Deep learning-aided joint DG-substation siting and sizing in distribution network stochastic expansion planning

The rapid growth of distributed generation (DG) and load has highlighted the necessity optimizing their ways integration, as siting sizing significantly impact distribution networks. However, little attention been paid to new substations which are be installed. This paper proposes deep learning-aided joint DG-substation in network stochastic expansion planning. First, model depends on an accurate forecast, Long Short-Term Memory (LSTM) neural is used forecast DG output load, where electricity rate, bidding capacity electric expansion, industrial difference all considered. Then, a two-stage mixed integer bilinear programming was established for under uncertainties, multiple objective functions comprehensively addressed. By using Fortuny-Amat McCarl Linearization, resultant equivalently transformed into linear program, can efficiently solved. Finally, power flow calculation IEEE 69-node system conducted analyze influence integration node voltage system. effectiveness proposed method also verified by simulation tests.