A Transformer Predictive Maintenance System Based On Agent-Oriented Programming

Transformers play a significant role in power systems. The traditional method of transformer maintenance is time-based or off-line-based, which is inefficient and costly. This paper demonstrates a methodology of using agent-oriented programming (AOP) for utilizing MATLAB based transformer thermal analysis algorithms. MATLAB is widely used for scientific research programming and system evaluation. AOP formalises interactions between multiple agents by means of exchanging their "mental states" through "agent speech acts". The practical implementation of MATLAB based algorithms in a multi-agent system is not straightforward. Therefore, several programming methods are embedded in this hybrid prototype system to predict preventative maintenance of transformers