Clarification of sugarcane juice is an important operation in sugar production process. The control of shock pH is usually done through LPU or control valve in which the variation in Shock pH ranges from (-) 0.5 to (+) 0.5 units. This shows that extra lime is being pumped into the mixed juice which requires extra SO_2 gas for neutralization. Besides this there may be localized alkaline and acidic pockets formation in the juice. In order to improve the same we have tried a machine learning time series model along with Artificial Intelligence to control shock pH within (+/-) 0.1 units. A data-driven modeling based on Vector Auto regression is proposed to predict the shock pH of juice based on input parameters. Various metrics are used for validation of the model. Lime dosing system was also changed from cylinder-controlled gate valve to VFD controlled pump.
展开▼