Abstract
The work develops the nonlinear state dependent parameter PID + control with the industrial implementation of the Bitumen system sited at INSUMAT Company, aiming to achieve better control performance for Bitumen temperature prior to the mixing process with the additives (polymers and fillers). The viability of the PID + control had been proved before inside laboratory, however, the current work presents its reliability for industrial implementation on-site for the first time. The challenges of nonlinearities and time delays that arise due to the nature of the Bitumen substance, the tank itself, and other issues such as the turbulent flow of the hot oil in the pipe surrounds the tank, the feedback of the PT100 temperature sensor, and the three-way actuating valve, persuade the authors to implement the time-variant PID + control system rather than the current discrete PID control to overcome this nonlinearity and counteract the high samples delay arises from the dynamics of this industrial system. In this regard, this work introduces the use of the nonlinear SDP-PID + control by means of discrete time transfer function (TF) for which its parameters are themselves state dependent (SDP). Here, SDP-TF model acts as a nonlinear description of the industrial Bitumen system. Regarding PID + control, the plus denotes extra input compensators over the three conventional PID gains to facilitate the use of the state variable feedback (SVF) control law and neutralize the influence of the samples delay arises in the nonlinear discrete time SDP-TF of the industrial Bitumen system. The work shows the first 'onsite' successful industrial implementation for the nonlinear SDP-PID + control for the industrial Bitumen system, located in INSUMAT Company, Giza-Egypt. The approach utilizes the Linear Quadratic (LQ) cost function which exploits the non-minimal state space (NMSS) form of the SDP-TF model for tuning the nonlinear PID + compensators. It is worth noting that all criteria for control design are verified in terms of appropriate closed loop performance and rejection of input/output disturbances together with retrieving the tracking to the set point in an acceptable time.