Understand Electrical Equipment Failures
EP Editorial Staff | March 19, 2018
New technologies and techniques are improving detection of damaging heat and failure-predicting partial discharge.
Most electrical system failures are current- or voltage-driven events that can lead to substantial business losses in plants, including loss of production, significant equipment damage, even personnel injury or death. Two of the most common causes of electrical-equipment failure involve bad connections and deteriorated insulation. Addressing these issues is paramount for personnel tasked with managing and/or maintaining plant equipment. The path to success starts with an understanding of what specifically causes these trouble spots—and how to detect them.
Bad connections can be the result of poor contact pressure (inadequate torquing of bolts, insufficient pressure in compression lugs, poor mechanism pressure), deficient contact surface, or dirt and corrosion. Bad connections fail because of heat produced by the current passing through the highly resistive connection, which is a current event. Thermographic or infrared (IR) surveys have been effectively used for decades to detect abnormal heating of electrical connections. Today’s technology allows equipment-temperature differences of one degree Celsius to be detected.
Insulation, on the other hand, deteriorates with age. Factors such as heat, moisture, and contamination, however, can lead to premature failure. Deteriorated insulation fails because the electric field stresses overcome the insulation properties and, eventually, a fault occurs. This is a voltage event. Deteriorated insulation can be detected by a partial discharge (PD) survey, which detects signals emitted from the partial discharges occurring in deteriorated or stressed insulation.
IR and PD surveys can be partnered and performed online during normal business hours without interrupting a facility’s electrical service. This is a very effective way to detect impending failure.
With regard to PD, keep in mind that signal detection is just the first step in executing an online predictive-testing program. Understanding the causes and how to best address them based on survey results is much more complicated—and may require expert assistance.
IDENTIFYING PD TYPES
Like developments in IR equipment and technology, PD testing equipment and technology for field applications has come a long way in the past 20 years. That said, it’s not just the equipment that makes a PD survey effective. Partial-discharge-signal detection is one of the more complex technologies applied to electrical-distribution assets. It requires expertise and resources to be able to interpret information gathered during a survey. The process can prove challenging for individuals that don’t have experience diagnosing electrical-equipment problems.
Once a PD signal is detected, knowing the exact type of PD will help determine steps for dealing with the issue. The five PD
• surface discharge
• corona discharge
• void discharge
• floating electrode
• particle discharge.
Proper PD classification is essential for conveying risk and urgency for remediation. PD signals should not be classified by magnitude alone. Some discharges, such as a floating electrode, can produce a large magnitude of PD signal but, because of its location, may pose little or no threat to equipment failure. In contrast, a void discharge buried deep within a cable’s insulation may generate a small signal but put the asset at high risk of failure.
Identifying an asset that is emitting a PD signal may be all the information that operating and/or maintenance personnel need. The asset can then be taken offline for an inspection to identify the cause of the PD and determine how to deal with it. Visible evidence of PD can be a deposit of white powder (nitric acid) or green patina (verdigris) of copper. Simply wiping away the powder or cleaning the patina is not a solution. PD will not simply go away. This evidence is the result of PD, not the cause.
If PD is found to be coming from an asset during an initial survey—but more information is needed to determine the cause—advanced techniques and technologies may be required to pinpoint the source. These techniques may include time-of-flight
This technology compares the speed of the radio-frequency signal to the soundwave of the PD source, or may involve triangulating the sound signal by examining how long the sound takes to travel to different sensors.
These advanced technologies can often narrow the PD location to a very small area. Having this level of detail makes it easier to acquire the correct spare parts prior to a scheduled outage. Personnel, in turn, will know exactly what needs to be replaced before even opening equipment, thus minimizing downtime.
APPLYING PD SENSORS
When it comes to PD technology, there are generally two areas of focus: cables and everything else.
For cables, PD testing helps pinpoint defects in insulation, splices, and terminations. The most common method is to place a high-frequency current transformer (HFCT) sensor around the cable and/or drain shield before it is earthed. This sensor is designed to transmit signals in a pre-determined frequency band to the PD-detection
There’s some debate in the industry regarding how far down the cable a sensor can detect PD. The ability to detect it down the length of a cable depends on several variables, including:
size of the PD signal at the source
• amount of background noise
• distance between the PD signal source and the sensor
• type of cable shield
• type of cable insulation
• condition of the shield.
For everything else, PD testing is a valuable, predictive tool. In fact, today’s technology is very effective at detecting PD in the vast majority of medium- and high-voltage electrical-distribution equipment, including:
• insulating boards
• dry-type and liquid-filled transformers
• air-insulated and gas-insulated circuit breakers
• potential, current, and control power transformers
• insulators of all types
• motors and generators.
No matter the location, effectively detecting PD signals requires the right sensors. Determining which sensor to use depends on the type of signal created and its frequency, something that can range from tens of kilohertz to a few gigahertz. The various types include acoustic, ultrasonic, transient-earth voltage (TEV), ultra-high frequency (UHF), and high-frequency current transformer (HFCT). While PD sometimes can be detected with just one or two types of sensors, the best approach is to apply all applicable sensors to a chosen asset and correlate the information that’s gathered.
Although there’s no such thing as good PD in an electrical-distribution system, some equipment may have low levels of it with no signs of damage. In these cases, it’s not the level of a PD signal that should cause concern, but, rather, any changes in it. An initial PD investigation will establish a baseline that can be compared with future testing results.
Annual or semiannual preventive maintenance and testing is recommended to maximize system reliability and availability of electrical systems. But, with so many sites operating around the clock, maintenance schedules can stretch beyond the comfort level of plant managers. That’s why online predictive tools such as IR surveys and PD testing are becoming increasingly important. These techniques can uncover areas of overheating and insulation deterioration—two of the leading causes of equipment failure—sooner than later. Leveraging multiple sensor technologies in these procedures helps ensure accurate data for improved diagnostics. Accurate interpretation of test data will provide operations with actionable information and, ultimately, lead to more-informed risk-management decisions. EP
Steve Park is the manager of Technical Services for Vertiv’s (Columbus, OH, Vertivco.com) Electrical Reliability Services business. His 35+ years of experience with power systems includes 14 years with the U.S. Air Force, during which he served as a high-voltage lineman, electrical-power-distribution engineer, and instructor. He holds Bachelor’s and Master’s degrees from Purdue Univ. and an MBA from Indiana Wesleyan Univ. Contact him directly at Steve.Park@Vertivco.com.