Design and Development of an Intelligent Advanced Process Control (APC) System Integrating Machine Learning, Model Predictive Control, and Fuzzy Logic

Department

Advanced Solutions department — Project 01/03 of the Integration Objects PFE Book 2026–2027. The topic is built around the SIOTH® Industrial IT-OT-IoT platform.

Project

Develop an intelligent Advanced Process Control (APC) approach integrating process modeling, Machine Learning, soft sensors, Model Predictive Control (MPC), and Fuzzy Logic to improve process stability, control performance, and operational efficiency.

Objectives and deliverables

• Collect, preprocess, and explore industrial process data to identify relevant variables, operating patterns, correlations, and disturbances.
• Develop mathematical and data-driven process models using Machine Learning to characterize process dynamics, establish relationships between process variables, and predict future process behavior and controlled-variable values.
• Develop and validate soft sensors for important process parameters that are difficult, costly, or impractical to measure directly.
• Evaluate MPC, Fuzzy Logic, and hybrid control strategies considering Controlled Variables, Manipulated Variables, Setpoints, Disturbance Variables, and operating constraints.
• Evaluate predictive models, soft sensors, and control strategies using KPIs related to prediction accuracy, estimation reliability, control performance, stability, robustness, disturbance rejection, and operational efficiency.
• Develop an interactive dashboard displaying process data, predictions, soft sensor estimates, control performance, and KPIs.
• Produce technical documentation, a final report, and a presentation covering the methodology, model development, control strategy evaluation, results, and potential business impact.

Keywords

APC, process modeling, system identification, Python, Model Predictive Control, Machine Learning, Fuzzy Logic, soft sensors, industrial data analytics, process optimization, multivariable control.