While refilling a mobile tyre handler at the vehicle refuelling bay of a mine, an operator suffered serious injuries when diesel fuel spilled and caught fire. A scientific investigation into the immediate and root causes of the incident was carried out on behalf of the relevant government department.

 

The vehicle refuelling bay was equipped with three delivery hoses rated at 300 litres per minute (lpm), 800 lpm and 2000 lpm respectively. These were fed by pumps located at the nearby fuel farm and switched on from a control room at the refuelling bay. The three delivery rates were set for different types of vehicle and fuel tank sizes.

 

Each of the hoses were equipped with a proprietary coupling designed to lock with a corresponding coupling on the vehicle fuel tank being filled. This coupling had a manually operated valve and an automatic shut-off mechanism, and was designed to prevent over-filling, spillage and spatter of fuel around the tank and operator.

 

All the mine operator’s vehicles and most of the contractor vehicles used on site were fitted with the appropriate fuel tank couplings to connect with the relevant dispensing hose for that vehicle type. 

 

However, there were some occasional rental and/or replacement vehicles and mobile plants, which had the usual open fuel tank filler necks with removable caps. To refill these at the refuelling bay, a ‘splash nozzle’ had been made up on site. This comprised a proprietary connector at one end which locked on to the proprietary at the hose end coupling while the other end was an angled open pipe, similar to a retail fuel delivery outlet in shape. In use, the open end was placed in the fuel tank filler neck and the couplings locked together. 

 

From the control room, the pump was then set to run for a specified time calculated from the flow rate and fuel tank capacity so that it would stop filling before overflowing.  Operators and contractors were given instruction on the appropriate delivery hose, and hence flow rate, to use for the various vehicle types.

 

In this case, the operator inadvertently attached the splash nozzle connector to the 800 lpm hose outlet coupling rather than the 300 lpm hose, which was the right one for the 305-litre fuel tank volume of the tyre handler. He placed the splash nozzle in the filler neck, went to the control room and switched on the pump.  When he came out, he found the splash nozzle on the ground and diesel fuel gushing out.  He tried to pick up the splash nozzle and hose to move it away from the vehicle and, in doing so, his clothing became partly covered in fuel.  He was standing alongside the vehicle at the time.

 

At that moment, an ignition occurred. The vehicle, his clothing and the ground around him were all on fire. He dropped the splash nozzle and ran to a nearby water source, where he jumped in and pulled off his burnt clothing.  The operator sustained significant thermal injuries but was eventually able to return to work.

 

Calculations showed that the reaction force at the splash nozzle resulting from the 800 lpm discharge overcame the weight (gravitational force on the mass) of the splash nozzle / hose coupling assembly which had been left sitting in the filler neck while the operator went to the pump room. This force caused the nozzle to be pushed out of the filler neck and fall to the ground. In the process, fuel would have been splashed or sprayed over the side of the tyre handler and specifically around the left side engine cover.

 

Potential ignition sources for the escaping diesel fuel were considered and eliminated, including electrical activity in the adjacent pump room and electrical failures within the vehicle.  Static electrical discharge was unlikely in the prevailing weather conditions. Open flame use and smoking activities were ruled out from the CCTV recordings of the incident.

 

The operator said he had carried out a correct engine shutdown prior to refuelling and the relevant switches were found in the off position, confirming his account. This left something within the engine compartment as the only realistic ignition source, and further investigations were needed to work out how this occurred. The machine had previously suffered a turbo timer failure, but this had been replaced before the incident.

 

Examination of the engine cover and side of the vehicle showed that the sides of the cover were fitted with rubber seals but there was an unsealed gap along the lower edge of the cover, through which vapours and droplets could enter the engine compartment.  This space was over a metre long and approximately 14mm wide, and directly beside the fuel tank filler neck.  Directly inside this gap was the turbocharger assembly and associated exhaust components.

 

Temperature measurements taken on an identical machine, under simulated conditions matching the described work history before refuelling, showed that the exhaust and turbocharger surface temperatures continued to climb for several minutes after shutting off the engine, even with the turbo cool-down timer operating correctly. The measured temperatures were well above the minimum required for hot surface ignition of diesel fuel spray or droplets.

 

It was concluded that escaping diesel fuel from the displaced nozzle entered the engine compartment via the unsealed edge of the engine cover, contacted a hot surface and ignited.  The resulting rapidly developing flame front (‘fireball’) extended out from the machine to surround the operator, igniting the liquid diesel fuel on his clothing and the surrounding ground surface.

 

Changes were later made in the refuelling system at the site, including discontinued use of the splash nozzle and fitting of proprietary couplings to all vehicle fuel tanks used there.  Flow rate identification of the hoses was improved. Improved training was implemented to remind operators that some engine surfaces remain hot enough to ignite spilled fuel for several minutes after being shut down.  

​

​