Future energy conversion systems are required to have high reliability and efficiency. This is particularly important in offshore wind due to the increasingly remote locations and complicated and costly maintenance for these systems, as well as the increasing share of offshore wind in UK's future energy scenarios. However, power converters which are the key enablers for producing a controlled electrical power in these applications, are proven by numerous studies to be one of the most fragile parts of these systems and a major source of maintenance cost and downtime. To improve the reliability of power converters in offshore wind farms, it is vital to be able to model the reliability of power converters at wind farm level and optimise it.
This work's key hypothesis is that by active thermal management of power converters of an offshore wind farm, incorporating not only the individual converters but the interactions between them due to the wake effect, it is possible to improve the overall reliability of the system and manage it actively. For example, by intelligent derating of a component (e.g., a switch) or derating of a converter (turbine), its lifetime consumption will decrease, therefore by proper control of the system it will be possible to shift the stress from the more-stressed converter to the less-stressed ones and optimise the overall reliability of the system and manage it actively. Moreover, it would be possible to coordinate the operation of power converters for an optimal maintenance plan, for example, by increase the likelihood that more converters can continue their operation until the next planned preventive maintenance. This project therefore proposes solutions for (1) real-time wind farm level assessment and monitoring of reliability, and (2) active reliability enhancement and lifetime management of power converters in offshore wind farms. It is important to note that given the interlinked operation of wind turbines due to wake effect, a reduction in the power handling capability of a particular turbine does not necessarily mean that the overall power generated will decrease, and in fact it may even increase. Therefore, this project considers the optimisation of both reliability and power generation in a multi-objective optimisation problem.