Partial status code calculation|OPC UA Standard|Forum|OPC Foundation

Avatar
Search
Forum Scope


Match



Forum Options



Minimum search word length is 3 characters - maximum search word length is 84 characters
Lost password?
sp_Feed sp_PrintTopic sp_TopicIcon
Partial status code calculation
Avatar
Siddhartha Ray
Member
Members
Forum Posts: 8
Member Since:
05/09/2016
sp_UserOfflineSmall Offline
1
05/09/2016 - 09:09
sp_Permalink sp_Print sp_EditHistory

Hi all.

Lets consider a scenario for partial status code calculation based on aggregate, Count.

Say the input data is as follows (provided in OPC UA doc for historian 1) with the parameters Start time: 12:00:00 End time: 12:01:40 Interval: 00:00:16

TimeStamp - value - statusCode

12:00:00        _       BadNoData

12:00:05        _       BadNoData

12:00:10        10     Good

12:00:15        15     Good,Interpolated

12:00:20        20     Good

12:00:25        25     Good,Interpolated

12:00:30        30      Good

12:00:35        35      UncertainDataSubNormal, Interpolated

... (end time stamps concerned with  partial calculation)

12:01:15        75     UncertainDataSubNormal, Interpolated

12:01:20        80      Good

12:01:25        85     Good,Interpolated

12:01:30        90      Good

12:01:35        90      UncertainDataSubNormal, Interpolated

The Partial status bit calculation says: SET SOMETIMES, If an interval is not a complete interval.

The ouput is:

TimeStamp - value - statusCode

 12:00:00       1       Good, Calculated,Partial

12:00:16        2       Good, Calculated

... (end 2 time stamps concerned with partial calculation)

12:01:20        2       Good, Calculated,Partial

12:01:36        2       BadNoData

 

As per the above output given, the partial bit is considered for calculation based on Start time and End time by checking if the data is available at the start time and the end time.

But as the definition and calculation of partial bit is concerned (SET SOMETIMES, If an interval is not a complete interval), partial bit calculation should be taken care of in every processing Interval.

So if we go as per the definition of partial calculation; For the interval 12:00:16, the output should be;

12:00:16        2       Good, Calculated, Partial  

because our input has no data at 12:00:16. (Note 1st interval is from 12:00:00 to 12:00:16 and the next from 12:00:16 to 12:00:32).

Please explain how the output calculation given in the OPC UA document holds correct or am I missing anything here.

 

                               

Avatar
Paul Hunkar
Cleveland, Ohio, USA
Moderator
Members

Moderators-Specifications

Moderators-Companion

Moderators-Implementation

Moderators-Certification

Moderators-COM
Forum Posts: 112
Member Since:
02/24/2014
sp_UserOfflineSmall Offline
2
05/12/2016 - 01:00
sp_Permalink sp_Print

Siddhartha,

"Partial bit is used to indicate that the interval is not a complete interval and that a client may receive a different value for the Aggregate if it re-fetches the interval with the same parameters." 

  Partial is used if the complete interval does not have data available, the first interval is for 12:00:00 - 12:00:16 which is missing values, in that the historian was not running prior to 12:00:10.  Someone could fill in data from before the historian was running, thus it is considered partial.  The second interval is from 12:00:16- 12:00:32 - in this interval there is data for the entire interval, the historian was running and collecting data. The result would not change if it is being recalculated later (historian filled in more data at the end - or maybe earlier data is added).  It is important to note that the time stamp is the timestamp of the start of the interval, for 12:00:16 is the second interval.  

Partial is much like bounding in that if there is a value before the interval and one after the interval than the interval is covered and it would not be partial.  

Paul

Paul Hunkar - DSInteroperability

Avatar
Siddhartha Ray
Member
Members
Forum Posts: 8
Member Since:
05/09/2016
sp_UserOfflineSmall Offline
3
05/15/2016 - 23:36
sp_Permalink sp_Print

Hi Paul.

Thanks for clearing the point.

Regards,

Sid

Avatar
SambsivaReddy Appireddygari
Member
Members
Forum Posts: 24
Member Since:
02/12/2016
sp_UserOfflineSmall Offline
4
05/26/2016 - 00:31
sp_Permalink sp_Print

Hello Paul, 

I have an extended question on the same topic.

Your explanation of Partial is perfectly fine if i have the visibility of entire historian data.

As an aggregate library developer, I will be processing one interval at a time. Some times I might not have knowledge about the data of previous interval and later interval . In this case how can I evaluate the partial bit for the current interval if it does not have start and end bounding values?

 

Thanks & Regards

Samba

Avatar
Paul Hunkar
Cleveland, Ohio, USA
Moderator
Members

Moderators-Specifications

Moderators-Companion

Moderators-Implementation

Moderators-Certification

Moderators-COM
Forum Posts: 112
Member Since:
02/24/2014
sp_UserOfflineSmall Offline
5
06/07/2016 - 00:26
sp_Permalink sp_Print

Samba,

  If you are generating a generic library for calculating aggregates - one in which the data is also provided to the library call, then you can not determine "partial" unless the caller provides the information (i.e. was the data source complete for the interval).  Most implementation I have seen are part of a historian or subscription service.  The historian has knowledge as to when a collection was running so the partial bit is easy to set.  A subscription does not return partial ever, since it starts on a boundary and end at the end of the boundary and thus always has data for the entire interval.

Paul

Paul Hunkar - DSInteroperability

Forum Timezone: America/Phoenix
Most Users Ever Online: 510
Currently Online:
Guest(s) 36
Currently Browsing this Page:
1 Guest(s)
Top Posters:
Forum Stats:
Groups: 2
Forums: 10
Topics: 1435
Posts: 4855