The old adage ‘you can lead a horse to water but you can’t make it drink’ is especially pertinent when it comes to instrumentation. Whilst manufacturers put time, effort and ingenuity into developing new ways to enable instruments to gather and display an expanding range of useful information, it would seem it doesn’t necessarily follow that people will use them or pay attention to them.
When it comes to the operation of industrial and municipal processes, it is true to say that the valuable contribution that instrumentation makes often goes un-noticed. As a valuable source of data on what’s happening, instruments are the frontline for ensuring that processes are running in line with expectations. Not only that, but with the inclusion of alarm and diagnostic functions, they are also the prime indicator for warning when things are going wrong.
Despite this, incidents continue to occur problems have arisen from instrument readings either being ignored or misunderstood.
In a recent example, misinterpretation of an alarm reading was found to be the main factor behind the discharge of untreated sewage by a water company. Alarms indicating a potential problem were ignored and not passed on for action, instead being regarded as a momentary blip most likely caused by a probable fault with the instrument. Unfortunately, the alarm was genuine. As a consequence, untreated sewage ended up spilling into a local waterway, which was observed and reported to an environmental inspector.
Luckily, in this example, the outcome of the ignored alarm – though undesirable – was not catastrophic, resulting in just a minimum fine being issued to the company concerned.
In situations where there is the potential for serious injury or endangerment to human life, the ramifications of ignoring what your instrument is telling you can be far more serious.
Complacency over safety was identified as one of the major contributing factors behind the Buncefield disaster in 2005, where failure of an automatic tank gauging system led to a fuel storage tank being overfilled. Problems with this system, caused by sticking of the gauge mechanism, had been known for a while, with 14 reported instances having taken place in the four months leading up to the disaster.
There were also problems with an independent high level system on one of the tanks that should have shut off the tank filling process and sounded an alarm. Although the system itself worked properly, a lack of knowledge over how it worked had rendered it ineffective.
The cumulative impact of this was that in the early hours of December 11th 2005, 250,000 litres of petrol seeped from the overfilled fuel tank, causing a vapour cloud which drifted to other tanks before then igniting. The subsequent chain reaction caused other tanks to explode, resulting in a fire that lasted five days and caused extensive damage to the facility and the surrounding areas as well as significant long-term contamination of the surrounding environment.
Luckily, although around 40 people were injured, no-one was killed. Nevertheless, the financial penalties imposed on the companies deemed responsible were substantial, totalling almost £10m.
In tackling the problem, it helps to first understand why instruments often get ignored. In many cases, there is a mismatch between the perception and reality of a problem, where the consequences of a potential failure are not fully considered.
Taking the example of an unreliable instrument, it is perhaps understandable that persistent nuisance trips might lead someone to eventually ignore them altogether. However, the instrument is there for a reason and the fault may actually be indicative of problems elsewhere other than within the instrument itself. For this reason, the adage ‘if in doubt, check it out’ should always be applied.
The same needs to apply consistently in an industrial context. If an instrument is deemed to be unreliable, then good practice dictates that it should either be repaired or replaced, especially if it’s in a critical application. If an incentive to do this is needed, then consider that UK law allows both companies and their employees to be punished in the event of a serious failure.
Potential human failings also need to be addressed. You can have the best automated system in place with as many fail-safes as you can cram in, but somewhere along the line, someone is going to be needed to perform a task, whether it’s initiating an emergency shutdown or carrying out maintenance and checking on the plant. When this needs to happen, the people concerned need to be properly trained to do what’s needed, with the knowledge, skills and motivation to carry out their role. They also need to feel confident enough to question what an instrument is telling them if they have doubts about a reading or an alarm.
Summary – it pays to pay attention
Ultimately, a decision to ignore what an instrument is trying to tell you will never be risk-free. If the data is not what you expect to see or if there are persistent alarms, then this could be indicative of a problem with either the instrument or the process it is being used to measure. Either way, taking the time to check and confirm the problem will always prove to be the smart decision.