Design of controllers for critical processes in thermal and hydro power plants using evolutionary computation techniques

Abstract

The controllers play critical role in power plants to meet the dynamic load demands. It is always a highly challenging and complicated task to design a controller for power plants due to highly interactive nonlinear behavior offered by plants for a wide range of operating conditions. The present thesis investigates an application of evolutionary computation techniques for optimal controllers design for thermal and hydro electric power plants, which should ultimately enhance the performance of the plants. At first, the present work attempts to design and implement an Optima z nonlinear controller for boiler-turbine system of thermal power plant. The present work is the first one which attempts to design and implement recently developed nonlinear finite time convergent controller for the thirdorder nonlinear boiler-turbine system. The present work also explores the possibility of optimal tuning of nonlinear controller using single objective evolutionary computation techniques such as Real-coded Genetic Algorithm (RGA), Modified Particle Swarm Optimization (MPSO), Differential Evolution (DE) and Self-Adaptive Differential Evolution (SADE) with the objective of minimal Integral Square Error (ISE) and multi objective techniques such as Nondominated Sorting Genetic Algorithm II (NSGA II) and Modified Nondominated Sorting Genetic Algorithm II (MNSGA II) with the objective of newlinesimultaneous minimization of ISE and control energy of actuators. The proposed optimal nonlinear controller has been tested againstset points tracking and robustness during wide range of operations. Among single objective algorithms, SADE has offered better results and among multi objective algorithms, MNSGA-II has provided better results. newline newline