CITE - Innovative Informatics

Expertise and Current Investigations

Research into Control Techniques and Theory (Prof. Stephen Dodds; Prof. Jan Vittek; Prof. Teresa Orlowska Kowalska; Dr. Krzysztof Szabat)

Research has been in the general areas of a) robust control, which seeks to achieve a specified performance of an automatic control system in the presence of plant modelling uncertainties and unknown external disturbances, and b) nonlinear model-based control.  Notable is the time domain based forced dynamic control (FDC) method, which, through nonlinear state feedback, yields a specified closed-loop performance that may or may not be nonlinear and takes external disturbances into account. 

A new approach to robust control with model order uncertainty is now being pursued stemming from sliding mode theory but deliberately introducing higher output derivatives than for the conventional sliding mode control.  Preliminary simulation studies have revealed the remarkable property of this new controller in producing precisely the same prescribed dynamic response to a given reference input for a range of different plants spanning an order range from 1 to N, where N is the order of the highest output derivative fed back.  Since the aforementioned higher output derivatives create state variables that, in addition to the plant state variables, take part in a sliding mode, the new technique has been called hyper sliding mode control.

A further innovative approach to robust control is now in the process of maturing, based on the notion that plant uncertainties can be represented as time varying disturbances referred to the control input.  A special state observer is formed based on the nominal plant model but including disturbance estimation.  The disturbance estimate is used to counteract the disturbance component equivalent to the plant modelling error together with any external disturbances, thereby achieving the required robustness.

Another recent innovation is discrete time forced dynamic control in which the aforementioned continuous time FDC has been extended to be applicable using both linear and nonlinear discrete time state space models of the plant, thereby allowing digital control law implementation with longer sampling times than possible when implementing FDC based on the continuous time domain.  The method is named Observer Based Robust Control (OBRC).

Research into Electrical Drive Control (Prof. Stephen Dodds, Prof. Roy Perryman, Dr. Wada Hosny; Prof. Jan Vittek; Prof. Teresa Orlowska Kowalska; Dr. Krzysztof Szabat)

The ultimate aim of the research into new control techniques has been to establish throughout the industry a new generation of electrical drives offering superior performance regarding robustness against external disturbances and motor/load parameter uncertainties together with prescribed closed-loop dynamic performance without the need to undertake trial and error tuning of the controllers. 

Controlled induction motor drives without mechanical speed sensors mounted on the motor’s shaft have the attraction of low cost and high reliability.  To replace the sensor, the information on the rotor speed is extracted from measured stator currents and from the voltages at the motor terminals.  Vector controlled drives require the estimation of the magnitude and spatial orientation of the fundamental magnetic flux component in the stator or in the rotor. Various methods of estimation of rotor flux, rotor speed and load torque have been studied which are based on the induction motor state space model.  A sinusoidal flux density distribution in the air gap is assumed, neglecting the space harmonics and the other secondary effects.  In contrast with estimators with open loop structures, closed loop structures in the form of observers have been employed, where the error signals between measured and estimated quantities are exploited in order to increase the system robustness.  The drawback of this method is the limitation of the model validity at low and zero stator frequency.  An approach to overcoming this has been generated within CRG creating a rotationally oscillating stator magnetic field at a sufficiently high frequency for the inertia of the driven rotor and its mechanical load to act as a low-pass filter to the oscillating torque.

Mutual cooperation between academic staff at the University of East London, the University of Zilina (Slovak Republic) and Institute of Control Sciences of the Russian Academy of Sciences, has established a new generation of sensor-less speed controlled ac machine drives by extending existing vector control methods, while retaining the mutual orthogonality between the torque producing flux and current vectors, with the help of forced dynamic control (FDC) referred to above. This offers operation in a number of prescribed dynamic modes:

While FDC yields good robustness and prescribed closed-loop dynamics, the smoothness of the stator current waveforms is not yet sufficiently good with the current fed inverters to compete with well established industrial electric drives.  This has stimulated the forced dynamic control of electric drives with voltage fed inverters and space vector modulation.

