Wind turbines are leading the way in helping to reduce the dependency on fossil fuel energy sources. However to compete with other energy sources there is a need to reduce the cost of energy from wind turbines. It has been shown in the literature that as wind turbines increase in size their reliability decreases. As wind turbines move further offshore and into deeper water this becomes more of an issue as carrying out maintenance becomes more challenging and costly. One way of improving the reliability of wind turbines is through the use of condition monitoring systems (CMS) which can continually monitor the health of the machine and allow more optimised maintenance and repair scheduling.Although the benefits of using a CMS may seem evident, operators have been slow in the uptake of such systems. One reason for this is due to issues with the reliability of CMS themselves. As stated in the literature, CMS must accurately detect 60-80% of faults to be economically justifiable. Not detecting faults or the occurrence of false alarms is detrimental to the effectiveness of CMS. The work presented in this thesis aims to address the issue of CMS reliability.Through the installation of two CMS in operational wind turbines the author of this thesis has gained valuable insight into the design, build and installation of CMS which has facilitated the novel contributions from this work.The first contribution comes from the formulation of an engineering design process which incorporates five categories of robustness which were identified by the author through Failure-Mode Effects Analysis on a wind turbine CMS that was installed in an operational wind turbine. The engineering design process incorporating the robustness categories will allow wind turbine CMS to be designed which are capable of operating reliably in the harsh environment they are subjected to.The second contribution comes from the development of three techniques which will increase CMS reliability by reducing false alarms and introducing the ability to detect erroneous data.
|Date of Award||5 Jun 2017|
- University Of Strathclyde
|Sponsors||EPSRC (Engineering and Physical Sciences Research Council)|
|Supervisor||Victoria Catterson (Supervisor) & Bill Leithead (Supervisor)|