Assumptions are often made concerning the linearity, type of variable distribution, and interactions in unit operations which lead to limitations in the way they are controlled. Industrial data and a physical first principles model for a rotary sugar dryer commonly used in the sugar industry are used together with the incorporation of non-symmetrical distribution functions for various variables. The effect of non-normal distribution functions on product quality and throughput can be determined. In particular, the tuning of the previous unit operation to give an appropriate skew input distribution for temperature and moisture to the sugar dryer can result in higher throughputs and better quality.