Premium
A flare deflagration incident at Rohm and Haas
Author(s) -
Desai Vipul M.
Publication year - 1996
Publication title -
process safety progress
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.378
H-Index - 40
eISSN - 1547-5913
pISSN - 1066-8527
DOI - 10.1002/prs.680150310
Subject(s) - flare , deflagration , flashback , methane , mechanics , seal (emblem) , chemistry , detonation , explosive material , engineering , physics , art , combustor , organic chemistry , combustion , visual arts , aerospace engineering
One of the flares at Rohm & Haas plant experienced a series of three consecutive deflagrations resulting in severe damage to the flare water seal flame arrestor. The deflagrations occurred during process start up after a complete unit shutdown. This incident caused no personal injuries or off‐site impact. The incident was investigated using MCSOII (Multiple Cause System Oriented Incident Investigation) process. The root cause for the event was inadvertent addition of hot air to the flare. The hot air flow was such that the methane‐air mixture in the flare was combustible (about 11% methane). In addition, hot air raised the combustible methane‐air mixture temperature to about 300°C. The flashback velocity from the flare tip is a strong function of process gas temperature and is much higher at elevated temperatures. Thus, the downward flame front velocity was greater than the upward methane‐air mixture velocity and caused flashback. The first flashback possibly damaged the water seal such that the water seal flame arrestor lost its functionality and the connector pipe, which brings gases from knockout drum through the water seal, was open (See Figure 1). The next two flashbacks resulted in deflagrations in the knockout durm through the open connector pipe. Thus, the key recommendation to avoid a similar incident is to eliminate hot air flow to the flare. The key learning from the above incident is that flashbacks are possible when process gas is hot. We would like to extend this learning to other flares and urge a further study of all flares for possible flashback scenarios. A key action point from this incident is to perform a HAZOP on other possibilities of FLASHBACK in the flare. This will be done by performing flashback calculations on all possible flaring scenarios in the flare.