Brandon Perkins is the product marketing manager for Industrial Data Intelligence at GE Intelligent Platforms. He believes big data, predictive analytics and industrial Internet tools can help mining customers better manage margins and grow profitability.
In complex systems, many factors can affect overall performance. For example, a mining facility was experiencing problems with small system components that caused its pumps to run faster without increasing flow rate. These issues meant increasing costs without greater production. To resolve the issue, a predictive analytic software solution was implemented to help this facility diagnose the problem and regain system efficiency.
The remote monitoring and diagnostic (RM&D) software solution detected that a pump at a South African mining facility was not operating at the speed that was expected based on previous historical data. The software identified that the percentage of time a proportional-integral-derivative (PID) control loop for a pump had exceeded its operating limits was increasing above historic values.
While further analyzing the pump’s operational history, personnel found periods of time when the pump was operating at 100 percent speed when it was expected, given the loading conditions of the pump, to operate at 80 percent speed. During the time that the pump was run at full capacity, a suction level sensor was returning flat line values. The RM&D team investigated this issue and discussed it with the customer during the weekly maintenance call.
When the customer investigated this issue, a level sensor was found to be malfunctioning, causing it to return flat line values. This level indication, which is part of the pump control system, caused the pump to operate at a higher speed than previously required for the same flow rate.
The end result was that the customer had the opportunity to correct a faulty level indication before secondary damage was done to the equipment, thus preventing unnecessary maintenance costs. In addition, running the pump at full speed, when the operation did not require it, caused unnecessary production costs and could have resulted in avoidable wear and tear on the pump.
The same solution detected an increase in the speed of a pump at another mining facility. Given load and operational conditions, pump speed was expected to remain at about 82 percent of full speed operation. Actual pump speed was noted at as high as 90 percent of full speed operation.
At the same time, the estimated pump driver power was increasing from expected values of 120 horsepower (HP) or 89 kilowatts (kW) to values as high as 150 HP or 112 kW.
The RM&D team notified the end user and began tracking the issue on the regularly scheduled weekly maintenance calls.
When the end user completed their investigation, they confirmed that the liners in the pump had degraded and needed to be replaced. Facility personnel had the opportunity to proactively schedule a maintenance action on this pump at a time that minimized maintenance and lost production costs. Continuing to run this pump with degraded performance caused the customer to incur excess production costs.