T.K. Radhakrishnan

Design of self tuning fuzzy controllers for nonlinear systems

Nonlinearities present in the systems make their controller design a non-trivial task. The difficulty further increases in case of multi-input-multi-output (MIMO) systems with increased number of variables and interactions between them. In this paper, fuzzy based intelligent control schemes are designed for control of nonlinear single-input-single-output (SISO) and MIMO systems. The comparative study of the designed self tuning fuzzy controller with a standard Takagi-Sugeno (TS) fuzzy controller is discussed with application to a shell and tube heat exchanger (nonlinear SISO system) and a coupled two tank system (nonlinear MIMO system). Online tuning of the membership functions and control rules of fuzzy controller is carried out using simulated annealing (SA) to obtain improved performance by minimizing the error function. Experimental results demonstrate the effectiveness of the control scheme.

Development and tuning of fuzzy controller for a conical level system

A real-time control of liquid level in a conical tank is studied. System identification of this nonlinear process and tuning of the fuzzy system are undertaken. Genetic algorithms are used to tune the membership functions of the fuzzy system. The results of the tuned fuzzy controller are compared to a proportional integral controller to evaluate the performance of the genetic algorithm tuned fuzzy controller.

Particle Swarm Optimization based PI controller tuning for Fermentation Process

The activities related to fermentation process are uncertain and nonlinear in nature. It poses many challenging characteristics like multivariable interactions, unmeasured state variables, system and time varying parameters. Generally, PID controllers are used in these processes. Tuning of PID controller is necessary for the satisfactory operation of the system. The tuning of PID controllers depends upon how the process responds to the controllers corrective effort. The conventional tuning methods like Z-N method and IMC method takes more time, reduces stability and are more sluggish. In this paper, the problem of identifying the PI controller parameters is considered as an optimization problem. An attempt has been made to optimize the values of the PI controller parameters employing particle swarm optimization technique (PSO). From the simulation results it is observed that the PI controller designed with PSO yields better results when compared to the conventional methods

GOBF-ARMA based model predictive control for an ideal reactive distillation column

This paper discusses the control of an ideal reactive distillation column (RDC) using model predictive control (MPC) based on a combination of deterministic generalized orthonormal basis filter (GOBF) and stochastic autoregressive moving average (ARMA) models. Reactive distillation (RD) integrates reaction and distillation in a single process resulting in process and energy integration promoting green chemistry principles. Improved selectivity of products, increased conversion, better utilization and control of reaction heat, scope for difficult separations and the avoidance of azeotropes are some of the advantages that reactive distillation offers over conventional technique of distillation column after reactor. The introduction of an in situ separation in the reaction zone leads to complex interactions between vapor-liquid equilibrium, mass transfer rates, diffusion and chemical kinetics. RD with its high order and nonlinear dynamics, and multiple steady states is a good candidate for testing and verification of new control schemes. Here a combination of GOBF-ARMA models is used to catch and represent the dynamics of the RDC. This GOBF-ARMA model is then used to design an MPC scheme for the control of product purity of RDC under different operating constraints and conditions. The performance of proposed modeling and control using GOBF-ARMA based MPC is simulated and analyzed. The proposed controller is found to perform satisfactorily for reference tracking and disturbance rejection in RDC.

Neural Model Predictive Controller for Multivariable Process

In this paper, control of a non-minimal quadruple tank process, which is non linear and multivariable is reported. A nonlinear Model Predictive Controller (NMPC) is developed using a Recurrent Neural Network (RNN) as a predictor. The process data is obtained from the laboratory scale experimental setup, which is used in training the RNN. The network trained is used in controlling the quadruple tank, solving the least square optimization problem with a quadratic performance objective. The control system is implemented in real-time on a laboratory scale plant using dSPACE interface card and Matlab software. The quality of controller using NMPC is compared with dynamic matrix control (DMC) for reference tracking and external disturbance rejection.

Kalman Filter Based State Estimation of a Thermal Power Plant

Tangentially-fired furnaces (TFF) are vortex-combustion units and are widely used in steam generators of thermal power plants. Perfect modeling and simulation of furnace gas temperature is quite difficult, due to its complex aerodynamics of burning particles, flame stability and hot gas flow distribution throughout the furnace. The temperature of the furnace gas depends on many parameters such as the inclination angle (tilt angle), fuel quality, burn out percentage and the flow rates in the burners for each of the furnace corners. However, the measurements are not available in the existing furnace operated at Neyveli Lignite Corporation (NLC), Neyveli. Thus, state estimation of temperature is an important prerequisite for safe and economical process operations. It is an integral part of applications such as process monitoring, fault detection and diagnosis, process optimization, and model-based control. Because all the process variables are generally not measured, an observer can be designed to generate an estimate of the state by making use of the relevant process inputs, outputs, and process knowledge, in the form of a mathematical model. The aim is to design a good state estimator for the furnace. Linear Kalman Filter (LKF) and Extended Kalman Filter (EKF) algorithms are developed for this problem and simulation results are compared.

Predictive controller design for a shell and tube heat exchanger

Process industry requires now accurate, efficient and flexible operation of the plants. There is always a need for development of innovative technologies for process modeling, dynamic trajectory optimization and high performance industrial process control. The advanced control technology which made a significant impact on industrial control engineering is model predictive control (MPC). The objective of this work is the development of MPC for a shell and tube heat exchanger and to address the difficulties in tuning a MPC. The identification of the heat exchanger is done by using the black box model technique. Although the PI controller is widely used for these types of applications there is still a need for optimization of conservation of energy. In this paper a model based predictive algorithm is used for controlling a temperature of a fluid stream using the shell and tube heat exchanger and the associated difficulties in tuning are analyzed. The transients and steady state results obtained using a MPC is compared with a conventional PI Ziegler Nicholas (ZN) controller.

Hybrid GA Fuzzy Controller for pH Process

Control of a pH process is great difficulty due to time varying and nonlinear characteristics. Fuzzy logic has been successfully applied to many applications in control with uncertainties. An important consideration in designing any fuzzy control system is the formation of the fuzzy rules and the membership functions. Generally the rules and the membership functions are formed from the experience of the human experts. With an increasing number of variables, the possible number of rules for the system increases exponentially, which makes it difficult for experts to define a complete rule set for good system performance. Also the system performance can be improved by tuning the membership functions. In a fuzzy system the membership functions and rule set are codependent, they are encoded into the chromosome and evolved simultaneously using genetic algorithm. The performance of the proposed approach is demonstrated through development of fuzzy controller for a benchmark pH process. In both set point tracking and disturbance rejection, simulation results show the better performance when compared to fuzzy logic controller.

Design of Fractional Order Controller for Biochemical Reactor

Abstract This paper presents a simple procedure to design fractional order controller based on synthesis method for the biochemical reactor. The biochemical reactor process exhibits high degree of non linearity. The process parameters will be be varying during the process of fermentation. Hence an attempt is made to design robust PI controller for the biochemical reactor to achieve high steady state productivity.

Design and Implementation of Controllers for MIMO Process

Most of the large and complex industrial processes are naturally Multi Input Multi Output systems. Compared with single-input single-output (SISO) counterparts, MIMO systems are more complex to control due to inherent nonlinearity and due to the existence of interactions among input and output variables. Control of nonlinear MIMO process is a challenging task because nonlinear processes do not share many properties. This paper presents a real-time implementation of conventional controller and multi model controller for a laboratory scale non linear MIMO process. Experimental results shows that control strategy based on the multi-linear models can effectively handle control problems in complex nonlinear plants.