Robust Real-Time Optimization for Refinery Process Operations

点击翻译按钮获取中文摘要

The chemical process industry is facing challenges with the intensification of global market competition, more stringent limits in product specifications, pricing pressure and environmental problems. To address those challenges, many large manufacturers are starting to use data analysis to optimize plant operations, improve equipment efficiency and product quality, and reduce energy consumption. Machine learning techniques and new digital technology is driving the process industry companies to collect and analyze the growth of the amount of data. As an integrated energy company and Canada's largest refiner of petroleum products, Imperial Oil is facing challenges in improving overall profitability. One of those challenges is related to making robust process operations decisions under uncertainty in real-time. The present proposal aims to address the above challenge by developing a data-driven robust real-time optimization algorithm for online application in manufacturing facilities. Real-time optimization is the process of optimizing the plant operating conditions online and represents the first level in a typical control hierarchy where the economics of the plant is addressed explicitly. The proposed research project will use big data, machine learning, and robust optimization technology to meet the industry needs. The proposed project will help Imperial Oil optimize the utilization of resources through process systems engineering techniques, so as to improve profitability. Imperial Oil will implement the developed data driven optimization algorithm to its plethora of applications and benefit economically from it. The proposed research will not only benefit Imperial Oil but also benefit Canada through technology transfer to other Canadian manufacturing companies, and through contribution to training HQPs with technique backgrounds that are needed by the oil and gas industry.