Air Start Explosion On The Capetown Castle
Built in 1938, she was powered by two 10 cylinder Harland and Wolff double acting two stroke cycle diesel engines 660mm bore and 1500mm stroke.
As normal with large marine diesel engines, compressed air was used to start the engines by admitting it into the cylinders in sequence using main air start valves, the timing being controlled by camshaft operated pilot valves (see how does an engine start?). After starting the system was vented through 3/8 " (10mm) copper drain lines.
Because some of the air start valves were defective and not seating properly, products of combustion and unburnt carbon leaked past the valves and used to choke the drains on the air start venting system. This had led to a routine being established to clear the choked drains before arrival in port. However, no instructions had been issued on how this work was to have been carried out, and was generally left to the junior engineers.
A practice had grown up to connect the hydraulic pump used for the tie bolt jacks to the drain lines and force lubricating oil up the drain lines to clear the blockage. If the drain lines had been disconnected, then no harm would have been done, but on the occasion in question some oil must have been forced into the air start line.
On the 17th October 1960 at 04:41 the engines were rung to stand by for arrival Las Palmas. The Chief Engineer was at the controls of the starboard engine and the 2nd Engineer at the controls of the port engine. (This ship was built in the days before control rooms; the engines were manoeuvered locally, and extra engineers were present in the engine room to operate the "handamatic" equipment).
After slowing down, "Stop" was rung at 04:48, followed by "Half Astern" at 04:49. Seconds later an explosion occurred and a sheet of flame swept through the engine room killing the Chief, First, Snr and Jnr 2nd, two Junior Engineers and a greaser.
At the formal investigation which followed it was established that the explosion started in the air start line in the port engine. For an explosion to occur, that together with the air, there must have been oil and a source of ignition. Expert witnesses agreed that the initial explosion was due to the presence of about 4 fluid oz (110cc) of oil - about half a teacup. It was agreed that this oil probably came from the hydraulic pump used to clear the drain lines.
The Air start pipe lines were coated with a film of oil carried over from the compressors. It was agreed that this film of oil did not cause the initial explosion, but was contributory to the escalation of the incident as described below.
"Oil, in quantity of at least four fluid ounces from the force pump that was used for clearing drain pipes was retained in the air starting system of the port engine, in or near to an air starting valve. When the port manoeuvring lever was moved to "start" the air start valve opened and highly turbulent air entrained the oil, sweeping it into the cylinder, acting to some extent like an air blast atomizer. The mixture of oil and air ignited as a result of high temperatures in the cylinder at some point remote from the starting air valve. This ignition first expelled a mixture of unburnt oil and air into the open line through the valve, then the explosion flame propagated from the source of ignition through the cylinder and out through the starting valve into the pipeline, there consuming the explosive mixture that had just been expelled. The flame accelerated and initiated a film detonation involving compressor oil in the main air pipelines. The detonation waves when reflected at the extremities of the system, or at T junctions, produced very high instantaneous pressures (in some way akin to a "water hammer" effect) and caused severe damage at these places. Rupture of the connections to the port aft air receiver caused the air in the receiver to be discharged directly at the back of the port engine, passing between the cylinders and swirling down over the manoeuvring platform. Flames associated with this discharge caused the casualties in the engine room. Fires in the generator room were likewise caused when the port forward air receiver connections were ruptured." (report of court No 8022)
The fires were eventually extinguished by the use of CO2 gas.
Following the formal investigation an M notice was issued (No M474) which recommended that: