Hello to all,
I'm doing an analysis of a specific event that includes a trip of a small plant. The control system is PCS 7 v 9.0. I don't have much experience with this particular system even though I know, it is specifically designed for process control in industry and it has all the feature of a modern DCS system.

I am in one particularly interesting and not easy explainable situation, where I can see a time difference (about 10 sec or more) between the timestamps recorded in the message lists (event, alarms, etc...) when the trip actually occurred and the timestamps I can see in the analog trends diagrams.
It is about the electrical measurements (currents, power etc), so when trip occurred and the circuit breaker is open, all these values must instantly go to zero, which is exactly what trends show. This is shown in the attached file and I can see the time instant when trip happened.
On the other hand, when I look event list in there are recorded messages such as Circuit breaker feedback OFF at about 10 sec earlier. This is a huge time difference, especially because of electrical values.
How to explain this?
Did anyone have similar issue?
I know this could be possible if for example a vale is shut down and there is a process quantity such as temperature or pressure that needs some time to rise or drop, but in case of electrical quantities such as current, it should be in the same sec.
According to the data from the PCS7, it seems hat currents flow 10 sec longer after circuit breaker is tripped. There are many binary signals in the event list such as "protection activated" cca 10 sec before analog trends show values suddenly go to zero.

 

What's the sampling/logging rate for trend data? It makes a difference.

Suppose it is a 10 second sampling/logging rate, not an unusual logging rate for analog data, in fact faster than many DCS historians.

The sampling/logging of a data point could occur just a couple mS before the breaker opens. But that data point will show current flowing, because current is flowing up until the breaker opens.

10 Seconds elapse before the next data point is sampled. This data point shows no current with a time stamp 10.0 seconds after data previous point, but 9.987 seconds after the breaker opened.

Adjust the timing in the example above for whatever the sampling/logging rate is for your system.

This is how an "apparent" delay can appear in sampled data,it reveals the fact that there is missing data in the unsampled time period.

This is not unique to PCS7; it's the nature of time sampled data.

 

Good point David,
do you know where in the settings I can check the sampling / logging rate?

 

No, I don't know. I've no exposure to PCS7, other than noting that PCS7 has its own "conference" on Siemens user forum hosted by the Siemens web site. Whereas most DCS's do not publically publish any documentation, there appears to be some documentation available for PCS7.

My impression for DCS's is that sampling rate for time-based, cyclical data logging or archiving is a "property" of the analog input tag in question. So the sampling rate is not part of the historian's configuration but part of the analog input's properties. Historians typically configure at a higher level than individual data point acquisition rate, but that's just my impression; in the end it all depends on how PCS7 works.