Pollution
Humain
Environnement
Economique

In a resin manufacturing plant, acrylic acid polymerised without a runaway reaction. The incident involved a polymerisation line consisting of a reactor close to a unit equipped with a pneumatic pump used for introducing a 1st catalyst and products in small quantities. The monomers, a 2nd catalyst and the other additives were prepared and pre-dosed in 3 relay tanks on spring balances, and their contents were transferred by pump into the reactor. This line was controlled from a central station in the control room adjacent to the workshop. This station was connected to a control plc and an independent safety plc. The technician had a worksheet listing the operations to be carried out and on which he had to record the data relating to the times of the basic operations, the weights introduced, and the temperatures, etc. Since September 2005, in response to a large proportion of non-compliant production, the company had added to this sheet a step, by the conformity verification officer for reagent preparation before introduction into the reactor. This operation had enabled the quantity of non-compliant products to be reduced. In addition, a trainer was present for the staff operating the installations. On 19/02, at around 1 am, the employee prepared the bottom of the tank (water + solvent), added the 1st catalyst, a transfer agent and started to heat the whole up to 83 °C without adding the 2nd transfer agent. He also prepared the relay tanks. At around 4 am, without validating the worksheet, the technician started the injection of reagents from the relay tanks. Polymerisation was immediate, the reactor cooled. After changing stations at 6 am, the temperature progressively rose (93 °C at 6.20 am). The technician then stopped the injection of reagents before the 1st plc safety level at 98 °C. A few minutes later, the agitator tripped due to an electrical overload. The technician diagnosed a non-compliant and viscous polymer and shut down the installation. The reactor was flooded with water.

The incident was caused by the formation of a polymer that was too viscous because of the absence of the 2nd transfer agent, which was the consequence of several uncoordinated actions, written reports not drafted, as revealed by the worksheet : the 1st catalyst introduction time had not been written down but had been recorded by the plc, the 2nd transfer agent introduction time had neither been noted, nor recorded and the sheet had not validated.