Solidification optimisation of electroplating sludge

Abstract

A systematic study was conducted for the treatment of electroplating sludge by solidification and stabilisation (S/S) with ordinary Portland cement, lime and magnesium oxide (MgO). Response surface methodology (RSM) and artificial neural networks (ANNs) in combination with central composite design were employed to develop the predictive models for simulation and optimisation of the S/S process. The independent variables were magnesium oxide, electroplating dried sludge, lime and distilled water, while the compressive strength and the concentrations of zinc and chromium in the toxicity characteristic leaching procedure leachate of the solidified waste were the response variables. Both the RSM and ANN models were developed based on the experimental designs. The generalisation and predictive capabilities of RSM and ANN were compared by unseen data (the data set that is not used for model training or validation). Both the RSM and ANN models determined the optimum S/S process. The results show that all independent variables had significant effects on the properties of the S/S products. The optimised method as determined by ANN or RSM can be used with confidence for determining response variables. However, the data predicted by the ANN model are more similar to the experimental results than that of RSM predicted results.