Model based predictive control with application to renewable energy systems
In the promotion and development of renewable energy systems, control engineering is one area which can directly affect the overall system performance and economics and thus help to make renewable energies more attractive and popular. For cost effectiveness, ideally the renewable energy industry requires a control design technique which is very effective yet simple with methods that are transparent enough to allow implementation by non-control engineers. The objective of this thesis is to determine if Model Based Predictive Control (MBPC) is a suitable control technique for use by the renewable energy industry. MBPC is chosen as it uses simple and fairly transparent methods yet claims to be powerful and can deal with issues, such as non linearities and controller constraints, which are important in renewable energy systems. MBPC is applied to a solar power parabolic trough system and a variable speed wind turbine to enable the general applicability of MBPC to renewable energy systems to be tested and the possible benefits to the industry to be assessed. Also by applying the MBPC technique to these two strongly contrasting systems much experience is gained about the MBPC technique itself, and its strengths and weaknesses and ease of application are assessed. The investigation into the performance of Model Based Predictive Control and in particular its application in the renewable energy industry leads to two contrasting conclusions. For simple systems with non-demanding dynamics and having a good model of the system, MBPC provides a very good and effective solution. However for more demanding systems with complex dynamics and strong non-linearities, a basic MBPC controller, applied by a non-control engineer, cannot be recommended.