We use cookies to ensure that we give you the best experience on our website. You can change your cookie settings at any time. Otherwise, we'll assume you're OK to continue.

Durham University

Research & business

View Profile

Dr Peter Matthews, MA DipCS PhD MIET

Associate Professor in the Department of Engineering
Telephone: +44 (0) 191 33 42538
Room number: E237 (Christopherson)

(email at


Peter Matthews is an Associate Professor in Design Informatics at the Department of Engineering. His research in Design Informatics utilises data mining and machine learning tools to critically appraise technical data sets, such as operational sensor data from wind turbines (typically from SCADA systems). Gaining insights to the underlying processes governing these systems gives us a deeper understanding, which ultimately leads to improved design for the next generation of system.

The primary goal for machine learning with technical data is to be able to predict with sufficient warning when a machine is likely to fail. This prognostic ability has the potential of providing significant cost savings to industry: maintenance can be performed at the optimal time, allowing better planning, and the risk of incurring secondary damage is mitigated. Dr Matthews has led several successful projects which have accurately predicted failures for specific sub-systems (see results with Chen (2015), Godwin (2013), and Smith (2015)).

Another important component of Dr Matthews’ research is the production of tractable and humanly-understandable rules. Tractable rules require a much simpler validation process and are therefore more useable by system operators and designers when seeking to improve a system’s performance.

Wind Energy

Dr Matthews' wind energy research is primarily in data mining SCADA and other wind turbine operational data. This research is primarily aimed at developing diagnostic and prognostic measures for individual wind turbine health. The approach taken is based around statistical modelling of healthy wind turbines, and then comparing live wind turbines against this healthy model. Other methods (eg physics based) are under development as well, again using ‘big data’ approaches to validate.

In addition to SCADA analysis, Dr Matthews has directed research in wake optimisation and maintenance strategy simulation. The wake optimisation research has delivered a workable dynamic wind farm controller that can minimise the effect of in-farm wakes on total production. The maintenance strategy simulation provided a Monte Carlo based approach for developing and testing alternative off-shore wind farm maintenance strategies.

Much of the Wind Energy research is undertaken with industrial partners Ørsted Energy and Maia Eolis (now Engie Green).

Energy Distribution

The energy distribution sector has a broad range of customers, from domestic through to large industrial customers. All these customers use electricity in different ways, and their consumption is recorded using SCADA systems. Dr Matthews’ research in the Energy Distribution sector centres around data mining these SCADA databases of thousands of customers, as well as hundreds of electrical substations, to gain better understanding of the overall picture of electricity use. Dr Matthews has also directed research to forecast the demand increases at substation level using substation demographic customer profiles.

Much of this research is undertaken with Northern Powergrid.

Design Analysis

The Design Analysis research is based on data mining, but with considerably smaller datasets. Here, the aim is to extract the tacit rules the human designers have applied, and gain better understanding of the design domain through making these rules explicit. Design data often contains greater textual information, and so text mining approaches have also been applied with interesting results. Other techniques that have been used include Bayesian Belief Network and p-boxes. These techniques have been used to mitigate against the greater uncertainty levels that can be associated with early designs.

This research has been undertaken with Rolls-Royce (Aerospace) and BAE Systems.

Research Groups

Department of Engineering

Research Interests

  • Artificial Intelligence and Machine Learning
  • Data mining
  • Wind Energy
  • Monte Carlo methods
  • Engineering Uncertainty modelling and management
  • Knowledge Management
  • Engineering Design
  • Design process
  • Game theory

Selected Publications

Chapter in book

Conference Paper

Journal Article


Show all publications


Selected Grants

  • 2017: A New Partnership in Offshore Wind (£767499.35 from Engineering and Physical Sciences Research Council)
  • 2012: KTP - Icona Solutions Ltd (£91146.00 from Icona Solutions Ltd)
  • 2011: EPSRC Centre for Innovative Manufacturing in Through-life Engineering Services
  • 2011: Machine learning of process production monitoring (£25859.00 from Epsrc)
  • 2010: EPSRC Summer studentship (in collaboration with BioInnovel)
  • 2008: Agent meta-learning for uncertain domains (£18620.09 from Royal Academy of Engineering)
  • 2008: Durham University Institute of Advanced Study
  • 2004: Framework 6: Virtual Reseach Lab for a Knowledge Community in Production (VRL-KCiP: FP6-507487-2)
  • 2004: MACHINE LEARNING OF PROBABILISTIC (£4500.00 from The Nuffield Foundation)
  • 2004: NDI RESEARCH (£13000.00 from Northern Defence Industries Ltd)

Teaching Areas

  • L1 Computer Aided Drawing

    (20 hours/year.)
  • L1 Manufacture

    (11 hours/year.)
  • L3 BEng Mechanical CAD

    (8 hours/year.)
  • L3 Management (Product Life Cycle and Game Theory for Engineers)

    (9 hours/year.)
  • L4 Advanced Engineering Design

    (10 hours/year.)