NATTECH HAZARDS - ARE YOU PREPARED FOR A FLOOD EVENT?

 Feedback on the consequences of heavy rain-induced flooding allows us to draw the following lessons:

  Flood risks must be identified: 

Flood risk maps, particularly those in flood risk prevention plans, must be consulted. It should be remembered that flooding — particularly from heavy rains — of sites located in areas deemed not subject to flooding is not an exceptional event. Identifying this risk entails determining its potential intensity, i.e., the flow rate and velocity (which can have an impact on which objects are carried away by the current, debris jams, bank erosion, etc.), the height and duration of submersion (which may result in damage either directly from the water or indirectly due to the static pressure exerted by the water) and the rate of water rise. 

Potential difficulties must be anticipated: 

- Impact of water: on buildings and facility structures (particularly tanks and pipes), stored products or waste (especially if they are hazardous, polluting or react with water), electrical and IT equipment and manufacturing processes; 

- Power infrastructure disruptions: whether on or off site (e.g., flooding of a transformer serving an industrial facility) during a rainfall event and sometimes several days afterwards. The operation of facilities that must be shut down progressively must be anticipated. Preventive shutdowns of utilities are often recommended; 

- Internal and external means of communication: PSTN services are frequently shut down during intense rainy episodes. GSM services may also be shut down; 

- Site access: automatic gate systems may be out of service (due to power outages, for example), there may be debris jams, access roads may be cut off, a site’s internal roads may have to be cleared to allow emergency access or enable diesel generators to be set up; 

- Objects that could cause debris jams: objects such as empty pallets should not be stored outdoors and the number of objects that may be carried away by floodwaters should be kept to a minimum; 

- Surface water infiltration: special attention must be paid to boreholes that may make it easy for surface water to contaminate groundwater; 

- Ancillary systems at risk of being rendered inoperative: special attention must be paid to fire detection and extinguishing systems, wastewater treatment plants, etc.  

Mitigation measures adapted to the vulnerability of the facilities and the kinetics of the event must be implemented. Such measures can be of several types: resistance measures (preventing water from entering the site with flood barriers, check valves on water drainage networks, self-powered pumping systems, etc.), management measures intended to limit damage from the onrush of water (by raising or moving sensitive materials, hazardous products, IT equipment, etc.) and thus reduce the time needed to normal operations and, finally, an ultimate phase of transitioning the facilities to a safe fallback state. 

The appropriate equipment (waders, torches and head torches, lifting pumps, diesel generators, etc.) and the necessary human resources able to go to the site (especially when events occur at night or on non business days when travel is strongly discouraged) must be planned in advance. Staff training and awareness-raising in flood-risk management is essential, as is including this risk in a site’s ad hoc procedures (particularly regarding the maintenance of water networks, gutters, drainpipes, etc.). It is wise to designate a fallback area beforehand. It is important to bear in mind that flood events can occur in entire geographical areas that are of varying extents and home to multiple industrial facilities that could be affected. In addition, emergency services will prioritize providing assistance to the general public and difficult weather conditions at night complicate air rescue operations. Lastly, getting a site back to normal operations after pumping (which sometimes requires the use of large capacity equipment) and clean-up operations have been completed can take time. There may be many layoffs during this period. Anticipating flood events due to heavy rainfall is therefore vital and must be combined with monitoring of weather warnings and forecast maps that give 24-hour warnings of approaching dangerous weather.

Source:Aria database NATECH hazards

MALFUNCTION OF LEVEL SENSOR CAUSES LIQUID GASOLINE TO DISCHARGE FROM A SAFETY VALVE

Fire broke out in a refinery on an atmospheric distillation unit. The blaze was triggered by an ignited leak on the exhaust of a safety valve associated with a gasoline stabilisation column (i.e. a debutaniser) whose function was to separate gasoline from gas. The internal emergency plan was activated at 6:56 pm and the unit was shut down as per an emergency procedure. The refinery's internal responders extinguished the fire at 7:49 pm. 

The atmospheric distillation unit had been restarted a few days prior and its operations had not yet stabilised. Just a few hours before the incident, console operators and technicians had experienced difficulties on a vacuum distillation pump. The stabilisation column was re-boiled at its base by an exchanger fitted with a column base level setting device. As of 3:30 that afternoon, a drift appeared on the corresponding measurement, leading to the partial, then complete, closure of the bottom valve and hence to the column filling with gasoline. A mix of gasoline and gas spilled out via the overhead valve, formed a stream and ignited at a hotspot at the column base. The flame front rose to the outfall, thus sending a flame above the main atmospheric distillation column. The pressure sensor had been servo-controlled to allow turning off the heat, thus avoiding insufficient condensation at the top, which constitutes the typical cause of pressure surges. This incident underscored the need for a new sensor locked loop on the column bypass, in order for the safety system to incorporate the risk of column overfilling.

