Current Projects

Return to Stephen Duncan’s homepage

 

Design of Sustainable Policies for Greenhouse Gas Emissions

 

This project uses concepts from modern robust control theory to develop algorithms for determining the optimal policy that both achieves sustainable levels of emissions of CO2 (and other greenhouse gases) and minimises the impact on the economy, but also explicitly addresses the high levels of uncertainty associated with predictions of future emissions. The aim of the optimal policy is to adjust factors such as the mix of energy generation methods and policies for reducing emissions from housing, industry and transport, in order to achieve a rate of emissions that will allow the UK to achieve its emissions targets while maximising economic growth as measured by discounted GDP. A key difficulty in determining the optimal policy is handling the uncertainty associated with the effect that the policy changes will have on the rate at which is CO2 emitted. Concepts from robust control theory are used to develop tools for incorporating uncertainty directly into the design of the optimal emissions policy; the tools can then be applied to other existing models. Including uncertainty within the design quantifies the risk associated with the emissions policy, which allows policy makers and emitters of CO2 to incorporate risk within their strategic plans.

 

Participants: Bing Chu, Stephen Duncan, Antonis Papachristodoulou, Cameron Hepburn1, Simon Roberts2.

Partners: 1Smith School for Enterprise and the Environment, University of Oxford; 2Arup.

Duration: 2010-2012.

Funding: EPSRC.

 

 

Design of Fast Orbit Feedback Stabilisation System for a Synchrotron

 

Synchrotrons can produce very intense of X-rays and ultra violet light for a range of applications including protein crystallography, materials characterisation and high resolution imaging. The Diamond Light Source is a recently constructed synchrotron that produces a 3GeV electron beam in a ring of circumference 560m. A key component of the process is the fast beam stabilisation system that regulates the horizontal and vertical position of the beam in the presence of disturbances in the range 1 to 100Hz using 178 sensors and actuators positioned around the ring. Using concepts from the design of cross-directional control systems for web industries, a control system has been designed and successfully implemented on the synchrotron. Current research is focussing on implementing the control system on the booster synchrotron. 

 

Participants: Sandira Gayadeen, Ashley Napier, Stephen Duncan, James Rowland1, Michael Abbott1, Isa Uzun1, Günther Rehm1, Mark Heron1

Partners: 1Diamond Light Source.

Duration: 2007-2013

Funding: Diamond Light Source

Keywords: Synchrotron, Beam stabilisation, Control

Keyphrases: Synchrtron, Fast orbit beam stabilisation

 

 

Control of Incremental Sheet Forming

 

Incremental sheet forming is a novel small-scale, production process, which is relevant for prototyping, low volume production and customisation and which potentially may lead to environmental improvements, primarily through locating production nearer to customers.  The process is an alternative to the standard sheet pressing processing currently used in many manufacturing industries, in which a pair of specially-made dies press a flat sheet into a required geometry. This is efficient, but it is expensive for low volume manufacture, due to the cost of making the dies.  Incremental sheet forming uses a small indentor, which is dragged around the sheet, creating a small 'kinked' track.  As the tool continues to track over the sheet, a deformed shape is built up as the sheet undergoes plastic deformation. Any shape that can be made by pressing, can also be made by incremental forming, the difference being that incremental forming does not require any special tooling. 

 

The project is carried out in conjunction with the Production Processes group in the Institute for Manufacturing at the University of Cambridge, where the first dedicated rig outside Japan has been built to explore this process.  The Cambridge group are currently investigating the mechanics of this process.  This project works in parallel with the research at Cambridge to develop the control system associated with the process. The design of a controller is a special case of distributed parameter control known as 'mobile control'. The control challenge is that the actuator is a moving tool, whose effect on the sheet depends on its location, and on the past deformation undergone by the sheet.  The project will use the model of the plastic deformation of the sheet to determine the optimal path for the indentor and then develop and implement a feedback control system, using measurements from a vision system.

