AN EXPERT SYSTEM ON ASSESSING ROCKBURST RISKS FOR SOUTH AFRICAN DEEP GOLD MINES

Expert systems are methods that can cope with rock engineer's expertise effectively.An expert system on assessing rockburst risks for South African deep gold mines is described.Expertise was represented by rules, frames and mathematical models integratively. About 950 rules was built in the knowledge base, which can be learned by the suggested learning algorithm. A new uncertain reasoning algorithm was proposed. According to the features such as depth below surface, energy release rate (ERR), excess shear stress (ESS), geological structure, face angle with structure, distance from structure, extent of mining, mining type, width of bracket pillar, stope width, local support, regional support and gully support, rockburst risk is assessed either "Low", "Moderate", "High" or "Severe". A computer program system was developed by Turbo Prolog. The results of testing cases show its applicability of the system.