Research that is now underway on the control of electric drives with mechanisms containing vibration modes.  This cross fertilises with research that has already been carried out in CRG on the simultaneous control of the vibrational states and attitude coordinates of three axis stabilised spacecraft.

ent innovation is a drive system for a.c. machines in which the usual three phase distributed windings have been replaced by one coil per slot with relatively low power individual power electronic coil drive circuits, the potential advantages being a) software control of the magnetic flux distribution of the rotating field to maximise the efficiency and vary the number of pole pairs of the induction motor according to the operational conditions, b) minimise cogging torques for a permanent magnet synchronous motor (PMSM) while achieving a higher torque efficiency than attainable with the usual methods of stator skewing and pole shading, and c) automatic compensation by software control of the stator currents to compensate for coil drive failures, thereby enhancing the system stability.  The induction motor project is being sponsored by Brook Crompton Induction Motors and the PMSM project by Control Techniques Dynamics (a branch of the Emmerson Corporation).

Spacecraft Control Research (Prof. Stephen Dodds, Dr Wada Hosny)

rch is underway on the Miniature Reaction Sphere, a new momentum exchange actuator for three-axis attitude control for a new generation of micro-satellites.  A proposal has been made for a reaction sphere based on a dodecahedral stator winding configuration making use of the whole area of a levitated metallic sphere for torque generation, via a completely controlled rotating magnetic field velocity vector.  The multi-phase windings are fed at differing and variable frequencies for this purpose and also cater for drive failures through redundancy coupled with robust control techniques.  This was the subject of a recent invited lecture to the Institute of Electrical Machines at the Warsaw University of Technology, Poland.

Other work in this area, stemming from consultancy with QinetiQ, comprises novel three axis stabilised attitude control laws for simultaneous slewing about all three axes, using singularity free kinematic differential equations based on either quaternions or direction cosine parameters and near-time optimal performance with the aid of forced dynamic control referred to above.

Recently, research has commenced into a novel star sensor based attitude measurement system for three-axis stabilised spacecraft in which an artificial neural network recognises the star field image and enables this to directly produce a set of attitude coordinates, thereby circumventing the usual time consuming and computationally demanding spacecraft search manoeuvres.  A special time varying translational and rotational transformation is applied to

Design Methodology for Permanent Magnet Synchronous Motors (Prof. Roy Perryman, Prof, Stephen Dodds)

The design specification of these machines covers aspects such as the rated power range, cogging torque, efficiency, speed range, the rotor moment of inertia, and the manufacturing costs. During the design process, it is therefore necessary to arrive at a combination of several adjustable parameters such that the design specification is satisfied. This process is at present extremely time-consuming since all the adjustable parameters affect the overall design solution. This research programme is aimed at developing a computer-based procedure that will utilise tools such as genetic algorithms, artificial neural networks, and numerical solutions to nonlinear inequalities in establishing a design approach that provides an optimum solution.

High Power Density, High Speed Brushless DC Drives (Prof. Roy Perryman)

Recent developments in rare-earth permanent magnet materials have been a significant factor in the improved performance of brushless dc drives. Critical to the use of permanent magnets in electrical machines are the basic magnetic parameters of remanence, coercivity and the maximum energy product. The energy products of sintered SmCo and NdFeB magnets provide high remanence, which enables high air-gap flux density, reduced flux leakage and increased efficiency, resulting in improvements in torque/inertia and torque/weight ratios. Rotor geometry also plays an important part in the design and operation of dc motors. A new form of construction has been developed using a laminated rotor, which has advantages of improved power to weight ratio, high acceleration and better thermal performance. This work has concentrated on the optimum design of a high speed, high power density, brushless dc drive with particular relevance to aerospace applications. Computer modelling using finite element analysis has also been applied to optimise the motor design.