 

Participants: Hao Wang, Ankor Raithatha, Stephen Duncan, Kathryn Jackson2, Omer Music2, Julian Allwood2

Partners: 2Institute for Manufacturing, University of Cambridge

Duration: 2005-2011

Funding: EPSRC

Keywords: Manufacturing, Optimal control, Modelling

Key phrases: Incremental sheet forming, Flexible manufacturing

 

 

Control of Fluidised-bed Dryers using Process Tomography

The aim of the project is to develop new technologies for product quality control by combining inferential sensing, online process modelling, tomographic imaging of physical properties and distributed parameter control. To demonstrate the benefits of this strategy, a real-time closed-loop control system will be developed to regulate spatial distribution of gas-solids and moisture content distribution in a fluidised bed dryer. A typical application of a fluidised bed dryer is in the drying of tablets for the pharmaceutical industry. Air is blown from the bottom through a vessel containing the tablets, which are dried as they “float” in the air flow and a typical quality specification is to ensure that moisture content of the all tablets in the dryer is uniform. At present, fluidised bed dryers are "black boxes", i.e. it is difficult to know what is going on inside, but industrial process tomography allows the variations in the moisture of the tablets within the dryer to be measured. In the past, tomography has primarily been used solely for process monitoring, but this project offers the first opportunity of using tomography for product quality control. The proposed control scheme will be based on online modelling of desired gas-solids distribution and moisture content distribution, from which desired capacitance and loss-conductance profiles can be derived. Both capacitance and loss conductance measurements from tomographic sensors will be used as feedback signals for controlling the process in terms of the difference between the desired profile within the dryer and the measured profile. The system will control both the gas-solids distribution and the moisture content distribution in the dryer.

Participants: Javier Villegas, Mingzhong Li, Stephen Duncan, Haigang Wang3, Rambadi Ragavan4, Wuqiang Yang3, Peter Senior4

Partners: 3Department of Electrical Engineering and Electronics and 4Department of Chemical Engineering, University of Manchester, Astra Zeneca, Sensatech.

Duration: 2005-2008

Funding: EPSRC

Keywords: Sensing, Process control

Key phrases: Industrial process tomography, Distributed parameter control

 

 

Control of Temperature Profile during Sprayforming

 

The sprayform process creates tools and dies by spraying molten metal from arc spray guns onto a ceramic former. The key step in this technology is to ensure that as the metal cools, it undergoes a specific expansive phase transformation that offsets the thermal contraction experienced by the metal. This allows the formation of dimensionally accurate tools. In order that the metal undergoes the required transformation, the temperature profile of the surface has to be accurately controlled throughout the spraying process. This is achieved by measuring temperature variations in real time with a thermal imaging camera and then using this information to adjust the rate of spraying as the guns are moved over the surface by a robot. A full-scale spray cell has been built and an online control system has been successfully demonstrated along with offline path planning tools. A commercial version of the system has been developed. Discussions are currently being held with Ford Motor Company to continue this research for a further three years during which time, the process will be scaled up to allow larger tools to be sprayed. This will involve extending the cell to enable metal to be sprayed simultaneously form guns mounted on two robots. The ideas are also being used to control a plasma spraying process.

Participants: Paul Jones, Vassilios Tsachouridis, Pubudu Pathirana, Stephen Duncan, Tim Rayment5, Sarah Hoile5, Zoran Djuric5, Patrick Grant5

Partners: 5Department of Materials Science, University of Oxford; Ford Motor Company; Novarc Limited; Sulzer Metco; Kuka Robots.

Duration: 2000-2006

Funding: EPSRC

Keywords: Manufacturing, Spraying, Control

Key phrases: Spray forming, Optimal path planning, Tooling

 

Design of Cross-directional Control Systems for Web Processes

 

Cross-directional (CD) control systems have been common in the web processing industries (such as paper making, metal rolling and plastic film extrusion) since the 1980’s. In the past two decades, considerable research has taken place in the field of CD control and a range of alternative strategies have been proposed. This study will benchmark an existing CD controller installed on a plastic film extrusion line against state of the art controllers, in order to identify opportunities for improved profile control and extend modern control design methods to multi-array CD control systems.

 

Participants: Andrew Taylor6, Stephen Duncan, William Heath7.

Partners: 6DuPont Teijin Films (UK) Limited, 7Department of Electrical Engineering and Electronics, University of Manchester.

Duration: 2006-2010

Funding: Scottish Enterprise

Keywords: Process control, Performance monitoring, Bayesian methods

Key phrases: Cross-directional control, Web processes

 

Process Control of Vacuum Plasma Coating

 

Vacuum plasma spraying is an industrial technique used to coat substrates onto a wide range of materials. Historically, vacuum plasma spraying processes have been open-loop processes with, at the most, closed-loop control of the spray parameters. The key control variable for these processes is the coating quality and in particular, the microstructure of the coated material. It has been shown that control of surface temperature during spraying has a direct effect on the microstructure and this project is designing and implementing a system for controlling temperature online so that the improvements in coating quality and process repeatability can be assessed.

