation firstly and controls the tap changer and capacitor switchgear according to both the forecasting and current information.  It will minimize both the operation times of tap changer and capacitor switchgear and the outside time of voltage in one day.

14. Improvement in Energy-efficiency of Servo Pneumatic Actuator Systems

Dr J Wang and Miss J Ke
Pneumatic actuators have been widely used in industrial processes.  Compared with their counterparts - electrical and hydraulic actuators, pneumatic actuators have distinct advantages: cleanliness in the environment, high load-carrying capacity to size ratio, low cost and ease of maintenance.  However, energy efficiency of pneumatic actuator systems is low.  A report by the British Fluid Power Association (BFPA) indicates that in the UK and other European countries, an efficiency of 23%-30% is achieved.  So improvement of energy efficiency is of foremost importance.  The project is proposed to investigate the methods of improving energy efficiency of servo-controlled pneumatic actuator systems through optimisation of profiles.