Multi-Objective LPO by Fuzzy Gravitational Search Algorithm in WWER1000
Authors: M. Aghaiea,*, S.M. Mahmoudib
a Engineering Department, Shahid Beheshti University, G.C, P.O. Box: 1983963113, Tehran, Iran bFaculty of Electrical and Computer Engineering, Graduate University of Advanced Technology, Kerman, Iran.
Number of manuscript pages:
Number of figures:
Number of tables:
* Corresponding author Email: M_Aghaie@sbu.ac.ir Fax: +98 21 29902546
Abstract
The Multi-Objective Loading Pattern Optimization is one of the most important concerns for the incore design of nuclear reactors. Hence, different techniques have been presented for optimization of incore patterns for nuclear reactors. This paper presents …show more content…
The Ackley function is a well-known optimization problem that utilized for comparison the performance of the method in searching the best answer and its convergence speed. To obtain best reactor core configuration with FGSA, a fitness function that satisfies the goals and related constraints with NTH properties, is prepared. The NTH goals in multi objective optimization are to maximize (Keff, CHF) and minimizing (PPFs and fuel temperature) with flattening of power density. In this case, WWER1000, the FGSA is coupled with NTH codes. For NTH calculations, the PARCS and COBRA-EN are applied, respectively. In this study, a program is developed in MATLAB to optimizing core fuel loading pattern using GSA and …show more content…
1.Theposition vector, Xi, contains the agents (masses) data. All of forces acting each mass, fdij, could be calculated with regard to its mass, position (or distance) and gravitational constant. The mass, distance and gravitational constant for each agent could be calculated with Eqs. (1-3), respectively. (1) (2) (3)
Newton’s second low of motion presents acceleration of the agents, F=ma. The net of forces that affected agent i at iteration t, Fdi(t), deviding by mi(t), presents its acceleration, aid(t).New position and velocity of agents in each dimension could be found with calculated acceleation (Fig. 1).In this way, the GSA could present the heaviear mass or the best fitness value in the search space.
In this paper, a FLC is designed to control the convergence speed of a GSA in LPO. It is obvious that the acceleration of the agents depends on the α in G (Eq. (3)). The FLC tries to a modification in population diversity and approach progress with controlling the α.
Nomenclature t iteration number xid(t) dth position of the agent i at iteration