LOGBOEK  - JOURNAL DE BORD

Reliable shipboard systems

Most shipping accidents can be attributed to failure of shipboard
control systems. Not surprising considering that no-one checks
how well different systems work together in advance

FAILURE of shipboard computer systems and software, including oper­ating systems, control systems and alarms, are most likely the root cause of a fair proportion of shipping accidents.

Class society Bureau Veritas has calculated that 32 per cent of major oil spills / near misses are caused by machinery failure, 34 per cent by naviga­tion error and 15 per cent by fire and explosion.

Ships are fitted with var­ious electronic devices and control systems to make sure the engine keeps running, the steering works, there is enough water under the vessel, the ship is a safe distance from other vessels, there are no explo­sive vapour mixtures or sparks, no flooding in the engine room, and the water cooling and fuel oil systems are working properly.

So why are there so many failures?

The answer is probably that very little research has been done into how well different systems from different manufacturers integrate with each other - the result often being summarised times as "plug and pray" rather than "plug and play".

Shipyards source systems from separate manufacturers, without assessing the risks associated with integration of those systems.

Engineers are given the task of designing products which solve specific problems, but often do not make sure their products work with other suppliers' products.

Many equipment "type approval" procedures do not consider whether the equipment integrates with other equipment - it just tests if the equipment works well running by itself.

Meanwhile individual software-based systems are becoming more complex and more automated - bugs can occur when they are interfaced with other systems which would never show up when they are tested on their own.

When problems do occur, the expertise required to identify the spe­cific problem and fix it is considerable - often involving an engineer from each of the companies which might be involved in the problem, together with the ship superintendent to force them to sit down together and stop blaming each other.

Some of the problems can be attributed to seafar­ers expecting the equip­ment to work better together than it actually does, not appreciating that it has all been put together by engineers not talking to each other in different parts of the world.

For example, a seafarer might expect different equipment to have similar interfaces, or similar shut down procedures, or expect the ship alarm sys­tem to notify them ade­quately of any problem - as all readers of this magazine are too well aware, this is rarely the case.

Another cause of prob­lems is the ship using soft­ware or systems which are unlikely to be supported in 20 years time when prob­lems arise.

Examples of problems

One example of a failure of systems to work well together is the grounding of cruise ship Royal Majesty near Massachusetts in 1997 - the antenna connection into the GPS failed, sending the GPS into "estimation" (dead reckoning) mode - the DGPS set off an alarm because the data from the GPS was not right, but the alarm was not noticed by the bridge team.

The ship carried on what it thought was the right course, but which was actually 17 miles off it, eventually running aground several hours after the initial fault occurred.