Electrical Acceptance Testing

NOTE This standard governs the field acceptance test program for new electrical power distribution equipment rated 1000 V and below through 38 kV: the testing organization and its technicians, the instruments, the equipment and tests within the program, the sequence and the hold points, the acceptance criteria, the resolution of deficiencies, and the test report. (1.1)

NOTE Factory and production tests prove that the equipment left the factory in good order. They say nothing about damage in transit, errors in field assembly, contamination during construction, a bolted connection left loose, a shipping brace left in, or a trip unit set to values other than the study recommends. Acceptance testing is the independent check between a finished installation and the moment it is placed in service, and most of its tests cannot be performed, or cannot be performed safely, once the equipment is energized. (1.2)

NOTE The equipment standards and this standard divide the work in one way throughout the corpus. The equipment standard for a transformer, a switchboard, a motor control center, or a transfer switch names the field tests that equipment receives, in a checkbox its project selects from. This standard governs who performs those tests, what instruments and safety practices they use, the criteria a result is judged against, what happens when a result fails, and the report that records it all. A test an equipment standard requires is performed under this standard; a criterion this standard states applies to every equipment standard's test of that kind. (1.3)

NOTE The following are governed elsewhere and are outside this standard: (1.4)

NOTE The test plan is the document that settles scope disputes before they start. Equipment that is not on it tends not to be tested, and a panelboard, a transfer switch, or a drive left off the plan is energized without a test and discovered only when the report is indexed against the one-line diagram. (3.1.3)

NOTE An independent firm has no stake in whether the installation passes, and NETA accreditation adds a published audit of the firm's technician certification, its instrument calibration program, and its procedures, which is what an Owner's insurer or an Authority Having Jurisdiction asks for when the report is questioned; an independent firm without accreditation offers the same independence on the firm's own representation of its program; the manufacturer's field service organization knows its own product and its trip units best and is already on site where the equipment standard requires a start-up visit, and its report carries the maker's name rather than a third party's; the installing contractor's technicians cost least and work on the installer's schedule, with the report reading as self-certification; an Owner's in-house group is the arrangement of a utility, a campus, or an industrial owner that maintains its own testing staff and instruments. The program this standard describes is written around ANSI/NETA ATS, which assumes an independent tester, and that is the arrangement most projects that invoke this standard specify. (4.1.2)

NOTE ANSI/NETA ETT defines four levels. A Level II technician performs tests under supervision; a Level III technician performs and evaluates the full range of acceptance tests and directs lower-level technicians, which is the level ANSI/NETA ATS itself requires of a crew lead and the level most projects adopt without discussion; a Level IV technician plans programs and resolves anomalous results, which a project asks for where the program includes medium-voltage relaying, differential schemes, or a transfer scheme whose test the crew lead has to design on site. Level II as a crew lead is the arrangement a small program of low-voltage equipment is sometimes let under, with evaluation of the results reserved to a senior technician off site. (4.2.2)

NOTE The acceptance criteria of this standard are tight tolerances on absolute measurements. A micro-ohmmeter reading 20% high fails a sound bus joint; a test set reading 5% low passes a relay that trips late. The instrument is the only thing standing between a number that looks authoritative and a number that is right, which is why the calibration record is part of the report. (4.3.4)

NOTE The equipment classes are selected here so that the testing organization can price and plan the whole program from one document, and so that a class left out is left out on purpose. The classes present on a project are read from the one-line diagram; which of them are tested is the project's decision, and the equipment that is tested in nearly every program is the service and distribution equipment, the transformers, the medium-voltage cable, the protective devices and their relays, and the grounding electrode system, with panelboards, molded-case breakers, and motors the classes most often tested on a sample or excluded. (5.1.2)

NOTE An optional test is one whose result adds information the required tests do not give, at a cost in time and in instruments. A dielectric withstand test of a switchgear bus finds a crack in an insulator or a foreign object left on the bus that an insulation resistance test reads through, at the cost of a high-voltage set on site and a stress the bus sees once; a partial discharge survey finds a void or a contaminated termination before it tracks, and matters where the equipment is medium-voltage and the terminations were made in the field; insulation power factor on a dry-type winding grades contamination that resistance alone does not; excitation current, core ground, and sweep frequency response detect a shorted turn, an unintended core ground, and winding movement in shipment, and are the tests a large or critical transformer receives; timing and travel on a low-voltage power breaker finds a slow mechanism before it fails to clear; excitation and burden tests confirm a current transformer will saturate where its relay expects it to; a watts-loss test finds a degraded arrester block. (5.2.2)

