Applying PQ Measurements to Predictive Maintenance

Dec 1, 2007 12:00 PM, By Wade Thompson, Fluke Corp.

You may already be using PM techniques on your motors and drives, but what about the power to your equipment?

When it comes to predictive maintenance (PdM) on electrical equipment, the numbers don't lie. According to insurance claims data in NFPA 70B-2006, “Recommended Practice for Electrical Equipment Maintenance,” roughly half of the cost associated with electrical failures could be prevented by regular maintenance. Another study published in IEEE 493-1997, “Recommended Practice for the Design of Reliable Industrial and Commercial Power Systems” (The Gold Book), cites a similar scenario, stating that a poorly maintained system can attribute 49% of its failures to lack of maintenance.

To determine the actual cost of a failure, it helps to consider three key categories:

Lost income (gross margin) due to downtime;
Cost of labor to troubleshoot, patch, clean up, repair and restart; and
Cost of damaged equipment and materials, including repairs, replacements, and scrapped material.

Your PdM inspection route probably already includes any motors, generators, pumps, A/C units, fans, gearboxes, or chillers onsite. By adding basic power quality measurements to production equipment maintenance procedures, you can head off unexpected failures in both production equipment and your power system. Unlike a comprehensive electrical system survey, power quality predictive maintenance (PQ PdM) focuses on a small set of measurements (click here to see Table) that can predict power distribution or critical load failures. By checking the power quality at critical loads, you see the effect of the electrical system up to the load.

Voltage stability, harmonic distortion, and unbalance are good indicators of load and distribution system health and can be taken and recorded quickly with little incremental labor. Current measurements can identify changes in the way the load is drawing. All of these measurements can be taken without halting operations and generate numbers that can easily be entered into maintenance software and plotted over time. For each measurement point or piece of equipment, determine what limit should trigger corrective action. Limits should be set well below the point of failure. As time goes on, limits may be “tightened” or “loosened” by analyzing historical data. The appropriate limits depend somewhat on the ability of your loads to deal with power variation. But for most equipment, your maintenance team can devise a set of default “house limits” based on industry standards and experience.

Fig. 1. Recording all phase voltages and neutral-to-ground voltage is a good start for a PQ analysis of critical equipment and overall distribution system health.

The cost of 3-phase power analyzers is lower now than ever, and it should only take roughly 15 minutes to take the readings discussed in this article. (Storing voltage sag data will add more time because it requires picking up the data after a day or so.)

Voltage

Good voltage level and stability are fundamental requirements for reliable equipment operation. Running loads at overly high or low voltages causes reliability problems and failures. Here are a few rules of thumb to remember.

Verify that line voltage is within 10% of the nameplate rating (Fig. 1). As connections in your system deteriorate, the rising impedance will cause “insulation resistance drops” in voltage.
Added loads, especially those with high inrush, will also cause voltage decline over time. The loads farthest from the service entrance or transformer will show the lowest voltage.
Neutral-to-ground (N-G) voltage tells you how heavily your system is loaded and helps you track harmonic current. N-G voltage higher than 3% should trigger further investigation.

Voltage sag count

Taking a single voltage reading tells only part of the story. How is the voltage changing during an hour? What about during a day? Sags, swells, and transients are short-term variations in voltage. The voltage sag (or dip) is the most common and troublesome variety. Sags indicate that a system is having trouble responding to load requirements, and significant sags can interrupt production.

Voltage sags can cause spurious resets on electronic equipment, such as computers or controllers, and a sag on one phase can cause the other two to overcompensate, potentially tripping the circuit.

Sags have several dimensions: depth, duration, and time of day. Utilities use a special index to track the number of sags that occur over a period of time. To gauge the depth of the sags, utilities count how often voltage drops below various thresholds. The longer and larger the voltage variations, the more likely equipment is to malfunction. For example, the Information Technology Industry Council (ITIC) curve specifies that 120V computer equipment should be able to run as long as voltage does not drop below 96V for more than 10 seconds or below 84V for more than 0.5 seconds.