presentation of statistical models for predicting the ornamental stone sawability

rock sawability prediction and determination of effective parameters are important in the cost estimation and the planning of the stone factories. an accurate estimation of rock sawability and performance prediction of sawing machine helps to make the planning of the rock sawing projects more efficient. it is necessary to predict the rock sawability and determine the effective parameters to achieve an accurate planning in stone factories. in this paper, the relationships between rock characteristics and production rate were investigated by statistical analysis. for this purpose, 17 different ornamental stones (10 granite rocks and 7 carbonate rocks) in stone factories were studied. in the first step, the rock samples were collected for laboratory tests after proper preparation. then four major characteristics of rock; uniaxial compressive strength, schmiazek f-abrasivity factor, mohs hardness and young’s modulus were determined in the laboratory. in the next step, statistical analyses (such as simple regression and multiple regression methods) were performed on the collected date. since the sawability of ornamental stones is affected by many factors, it can not be analyzed by using only simple regression models. therefore, the analyses were carried out by using multiple regression methods. multiple regression methods can be divided into two types as linear and nonlinear. in this study, the twin-logarithmic model, which is one of the nonlinear methods, was used. variation of production rate was evaluated using the spss software, and five equations were developed for prediction of the production rate. validation of the models was carried out by considering the determination coefficient, t-test, f-test and the plots of observed production versus estimated production (the determination coefficient of all models was higher than 0.96). finally, the best model was proposed for predicting the production rate. the best model obtained by multiple regression analysis showed that the most significant parameters affecting the sawability of ornamental stones are uniaxial compressive strength, schmiazek f-abrasivity factor and young’s modulus. it was concluded that the production rate of ornamental stones can be reliably estimated using the developed model.