Journal of American Science, 2011;7(4) http://www.americanscience.org http://www.americanscience.org editor@americanscience.org 629 Design of optimal fuzzy controller for water level of U-Tube steam generator in nuclear power station Hamdi. M. Mousa * , Magdy. A. Koutb ** , Sayed. M. El-Araby *** , And Elsayed. H. M. Ali *** * Faculty of Computers and Information, Menoufia University, Egypt ** Industrial electronics and control department, Faculty of electronic engineering, Menoufia University, Egypt *** Engineering and scientific instruments department, nuclear research center, Atomic energy authority, Cairo, Egypt sayedmahdy@yahoo.com Abstract: The steam generator is a highly complex, nonlinear and time-varying system and its parameters vary with operating conditions. A method to improve the performance of nuclear steam generator in nuclear power station is introduced. Combination of genetic algorithm technique and fuzzy logic control is carried out. The optimal parameters of fuzzy logic controller are achieved. These parameters include; the membership functions of water level error and changes water level error, the rule base, and the input scaling gains. Steam generator model implemented using MATLAB/SIMULINK. The optimal controller was applied to control the water level of nuclear steam generator and it’s compared with conventional controller. Simulation results indicate that the optimal fuzzy controller greatly improves the performance of nuclear steam generator. Moreover the proposed controller is robust to any disturbance related to sudden changes in steam flow rate and water level. Moreover the proposed controller is robust to any disturbance related to load variations. [Hamdi. M. Mousa, Magdy. A. Koutb, Sayed. M. El-Araby, And Elsayed. H. M. Ali. Design of optimal fuzzy controller for water level of U-Tube steam generator in nuclear power station. Journal of American Science 2011;7(4):629-637]. (ISSN: 1545-1003). http://www.americanscience.org . Keywords: Fuzzy logic control, genetic algorithm, steam generator, nuclear power stations. 1. Introduction Due to growing electricity demands, many countries are investing in nuclear power stations. During the operation of the nuclear power station, different changes in the operating conditions may occur such as water level in steam generator. Steam generator is a very important component of nuclear plant. It is of significance for nuclear reactor normal operation to maintain it work safely and reliably. For the steam generator in a nuclear power station, the main goal of its control system is to maintain the steam generator water level at a desired value [1]. The water level of a nuclear steam generator is of great importance in order to secure the sufficient cooling water for removing the primary heat and to prevent the damage of turbine blades [2]. The water level control problem of steam generator has been a main cause of unexpected shutdowns of nuclear power plants. In France, nuclear energy provides about 80% of the whole electricity production. This production must be able to vary proportionally to the consumption [3]. Figure 1 shows the steam generator (SG) in a nuclear power station [4]. It is very important to keep the operation of nuclear power station in highly safe mode. Therefore, an intelligent control system is required to compensate the steam flow rate changes produced sudden changes in load variations in the operating condition of the nuclear power station. The conventional proportional– Integral (PI) controller [5] was used to control the water level of the nuclear power stations. Simulation results show that conventional PI controller has long time delay between actual water level and reference level, big value of overshoot and undershoot, and bad tracking. Fuzzy Logic Control (FLC) is one approach of the Artificial Intelligence (AI) techniques that is used to control the performance of nuclear power stations [6],[7]. Simulation results show that the output response is improved, but it still insufficient to get an optimum response. The rule -base of fuzzy logic controller reflects the human expert knowledge, expressed as linguistic variables, while the membership functions represent expert interpretation of those same variables. In the absence of such knowledge, trial and error is a common approach used to optimize these fuzzy logic controller parameters, with respect to the performance of the system for each Knowledge base formulated [8]. This approach becomes impractical and not accurate for adjusting the parameters of membership function and rule base of nonlinear systems. In this paper a genetic algorithm is proposed for optimal tuning parameters of fuzzy controller namely, rule base, membership function, and input scaling gains. The proposed genetic fuzzy controllers greatly improves the output response of the nuclear power station and dramatically decreases the overshoot and undershoot, the steady state error close to zero value, decreases the settling time and rise time. The paper organized as follows: Section one describe the mathematical model of nuclear steam generator.