How Random is random failure?

In safety assessment there are two types of failure assessed.? But which is more likely to happen?? There are multiple views on this and multiple definitions.? Let’s look at the IE 61508 versions:

• Random Hardware failure:? "failure, occurring at a random time, which results from one or more of the possible degradation mechanisms in the hardware."
• Systematic Failure: "related in a deterministic way to a certain cause, which can only be eliminated by a modification of the design or of the manufacturing process, operational procedures, documentation or other relevant factors."

(source: IEC 61508-4:2010 https://webstore.iec.ch/publication/5518)

Random Failures

Much focus is given to the prevention of random hardware failure.? But is this in vain.? Random is defined on the OED as:

"Having no definite aim or purpose; not sent or guided in a particular direction; made done, occurring, etc, without method or conscious choice..."

(source: https://www.oed.com/search/dictionary/?scope=Entries&q=random)

By this definition therefore random cannot be predicted.? Therefore can you reduce what cannot be predicted?

If a failure can be reduced it is by definition predictable to an extent.? Therefore at least a part of the failure is systematic.

Let’s take the example of a fault within a cast metal structure.? The fault could be seen as a statistical anomaly.? If could also caused by sub-standard foundry methodology or fault detection which is not suitable for the application.

The first is random.? It will happen given enough units.? However a fault such as this may be detectable in time.? Therefore, if the integrity of a safety system relies of this foreseeable failure the only way to reduce the impact is detection.? The failure rate cannot be reduced.

Systematic failures

Systematic failures are often complex and involve multiple levels of failure.? For instance, the casting above may have been due to a small foundry error, an error in installation and an error in maintenance and detection combined.? These are all systematic as the whole fault can be traced but are very complex.

Systematic failures may occur at any point and be cause by any or multiple people.? Failure to communicate between teams is a classic systematic failure.? As is failure to read the installation instructions…

Random or Systematic

In conclusion before classing a failure as random or systematic.? Ask a question, could this be avoided in the future?? If yes it is systematic.

In this case in the vast majority of failures are systematic in nature.? There is a great tendency to class complex systematic failures as random failures.? This is because the chain of events is difficult to understand, not because the failure is truly random.