A Computationally Efficient Technique for Separation Data Balancing

A computationally efficient technique for generating a balanced set of density or size separation data is developed. The technique proposed utilized minimization of a modified form of error squares to generate the optimum estimates of outputs mass yields (splits) and balanced direct mass distributions of feed and outputs. The estimated separation data set satisfies mass balance conditions of the streams mass distributions around the separator. The technique proposed avoids the generation of small negative values for the estimated feed and outputs direct mass distributions. Actual data are used to demonstrate the technique proposed.