A Single Point of Centrifugal Truth: Closing the Gap Between Operations and Maintenance Through Real-Time Performance Data

Timely Data Can End the Blame Game

What is needed is a way to put operations and maintenance back on the same team: a neutral source of timely data that the two departments can use together, to solve and prevent centrifugal pump breakdowns. With the same real-time data available to everyone, the patterns behind recurring breakdowns can be identified, and there would be no need to search for a human source of error.

High-Frequency Current and Voltage Data

A data-driven tool, such as a system to remotely monitor the health of centrifugal pumps and other rotating equipment in real time, can be beneficial. A system such as SAM4 can use learning algorithms to analyze high-frequency current and voltage data in order to accurately predict upcoming machine failures. Simple metrics like running time, energy use, and starts/stops are part of the main dashboard for every asset the system monitors. The use of current and voltage data (rather than vibration, temperature, or other physical signals) enables the system to provide complex metrics that help companies save energy, improve performance, and extend machine life. One of these metrics is a real-time pump performance curve.

A Real-Time Pump Performance Curve

The pump performance curve is based on the affinity laws, which relate a pump’s shaft speed and power consumption to its head and flow. Shaft speed and power consumption, in turn, are directly related to the instantaneous current and voltage drawn by the pump’s motor. As a result, the system can automatically plot a pump’s operation relative to its best efficiency point in real time, using the data it is already collecting to monitor for developing damage. Figure 1 shows how a pump’s real-time performance is plotted against its designed operation.

The calculations used to plot this curve are brand-independent; they can be made for any pump, as long as the pump owner provides the motor’s rated efficiency and RPM, and the suitable pump and power reference curves.

The Real-Time Pump Curve in Action

Figure 2 shows an eight-hour view of the performance data  for a centrifugal pump that moves oil from a storage tank to a tank truck. At the top left of the graph, the pump’s startup time, each of its three loading sessions, and the pump’s operation between sessions are clearly depicted.

In the graphs depicted on the right side of Figure 2, the gathered information is plotted on the pump curve, with the best efficiency point (BEP) marked by the black triangle. Each cluster of red dots reflects the pump’s actual operation in one of those three states: startup, loading, and waiting. Through these graphs, the pump owner can see precisely when the pump is operating well, and when it is operating poorly. This data instantly communicates two valuable pieces of information: the pump’s very off-BEP operation (top graph), and its power consumption (bottom graph).

From this data, both the maintenance and operations departments can make informed decisions on how to improve the pump’s efficiency. Looking at this pump’s operating points, it becomes clear that it is too large for its chosen task. If it continues to operate at this speed, seals and bearings will begin to break down more often and the pump will ultimately wear out faster. A solution might be to install a smaller pump. Alternatively, the operations department could add in a VFD to maintain greater pump capacity during different operational requirements. Either way, maintenance and operations are able to use hard data to understand the true source of this pump’s problems, and to find the best solution together.