Sensor fusion as a tool to monitor dynamic dairy processes

A system for monitoring milk and fat concentration in a dynamic milk/water system by fusing information from several sensors was investigated. Standard instrumentation for food production was used, the sensors were a conductivity meter, a density meter and an optical instrument used to measure backscattered light. The system was applied to a dynamic mixing situation. Prediction error did not exceed 2% in the milk concentration and 0.1% fat in the total fat concentration. The applicability of the sensor fusion approach in field conditions was demonstrated by mounting the sensors in a dairy plant and monitoring the start-up of a pasteurizer.