A Hybrid Heuristic Algorithm for Optimal Energy Management Problem in Micro Grids


BEKTAŞ TOPÇU Z. , Kayakutlu G.

International Symposium for Production Research 2017, Viyana, Austria, 13 - 15 September 2017, pp.75-84

  • Publication Type: Conference Paper / Full Text
  • City: Viyana
  • Country: Austria
  • Page Numbers: pp.75-84

Abstract

Micro grids have been widespread in order to have the control on the energy supply security. The micro grids (MG) are small-scaled and restricted energy systems using distributed energy sources and storages. They can be operated in two different ways; grid-connected or islanded modes. The shifting between the modes is done due to the demand. On the islanded mode, the MG works autonomously. The optimal energy management problem is a generally accepted issue for MGs. Using islanded mode is requested because it minimizes the amount of power bought from main grid. It is not always possible unless a fertile field is found. This study proposes a hybrid heuristic approach for optimal management of MG considering regional conditions and constraints. The problem is for a MG that generates power using the renewable resources in the field, as well as purchasing power from the main grid. MG is constructed in an Industrial Zone where the hourly power demand has to be responded. The aim is to schedule the power loads to minimize amount of power taken from main grid. To deal with this complex problem which contains power generation and consumption constraints, a versatile mathematical model must be established. The mathematical model needs to be integrated with a hybrid heuristic algorithm; only such an improved heuristic approach can be satisfying in solution of the model. Thus, a hybrid Genetic Algorithm (GA) – Simulated Annealing (SA) method has been proposed for solution. The schedule is programmed using GA, while, parameters are optimized by using SA. The structure of the proposed algorithm is suitable for scrutinizing the deficient points of the cases. It can be controlled that stepping in time of power resources. The proposed method can be used for evaluating several distributed generation MG systems to give a decision between two modes of operation. In the application stage, a MG in Gebze, Turkey is simulated with three factories as consumer, where, grid connection and a wind turbine and photovoltaic panels are assumed to be in use. Four different dates have been covered as case studies. If the only concern is minimizing the total cost of power use, dependence on only renewable resources are quite costly because of the abundance of wind and solar sources. To solve this problem, natural grid connection and storage are to be active. It is observed that establishing a standalone MG is not easy in spite of the power surplus in most of the day hours.