PI control based on parameter space approach supported metaheuristic optimization algorithms and ANFIS in a natural gas combined cycle power plant

• Published in: Optimal control applications & methods 2023-03, Vol.44 (2), p.846-865
• Main Authors: Bayhan, Nevra, Arıkuşu, Yılmaz Seryar, Tiryaki, Hasan
• Format: Article
• Language: English
• Subjects: Active control; ANFIS controller; Artificial intelligence; atomic search optimization based PI; Closed loops; Combined cycle power generation; Controllers; coronavirus herd immunity optimization based PI; Genetic algorithms; Heuristic methods; metaheuristic optimization algorithms; Natural gas; natural gas combined cycle power plant; Natural gas industry; Optimization algorithms; Parameters; Particle swarm optimization; Power control; Power flow; Power plants; power system; Steam turbines; Supply & demand; Swarm intelligence; Transient response
• ISSN: 0143-2087; 1099-1514
• DOI: 10.1002/oca.2823

In this article, proportional‐integral (PI) control to ensure stable operation of a steam turbine in a natural gas combined cycle power plant is investigated, since active power control is very important due to the constantly changing power flow differences between supply and demand in power systems. For this purpose, an approach combining stability and optimization in PI control of a steam turbine in a natural gas combined cycle power plant is proposed. First, the regions of the PI controller, which will stabilize this power plant system in closed loop, are obtained by parameter space approach method. In the next step of this article, it is aimed to find the best parameter values of the PI controller, which stabilizes the system in the parameter space, with artificial intelligence‐based control and metaheuristic optimization. Through parameter space approach, the proposed optimization algorithms limit the search space to a stable region. The controller parameters are examined with Particle Swarm Optimization based PI, artificial bee colony based PI, genetic algorithm based PI, gray wolf optimization based PI, equilibrium optimization based PI, atom search optimization based PI, coronavirus herd immunity optimization based PI, and adaptive neuro‐fuzzy inference system based PI (ANFIS‐PI) algorithms. The optimized PI controller parameters are applied to the system model, and the transient responses performances of the system output signals are compared. Comparison results of all these methods based on parameter space approach that guarantee stability for this power plant system are presented. According to the results, ANFIS‐ PI controller is better than other methods.