Prediction early age compressive strength of OPC-based geopolymers with different alkali activators and seashell powder by gene expression programming

In the present work, compressive strength of geopolymers produced by ordinary Portland cement, seashell powder and different alkali activators was predicted by gene expression programming design code. For constructing the models, 8 input parameters including the concentration of alkali solution, alkali binder to alkali solution weight ratio, alkali activator to cement+seashell powder weight ratio, oven curing temperature, water curing time, alkali solution type, alkali binder type and test trial number were considered. The output parameter was the compressive strength of the OPC+seashell powder based specimens. For collecting the data, 6 main factors each at 4 levels were considered for designing by the Taguchi method. A total of 32 series experiments (with 3 trials in each series) were conducted according to the L32 array proposed by the Taguchi method. According to the input parameters, the constructed models were trained and tested. From the total 96 obtained data, 70 sets were used for training and the remained 26 sets for testing the performance of the models. The results indicate that gene expression programming is a powerful tool for predicting the compressive strength of the geopolymers containing ordinary Portland cement, seashell powder and different alkali activators in the considered range with acceptable accuracy.