Estimation of super heater Steam temperature using regression models on MISO systems

Super heater Steam temperature is an important and critical parameter requiring continuous monitoring and control for efficient operation of a thermal power plant. The temperature control is achieved by spraying water and the amount of spray water is regulated and sprayed on the tubes carrying superheated steam. Traditional PID controllers used for this control application have an inherent drawback of not being able to adjust to the changing process dynamics as they do not have a plant model in the loop. Further, the temperature variations are dependent on many factors including process time delay, load and other plant disturbances. Adaptive & Feed Forward controllers are limited by the fact that they are all reactive in nature and can act only after the cause of deviation has occurred. Precise control of steam temperature requires a plant model capable of predicting the changes in advance so that it can modify the control action to control the steam temperature more precisely. This paper presents an overview of the performance of a few regression models in predicting the output of Multi-Input-Single-Output (MISO) models which can also be used in the context of predicting the super heater steam temperature.