Aspects concerning power distribution networks planning using artificial intelligence

This paper presents the application of AI tools for the on-line identification of load structure in distribution networks. The authors have considered artificial neural networks (ANN) which are known as valuable and fast tools for pattern identification or completion. This approach to the load model allows a more detailed analysis directed towards the optimization of system structure and working conditions. Traditional methods produce good results but raise the processing time problem, especially when applied to large systems. For such cases another approach uses genetic algorithms, which are frequently referenced in the literature concerned with power distribution systems (reconfiguration of open loop radial networks, optimal VAr-sources distribution, optimal selection of transformer tap position).