Scientific operation of production processes with quantitative evaluation of sensory characteristics is important for improving product quality in the food industry. In order to achieve the target quality in the production processes with controllable parameters, the set-point of process parameters must be controlled to suit the operating conditions. In food production processes where the quality is represented as sensory characteristics, it is particularly important to quantitatively evaluate the quality of products.
Set-point control methods include intelligent supervisory control, case-based reasoning, etc. In addition to rule-based reasoning, case-based reasoning that utilizes operation expertise and cases to infer similar solutions corresponding to similar conditions has been widely applied to the optimal operation of industrial processes.
One of the major challenges in applying case-based reasoning to the set-point control of industrial processes is to constantly adapt the case base to the changing characteristics of processes, thus increasing the accuracy of reasoning. There was a methodology to solve this problem, which failed to deal with updating the case base with operational experience and new knowledge obtained during the online operation of industrial processes.
Song Kwang Rim, a researcher at the Faculty of Automation Engineering, has proposed a new method for online updating and addition of case base and applied it to the set-point control of a rolled cake production process.
First, he presented an approach for quantitatively evaluating the rolled cake based on the sensory quality. Second, he proposed an interactive method of acquiring knowledge from operator’s experiences and adapting the case base according to the behavior of the process and the quantitatively evaluated quality level.
The comparative experiments for validation show that the proposed method results in 4.1% improvement in the ratio of good products.
You can find the details in his paper “Acquisition of adaptive knowledge in case-based reasoning for the online set-point control of industrial process” in “Third International Conference on Communications, Information System, and Data Science (CISDS 2024)” (EI).
© 2021 Kim Chaek University of Technology