Discharging contaminated process water into the environment is a major event that has caused mines to be shut down. December is the rainy season in the DRC and unexpected events can occur.
One of our client's water circuits have been designed in such a way that 90% of the catchment of process water on site either gets utilised back into the circuit or ends up in the process water pond, PWP for short. PWP in this instance is a medium sized pond that has a high risk of discharging into the environment if not managed correctly. PWP can receive a large, flash inflow as the catchment area of all the other process ponds, the catchment from the massive heaps and the nearby hydromining all collect here.
To prevent the PWP from overflowing, water is pumped to the nearby much larger tailings facility (TSF3). During a 100 year rain event (defined as 240 mm of rainfall in 3 days) this pump needs to be able to pump 250 m3/hr to ensure that the level of the PWP does not increase. It should also be noted that no flow meter exists on this line, nor is there any way of remotely monitoring the operation of this pump to ensure the modelled pond level matches the actual pond level.
Modelling the water circuit
WIRE (MMS's software solution) keeps track of two levels, namely the calculated level and the last available physical measurement. Every time a level reading is entered, the model is updated and re-baselined. The calculations involved in estimating a level are complex and take into account flows which are measured, or flows that have been estimated, calibrated rainfall and catchment areas, pond evaporation, drainage and any other forced evaporation that may be taking place. Given the amount of unmeasured variance, the calculated level becomes less accurate the further out one forecasts. A way to train ones model is to monitor the accuracy of the calculated level which allows one to see discrepancies or unknown flows within the system and make adjustments.
The build up
During the 2017 rainy season a pump was placed with a supposed capacity to pump 250 m3/hr to the nearby TSF3. In 2017’s rainy season it was noticed that the pump only effectively managed 100 m3/hr, which caused a near overflow, before the pump was replaced by one that could achieve the required 250 m3/hr.
One of the reasons that the issue became critical in 2017 was that the level sensor on the process water pond was poorly calibrated and often overestimated the level, frustrating staff with false alarms when they could observe that the pond was not overflowing. One of the actions made for 2018’s rainy season was then to rely more heavily on the daily pond level surveys performed by the mine's staff, instead of relying on a potentially faulty sensor to take action.
In 2018 it was known that closely monitoring the PWP was crucial to mitigating the water balance event risk of the mine as it is a bottleneck. As one needs to expect the unexpected, and bad things to happen at the most inconvenient times, PWP was getting fuller without anyone noticing.
During December 2018 the DRC was going through an election which had been delayed for two years. Many expats were offsite and the mine was on skeleton staff as local workers went home to vote. Nevertheless, a mine can’t cease production; and so more water was entering the circuit (hydro mining, tank leach etc).
Based on initial forecasts for this wet season, the mine had been instructed to run the pump from PWP to TSF3 at a minimum of 140 m3/hr continuously for the next two months to keep the PWP in check. The mine's staff had received the instruction, but instead of running the pump at 140 m3/hr they had decided to be prudent and run at full capacity (which we have estimated to be 200 m3/hr rather than the nameplate 250 m3/hr).
One of the problems with running a pump at its maximum capacity is that it is much more likely trip and that's exactly what started to happen here. As we all know, stopping and starting lowers the average speed drastically, which is what happened to this pump and thus it was effectively only pumping less than 100 m3/hr, which caused the pond level to continue to rise, with little oversight.
At this time there was no significant rainfall events, just the standard 12-20 mm per day as expected during that time of year which meant that people were also not expecting the PWP to be near capacity. To make things worse, the surveyors where taking advantage of the lack of management presence onsite and had not been as diligent in updating the surveys, which meant trouble was brewing.
MMS personnel onsite were taking leave a few hours from the mine, but thankfully had kept pestering the surveyors to provide the PWP level. As soon as our team member arrived back onsite on the 29 of December he noted that the WIRE water balance was in the danger zone and went directly to the PWP to check if the pond was indeed at capacity. Unfortunately it was and the calculated level matched reality at a fully loaded at 99.7%. Something urgent needed to be done.
In the nick of time mine personnel were notified, the pump was fixed and prevented from continually tripping by adjusting the operating level down to 80% of capacity, (which was estimated to be running at the minimum of 140 m3/hr required).
On the 30th the mine had a partial shutdown so that the staff could go vote, which played into the mines required remedial action, as no more water from hydromining was added to the circuit and for a short period of time on the 1st the mine was advised to temporarily stop hydromining to allow the PWP level to be dropped to an acceptable level, given the risk of the pump tripping again.
Now the big question is how did an uncontrolled discharge get so close to happening? As always issues occur when multiple causes stack up. In this case we have elections, surveyors and faulty pumps and sensors to blame. There will always be multiple issues on a site, and having MMS and WIRE involved in your operation has proved itself valuable once again.
Please contact email@example.com to understand more about our water balance services which we can offer to your operation.