Development of a Smart Monitoring System for Palm Oil Processing Using Internet of Things Technology

Abstract

Palm oil mills operate with limited real-time visibility of their own processes, so that deviations in temperature, level, throughput, and product quality are typically detected through periodic laboratory analysis rather than as they occur. This paper develops a system architecture for Internet of Things based monitoring of palm oil processing and specifies the design requirements that distinguish workable installations from those that fail after commissioning. The study followed a design research approach in which functional requirements were elicited from published mill-level monitoring implementations indexed in Scopus and SINTA, an architecture was synthesised, and evaluation criteria were derived and applied analytically to the proposed design. The resulting architecture comprises four tiers: instrumentation at the unit operation, edge acquisition and local control, a communication and integration layer, and an analytics and visualisation layer. The review of existing implementations indicates that sensing at the continuous settling tank, sterilizer, boiler, and effluent pond is technically straightforward with low-cost microcontroller hardware, but that sensor linearity outside a narrow measurement range, the absence of calibration procedures, and the lack of integration with existing supervisory control systems repeatedly limit operational value. The paper argues that the binding design constraint in this application is not connectivity but measurement traceability, and proposes that any deployment specify calibration intervals, acceptable measurement ranges, and a defined action for each monitored variable before hardware is selected.