Ignition Source
Electrical short

An employee noticed an unusual smell in a fuel cell laboratory. A shunt inside experimental equipment overheated and caused insulation on conductors to burn. Flames were approximately one inch high and very localized. The employee de-energized equipment and blew out the flames. No combustible material was in the vicinity of the experiment. The fire was contained within the fuel cell and resulted in no damage to equipment.

The employee was conducting work with a fuel cell supplied by oxygen gas. The hazard control plan (HCP) associated with the work was for use with fuel cells supplied by air or hydrogen, but not for oxygen, which yields a higher current density. The technician had set up the station wiring to handle a current of 100 amps and the shunt was configured to handle a current of 100 amps, but the station load box could handle up to 180 amps, and during the experimental run the increased current density applied 150 amps power to the wiring.

Electrical wires overheated and melted, causing a short and fire in the fuel cell.

Incident Date
Nov 09, 2001
  • Hydrogen Production/Use Systems
  • Fuel Cell
Damage and Injuries
When Incident Discovered
Lessons Learned

Procedures will be revised to include specific information about hydrogen sensors and alarm controls. Caution will be added about sparks and flames in the fuel cell operation area. Workers will be cautioned that shunt resistors may reach high temperatures, shunt resistors should be kept away from gas tubing and other electrical wiring but near the fan; and shunt resistors should be checked periodically to assure that a good connection is made, thus preventing overheating. The procedure will also be revised to specify the proper hierarchy of emergency notification.

Workers should think twice about blowing out a fire, even if it is perceived as small. Serious injury can result even if the fire is relatively small. Flammable gases such as hydrogen and oxygen could provide additional fuel. It is important to document changes to an experiment through a formal change control process. On-the-Job training should be provided in several modes of communication. An electronic mail notification should not be the only and primary method of communication.