Downtime of longwall machinery is expensive. There is the potential for large profits to be made by increasing the availability of longwall equipment.
The maintenance records from two longwall mines were processed using Pareto analysis and scatter log plots. The data from Mine 1 covered the period between 25/4/01 to 24/7/01. The data from Mine 2 covered the period between 1/1/02 to 28/2/02.
Maintenance priorities were established for each mine and standard probability distributions were fitted to the data for the most common failures. Minitab was used to identify the most appropriate distribution for each failure and this was used to provide recommendations on maintenance practises.
It was recommended that a condition based maintenance policy involving increased inspection frequency be mostly employed in both mines. Failure based policy was also suggested for some failures.
All the selected failures followed either a Weibull or lognormal distribution. It was recommended that these distributions be utilised to predict failure of components and produce efficient inspection schedules and supply of spares.
The failures that followed a Weibull distribution all displayed decreasing failure rates and as such it was recommended that repair/replacement procedures be reviewed and preventative maintenance kept to a minimum.
The results and recommendations were similar for both mines. However, there were a number of errors and assumptions associated with the data analysis that limited the validity of the results and recommendations.