Optimal nonlinear Fractional-Order Proportional-Integral-Derivative controller design using a novel hybrid atom search optimization for nonlinear Continuously stirred Tank reactor

• Published in: Thermal science and engineering progress Vol. 54; p. 102862
• Main Authors: Kamalakkannan, P.E., Vinoth kumar, B., Kalamani, M.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.09.2024
• Subjects: Artificial Bee Colony(ABC) Optimisation; Atom Search Optimization(ASO); CSTR; NL-FOPID; Atom Search Optimization(ASO); CSTR; NL-FOPID; Artificial Bee Colony(ABC) Optimisation
• ISSN: 2451-9049
• DOI: 10.1016/j.tsep.2024.102862

•Inspired by molecular dynamics, particularly the Lennard-Jones potential and bond length potentials.•The algorithm simulates the forces between atoms to find optimal solutions in a search space.•Enhances ASO by integrating the Artificial Bee Colony (ABC) algorithm.•ABC guides the replacement of atoms that move out of the search space, improving convergence and optimization efficiency.•A more complex controller than traditional PID, involving additional parameters for better control performance.•Initially optimized using Genetic Algorithm (GA) and then improved using HASO.•Operates on a population of potential solutions using crossover and mutation to evolve better solutions.•Applied to benchmark functions and controller optimization, with parameters like crossover rate, mutation rate, and population size carefully tuned.•Mimics the annealing process of metals to escape local minima by occasionally accepting worse solutions.•Parameters include initial temperature, reannealing interval, and maximum iterations.•Models the social behavior of particles to find optimal solutions, involving social and cognitive attraction factors. Atom search optimization (ASO) algorithm derived from physics molecular dynamics. Lennard-Jones(L-J) and bond length potential of molecules are used to derive the model for optimization. In this paper, ASO is used for developing a nonlinear Fractional Order Proportional Integral Derivative controller (NL-FOPID) for Continuously Stirred Tank Reactor (CSTR). The convergence characteristics of ASO was improved by proposing a novel hybridization approach. The proposed hybridization approach called Hybrid ASO(HASO) guides the Atom search algorithm to optimally replace the atoms that goes out of the boundary of the search space. The designed algorithm is implemented to optimize various unimodal and multi model standard benchmark functions. From results obtained from this extensive simulation, it is indicated that proposed approach increased the convergence rate and also improved the optimization effort of conventional ASO. The proposed algorithm also tested with controller design for nonlinear CSTR. The NLPID and NL FOPID designed by HASO was better than conventional controllers found in the literature.