Participants: Ewan Davis, Stephen Duncan, Patrick Grant8

Partner: 8Department of Materials Science, University of Oxford.

Funding: EPSRC

Keywords: Spraying, Control

Key phrases: Thermal spray processes, Vacuum plasma spraying

 

Control of Concentration in Fluid Flow

 

Industrial process tomography systems can measure the concentration profile of different components or phases of a fluid flow within a pipe. The tomographic system works by measuring the capacitance or impedance between successive pairs of electrodes within an array that is mounted around the pipe. A state estimation scheme is then used to determine the concentration profile. 2-dimensional simulation studies carried out in conjunction with the Department of Applied Physics at the University of Kuopio, Finland, have indicated that a tomographic sensor can be combined with an array of injectors in a feedback loop to regulate the concentration profile in the pipe. The results of this research have been used in the project to control fluidised-bed dryers.

Participants: Anna Ruuskanen9, Stephen Duncan, Jari Kaipio9

Partners: 9Department of Applied Physics, University of Kuopio, Finland.

Duration: 2003-2005

Funding: Finnish Academy

Keywords: Sensing, process control, flow control

Key phrases: Industrial process tomography, Distributed parameter control

 

Control of Ultra-sound Treatment for Cancer

 

Cancer can be treated by using ultra-sound to heat tumours. A key component of the treatment is the ability to control the temperature profile during heating as it is necessary to ensure that the temperature within the tumour is sufficiently high to kill the cancerous cells without affecting the surrounding healthy tissue. A feedforward strategy for determining the power that should be applied to an array of ultra-sound sources has been developed, together with a feedback mechanism that uses images from an MRI scanner to control the temperature profile within and around the tumour. An alternative system, that uses ultra-sound for both sensing and actuation, is currently being investigated.

Participants: Matti Malinen10, Stephen Duncan, Jari Kaipio10

Partner: 8Department of Applied Physics, University of Kuopio, Finland.

Duration: 2003-2005

Funding: Finnish Academy

Keywords: Ultra-sound, cancer

Key phrases: Distributed parameter control

 

Control of Gas Flows during Anaesthesia

 

Current systems for monitoring and regulating the flow of gases during anaesthesia are relatively crude and there is a considerable scope for using modern sensing technology to improve both the concentration of gases delivered to the patient and the overall amount of gas used during the procedure. In conjunction with the Nuffield Department of Anaesthetics, a prototype control system is being developed that will be tested. It is intended that the system will be used to determine the fraction of the blood that is not oxygenated as it passes through the lungs, directly from the difference between the composition of the inspired and expired air. Currently, this has to be inferred from the composition of the blood taken from a separate measurement.

Participants: Sara van der Hoeven, Stephen Duncan, Andrew Farmery11, Clive Hahn11

Partner: 11Nuffield Department of Anaesthetics, University of Oxford.

Duration: 2004-2008

Funding: EPSRC

Keywords: Anaesthesia, Control

Key phrases: Flow control, Concentration control

 

Modelling the Spread of Epidemics

 

The spread of a disease through a population can be modelled by as a dynamic spatio-temporal system. My work has concentrated on the spread of smallpox and interest in modelling the spread of this disease has increased recently, due to the possibility of smallpox being used in a bio-terrorist attack. The design of strategies for vaccination and quarantine and other public health measures to combat an outbreak of smallpox started by bio-terrorist action is critically dependent on an accurate measure of the transmissibility of the virus. By modelling the dynamics of the disease as a linear system with a non-linear feedback term, the transmissibility of the virus can be estimated from the frequency of outbreaks in populations where the disease is endemic. This approach has been used to derive an estimate of the value for the transmissibility for London in the 18th Century.  The value for the transmissibility of smallpox is much higher than previously assumed and has major implications for the health measures that need to be instituted following a terrorist attack.

Participants: Stephen Duncan, Susan Scott12

Partners: 12School of Biological Sciences, University of Liverpool.

Duration: 2001-2009

Keywords: Epidemics, Modelling, Transmissibility

Key phrases: Distributed parameter modelling, System Identification

 

Return to Stephen Duncan’s homepage