Honeywell Uniformance R300, Руководство пользователя

Страница: 54 / 60

5 Using the PHD OPC Server
5.4
PHD OPC Data Standards for Client Applications
54
•
Uniformance - PHD OPC Server User Guide
Data confidence versus data quality
Applications programs that insert data into PHD, for example, using PHD Manual Input tags,
should (but are not guaranteed to) follow these conventions. An application may choose to
assign any confidence between -1 and 100 to a PHD data value when it is stored.
A confidence of… Is assigned to this data value…
100% A data value known to be accurate.
0 A data value that is derived, extrapolated or otherwise
"assumed".
50% (degraded confidence) A data value which can be accurately extrapolated by
regression formula using values recently stored in the
PHD database that are known to be valid.
-1
Any error that prevents an RDI from presenting a data
value obtained from the DCS interface, for example, a
communications error, is a No Data condition. This has a
data value of 0.
Out of range and NaN values
By definition, an Out of Range data value is any value that exceeds the configured Range Hi
or Range Low for the PHD tag. With only one exception, all out of range values are
arbitrarily assigned a confidence of zero when inserted into PHD by collectors. In addition,
Out of Range values are clamped to the PHD configured range limit that was exceeded.
The only exception to this convention is NaN (not a number) values collected from the DCS
by the collector. PHD assigns NaN values a confidence factor of –1, the same as No Data.
Data for requested time of NOW
The PHD OPC Server, by default, returns as-stored (or raw) data values obtained from the
PHD Server. Process data values are periodically collected and stored in the history database.
This means that data values are periodic, and a data value returned to a client for a requested
time of NOW may actually be the most recent value, not a real-time current value.