RECENT EH&S INCIDENTS
July 2005
1. Caught in Robot Guard Door
Injury: Bruising to Forearms
Time Lost: 3 hrs
Description
An operator was working on a packing machine fitted with a small robotic arm to load materials. The robot is protected by a see through plastic panelled door which can be raised to allow pallets of materials to be loaded (Fig 1) raising the door disables the robot.
The operator had operated the switch to raise the door for a material change; at the same time he noticed that the robot arm was about to knock over a stack of materials it was loading into the packing machine. Knowing this would take time and probably lost production to correct, he lifted one of the door slats (which were loose fitting) and put his arms inside to catch the stack of materials (Fig 2 ) at the same time the robot door started to lift following the material change instruction. The operator’s arms were trapped in the door slats causing some bruising although he was back at work the next day.
Comment
During the incident investigation it was found all similar machines had slats on the guards that allowed them to be lifted and access the robot area without it stopping. One machine had a handwritten notice warning people not to do this, suggesting it was a known practice.
The operator was involved in the investigation and whilst it was accepted they should clearly not have moved the guard panel to access an area it was meant to protect, they were only trying to keep the machine running - and should to be applauded for reporting to work the next day. The guard should have been designed so it could not be so easily defeated by people trying to maintain production. A better risk assessment either at machine design or installation would have identified such a potential; latter machines do in fact have a different design of door which does not allow access whilst the robot is operational.
Action
All relevant personnel being briefed on the hazard in this area. Managers and cell coordinators to ensure safe practice is followed; rather than simply putting signs up which can quickly become ‘part of the furniture’.
Investigations are ongoing to fit brackets so the guard panels cannot be lifted out of place.
Increase awareness during risk assessment process to ensure that issues like these which are less easy to anticipate are identified.
Circulate report to machine purchasers and other plants.
2.Tank Overflow
Injury: None
Other effects: Some spillage – cleared up
Time Lost: None
Description
A tanker arrived with a delivery of chemicals used in the manufacturing process to fill a bulk receiving tank. When the tanker was initially connected the tank control system refused to allow it to fill. It was assumed this was due to some fault in the system as the mimic panel in the factory control room showed the tank was empty. There was a concern that the factory might run short of chemical if the delivery was not completed, so a plant technician was called to see if he could correct the problem. He was told that the tank was empty according to the factory mimic, he therefore assumed the loadcell readout adjacent to the tank - which was showing 20 tonnes in the tank - was faulty. He managed to get the delivery started which continued for about 20 minutes until the tank’s manual overflow valve lifted and the chemical started to escape. The delivery was immediately stopped and the spilled chemical which had been caught in the tank and delivery area bund was collected.
Comment
The tank being filled is fitted with a high level probe which should have stopped the delivery before it overflowed, this did not operate. When the other tanks in the small tank farm were checked it was found that the overflow detectors did not appear to be operating correctly either. The tanks had been installed for nearly 10 yrs but because the chemical delivery was always made in a regular fixed amount when the receiving tank was empty, the high level probes had never been required to operate. It would appear they had never been fully commissioned when the tanks were first installed.
Investigations are still ongoing as to why the factory mimic was showing the tank empty when the loadcell was showing it full at the time of the incident (the system electronic archives have confirmed this) yet the system is now working correctly. The load cell was checked during the investigation and found to be operating. The tank’s control system which is designed to prevent filling of a full, or partially full, tanks was found to be working correctly but that it had been overridden in the past to allow deliveries to be started when a small reading on the load cell (often residue from a previous delivery) would cause the control system to inhibit a new delivery.
The clean-up operation by the maintenance group had worked very efficiently.
Actions
Overflow detectors commissioned and a regime put in place to regularly check that overflow detectors on all site tanks are operating. Content level at which the load cell prevents a delivery to be established and adjusted to ensure the system does not have to be overridden if there is a small residue of liquid in the tank. Investigate fault in either mimic or tank control system that caused a full tank to be denoted as empty. Managers to ensure this and other control systems only over-ridden as a last resort when the nature of any fault and the consequences of a system over-ride have been properly established. Project engineering management has been developed over a long period with the adoption of management processes such as Prince 2 and formal project files; however this incident is a reminder of the need for a thorough commissioning process particularly of shut off and safety devices. Although not directly related to this incident a regime was also started to check all tank vents regularly. It is worth noting a large ambient pressure tank can be exploded or collapsed by a pressure as low as 200mm of water gauge - the pressure at the bottom of a pint of beer. Hence a blocked, or otherwise inoperable, vent on a tank being emptied or filled can split the tank and result in a major spillage or worse. (Another way to look at this is to think of a baked bean tin and how easy it is to crush – the thickness of a large storage tank’s walls relative to its size is several times less than that of the baked bean tin)
Safety Incidents YTD in UK 2005
| LWC* | LWCIR** | No Lost Time | Near Miss |
| 4 | 0.14 | 268 | 91 |
* Incident resulting in a day or more of lost time
** Number of LWCs per 100,000 hrs worked
Lost Time Accidents YTD in UK 2005
| Month | Incident | Corrective Action | Days Lost |
| Jan | Operative caught hand whilst shutting embossing head. Bruised fingers |
Latch mechanism on embossing head altered to move hand away from joint area | 2 |
| Mar | Operative slipped off steps whilst clearing cylinder blockage. Broke bones in hand | Installed access steps and platform to all cylinders for safe access | 32 |
| Mar | Fitter caught finger with knife whilst cutting conveyor band when knife slipped. Cut Hand |
Advised on use of clamps and/or hand protection when risk of being cut | 5 |
| Apr | Operative overreached when using foot stool. Hurt back. | Advised on safe use of stool. No other problems found. | 7 |
“Experience is the best of schoolmasters, only the school fees are heavy.”
Thomas Carlyle