by Mike Pemberton
January 25, 2016

Pump users may have a reasonable expectation that a properly installed fixed-speed pump will run reliably at any point on the manufacturer's published head-capacity curve. While this is true for most centrifugal pump applications running at a nominal speed of 900 revolutions per minute (rpm) or less, this is not the case for many running faster, in the 1,200 to 3,600 rpm range. The top predictor of pump system reliability is rotational speed. The second predictor is the distance between the actual flow rate and the best efficiency point (BEP) flow. The faster an impeller turns, the narrower the range that the pump can operate and be reliable.

This relationship is often called the Allowable Operating Range (AOR), defined by the Hydraulic Institute (HI) guidebook Optimizing Pumping Systems: A Guide for Improved Energy Efficiency, Reliability & Profitability as "that range of rates of flow recommended by the pump manufacturer over which the service life of the pump is not seriously compromised." The definition could also say the AOR is a function of rotational speed. The HI guidebook defines pump efficiency as "the ratio of the pump output power to the pump input power; that is, the ratio of the water horsepower to the brake horsepower, expressed in percent." Or, the further the pump operates from its BEP, the less effective the pump is in transferring motor input power to fluid output power. The excess energy imparted into the pump that is not used to move fluid is transformed into vibration, heat and noise—a core issue of pump reliability.

Traditional reliability engineering is a well-established science and an essential practice to keep rotating assets healthy, but if the root cause of unreliability is an inefficient machine, all the techniques and best practices will not completely mitigate the core issue. Inefficiency is inherent in the practice of throttling flow with a control valve. This mechanical approach is the standard for fluid-handling systems across industrial, commercial and residential facilities. Yet, the side effects of operating away from the BEP are higher energy costs and lower reliability, as well as the deleterious effects on the process control loop. Poor valve performance, primarily caused by stiction and backlash from non-optimal sizing, degrades loop performance for increased process variability and can result in manual operation.

Poorly performing loops increase operating costs; therefore, energy, reliability and process control are intertwined. In some respects, resizing the fluid-handling components represents a wormhole to plant sustainability. Unreliability and poor financial performance have been inadvertently designed into the process. A pump and all of its internal components are highly engineered products that, when designed to operate near the BEP, typically provide smooth flow and extended mean time between failures (MTBF).

Sizing Matters

Greenfield plants and brownfield modernizations are predominantly based on the same process designs as past decades. Traditional design methods have been based on sizing the specialized production equipment to meet production goals, then the associated vessels, motors, pumps, pipes and valves are selected to ensure these goals are met.

But the missing link is that they are often not sized in an optimal manner. The pump and motor are oversized via a safety margin, and the control valve is over or undersized based on anticipation of better controllability or first-cost savings.

Pipe modeling software, which offers concurrent design capability, will calculate thousands of dimensional combinations before settling on the best one. These sophisticated modeling tools are growing in use by pipe designers but are still used infrequently. Most industrial design groups use spreadsheets for calculating pipe friction loss. While the calculations are accurate, the tool limitation is that the optimum pump, pipe and valve combination—typically the lowest life-cycle cost—is not selected.

As a result, optimal process control is seldom achieved during plant startup. Further, as control loop performance decays over time, the performance gap grows over the facility's life.

Optimization Research

Some studies claim that around 25 percent of installed industrial pump systems can be cost-justified for study with most offering optimization projects with paybacks from a few months to three years. These studies also show that inefficient designs result in many control loops operating in manual mode in the range of 20 to 40 percent. Knowing how much of a plant's operating cost is due to pump system inefficiency should accelerate the market for pump systems assessments and project implementation. Over the past decade, the number of pump system efficiency studies has grown and the implementation rate, after a certain amount of lag time, has begun to increase.

The U.S. Department of Energy in 2002 published a study titled United States Electric Motor Systems Opportunity Report that estimated that 3,583 plants were included in the top 10 percent of energy-intensive manufacturers by market segment. It showed opportunities for optimization. An updated version of the study, expected sometime in 2017, should answer some important questions about the adoption and use of pump optimization techniques and technologies over the past decade, in addition to many more case study results.

Pages