Prediction of high furnace slag pressure activity by using cultural algorithm-based on artificial neural network

Scientific methods must be applied in order to produce at an optimal level in the industry. However, in some sectors, the tradition of a production model based on the experience of the employees still continues in determining the product quality. This situation causes significant loss of time and quality. These losses have been reduced to a minimum with the application of prediction methods to industry in recent years. In this thesis study, the most suitable model for cement sector was determined and adapted by researching the prediction models used in the literature. In this context, the model in which Artificial Neural Network and Cultural Algorithm work as a hybrid is studied for the first time in this thesis. Real time data of a cement factory were used in the study. These data constitute the inputs of the developed algorithm. As an output, the 28-day pressure activity index value of Ground Blast Furnace Slag was estimated. According to the results obtained, the proposed hybrid model was compared with different estimation models and managed to make predictions with an average absolute percent error value of 3.66% on test data.