Fuzzy Decision Support System for the Calibration of Laboratory-Scale Mill Press Parameters

Abstract

Chemical analyses of iron ore samples are crucial to identifying the characteristics of the ore being mined and processed by mining industries. Moreover, reliable ore content results are fundamental to optimizing the mining plant operation and correctly pricing the final product. An industrial laboratory usually performs hundreds of analyses per day and requires a certain amount of equipment to increase the laboratory capacity. However, the parallel chemical analyses approach demands proper equipment calibration to avoid result variability due to the use of different equipment. This paper proposes a method to help laboratory operators select the correct mill press equipment parameters during calibration. First, we apply the design of experiments (DoE) technique to determine the calibration parameter that affects the final content result. Then, a decision support system (DSS) is developed based on the fuzzy logic approach to adjust the equipment parameters. Results with real data show that the proposed system reduces the standard deviation of the concentrate grade of iron (%Fe) by up to 65.01% and that of the silica concentrate grade (%SiO2) by up to 60.81%. INDEX TERMS Iron content, design of experiments, fuzzy logic, laboratory, chemical analysis.