NOTE A single mobilization tests everything in one visit at the lowest cost per test and holds every energization until the last item is installed; phased mobilizations follow the construction sequence so that a service, a substation, or a wing is tested and energized while the rest is still being installed, at the cost of repeated mobilizations and a test plan that tracks which items have been released. (5.3.3)

NOTE The Engineer of Record is the author of the studies the tests verify against and the party most projects have review the field data before energization; a commissioning authority holds the release where the project runs a commissioning program under CommissioningTotal Building CommissioningResolves to the current adopted revision.sync/commissioning and the Owner wants one party to control every system release; an Owner's representative holds it where the Owner maintains its own electrical staff and the review is performed in house. (5.4.2)

NOTE A hold point that is not contractual is not enforced. Without a written release tied to the field data, a system is energized the day the utility is available and the de-energized tests that remain are deferred, then waived. The release record in the closeout submittals is what proves each item was tested before it was placed in service. (5.4.7)

NOTE Acceptance testing proves the equipment is electrically sound and set as the studies direct; functional performance testing proves the energized system does what the sequence of operations says. The two are sequential, and a transfer switch timing test, a generator load test, or a ground-fault performance test belongs to whichever program the test plan assigns it to, as long as it belongs to exactly one. (5.5.3)

NOTE The manufacturer's published value is the one the equipment was designed and factory-tested to, and the ANSI/NETA ATS tables state their values for use where the manufacturer publishes none, so a project that cites ATS and says nothing else is already under this order; the ATS table value as the governing criterion gives one published number for every make on the project, which is what a reviewer comparing results across equipment wants, at the cost of failing a product whose maker allows more; the more stringent of the two gives the tightest acceptance and the most retests. (6.1.2)

NOTE Insulation resistance readings are compared to a minimum, and the minimum is set by the test voltage, so a test at a voltage below the one the table assigns yields a number that cannot be judged against the table. The temperature correction matters because resistance halves for roughly every 10 °C of winding temperature, and a transformer tested warm from a dry-out reads a fraction of what the same winding reads at 20 °C. (6.2.4)

NOTE ANSI/NETA ATS admits three ways to prove a bolted joint. A torque check with a calibrated wrench, to the manufacturer's value or to the ANSI/NETA ATS bolt-torque table where the manufacturer publishes none, is the method most programs use, because it is done at the same time as the visual inspection with no instrument beyond the wrench; a low-resistance reading across each joint measures the joint rather than the bolt and finds a contaminated or mis-seated contact surface that is torqued correctly, at the cost of a micro-ohmmeter reading per joint; a thermographic survey proves the joint under load after energization, finds the joint that carries current badly for whatever reason, and leaves nothing checked before the equipment is energized. (6.3.2)

NOTE Contact resistance is the most sensitive indicator of a connection that will fail under load. A loose bolt, an oxidized contact face, a conductor not fully seated in its lug, and a vacuum interrupter with a damaged contact all read as elevated resistance long before they heat visibly, and a pole that reads half again what its neighbors read on the same breaker is the one that will overheat at full load. (6.3.7)

NOTE The testing organization applying the settings puts the party with the test set and the study in hand at the trip unit, and makes the setting and its verification one visit; the installing contractor applying them is the lowest-cost arrangement, with the risk that a setting is transcribed from the wrong document and the verification visit becomes a correction visit; the manufacturer's field service organization applies them where the equipment standard requires a start-up visit and the trip units need the maker's software to program, as with a networked trip unit or a transfer scheme, and the testing organization then verifies what the manufacturer left. (6.4.3)

NOTE A relay that tests perfectly against the wrong settings is a coordination failure waiting for its first fault. The study is the only authority for a setting, which is why the reviewed study is an informational submittal to the testing organization and why the as-left settings record in the closeout submittals is cross-referenced to it line by line. (6.4.7)

NOTE Primary current injection passes test current through the breaker poles and proves the current sensors, the wiring between them and the trip unit, the trip unit, and the mechanism as one path, and needs a high-current test set, a de-energized bus, and time per breaker that grows with the frame size; secondary injection feeds a signal into the trip unit's test port and proves the trip unit's logic and settings and the trip actuator, in minutes per breaker with a portable kit, and proves nothing about the sensors or the wiring to them; secondary injection with a primary injection check of each sensor proves the trip unit by the fast method and the sensor path by one primary injection at a single point per pole, which is the arrangement that catches a swapped sensor lead or a sensor of the wrong ratio without the full primary injection of every function. (7.1.2)

