Abstract
The need to meet global mineral demand by using underground loaders to improve productivity, has significantly increased the risk of collision accidents involving loaders and miners. A platinum mine in South Africa deployed a radio-frequency identification collision avoidance system on its underground fleet to reduce this risk. Productivity decreased by 13.28% per month after the deployment. This research study focusses on identifying design constraints and environmental elements that affect the operational efficiency of a collision avoidance system for underground loaders and to evaluate the system’s ranging and detection accuracy.
To assess and evaluate the system, 20 field trials on surface and 30 underground trials were performed. The trials consisted static, in which a mobile miner with a tag affixed to the cap lamp approached a stationary loader, and dynamic trials, where both the miner and the loader were mobile and moving towards each other. The system was configured to give a caution alarm when the miner was between 9 and 20 metres, with a permitted tolerance of ±0.5 metres, to crawl between 5 and 9 (±0.3) metres and to stop when the miner is within a 5-metre range with a tolerance of ±0.2 metres. The actual alarm activation distances were measured and compared to the configured distances and commands. The data was computed to give the level of ranging accuracy which was examined to determine its effect on productivity. Underground ranging accuracy was 96.04% against 90.88% for surface trials. Caution mode had 97.82%, crawl had 91.04% and stop mode scored 91.52% against 97.5%, 96.67% and 96% target accuracies, respectively. Rear direction was safest amongst all orientations with a ranging accuracy of 96.92%, front-left direction was the most dangerous with an accuracy of 89.32%.
The loader's bucket and other metallic parts that were in line of sight weakened the signal, resulting in inaccurate distance estimation which fell outside the permitted tolerances. Because of this, the system also could not detect a tag approaching the loader from the front direction, during surface trials. Signal bouncing on the hanging and side walls improved underground detection and ranging accuracy. Metallic objects and the system’s low ranging accuracy resulted
iv
in low distance estimation accuracy. The ranging accuracies for crawl and stop modes which were below target accuracies resulted in early crawl or stop alarm, causing the loader to reduce its speed or stop earlier than it should. This prolongs the travelling time from the loading zone to the tippler, effectively, reducing the number of loads per shift. The evaluated system utilises Return Time of Flight technology for distance estimation. Replacing it with Received Signal Strength technology would improve ranging and detection accuracy. Received Signal Strength technology due to its magnetic distribution in the presence of metallic objects, is less susceptible to signal interference from metallic objects compared to Return Time of Flight technology.
Keywords: Radio frequency identification; tag; collision avoidance system; underground mine; loader.