Source:Aria database

ARE YOU CARRYING OUT POSITIVE MATERIAL IDENTIFICATION (PMI) ON CRITICAL VALVES?

 In a fertilizer plant, 62% nitric acid leaked around 8.10 am from a sectional valve at the inlet of calcium ammonium nitrate saturator in stable operating mode. The 20 min leak resulting in the loss of 1 tonne of acid led to the release of significant quantities of nitrous vapours in the unit upon contact with the insulation of a steam pipe below. The alert was sounded by an employee in the vicinity. The internal emergency plan was triggered at 8.12 am and lifted at 9.47 am. The operations in the unit were stopped. The released acid, as well as the 2 m³ of water used to cool the facilities and reduce acid evaporation were collected in a retention tank and channelled to a pre-isolated recovery tank. 

The leak occurred due to the premature corrosion of the metal joints of the sectional valve. According to the operator, the material of the joint assembled a few days back did not match the specifications indicated in the purchase order of the joint. No other safety device alerted the technicians of the leak. The unit was in fact not equipped with a sensor likely to detect such a leak. Further to this accident, and before re-starting the unit, the operator had to carry out an exhaustive inspection of similar joints likely to have been replaced during recent maintenance operations. In addition, the operator also had to conduct an inspection and overhaul of all facilities likely to have been damaged due to the acid spill.

Source:Aria database

PROCEDURES MUST BE FOLLOWED STEP BY STEP

At around 6.00 am, a flash was produced during operations in a chemical plant. The accident took place between two production campaigns. On the 7 and 9 February, the hydrogenation reactor was cleaned, deodorised and dried to eliminate residual waste from the previous production cycle and prepare the reactor for the following one. On 12/02 at 5.00 am, the production supervisor issued specific instruction on a new production cycle to a technician. The technician did not follow them in the right sequence as indicated in the instructions before staring production: to save time when pressurising the device at 9 bar of nitrogen, he simultaneously made several adjustments and carried out several checks The reactor was pressurised using nitrogen. The operator observed N2 leaking from a manhole. He decompressed the reactor and removed the fastening bolts from the lid to change the gasket. During this operation, he heard a leaking noise at the joint. The two technicians were projected backwards in an explosion that ensued. Both the technicians sustained burns on their faces and hands and were initially hospitalised in Nice and subsequently transferred to a specialised unit in Toulon. 

The investigation report explained that the explosion occurred as a small quantity of H2 ignited upon contact with the catalyst in presence of oxygen coming from the manhole. It also noted that the tightness of the reactor was checked and the pipe pressurised at the same time without following the guidelines

The drying conditions may have contributed to have increased the pyrophoric nature of the catalyst. Changes have been planned in the hydrogenation equipment (installing pressure sensors, flame guards on regulator vents, etc.) and procedures (risk analysis, verification of H2 pipes, de-commissioning the H2 pipe before opening the manhole.

Source: Aria database

ARE YOUR FLANGE TIGHTENING TORQUES PROPERLY SET?

While re-starting an ammonia production unit in a chemical plant following a technical breakdown that lasted one hour and 30 minutes, a synthesis gas leak (50% hydrogen, methane, nitrogen and 17% ammonia) ignited on a flange (O-ring joint) located just after the synthesis reactor. The operator triggered the emergency shutdown after the unit’s alarms were triggered, isolating and depressurising the synthesis cycle. Steam was sprayed on to the leak using mobile fire nozzles to extinguish the fire that was brought under control 55 minutes following its start. Material damage was estimated at 60 K euro The blaze did not damage the reactor that was protected by a deflector. Operating losses stood at 300 k euro. 

The tightening torques of the bolt on flange where the leak occurred is responsible for the accident: they were not adapted to the exceptional conditions prevailing during the accident (significant temperature difference between the bolt and the flange resulting from the abrupt temperature variation due to the short technical breakdown.) The implemented corrective actions involved redefining the tightening torques, improving the tightness of the drip collar on the flanges (defective at the time of the accident), designing a nitrogen injection system in drips collars and installing a steam injection crown to protect the bottom of the synthesis reactor.

Source:Aria database