Design and development of an SDN robotic system with intelligent openflow IOT testbeds for power assessment, prediction and fault management

Abstract

Current wind turbine and power grid industry have relatively little research and development with regards to implementing novel communication network and intel- ligent system to overcome issues that pertain to network failure and lack of monitor- ing. Wind turbine location could be a big concern when it comes to identifying an efficient location for future wind turbine and the impact of a site with non-efficient meteorological parameters can result in relocation of a wind turbine and revenue- loss. Unplanned wind turbine shutdowns that are considered to be one of the major revenue-loss factors of a modern wind farm business. Typically, the unplanned wind turbine shutdown is a result of sensors fail due to harsh environment challenges that prevent hardware status from being available on the monitoring system. The above mentioned research problems pertain to wind turbine site assessment and predic- tion of power. In this thesis, a novel programmable software-defined robotics and IoT testbeds are proposed with the fusion of Artificial Intelligence and optimiza- tion methods to solve specific problems related to wind turbine site assessment and fault management. The site selection process is implemented using proposed aerial and ground robotic systems that are incorporated with Software-Defined Networks and OpenFlow switching capabilities. A second stage development of the system is proposing a prediction platform that run on the aerial robot cluster using neural net- works optimization regression techniques. To overcome the unplanned wind turbine network outage, an IoT micro cloud cluster system is proposed that act as immedi- ate fail-over platform to provide continuous health readings of the wind turbine to ensure the turbine in question will not get shutdown unnecessarily. The proposed system help in minimizing revenue-loss caused by stopping a wind turbine from op- eration and help maintain generated power stability on the grid. Additionally, since large wind farms require an agile and scalable management of selecting the most efficient wind turbine location install. Thus, a softwarized cognitive routing proto- col is proposed. The group of quadcopters is a redundant failover Software-Defined Network/OpenFlow system that can cover every single way point of the farm land. Although, power consumption is essential for the continuity the service, a Software- Defined charging system testbed is proposed that uses inductive power transfer with