Generation Of Random Networks For Evaluation Of Mine Ventilation Network Methods

The simulation of randomness in a system of components is important in certain types of systems analyses. This paper presents procedures for generating ventilation networks that are random in many of their characteristics and which can be used to study the performance characteristics of various mine ventilation network analysis procedures without introducing undue bias as a result of the nature of the network. Among the parameters of a network that affect the efficiency of a network analysis procedure are: (1) the interconnection pattern of airways, (2) the distribution of resistances in the network, (3) the number and location of fans, (4) the fan characteristic curve, (5) the size of the network (number of branches), and (6) the density of the network (number of nodes compared with the number of branches). In the procedures proposed here, a single fan with a fixed characteristic is used to simplify the problem and eliminate the fan as a source of variation. The resistances of the airways, the pattern of interconnection of the airways, and the fan location are randomly chosen in an attempt to eliminate the bias in the performance characteristics of the network analysis methods caused by these parameters. A method of generating biconnected networks while keeping the "randomness" of these other parameters is suggested. At the same time, the size and the density of networks are controlled to investigate their impacts on the performances of the network analysis procedures. The combination of these two variables has been termed the "level of complexity" of the network in this paper. Finally, a summary and conclusions section outlines some characteristics of the networks generated, using the random generation procedures outlined in the paper.