NOTE Trip testing every molded-case breaker proves every device at the cost of a primary injection per breaker, which on a project with hundreds of branch breakers is the largest single item in the testing bid; limiting the trip test to electronic and adjustable trip units tests the devices whose settings can be wrong and leaves the fixed thermal-magnetic devices, which have no settings to verify, to inspection; a frame size threshold tests the feeder and main devices whose failure to clear would be consequential and leaves the small branch devices untested; a sample of each frame size and type proves the product family rather than each device, and expands to the whole family where a sample fails; inspection alone is the program a project adopts where the molded-case breakers are small, factory-tested, and protected by a tested device upstream. (7.2.2)

NOTE A breaker that carries normal current correctly and opens slowly is a breaker that clears a fault late, and the timing test is the only acceptance test that sees the mechanism rather than the contacts. On a medium-voltage breaker it is a required test; on a low-voltage power breaker it is the optional test selected under this standard where the breaker is the main of a selectively coordinated system. (7.3.4)

NOTE A functional trip check proves that the relay's output contact trips its breaker and nothing more, and is the scope of a project whose relays were bench-tested and settings-verified off site; setting verification injects each element at its applied pickup and at one or two multiples of it, proves that the relay picks up and times out where the study says, and is the scope ANSI/NETA ATS requires and the one most projects adopt; full characteristic verification injects enough points on each element to plot its curve against the published characteristic, proves the relay rather than its settings, and is the scope a utility or an industrial owner specifies for a differential, a distance, or a generator protection relay whose misoperation would be costly. (8.1.2)

NOTE Transformer acceptance testing confirms three things: that the windings are electrically continuous and connected as the nameplate says, which the turns ratio, the winding resistance, and the phase relationship prove; that the insulation is dry and clean, which the insulation resistance and the power factor prove; and that the unit was not damaged in shipment, which the sum of those tests and the liquid sample prove. The test methods are those of IEEE C57.12.90 for a liquid-immersed unit and IEEE C57.12.91 for a dry-type unit, applied in the field with portable instruments. (9.7)

NOTE A dielectric withstand test stresses the bus insulation at a voltage above operating voltage and below the factory test voltage, so that a sound bus is not damaged and a cracked insulator, a foreign object, or an insufficient clearance left by field assembly breaks down under test rather than in service. The 75% ceiling exists because the factory test has already been applied once and insulation is not meant to see that voltage twice. (10.6)

NOTE A current transformer with reversed polarity or a swapped secondary lead makes every relay and meter downstream of it wrong while every one of them tests perfectly on secondary injection, which is why the instrument transformer circuit is verified end to end before the relay test is counted. (11.5)

NOTE These classes carry their own test lists in their own standards because their functional tests, the transfer sequence, the load test, the drive start, and the battery discharge, are tests of what the equipment does rather than of its insulation and connections. This standard adds the tester, the instruments, the criteria for the electrical tests among them, and the report. (12.6)

NOTE An infrared survey finds the connection that passed its torque check and heats under load anyway, and it is the only acceptance test performed on energized equipment under load; one survey within 90 days finds the connection that was wrong from the start; a second survey near the end of the first year finds the termination that loosened through the first full cycle of seasonal load and files the warranty claim before the warranty ends, at the cost of a second visit under load; a program that omits the survey relies on the de-energized tests and on the Owner's maintenance program under NFPA 70B to find a heating joint. (13.2)

NOTE A joint heats in proportion to the square of the current through it, so a survey at light load sees a fraction of the rise the same joint shows at full load; 40% of rated load is the minimum ANSI/NETA ATS states and the level most programs survey at, a lower value is accepted where the facility cannot be loaded further before the survey window closes, and a higher value is specified where the Owner can apply load banks or the facility reaches full load early. (13.5)

NOTE The test report is the permanent record of the as-installed condition of the electrical system and the baseline every later maintenance test is judged against under NFPA 70B. A report that records pass and fail without the measured values is not a baseline, and a report that cannot be indexed against the one-line diagram cannot show what was never tested. (14.1.5)

NOTE A short delivery period gets the baseline into the Owner's hands while the testing crew is still available to answer questions and before the warranty clock has run far, at the cost of a report compiled under schedule pressure; a longer period lets the testing organization compile, check, and index a large program's data, and leaves the Owner operating for that period on the field data sheets alone. (14.2.4)