Sandeep Kumar Sunori

Comparative analysis of controllers designed for pure integral complex delayed process model

In present analysis, controllers based on different tuning techniques for a pure integral process with time delay are designed for different approximation of the time delay. The controllers are compared for the approximated process for the set point tracking capability. The set point tracking capability is determined on the basis of transient as well as the steady state analysis of a step response.

Neuro-Fuzzy Controller Design for MIMO Boiler Turbine Process

The plant selected for control system analysis is a 2 × 2 MIMO system with high multivariable interaction having two manipulated variables, the position of the governor valve and the fuel flow rate and two controlled variables, generated electric power and the steam pressure. A neuro-fuzzy controller has been designed using ANFIS in MATLAB for a boiler turbine plant. This technique is a hybrid of neural network and fuzzy logic techniques.

Dead Time Compensation in Sugar Crystallization Process

The input to the sugar factory is the sugarcane billets and the output is the crystal sugar. There are many subprocesses with significant multivariable interaction involved within this process which is cane preparation, juice extraction by crushing mill, heating, clarification and filtration, evaporation, and crystallization. In the present paper, the Smith predictor is designed using MATLAB in order to compensate the dead time present in heat exchanger system of the crystallization process and its performance is compared to that of a conventional PI controller with no dead time compensation.

Model Order Reduction of a Higher Order Model of pH Neutralizer of Sugar Mill

Several industrial chemical processes are found tobe FOPDT (first order plus dead time). The presence of deadtime makes the transfer function to be irrational and makesthe simulation and controller design very complex. In thepresent work an FOPDT model of the pH neutralizationprocess which is found in sugar mills is considered as a subjectof study and analysis. Having obtained its Padéapproximation, its order has been reduced using Square RootBalance Truncation Method and then for this reduced ordermodel, Ziegler-Nichols(ZN), Internal model control (IMC)and optimized IMC using Genetic Algorithm(GA-IMC)based controllers are designed and their performance is compared.

Rationalizing the Mathematical Model of Boiler Turbine Process with Dead Time

More or less every chemical industrial process suffers from some amount of inherent time delay which is highly undesirable. The presence of this dead time term makes the transfer function model irrational. In the present work, a 2x2 transfer function model of an industrial boiler turbine process with dead time is considered, which after decoupling has been rationalized using Padé approximation. The first, second and third order approximants are determined and validated by comparing open loop step responses.

Design and Analysis of Control Systems for Heat Exchanger System of Sugar Mill

In sugar factory, after clarification andevaporation, the extracted juice from the crushing mill issubjected to the crystallization process. Mass transfer andevaporation process are involved in the crystallizationprocess. Vacuum keeps the temperature at low value forminimizing the formation of colour and sucrose inversion. In the present work, the heat exchanger system, deployed incrystallization process of sugar manufacturing industry, isconsidered and control system is designed for it. For controlsystem design, Ziegler-Nichols (ZN), Internal model control(IMC) and Linear quadratic Gaussian (LQG) tuning methodsare adopted using MATLAB and finally their controlperformances are compared.

Investigation of Parameters Affecting Predictive Controller Performance for Boiler Turbine

The predictive control system decides the discrete moves in manipulated variables on basis of the prediction of output of the process upto the prediction horizon. The predictive controllers show an excellent performance as compared to conventional controllers. In the present work, a boiler turbine process with two manipulated variables and two controlled variables has been considered and a predictive controller has been designed for it. The impact of variation in the value of sampling interval, prediction horizon and control horizon on the setpoint tracking performance and the disturbance rejection performance has been investigated.

Disturbance Rejection Performance Analysis of Predictive Controller for a Lime Kiln Process

Every chemical industrial process is affected by the presence of measured and unmeasured input and output disturbances. A good controller is one which cancels the effect of disturbance as soon as possible else the performance of the designed controller will suffer from a severe degradation. In the present paper a complex multivariable lime kiln process has been considered where the occurrence of disturbances is very common. A prediction based controller has been designed for this process and its disturbance rejection performance has been investigated under various values of control horizon, prediction horizon and sampling interval.

Predictive Control System Design for Lime Kiln Process

MPC is a computer-based technique that requires the process model to anticipate the future outputs of that process. An optimal control action is taken by MPC based on this prediction. The MPC is so popular since its control performance has been reported to be best among other conventional techniques to control the multivariable dynamical plants with various inputs and outputs constraints. In this work, the control of lime kiln process with two manipulated variables namely the fuel gas flowrate, and the percent opening of the induced draft damper and two controlled variables namely front-end temperature and back-end temperature has been attempted using MPC technique. Lime kiln process is very complex and nonlinear multivariable process. A linearized model obtained using Taylor series expansion around operating point has been used.

Multiloop and prediction based controller design for sugarcane crushing mill process

In the present work a sugarcane crushing mill is presented as a MIMO system with high multivariable interaction. A linear model of the plant is taken with flap position and turbine speed as manipulated variables and mill torque and buffer chute height as controlled variables. The multiloop PI controller has been designed for this plant by first investigating the RGA and the value of Niederlinski index of this plant. The decoupling of this system is done and the respective open loop and closed loop step responses are observed and compared with those of the composite MIMO system. At last, the performance of multiloop controller is compared with that of a model predictive control system for this plant.