About the book

Industrial processes can be broadly classified into stable, unstable, and integrating types. Time delay is inherent in many such industrial processes due to distance velocity lags, transportation and recycle loops, composition analysis loops, etc. The majority of the available literature is pertaining to controller design for stable processes with dead time. Integrating plants are non-self-regulating, and in case of any disturbance or change in plant input, the process output becomes unbounded. Models of these processes have at least one pole at the origin. Unstable processes have at least one positive pole. These processes are often sighted in dynamics associated with chemical industries like chemical reactors, pressure boilers, heat exchangers, etc. Dynamics of plants like storing tanks, drying cans, and distillation columns are of integrating type. Controller design for these plants involving dead time is a formidable exercise as it contributes to additional phase lag which destabilizes the closed-loop system. Also, processes such as boiler steam drums, adiabatic tubular reactors, and reboilers in distillation columns exhibit inverse response. Transfer functions of inverse response processes are characterized by an odd number of zeros in the right half-plane. The control of unstable and integrating processes becomes all the more difficult in the presence of time delay and right half-plane zeros. Hence, unstable and integrating processes require specific controller design methods.

PID controllers are widely employed in a unity feedback configuration to control the aforesaid time delayed stable, unstable, and integrating processes with/without inverse response behavior. However, researchers have found that PID-based double-loop control schemes can serve as a better alternative for unity feedback schemes while controlling unstable and integrating processes. Moreover, to improve the servo response of processes with large dead-time, PID-based dead-time compensators (Smith predictors) are preferred. These Smith predictors need to be further modified for controlling unstable and integrating processes with large dead-time. In recent times, fractional order controllers have become increasingly popular due to their flexibility in tuning and enhanced performance.