Répartition économique de l'énergie électrique utilisant les techniques d'intelligence artificielle

Abstract

The development of computer software has contributed to the innovation of power systems to becomesmart grids. This development is necessarily linked to several concerns: energetic, economic, environmental, etc. The introduction of techniques of artificial intelligence in software of control and decisionis an essential element in research and development of tomorrow’snetworks. Neural networks are among the techniques most usedin the field of artificial intelligence. The economic dispatch is a key sector in electricity network, where it must generate less energy for the same demand with good economic operation reducingrepartition grid losses to have the least cost of kWh possible.In this thesis, we will opt for a quicker economic dispatch. Beginning with programmingradial network economic dispatch with and without losses using traditional programanderror backpropagation learning neural network program, then, we will program a mesh network of 9 busses including 3 production units using traditional program and error backpropagation neural network program, finally; we will compare the two programs in terms of speed and reliability.