Utilizing Nonlinear Autoregressive with Exogenous Input Neural Networks to Evaluate the Thermal Flywheel Effect Along Intake Shafts at Nevada Mines

Abstract

Understanding the climatic conditions in underground mines is necessary for efficient ventilation design, cost savings, and to ensure the health and safety of mine workers. Large volumes of ventilation and climatic data including air volume, barometric pressure, dry bulb temperature, and relative humidity were collected at active underground precious metal mines in Nevada, which allows for the determination of wet bulb temperature and other key parameters. Through the utilization of neural networks, the wet bulb temperature at the bottom of the intake shafts is predicted, while taking into account the “thermal flywheel effect” (TFE). Wet bulb temperature is one of the most important climatic parameters to model and understand because it significantly affects the work conditions and the cooling capacity of the ventilating air. The accurate prediction of the dry bulb and the wet bulb temperatures at the bottom of intake shafts is critical when assessing the climatic conditions in future underground mines and deciding on whether a cooling system is needed to assure adequate working conditions throughout the mine. By utilizing accurate predictions of wet bulb temperatures and other climatic parameters, mine personnel will be safer as reported by Bluhm et al. (2003), and a more accurate ventilation design can be achieved resulting in major cost savings for underground mines.