Digitization of information
On-site work efficiency
More sophisticated decision making
Information visualization and browsing
Utilization of knowledge database
From anomaly detection Prevention of accidents
O&M improvement by predictive detection
Automation of operation and inspection
･ AI Prediction: AI prediction and visualization of important instruments and quality data of products
･ Visualization of operating status: Visualization of tray load, heat exchanger dirtiness, etc.
･ Changeover evaluation: Technology transfer of advanced operation knowledge and skills to junior operators
･ Optimization of operating parameters: Increase of middle distillate amount (minimize residual oil), shortening of crude oil changeover time, optimization of heat balance
･ Operation optimization: Application for the crude distillation unit that performs crude oil switching operation about 100 times/year (advanced control application is difficult and forced to manual operation)
･ Operation optimization: Improving yield of products and energy loss by optimum operating parameters
･ Technology transfer: Supporting advanced operation knowledge and skills by switching operation evaluation
The CDU Operations Optimizer is an AI application that learnt the correlation of operating parameters using
plant operation data, dynamic simulators, and AI technology (deep reinforcement learning).
The CDU Operations Optimizer provides optimal operating parameters in real time during the switching operation, and contributes the optimal operation in terms of energy saving, product loss minimization, reducing operation restriction, and early switching completion etc.
Achievements and examples
Japanese Domestic refineries are doing refining business by using various types of crude oil as raw materials. The switching of crude oil types occurs at a frequency of once every few days, and the oil type switching operation of the crude oil distillation column is carried out each time.
In order to prioritize product quality while crude oil and product properties fluctuate in the oil type switching operation, the yield of kerosene, gas oil, etc. from the crude oil distillation column is reduced and product loss occurs.
In addition, it is necessary to operate while considering the wide range of effects such as energy saving, product loss minimization, operation restrictions such as corrosion, and early switching operation while the flow rate balance is avoided. Because of the complexity of operation, a single control model is not adopted nor automation has not been reached.
Furthermore, due to the aging of the operator and the retirement of skilled operators, and the range of support for each operator is expanding, technology transfer has become a difficult problem.
It is expected that the CDU Operations Optimizer is to reduce the work load on operators, improve the yield of kerosene and gas oil, reduce utility costs in the oil type switching operation by the indication of the best parameters of the optimum operation by using AI technology. And the system contributes the transfer of skill and knowledge of skilled operators to junior operators.
Copyright Ministry of Economy, Trade and Industry. All rights reserved.