Evaluation of Stray Losses in Servomotors (Prof. Roy Perryman)

Global concerns about the increasing levels of greenhouse gases in the atmosphere has led to governments committing to programmes aimed at reducing the quantities of these gases that they emit each year. If the efficiency of electrical machines can be increased then less electrical energy will be required to produce the same motive power, with a knock-on effect on the release of pollutants from power stations. In Europe a voluntary efficiency-labelling scheme has been agreed between the EU and CEMEP, the organisation of European Electric Motor Manufacturers’ Associations. Under this agreement 2- and 4-pole motors rated at between 1.1 and 90kW and sold within the EU will carry labels indicating that they are eff1, eff2 or eff3, according to the efficiency band in which they lie. There is now strong pressure to make this voluntary agreement mandatory and to provide strong financial incentives to promote eff1 motor sales.

Of the loss components, the stray loss is least understood in terms of its subcomponents and this provides the focus of the research programme. In order to examine the dependency of stray loss on load level and the frequency of the applied load, a series of test will be undertaken on a selection of servomotors over a range of frequencies. Correlation will be made with theoretical analysis and further study will be undertaken in relation to inverter fed variable speed drives, and the control of harmonic distortion related to efficiency/loss analysis.

Mobile Servo-robots. (Dr. Nic Snailum, Dr. Wada Hosny)

An investigation is being carried out into the automation of taking orders and serving dishes to customers in a fast food restaurant. One component of the investigation is image processing using neural networks and the communication link and control electronic systems needed. The other component of the investigation is the mechanical design of the conveyor belt, manipulators and so forth. The set of objectives for this investigation are summarised as taking the customers order, sending the order to the kitchen for processing using a wireless communication link, fetching the order when it is ready, delivery to customer/customers and eventually picking up the empty plates and taking them to the serving hole in the kitchen. This process is then repeated for other customers.

Control Strategy for single phase Active power filters using Orthogonal Transformation Technique ( Dr Wada Hosny, Professor B Dobrucky)

The othogonal transformation technique instigated by Akagi et al for a three phase power system and extended by Professor Bobrucky ( University of zilina) for a single phase power system, is used to drive novel compensating current expression for a single phase, shunt active power filter. This expression has been implemented using a digital signal processing and the experimental results conducted on a single phase power system feeding a non-linear load confirmed the simulation results and the effectiveness of this novel control strategy for the filter under investigation.

Active Power filters for 25 KV Railway Power supply ( Dr Wada Hosny, Professor Joongho Song, National University of Soeul, Korea)

Professor Song spent 6 months as part of his sabbatical leave from his university in South Korea at UEL working on this collaborative research with Dr Hosny, investigations are underway using novel compensating current algorithms for the active power filter dedicated mainly for 25 kV railway supply and the results will be published shortly.

Control Strategy for Three Phase Active Power Filters Using Neural Networks (Dr. Wada Hosny)

Novel control algorithms for an active power filter of a 3 phase, 4 wire supply network feeding a solid power converter are investigated. The implementation of these algorithms leads to an improvement in the harmonic distortion and the power factor of the line current fundamental component. Thus, the neutral wire current can be eliminated.

The control strategy is based on the extraction of the line current harmonic components by means of an adaptive linear neutral network based on the Widrow-Hoff learning algorithm.
Computer simulations have been generated for a specified power system which demonstrated the excellent steady state and dynamic performance of the proposed active filter control strategy. Experimental verification of the simulation results is underway.

Seismic Data Analysis (Andrew Chanerley, Dr A Alexander)

Research into using adaptive digital signal processing algorithms and wavelet transforms for the correction of raw data regarding the earth movement is carried out in collaboration with the University of Bristol, other universities and research institutions in Iceland, Finland and Germany.

Communications Engineering ( Dr K Yeo and Dr Jaswinder Lota)

The research in this area focuses on microwave electronics with tunable dielectric layers,  high temperature superconducting filters, micro-machined electromechanical systems for agile microwave and millimetrer wave components, delta-sigma analog to digital converters and software defined and cognitive radio networks

---

© 2010