05/09/2016
Hello team.
The OPC UA document states the configuration for Historian 2 as provided below;
For the purposes of all Historian 2 examples:
- TreatUncertainAsBad = True. Therefore Uncertain values are treated as Bad, and not included in the Aggregate
- Stepped Attribute = False. Therefore SlopedInterpolation is used between data points.
- UseSlopedExtrapolation = False. Therefore SteppedExtrapolation is used at end boundary conditions.
- PercentBad = 100, PercentGood = 100. Therefore unless if all values are Good then the quality will be Good, or if all values are Bad then the quality will be Bad, but if there is some Good and some Bad then the quality will be Uncertain.
In the aggregate tester utility of OPC UA, the same configuration has been taken with 1 deviation, i.e. UseSlopedExtrapolation = True.
Consider Historian 2 relevant datapoints near to interval 1:20 to 1:36;
12:01:17 | 70 | Raw, Uncertain | Value is flagged as questionable |
12:01:23 | 70 | Raw, Good | |
12:01:26 | 80 | Raw, Good | |
12:01:30 | 90 | Raw, Good | |
- | No Data | No more entries, awaiting next Value |
The output for 1:20 to 1:36 interval is given below;
12:01:20.000 | 7001 | Good, Calculated, Partial | |
12:01:36.000 | BadNoData |
So if we consider UseSlopedExtrapolation = True, and decide on the last interval 1:20 to 1:36 of DurationGood Aggregate, following is the observation;
- The value interpolated at 1:20 is Bad.
- The value at 1:30 is however should be extrapolated using 1:26 Good and 1:30 Good. Thus if extrapolated value is Good, the region calculation would give duration good value for this interval to be 13001.
So the doubts are;
1. What is the parameter value of UseSlopedExtrapolation for Historian 2.
2. How to do slopped extrapolation for Simple bounding. And what's its quality when extrapolated using 2 good points? Is the above 13001 value correct?
Moderators-Specifications
Moderators-Companion
Moderators-Implementation
Moderators-Certification
Moderators-COM
02/24/2014
OK lets start with the aggregate you chose - DurationGood does not matter if you are using stepped or slopped extrapolation, since it looks at the status/quality of the values only - it is measuring the time (in milli-seconds) that the value was good. So changing the setting does of the extrapolation flag does not affect the results 7001 milli-seconds is all that the value was good in the interval. The definition of the aggregate describes what is use or not used.
1) For all of the sample calculation the value of UseSlopedExtrapolation=false is used for historian #2.
2) If using SlopedExtrapolation for the two points and trying to find the value at the end of the interval =>V2+ (v2-v1)/(T2-T1)*(Te-T2) => 90+ (10/4000 * 6000) = > 105
Paul
Paul Hunkar - DSInteroperability
The other issue you bring up has already been looked at for Historian2 in the Aggregate Test Utility. Next update will have UseSlopedExtrapolation=false.
Regards,
Rod
Rod Stein Manager of Technology Matrikon OPC http://www.matrikonopc.com
05/09/2016
paul.hunkar@dsinteroperability.com said
OK lets start with the aggregate you chose - DurationGood does not matter if you are using stepped or slopped extrapolation, since it looks at the status/quality of the values only - it is measuring the time (in milli-seconds) that the value was good. So changing the setting does of the extrapolation flag does not affect the results 7001 milli-seconds is all that the value was good in the interval. The definition of the aggregate describes what is use or not used.1) For all of the sample calculation the value of UseSlopedExtrapolation=false is used for historian #2.
2) If using SlopedExtrapolation for the two points and trying to find the value at the end of the interval =>V2+ (v2-v1)/(T2-T1)*(Te-T2) => 90+ (10/4000 * 6000) = > 105
Paul
Thanks Paul for the explanation. So we can safely assume that all the outputs of different aggregates related to Historian 2 in OPC UA specification document follow the calculation on the basis of UseSlopedExtrapolation=false. Am I Right?
Regards,
Sid
1 Guest(s)