The facility involved: A CO2 liquefaction plant. Gaseous CO2, a by-product during the fabrication of NH3, which is a key input in nitrogenous fertilisers, was separated from the other input gases (N2 and H2) for subsequent reuse. It was being piped to the plant’s compression / liquefaction installation, which features an ammonia cooling circuit intended to liquefy the CO2, consisting of: compressor, condenser, and exchanger/evaporator.
Around 5:20 pm, a short-lived atmospheric release of ammonia (NH3, a toxic gas) occurred on the CO2 liquefaction circuit of a chemical plant specialised in producing nitrogenous fertilisers.
Circumstances surrounding the leak
Upon restarting the liquefaction installation, the compressor’s feed valve opened as programmed, but pressure in the compressor exceeded the set pressure of its relief valve (i.e. 4.8 bar): the relief valve then opened to the atmosphere at an elevation of 6 m above the premises. This opening created a pressure drop in the installation circuit, followed by sudden evaporation of a portion of the liquefied NH3 in the downstream exchanger (flash point). A high-level safety device detected the rapid pressure rise in the circuit and automatically closed the compressor’s inlet valve. Its relief valve closed after a few seconds. A detector, positioned in the downwind direction adjacent to a neighbouring building 100 m away, recorded a very brief NH3 spike (like a puff). The operator estimated that the quantity of NH3 released into the atmosphere was less than 50 kg. When a local resident called after noticing an ammonia smell, fire-fighters visited the site around 6 pm to verify the absence of ammonia in the vicinity of the site and discussed the circumstances surrounding this incident with the operator. The ammonia contents measured at and around the site read zero.
Hypothesis regarding the causes of this leak
The exchanger containing NH3 in a liquid/gaseous equilibrium state was not drained during the (1-month) maintenance period. These works had been scheduled for the summer; when the heat wave struck, the liquid phase temperature inside the exchanger (despite its heat insulation) slowly rose from -30° to +15°C. The ensuing evaporation caused the pressure to reach 8 bar. This pressure did not pose a problem for the exchanger, which was capable of withstanding considerably higher pressures (the relief valve had been calibrated at 15 bar). Upon opening the compressor valve, the exchanger pressure filled the compressor via the circuit. This pressure level, exceeding that of compressor operations, triggered opening of the compressor relief valve, which had been calibrated at 4.8 bar.
