Electrical insulation is designed to keep current from flowing where it shouldn’t. Insulation integrity is therefore critical and is assessed via a battery of tests. Many things can compromise the results of these tests, including temperature, humidity, erroneous test procedures, etc. Previously, I wrote about the effects of humidity on insulation integrity tests. This time I focus on erroneous test procedures, specifically using additional insulation to aid in setting up the test.
I was recently involved in completing routine testing on a 1971 121 kV GE oil blast circuit breaker (OCB). Utilities are well on their way to replacing these aging hulks, but I see it as a privilege to be involved with transmission OCBs, particularly since they and their oil-impregnated lead cable counterparts seem to have longevities that best their more modern counterparts — but I digress. We will see if SF6 units can last 50-plus years of fault interruptions on minimal maintenance.
Prior to testing, proper de-energization procedures were followed. The breaker was opened, air disconnects on either side had been opened and locked out/tagged out (LOTO), and two separate sets of personal protective grounding (PPG) conductors were installed on the intermediate buswork between the disconnects and the OCB bushings (Figure 1). Finally, the intermediate buswork, composed of hollow pipe conductors, was unbolted from each OCB bushing.
Unfortunately, the buswork proved to be rigid and uncooperative, so to maintain the gap between the buswork and OCB bushings upon disconnection for testing purposes, discarded rubber blanket material was wedged between the buswork and the bushings (Figure 2). Note that serviceable PPE blankets intended for personal safety are not used for this purpose. This isolates the breaker components under test from the grounded buswork, which wasn’t on the testing menu. However, there are holes in this claim.
THEORY REVISITED
As I’ve shared before, power factor/capacitance testing involves applying an AC voltage across insulation, calculating capacitance, and comparing the resultant resistive (IR) and capacitive (IC) currents that flow through the insulation. An insulator’s power factor is the ratio of resistive current flowing through it to total current flowing through it when energized, or IR/IT, where total current IT = √(IR2 + IC2). In perfect insulation under electric stress, IR = 0 and IT = IC. Unfortunately, no insulator is perfect, so some IR always exists upon energization. But the lower the power factor, the better the insulator.
Closely associated with power factor is capacitance C. To reiterate, all real-life insulation systems can be conceived of as capacitors, where capacitance C = ε(A/d) and refers to the ability of a material to store a charge in an electric field. When we view insulation systems as capacitors, ε is the relative permittivity (or dielectric constant) of the insulator material, A is the surface area of the conductors opposite the insulator, and d is the distance between the conductors within which the insulator exists.
Several insulating media stand between the high voltage that exists from an OCB bushing terminal to ground, but the two major insulators are oil and porcelain. These dielectrics prevent appreciable current in the normally current-carrying elements of the breaker such as the terminals and the contacts from reaching the grounded frame. The resistive and capacitive currents that flow through the oil and porcelain comprise the power factor that is calculated during testing.
By wedging insulating blanket material between the ungrounded terminals under test and the grounded intermediate buswork in an attempt to separate the two, a third major insulator is introduced into the circuit. Now all power factor tests between a conductor and ground yield results that will include — and therefore be affected by — resistive and capacitive currents flowing through the blanket material. But we don’t care about the power factor of the blanket material; they get their own overvoltage testing. We only care about the power factor of the oil breaker. Hence, the conundrum.
Figure 3 illustrates schematically what happens when the blanket material is added as an additional dielectric in the power factor test circuit.
Without the influence of the blanket material, the total current IT = √((IRO + IRP )2 + (ICO + ICP )2 ), where IRO and IRP represent the respective resistive currents through the oil and porcelain, and ICO and ICP represent the respective capacitive currents through the oil and porcelain. When a parallel blanket-material circuit is added, resistive current through the blanket material IRB and capacitive current through the blanket material ICB emerge. Now, the total current IT = √((IRO + IRP + IRB)2 +
(ICO + ICP + ICB)2). ICB is a function of the capacitive reactance XCB of the blanket material and the test voltage applied VT, where VT/ XCB = ICB. The test voltage is governed by the technician but the capacitive reactance XCB is itself a function of test voltage frequency f and the actual capacitance of the insulator/dielectric under test C, where XCB = 1/(2πfC). And if you recall, C = ε(A/d).
The electric properties of insulating blanket material are such that they can adversely affect overall OCB power factor. One, since insulating blanket material is a dedicated insulator, IRB will be negligible. Two, insulating blanket material possesses properties that can generate considerable capacitive current ICB. They are made of various rubbers, e.g., latex, EPDM, etc. EPDM rubber has a relative permittivity ε of 3.3. This would raise the capacitance compared to a capacitive circuit where the dielectric was simply air.
The surface area A of an OCB terminal and a grounded conductor squeezing the blanket material is admittedly constrained. However, the thickness of the insulating blanket material is minimal, less than an inch, thereby keeping the distance d of the capacitive circuit low (folding or adding blanket material will increase the distance between conductors).
A high permittivity ε, a largesurface area A, and a small distance d will increase capacitance C. As capacitance C increases, capacitive reactance XCB decreases. As capacitive reactance XCB decreases, capacitive current ICB increases. As capacitive current ICB increases, total current IT increases. As total current IT increases, power factor IR/IT decreases. Using insulating materials to isolate the equipment under test can reduce the overall power factor and yield results that are lower/better than actual for the OCB insulation system.
LESSONS LEARNED
I don’t fault the technicians involved in this oil breaker testing. They probably couldn’t have removed the intermediate buswork even if they wanted to, so adding insulating blanket material was the best of an imperfect scenario. The field is emphatically unlike the lab, where all variables can be scrupulously controlled. We make do with what we have.
However, if variables can be controlled, they should be. While we cannot recreate lab conditions, we try to approximate them as much as humanly possible. This means controlling for things that may corrupt test data such as temperature, humidity, pressure, and ill-advised testing procedures.
Michael Labeitis a Prime Power Production Specialist, Lineman, Substation Test Technician for Service Electric, and a NETA Level 3 Technician in the 249th Engineer Battalion, U.S. Army Corps of Engineers. He has operated and maintained medium-voltage power plants in Turkey and Saudi Arabia as well as at Ft. Leonard Wood, Missouri, and Ft. Liberty, North Carolina. His team won 1st place in the military division at the 2021 International Lineman’s Rodeo in Kansas City. Labeit graduated from Prime Power School in 2018 and has an AAS from Excelsior College.