Operations at a remote cellular telephone relaying facility came to an abrupt halt one afternoon in 1992 when a 12V lead-acid automotive battery used to start and run a 50kW emergency engine-generator set (gen-set) exploded and emitted damaging acid fumes throughout the building. The gen-set served the facility's electronic equipment, HVAC, and lighting loads, and was located in a separate room from the telephone electronic equipment, along with a 200A automatic transfer switch that included the starting battery float charger. Photo 1 shows the battery with its clearly expanded-from-inside-outward sides, missing ends, and missing plate inspection covers over the four plate groups.

The owner/operator of the cellular telephone facility filed suit against the manufacturer of the emergency gen-set and maintenance company, claiming, among other things, that the gen-set battery charger and regulator were improperly designed, manufactured, and maintained. The plaintiffs alleged their equipment and facility suffered more than $600,000 in damages.

The evidence. The owners of the facility began investigating the incident immediately after it happened, but it wasn't until four years later that the insurer for the gen-set manufacturer hired me to conduct a separate investigation. The scene of the accident had long since been repaired, but significant evidence taken at the point of origin for the explosion and acid expulsion had been retained for examination. In addition, a remote cellular telephone relaying facility that was configured in a similar fashion to the incident site was made available for my inspection in 1996.

A visual examination of this similar site revealed first that the facility was improperly ventilated. Air was able to move freely between the electronic equipment room and the generator room where the lead-acid starting battery had been stored. Each of these rooms should have had its own ventilation system, and at a minimum, the air in the facility shouldn't have been permitted to recirculate. This facility design allowed the acid fumes that were released in the explosion to reach the electronic equipment room, thereby increasing the severity of the damage.

When I picked up the case, the original battery and its charger and regulator were still available for me to examine and test. An examination of the regulator's charge rates, electrical tests of the subject charger, and measurements of the specific gravity during charging all indicated that the regulator was continuously overcharging a fully charged battery. A second test with a new regulator showed the charger to perform properly as a “oat” type charger.

An investigation of the service records for the subject gen-set disclosed that approximately 10 weeks before the explosion, a field service technician who worked for the gen-set manufacturer's local distributor performed a scheduled semi-annual maintenance during which he found and replaced a dead battery and set the adjustable voltage regulator to what he thought was its lowest setting. Photo 2 shows the charger, regulator, and adjustable potentiometer. In his deposition, the service technician described the lowest setting as being a fully counterclockwise setting of the adjusting potentiometer. The manufacturer's specifications, however, state that the potentiometer is rotated counterclockwise to increase the oat voltage and clockwise to decrease the oat voltage.

Next I conducted a third laboratory test with the subject charger and regulator, but with the regulator adjusted following the manufacturer's specifications. This test confirmed that a properly adjusted regulator caused the charger to perform correctly as a oat type charger.

Normally a properly operating oat charger for this application delivers 100mA and maintains a battery open-circuit terminal voltage of 13.3V. The subject charger as adjusted by the field technician was delivering 1.4A to 1.6A into a partially discharged (12.04V) battery and stabilizing the battery open-circuit terminal voltage at 14.4V. From the service records I also discovered that an automatic weekly test was programmed to start the gen-set, run it for 10 minutes, and then shut it down.

I also determined that the electrolyte level in the battery was being checked semi-annually. The manufacturer's recommended interval for electrolyte monitoring was monthly, in accordance with NFPA 110, “Standard for Emergency and Standby Power Systems,” 1996 Ed., Table A-6.3.1 (a), which suggests that the electrolyte level, specific gravity or state of charge, and charger and charge rate should be scheduled for monthly maintenance.

In addition, IEEE Std. 446-1987, the Orange Book, “Recommended Practice for Emergency and Standby Power Systems for Industrial and Commercial Applications,” 8.6.4, Typical Maintenance Schedules, are as follows:

  • Daily — Check bus voltage, pilot cell hydrometer, and charge or discharge current.

  • Monthly — Check cell level and filter cap vent holes for plugs, inspect electrode plates for color, particles, damage, or swelling, inspect bottom of cell for residue, and inspect exterior of charger cabinet for signs of water entry.

  • Quarterly — Perform equalizing charge, and after completion take hydrometer readings of all cells; read and record cell voltages 10 to 20 minutes after starting equalizing charge; clean off cell tops with soda solution; check condition of cell terminations and straps, and check torque on connection bolts.

However, the maintenance schedule and work practices at this telephone facility didn't follow these suggestions or recommendations. In fact, the owner opted to reduce the maintenance schedule significantly in order to cut costs.

The ruling. My investigation revealed that the high continuous charging level set by the field technician, the frequent (weekly) exercising of the engine (starting), and the lack of regular monthly monitoring of the electrolyte level caused a boiling liquid expanding vapor explosion (BLEVE) condition. Such a condition can result from the superheating of the liquid electrolyte in the battery case to the point where it suddenly flashes into vapor or steam with an attendant large increase in volume; in the case of water, this increase is approximately 1,500 times. This large volumetric increase within the battery can produce forces that cause the case to explode. Several of the experts in this case postulated that a low electrolyte level in the battery would be further diminished during a heavy current discharge of engine starting and that the battery “BLEVEd” during such an operation, causing the explosion.

Each repetitive starting of the engine produced current heating of the battery plates and electrolyte. On discharge, both the peroxide (PbO2) on the positive plate and the “sponge” lead on the negative plate are quantitatively converted into PbSO4, which increases plate heating during current flow. This was followed by a short (one-week) period of relatively high current recharging that would have accelerated the loss of electrolyte from the battery case. As the cells approached full charge, they couldn't absorb all the energy from the charging current. The excess energy would have electrolyzed the water into free hydrogen and oxygen liberated from the cells as gases and decreased the water level of the cells. The BLEVE occurred when this combination of events overheated the diminished electrolyte level in the closed battery case and caused the liquid electrolyte to boil and expand catastrophically.

In the end, I concluded that the accident wasn't caused by the design or manufacture of the battery charger and regulator, but rather by the field maintenance — or lack thereof — of the battery charging system and battery. Developing this maintenance program was the joint responsibility of the gen-set manufacturer and the owner of the facility who was paying the bill for this service. The facility was located more than an hour from the distributor's office, making it impractical to perform daily inspections, but weekly inspections would have been feasible.

Because the facility owner went against manufacturer and IEEE recommendations for maintenance scheduling, the jury reached a defendant's verdict in favor of the gen-set manufacturer, who was awarded $7,500 in costs from the facility owner. However, the service company was ordered to pay the facility owner damages and costs totaling $140,000 due to the technician incorrectly setting the voltage regulator dial. Regardless of who was at fault, though, the case demonstrates the importance of regular, proper maintenance and what can happen when it's ignored.

Nabours is president of RENCEE (Robert E. Nabours Consulting Electrical Engineers) in Tucson, Ariz.