System Architecture Modelling of Sugarcane Agroindustry Machinery Anomaly Detection Through Machine Learning

Anomaly detection in sugar agroindustry machinery is important to improve operational robustness. One of the machines that have a crucial role in the sugar agroindustry process is the sugarcane milling machine. A disturbed and abnormal sugarcane milling machine will cause various losses including a decrease in production efficiency, deterioration of product quality, and excessive consumption. Based on this, anomaly detection that can be done quickly, one of which uses machine learning, will help the agroindustry in making decisions. The purpose of the research is to design an early anomaly detection system in sugarcane milling machines with a model-based system engineering approach. The method used in this research is a model-based system engineering framework that is used to map the machine learning-based anomaly detection system to be built. Through model based system engineering, functional architecture, non-functional architecture, logical architecture, and physical architecture will be known. The architecture involved in the system will be described in the form of block definition diagram (BDD) and internal block diagram (IBD). This research is expected to describe the behavioral needs of the machine learning-based anomaly detection system in sugarcane milling machines. This research is expected to be a reference for the sugar cane agroindustry to increase operational robustness and carry out digital transformation to increase production efficiency and improve the quality of the final product.