Fuzzy Sensor Fuzion and Adaptive Threshold Classifier in Refiner Monitoring

J. Hiltunen and M. Tervaskanto (Finland)


fuzzy logic, sensor fusion, classification, monitoring, pulp refining


Quality properties in production may vary due to alterations in raw materials and/or conditions in the process. These quality variations are often hard to observe and control by basic control and monitoring systems. The paper presents some useful features of the quality control system, which is successfully used to detect and stabilise disturbances in the reject refining and even quality variations in the pulp entering the paper machine. The methods used are based on fuzzy sensor fusion and adaptive threshold classifier. The presented results are based on the experiments made in a full-scale plant in Finland.

Important Links:

Go Back