Fire destroys equipment. Strangely, additional electrical equipment not consumed by fire is also ruined. What was the cause of this secondary damage?
While using an acetylene torch during renovation work, a worker inadvertently starts a fire in the main building of a West Coast research center. Individual feeder conductor insulation melts, shorting out the power to various loads. The fire shorts the facility's panelboard feeder circuit breakers. A separate molded case feeder circuit breaker and individual electricalfeeder riser individually feed each building's electrical power panelboard. When the shorts occur, clocks stop at respective buildings.
This is the scenario we faced when starting our investigation. With the clock stoppage, we at least knew the time each distribution circuit shorted.
At approximately 2:02 p.m., the main building lost partial electrical power when a single-phase primary feeder failed because the main building's fuse blew on the utility's primary power pole. Smoke and water vapor caused an ionizing conductivity path between Phase B of the 3-phase primary electrical load break switch and ground.
The blown Phase B fuse caused a single-phase condition in the 12.47kV, 3-phase, primary electrical system. Two of three phases were operational, causing a single-phase, 12.47kV line-to-line condition, as opposed to the 12.47kV, 3-phase electrical service, composed of three 7.19kV (phase-to-ground) conductors.
At the facility, the 12.47Y/7.19kV to 480/277V, 3-phase, 4-wire transformers are wired wye-wye. And all the 480V to 208/120V, 3-phase, 4-wire transformers (which are the predominant type on the site) are delta-wye.
At approximately 2:05 p.m., the pump motors in the H building's mechanical equipment room burn out. These motors include three 10 hp, one 25 hp, and two 40 hp units. All the motors are 3-phase induction machines. A 20-hp motor serving incremental air handling equipment on the building's roof is also ruined. These motors are powered from a motor control center (MCC). Some are fed from fused disconnects, while others are fed from molded case circuit breakers. How were these motors damaged? They were nowhere near the fire, which occurred in the main building about 400 ft away.
We inspected the overcurrent protection devices in the MCC and found the fuses had not blown nor had the circuit breakers tripped. We suspected that since one blown fuse (on the utility's pole) of the primary feeder caused a single-phase condition, the 3-phase motors lost their rotating magnetic field. These units have motor starters equipped with thermal overloads, to prevent motor burn out due to excessive phase current. Since the motors burned out under a single-phase condition, it was obvious the fused disconnects and the molded case circuit breakers hadn't functioned as they should.
The MCC in this remote building had dust and dirt on it, indicating no recent preventive maintenance. The inside of the cabinet was dusty, as were the bus bars and over current protection devices. Were electrical feeders torqued down to manufacturers' specifications? It's doubtful anyone had electrically tested the circuit breakers.
We looked at the ratings of the overload devices and associated protecting items. Here, the heaters in the thermal overloads (with one exception) were so oversized they could not provide protection, while the fuses in the disconnects had too high a rating.
Protection features should have opened the circuits to the motors and prevented damage. Magnetic holding coil contacts for the thermal overload and motor control circuit should have been tested for proper operation within manufacturers' tolerances for current draw. This would reveal any shorts, grounds, shorted windings, or open circuits.
Maintenance personnel should have tested the bimetallic thermal overloads to verify they would open the circuit within the rated overcurrent levels published by the manufacturer for the type, frame, and horsepower rating of the associated motor associated. This MCC's breakers needed cleaning with a solvent, and all main feeder circuit breakers should have been tested for short circuit responsiveness of their armatures and long term thermal overload calibration of their bimetallic thermal sensors.
We investigated the electrical systems in the other buildings and found 40% of the existing electrical equipment would have failed to function properly upon overload. We also conducted an inspection of a remote structure from the research center (but receiving its power from the main building). It had a final inspection and punch list prepared a week preceding the fire. We found that some of the duplex convenience receptacle branch circuits had no voltage. The circuit breakers feeding these circuits were closed, and we also checked them for operation. We used a multimeter to measure voltage and power. The circuit breakers in the building checked out correctly.
After opening some receptacle outlet boxes, we found, in many cases, the first duplex convenience outlet in the circuit (the receptacle closest to the panelboard from which it receives its electrical feed) had loose connections or no connections to the outlet. In several cases a junction box that received the panelboard home run (and subsequently fed the downstream receptacles on its circuit) did not have the phase and neutral conductor connections. This resulted in an open circuit.
No matter how good the electrical design, there will be problems and failures if its not properly installed and calibrated. The electrical work suffered from a combination of poor design and bad installation techniques. Poor installation techniques, including absence of protection coordination, installation of improper-sized fuses, and circuit breakers with improperly-sized bimetallic thermal motor overloads caused the majority of the secondary electrical damage from the fire.
MCCs house devices (starters) that start, stop and/or automatically control motors while protecting them from undervoltage and high short circuit currents. The starters also have bimetallic thermal overload protectors in the incoming phase conductors to protect the motor should the full load current exceed its nameplate, full-load capacity. This is important when starting a motor as the initial current can substantially exceed normal rating. By the time the thermal overload protector heats when starting a motor, the current is back to normal, and this protector doesn't have time to function, thus current continues to flow to the motor.