NEURAL NETWORKS MODEL OF MOTHER LIQUOR PURITY OF C MASSECUITE I
暂无分享,去创建一个
I Crystallisation is an important stage in the manufacturing process of sugar cane. In Bois Rouge sugar mill, the crystallisation process is performed in three stages. Our work deals with C crystallisation. During crystallisation, purity of mother liquor is a parameter of importance as it indicates the concentration of sucrose remaining in solution. The final productivity of the sugar cane conversion is dependent on the efficiency of C crystallisation. In Bois Rouge sugar mill, no physical measure is available concerning exhaustion on-line in C crystallisation. To observe the dynamic exhaustion of mother liquor during the growing stage, we took samples of molasses and massecuite every ten minutes. These samples were analyzed to determine purity of mother liquor. Then, we built a multi- step predictor of mother liquor purity using physical measures on-line and off-line. First results are very encouraging.
[1] Kurt Hornik,et al. Multilayer feedforward networks are universal approximators , 1989, Neural Networks.
[2] Martin Pottman,et al. Identification of non-linear processes using reciprocal multiquadric functions , 1992 .
[3] S. Billings,et al. Correlation based model validity tests for non-linear models , 1986 .