A hybrid optimization technique for improvement of P-recovery in a pellet reactor

Abstract

Emphasis in recent years has been focused on improving processes which lead to enhanced phosphate recovery. This paper studies the precipitation features of calcium phosphate in a fluidized bed reactor in a concentration range between 50 et 4mg/L and establishes the conditions for optimum phosphate removal efficiency. For this purpose, a hybrid optimization technique based on Simulated Annealing (SA) and Quadratic Programming (QP) is used to optimize the efficiency of the pellet reactor. The efficiency is computed by coupling a simple agglomeration model with a combination of elementary systems representing basic ideal flow patterns (perfect mixed flow, plug flow,…). More precisely, the superstructure represents the hydrodynamics conditions in the fluidized bed. The “kinetic” constant is obtained for each combination. The two levels of the resolution procedure are as follows: at the upper level, SA generates different combinations and at the lower level, the set of parameters is identified by a QP method for each combination. The observed results show that a simple combination of ideal flow patterns is involved in the pellet reactor modeling, which seems interesting for a future control.