Feasibility study of the implementation of A.I. automation techniques in modern power distribution networks

Contemporary power distribution networks are no longer regarded as passive power system elements. On the contrary, novel control systems are being constantly developed over the last decade, aiming to achieve reliability improvement and operational optimization by means of power loss reduction, prompt fault detection and power restoration etc. A crucial aspect of the systems developed to achieve these goals will inevitably be their ability to integrate new functions without the need for further investment. In this paper, a multi agent system (MAS) initially developed for fault detection and power restoration is studied with respect to these issues. More specifically, a feasibility analysis is conducted regarding the implementation of the MAS on a segment of the underground 20 kV power distribution network of the city of Thessaloniki, Greece. The analysis focuses on the initial investment cost and the payback of the application, as well as on the additional benefits for the power distribution system operator due to the system reliability improvement. The ability of the MAS to incorporate loss reduction algorithms without further investment is also studied, and the respective benefits of the power distribution system operator are analyzed. Moreover, the feasibility analysis is generalized so as to be able to be applied to any power distribution automation implementation with similar attributes.