Digital Protocol Diagnostics and Risk Analysis

In an earlier post, Wired Versus Wireless Risk Analysis for Process Instrumentation Measurements, we highlighted the fault tree methodology for performing risk analysis on wired and wireless instrumentation.

Emerson’s Jonas Berge added a thoughtful comment which I wanted to bring out and share in this post.

Jonas Berge Director, Applied Technology

Jonas Berge
Director, Applied Technology

I attended their workshop. It was great. At 9:30 in the video Ed explains: “failures in communication systems are generally overt” which, as shown in the slide is generally safe, and he continues “the same is actually quite true of a wireless system because with wireless systems the safety, if you will, is not in the medium by which the signal is sent, it is in the protocol for sending and receiving that signal”. He goes on: “if the data received by a final element or any receiving device is not timely, and is not valid, it is very easy to detect”. The ability to diagnose the digital communication to detect errors is an important advantage of digital protocols (fieldbus or wireless) over hardwired systems (4-20 mA or on-off).

At 11:00 Ed explains partial circuit failures: there are also partial circuit failures which is kind of a wrong measurement as opposed to loss of measurement, this is going to be the result of loose or corroded screw terminals or ground faults that are partially dragging down the circuit, so for an analog circuit you might have a 4-20 signal that because of some sort of impedance on the circuit the transmitter might be trying to send 20 mA but as a result of excessive impedance you might only get 16 or 18 mA at the receiver.

Fault-tree-analysisThe slide shows partial circuit failures are undetected (covert) which as explained in the earlier slide is generally dangerous. Such analog signal problems are hard to detect because they are ‘on-scale’ meaning they are wrong, but still within the range of 4 mA and 20 mA so they appear to be correct to the system even though they are invalid. The slide shows such increased impedance may be due to loose or corroded screw terminals, or it could be ground fault. Water ingress creating a partial short is another common cause.

That is, it is easy to detect a problem with a digital signal whereas a problem with an analog signal may go undetected. Fortunately logic solvers internally use digital communication with the I/O cards. And now we are talking about digital communication with sensors and actuators.

Another great workshop along similar lines was the 2-1810 “On-Scale Failures – What You Don’t Know Can Hurt You” which also talks about the dangers of ‘on-scale’ failures such as Ground loop that causes the analog signal to wander, Water in the electrical house creating a partial short, and Loop over loaded – exceeding required maximum loop load etc. for which they have 3-4 great short little video clips.

This includes voltage starvation due to too much loop load or unexpected loop load increase from poor electrical termination or corrosion, cable damage. This may result in that the transmitter won’t be able to go into alarm or higher mA value – which could be dangerous. They also note that this is the nature of analog and that use of a digital or wireless protocol can prevent the problem. The workshop also suggests diagnostics for electrical integrity such as “power advisory”, or it could be through digital communication statistics. It concludes that without diagnostics, On-Scale failures result in an undetected and inappropriate control actions.

You can connect and interact with other process safety professionals in the Safety Instrumented Systems group of the Emerson Exchange 365 community.

One comment so far

Leave a Reply