A Control Evaluation on a Slurry-Filled Sump

Sump can be definedas a reservoir which is normally located at the downstream of
a process, from which water is pumped where solids accumulates. Sump is a
system which faces instability due to its dynamic behaviors. Application of control
strategies to the sump might add some advantages in handling the sump, where in
some process, sumpplays a significant role especiallywhen it involves in a process
that have highly hazardous stage. This project studies generally observe the
behavior of the sump by adapting control type such as the feedback, Smith
predictor, feedforward, and cascade control. In addition, the sump is adaptedwith a
neurocontrol from Neural Network Control strategies. The neurocontrol used is the
NARMA-L2 controller. The variation of simulation have resulted in a way that
majority of the controller are at highest performance when the percentage of solids
in is 80%. The application of neurocontrol to the system somehow does not meet
the target due to high error that the controller sustained. However, the neurocontrol
can be enhanced more by data training and this is recommended for further
researchby using other neurocontrol that is available in the Neural Networkcontrol
strategies.