Performance Modelling and Analysis of Olympic Class Sailing Boats by Alexander S. Reid A thesis submitted for the degree of Doctor of Philosophy School of Marine Science and Technology The University of Newcastle upon Tyne 2011 ii Acknowledgements I would like to thank the following people who helped me during this project and made it possible. Professor Martin Downie of the School of Marine Science and Technology, University of Newcastle upon Tyne for being my main supervisor and helping me develop various ideas and providing continuous invaluable guidance. Dr. Peter Wright also of the School of Marine Science and Technology, University of Newcastle upon Tyne for being my second supervisor. With his input it was possible to develop multiple forward thinking methods for solving the myriad of problems experienced. I would like also to thank Peter Bentley from the RYA for being my industrial supervisor. Peter was the main driving force behind the project and through his enthusiasm and expertise a successful project was produced. His attitude of ‘getting the job done whatever the cost’ was also a great inspiration. Many thanks to Scott Drawer and Naomi Siddall from UK Sport for providing the funding and also the various UK Sport training programs that proved invaluable. I would also like to thank them for various pieces of equipment provided including multiple laptops. An important part of the project was made possible by Dr Martin Smith and his assistant Ahmad Taha from the Institute of Engineering Surveying and Space Geodesy at Nottingham University who provided valuable knowledge in more than one area. A special thanks to Dr. Martyn Prince of the Wolfson Unit at Southampton University whose expertise enabled a significant section of the project to become a success. I would also like to thank all my family and friends who have supported me for the past three years as well as the many I have made through working with the RYA. iii Authors Declaration I declare that all the work in this project is that of the author. Where information has been taken from other sources it has been referenced appropriately. Alexander Stanley Reid iv Preface The work in this thesis is preceded by a Master of Research in Marine Technology project between September 2004 and October 2005. The project was supervised by Professor Martin Downie and was carried out with significant time present in the field, working closely with Olympic sailors from multiple different classes. This project was funded by UK Sport and considered a pilot project to investigate the feasibility of using data logging equipment with GPS in the marine Olympic environment. A series of prototype systems were engineered to meet the requirements specified by the Royal Yachting Association. The engineering and validation of the software and hardware formed a key part of the project to ensure that the results obtained were accurate and repeatable. This included software design within two different software platforms as well as embedded hardware developments. Significant testing and development were implemented in the laboratory as well as on the water during the beginning of the project and as a continuous background task throughout the project. Over eighty days were spent in the field developing and testing hardware and software as well as determining the optimum performance analysis methods. Data loggers were fitted to several Olympic class boats during the evaluation process to ascertain the performance of the data logging system as well as the performance of the boat and crew. Data was logged from the onboard GPS and accelerometers and analysed post training. Later in the project, wind information was also collected and fused together with the onboard data post training. The hypothesis was to demonstrate performance gains in the participating classes through the means of quantitative analysis. Prior to the project the performance analysis had been almost entirely qualitative. Through the course of the project various techniques were developed allowing quantitative performance analysis to supplement the efforts of the training group and coach. Key performance factors were determined by data analysis techniques developed during the project. One of the significant tools developed was a tacking performance analysis routine which analysed multiple different styles of tacks, calculating the distance lost with respect to wind strength and course length resulting in an important strategic tool. Other tools relating to starting performance and straight line speed were also developed in custom software allowing rapid analysis of the data to feed back to the teams in the debrief. v At the close of the project, when wind measurement was used for the first time with data collected from the target boats, the truly dynamic problem of modelling high performance dinghies was apparent. This resulted in a search for a method that would be able to cope with the highly dynamic environment with temporal datasets. An extensive search in future methods was conducted primarily focussing on creating a VPP for dinghies, or using Neural Networks to model the performance of the boat. After the pilot project was concluded, sufficient evidence had been demonstrated to UK Sport to ensure that further research could improve the performance of the Olympic classes, leading on to a deeper investigation and a PhD in Marine Technology at the University of Newcastle upon Tyne. vi Table of Contents Contents Acknowledgements ............................................................................................................. ii Preface .............................................................................................................................. iv Table of Contents ............................................................................................................... vi List of Figures ..................................................................................................................... x List of Tables .................................................................................................................... xiii Abstract ............................................................................................................................ xiv Nomenclature ..................................................................................................................... xv Chapter 1 ............................................................................................................................ 1 Introduction ........................................................................................................................ 1 1.1 Context ...................................................................................................................... 1 1.2 Aims and Objectives .................................................................................................. 2 1.3 Chapter Outline ......................................................................................................... 3 Chapter 2 ............................................................................................................................ 5 Methods of Yacht Performance Analysis ............................................................................. 5 2.1 Introduction ............................................................................................................... 5 2.2 Velocity Prediction Programs .................................................................................... 5 2.2.1 VPP Background ................................................................................................ 5 2.2.2 VPP Methods ...................................................................................................... 8 2.3 Performance Data Collection ................................................................................... 16 2.3.1 Yacht Data Collection Methods ........................................................................ 16 2.3.2 Dinghy Performance Analysis ........................................................................... 20 Chapter 3 .......................................................................................................................... 26 Artificial Neural Networks ................................................................................................ 26 3.1 Introduction ............................................................................................................. 26 3.2 The Development of Artificial Neural Networks ...................................................... 26 3.3 Implementation of Artificial Neural Networks ......................................................... 28 3.4 Assessing ANN Performance ................................................................................... 29 3.4.1 Goodness of Fit ................................................................................................. 31 3.5 Training Neural Networks ........................................................................................ 32 vii 3.5.1 Error Minimisation ............................................................................................ 32 3.5.2 Rattling and Step Size ....................................................................................... 34 3.5.3 Momentum ....................................................................................................... 35 3.6 Neural Network Designs .........................................................................................