Soft sensing of sodium aluminate solution component concentrations via on-line clustering and fuzzy modeling

The component concentrations measurement of sodium aluminate solution are critical to the process of alumina production, they affect the product quality. However, they can not be measured online at present, thus the control and optimal operation is hardly to be achieved. This paper presents an on-line fuzzy modeling method to predict the component concentrations. It includes an on-line clustering approach which can be applied in a general class of fuzzy TKS models. Stable learning algorithms for the premise and the consequence parts of fuzzy rules are also given. A measuring device is developed to achieve the proposed method and industry experiments are conducted in the alumina production process, the predicted results show the effectiveness of the proposed method.