NEAR-MISS INCIDENTS IN DRYERS
Drying operations in the Chemical industry frequently cause concern over fire safety. This article reports two cases where investigation of near-misses has positively contributed to reducing the risk of serious incidents. An improvement in productivity also resulted in one case.
In the first incident, an unexpected delay in servicing some ancillary plant resulted in some material charring. In the second, a drying operation had been run without problems in a number of individual dryers for a period of years. A spate of three apparently unconnected incidents occurred in a short period of time.
Case 1: Charred solids in a solvent recovery dryer.
1.1. The Plant.
A thin film evaporator was being used to recover solvents from a slurry of variable composition. The operation was not central to the manufacturing operation and the materials concerned were known to be chemically inert. For these reasons, only a cursory safety study had been carried out prior to commissioning.
The plant was fed with a slurry from the waste streams of another plant. The feed consisted of a polymer solid dispersed in a hydrocarbon solvent. The plant was designed to recover the solvent by evaporation from a thin film under vacuum, contributing to reduction in costs of both raw materials and waste disposal. It was operated by a team whose primary interest was the plant producing the feed. The composition of the feed varied at around 10-20% solids, and the flow through the dryer was adjusted manually to allow for this variation. The dried powder waste was dropped periodically from the bottom of the dryer into a bottoms vessel, where the vacuum was let up with nitrogen. This arrangement allowed the drying process to continue whilst each slug of solid was dropped into a drum under the dryer. The sequence controller dropped solids out through the bottom of the dryer every 5 minutes. Blockages in the flow of dried powder waste were not uncommon , because of the control difficulties caused by the changes in the composition of the feed.
1.2. The Incident
On the night to the incident, a blockage occurred at the base of the dryer or in the bottoms vessel. The shift team decided that the bottom section of the plant had to be removed in order to allow the blockage to be cleared, as occasionally happened. The fitter was notified and the plant left idle in readiness.
The fitter was delayed on more urgent work, for around 5 hours. The permit to allow the work to proceed was issued accordingly to the local rules, and the dryer was split at the bottom slide valve. Whilst the blockage was being cleared, a quantity of charred material was seen to fall from the bottom of the dryer, and the Company Fire and Emergency Crew summoned. No flames were observed.
1.3. The Inquiry
The inquiry found that the steam supply to the dryer had been left switched on during the time that the fitter was delayed. Some sections of the dryer which were not swept by the dryer blade were heated by steam. The dry material, normally inert polymer, was being heated for the extended period. It is also found that the slide valves fitted to the bottom of the dryer bottoms vessel were probably leaking slightly, and that air was being drawn through the solids during this time, A laboratory investigation showed that the solids were stable up to very high temperatures in anaerobic conditions. In a stream of fresh air the solids were susceptible to oxidative degradation. In the original design, this possibility was not considered because the dryer operated under vacuum, with the bottoms vessel being relieved with nitrogen every 5 minutes during normal operation.
1.4. The Lessons
The composition of the feed stream was to be checked prior to each run and adjusted by addition of solids/solvents to some consistent value to allow easier control. It was also considered that addition of an inert powder (such as talc) might assist flow of the solids at the bottom of the dryer.
Operating instructions and training were to refer to the need to cut steam supplies to the plant whilst it was awaiting attention from the fitters.
Steam heating on unswept sections of dryer was to be removed.
The inquiry also identified potential problems which could arise due to evaporation of an inflammable liquid, albeit in a correctly zoned plant area.
Case 2: Glowing Embers in Fluidised Bed Drying Operations.
2.1. The Plant
The drying operation occurred in two stages. In the first the moist product was broken down into lumps of around 1cm diameter, and dried in a continuous process in a fluidised bed dryer. To assist the flow of the chips along the bed, a number of additives (totalling much less than 1% of the output from the dryer) in a solution were mixed with the chips before they went into the dryer. In the hazards assessment for the process, no adverse interactions between any ingredients in the solution or between the product and any other ingredient were found.
The dried chips were further processed before being sprayed with another set of additives (also amounting to much less than 1% of the product) and then dried again, this time with a smaller chip size than previously. Again, no adverse chemical interactions were found prior to commissioning.
2.2. The Incidents.
The first incident occurred when the filter elements in a polishing filter on the output side of the primary dust collectors were changed. Surprisingly, the embers were not seen whilst the filter elements were being changed, but over eight hours later when the discarded cartridges were found to be smoldering in the skip outside the factory building.
The second incident occurred whilst drying fine material recovered from the primary dust filters. This resulted in a lower grade of product, which was both salable and helped to reduce waste. Due to an oversight, the temperature of the fluidised bed was not reduced for the different feed, and glowing embers were observed by operators halfway down the bed a short time after processing began.
The third lot of charred powders were found by a commissioning engineer whilst a new dryer was being brought on-line. The first runs were conducted with (expendable) recovered fines. A blockage had occurred in a hopper underneath a dust filter, and some (cold) charred material was found in the blockage.
2.3. The Inquiry
The three incidents were considered to be linked after a table of recent changes to the process was drawn up. The industry-standard aerobic and anaerobic tests on samples taken from the plant failed to show any signs that any part of the process could have caused the fire at the process temperatures.
However, highly speculative tests demonstrated that one of the ingredients from the first dryer process could interact with another ingredient from the second dryer. The reaction gave ultimate temperatures of 700°C in the presence of inert organic materials such as the filter medium and the product, thought to be necessary as fuel.
The standard tests missed the problem because of the diluting effect of the inert product, which masked the tiny amount of reaction which must have occurred during tests. For the reaction to have become significant, clearly some concentration was necessary. This concentration which was thought to have occurred in the three incidents as described below.
One of the problem components tended to break down into very small particles by attrition in the dryers. This effect allowed the component to concentrate in the various dust filters around the plant. The other component was a liquid at the dryer temperatures, and could have traveled either with the fine solid component or it could have evaporated in the dryers to condense later on the other component in the dust filters. Air flow through the filters kept the reaction matrix cool until the filter elements were removed and dumped in close-fitting cardboard boxes.
Incidents 2 and 3:
Incidents 2 and 3 both involved recycling fines taken from the plant's primary dust extraction system. Chemical analysis showed that this material had very high levels of the solid component (10 to 100 times as much as normal). The second component was added directly to the fines on entry into the dryer. The slightly higher temperatures used at the time of the second incident reduced the induction time for the reaction from two hours (four times the normal residence time) to 15 minutes (around half the residence time). The charred material in the third incident had collected in a discharge hopper beneath the filter bags. The mixture would have been out of any cooling air-flows once the blockage had built up.
- The standard battery of tests (described by J. Abbott1) did not spot the problem, because all materials tested were plant samples, and the principle components of the problem reaction were massively diluted and apparently separated in the process.
- A new testing regime was needed which tested not only the mixtures expected to arise on plant, but also all those which could arise, even if they contain materials which are intended to form only a tiny fraction of the product.
- Even tiny changes in formulation can give rise to wholesale changes in the properties of the product. The effects of formulation changes on the hazardous properties of materials were to be checked to ensure that the Basis of Safety for the process remains valid.
John Abbott, "Prevention of Fires and Explosions in Dryers", 2nd edition, I.Chem.E.,Rugby, UK (1990)