Enhancing Realtime Supervision and Control of Industrial Processes over Wireless Network Architecture Using Model Predictive Controller

This paper presents enhancing real-time supervision and control of industrial processes over wireless network architecture using model predictive controller. The research reviewed various related literatures on Real Time Operating Center (RTOC) and their importance on industrial control systems. From the review it was observed that one of the major components of RTOC is the Remote telemetry Unit (RTU) or Programmable Logic Controller (PLC). These systems are embedded with Proportional Integral Differential (PID) controllers for processing of data collected and transmitted to the RTOC monitor via the communication bus; however the delay response time of the PID controllers induce latency on the data transmitted, thus affecting the quality of RTOC analysis and as a result has remained a major problem all over the world. This problem was addressed in this research using artificial neural network (ANN) based model predictive controller. The ANN was trained using data collected from an oil and gas drilling process to develop a predictive model which was used to collect time series data of the plant and send to the RTOC monitor in real-time. The system was implemented with Simulink and the performance was evaluated. The result showed that the predictive controller was able to collect data and transmit to the RTOC at 22.5ms, which according to IEC 60870-6 and IEC 62591 Standard for RTOC satisfy the requirement for real-time and better then the 40ms achieved in the conventional system.