Optimización del proceso de granulación húmeda para la obtención de granulados de lactosa monohidratada utilizando métodos de programación no lineal

A wet granulation optimization was done in this work to obtain monohydrated lactose granulates using a two phasescheme. The response surfaces were built in the first phase. In the second phase non linear optimization algorithmswere used. For the construction of the response surfaces a central compose two factors experimental design wasapplied in order to determine the effect of polyvinyl pyrrolidone (PVP) concentration and the solid -liquid mixing timeover flow properties, particle size, friability and the percentage of fine powder on the obtained granulates. A multi-ple regression analysis was applied to the experimental results to obtain the mathematical equations that representeach one of the dependent variables studied. Graphic representations of the response surfaces and contour curveswere obtained. The optimization of the wet granulation process was done by establishing the optimal values range using five diffe-rent methods of nonlinear programming (gradient, uniform, golden section, bisection and Hooke and Jeeves). TheHooke and Jeeves method was the most promising due to its effectivity for obtaining an optimal value range for eachgranulates properties studied. An optimal result for a value for each independent variable within this optimal rangewas verified experimentally. The optimal values were 0.917% (dried base) polyvinyl pirrolidone concentration and asolid-liquid mixing time of 7 minutes.