Multiple period mine ventilation and fan selection optimization .

摘要

This thesis presents the results of a research investigation that has studied the application of genetic algorithms (GAs) to optimize single period main ventilation systems and the application of linear mixed integer programming (MIP) to optimize the selection of fans for multiple period auxiliary mine ventilation systems of a mine. Both the GA and MIP were then integrated to optimized and select the fans of a multiple period main ventilation system Although these optimization techniques were developed in the context of mine ventilation networks, they could also be applied to other network flow problems.;Currently available options, in the mine industry, are based on the manual use of a ventilation solver to calculate the expected airflows of a mine ventilation network and then generate a manual fan selection The ventilation solver determines the pressure and airflow distribution of the network based on the configurations of the different ventilation controls of the mine: fans or regulators. Then the ventilation practitioners use these results to make their decisions. The techniques presented in this thesis offer the opportunity to analyze a higher number of potential solutions, and to determine the best one in terms of capital and energy costs. Although the technique does not certify optimally, as far as is currently known there is no available technique that can guarantee optimally for the problem targeted because of its mathematical properties.;The first objective of the project was to use a genetic algorithm integrated with a ventilation solver to evolve a population of potential solutions for the single and multiple period mine ventilation network problem This operation determines the optimum location and the combination of pressure and airflow requirements for the fans that can be installed in the main ventilation network. At the following stage, linear mixed integer programming (MIP) models are used to select the practical optimum fan unit that will provide the required pressure and airflow. This solution approach is then extended to generate multiple period solutions for the main ventilation system The second objective, at the auxiliary fan level was to select the fan units to supply the pressure and airflow requirement to the dead end development tunnels for multiple periods. The same MIP models can be used in this case to generate the fan selection The thesis presents a discussion of the model simulations obtained on the application of these fan optimization and selection techniques to small, medium. and large representative mine ventilation networks. It was concluded that in comparison to previously published research studies, the developed techniques demonstrated improved computational benefits when applied to large representative mine ventilation networks.