Avaliação de plano de fogo para desmonte de rochas em uma mina de calcário baseada na previsão da fragmentação pelos métodos de Kuz-Ram e KCO

Abstract

The blasting of rocks using explosives in limestone quarries dedicated to the cement industry holds strategic importance, as it ensures efficient fragmentation of rock masses and sustains the sector’s production chain. This technique enables the processing of large volumes of rock within a short time frame, ensuring optimization and productivity in the unit operations of loading, transportation, and comminution. However, when the blast design is inadequate, several negative consequences may occur, such as the formation of boulders, the presence of toes, and excessive flyrock—factors that compromise operational safety and increase production costs. In this context, it becomes essential to develop studies that allow understanding and controlling the parameters that directly influence the fragmentation of rock masses. This research aims to analyze the granulometric curves generated by the empirical Kuz-Ram and KCO methods in a quarry of the cement sector, comparing them with data processed in the WipFrag software from images of a rock blast. To achieve this goal, a methodology based on granulometric prediction through the mathematical fragmentation models Kuz-Ram and KCO will be employed. The choice of these two models is justified by the fact that Kuz-Ram, although widely used, presents limitations in describing finer particle sizes, whereas KCO, based on the Swebrec equation, proves more efficient in representing the complete granulometric behavior. Through their combined application, it was possible to perform a comparative study between theoretical granulometric curves and the actual curve obtained in the field, allowing the evaluation of model adherence to real conditions, identification of deviations, and assessment of the need for adjustments in the blast design to optimize fragmentation. As a result, it was found that the curve generated by the KCO method was closer to the curve produced by WipFrag than that generated by Kuz-Ram. Consequently, the application of the KCO method calculated a D80 = 56.23 cm, proving more representative compared to the institutional target (D80 = 60 cm), while the Kuz-Ram method predicted D80 = 52.42 cm. It was also observed that the granulometric curve generated by WipFrag did not identify the presence of boulders, highlighting the need to collect more representative images of the blasted pile. Finally, it is evident that parameters of the rock mass and the explosive must be considered for a more effective blast design, predicting different geometric parameter values for dissimilar rock masses.