Learnings from Monitoring System Toolkit - MOST

[Robert Zach <robert.zach@tuwien.ac.at>]

Dear oBIX list,

here is some information about our work (http://most.bpi.tuwien.ac.at/) and the relation to oBIX.

In the last 4 years we implemented something we call the Monitoring System Toolkit - MOST.

Its about collecting, storing and preprocessing building related data on the management level.

Some code is stable and tested, some code is more like a prototype. See "git clone http://projects.bpi.tuwien.ac.at/git/most-open-source.git".

At present, we implemented a oBIX, a OPC Unified Architecture, and a custom XML/REST interface to data preprocessed in MOST.

Data can be collected from various building technologies or calculated within in a "virtual datapoint" on the fly.

Please see the attached paper for the "big picture".

I think the following "learnings" are of interest for future oBIX development.

A) "SLIM BIM" in oBIX

We use a kind of "SLIM BIM" to group sensors/actors into zones. If a BIMserver is available, we plan to load this information with BIMSie.

We implemented the "browsing" of zone and points with oBIX in the following way:

user: "mostsoc"

pass: "demo12"

listing "head"/"top" zones:

http://demo-most.bpi.tuwien.ac.at:8080/most-obix-1.0.0-SNAPSHOT/obix/zones/

reading one specific zone:

http://demo-most.bpi.tuwien.ac.at:8080/most-obix-1.0.0-SNAPSHOT/obix/zones/10/

getting points within a zone (change "level" to get points in subzones too):

http://demo-most.bpi.tuwien.ac.at:8080/most-obix-1.0.0-SNAPSHOT/obix/zones/10/data?level=1

B) HistoryRollup - timely weight values of a HistoryRollup

It often makes sense to timely weight average values within a HistoryRollup.

Example: interval 1hour -->

In the first 5 minutes the temperature was 20C. The next 55 minutes the temperature was 23C.

Therefore the timely weighted average should be:

(20*5 + 23*55)/60 = 22,75C

C) "EnterpriseHistoryRollup" (HistoryRollup with support for modes in requests)

We extended the HistoryRollup contract with support for requests with different "modes". See chapter "DATA PREPROCESSING" in the attachment.

Unfortunately, our oBIX implementation of HistoryRollup with modes is currently broken.

We still need to refactor the oBIX implementation and bring our implemented features into additional contracts.

Here is a link to our "proprietary" REST/XML interface to get a idea of what it is all about.

mode 1:

http://demo-most.bpi.tuwien.ac.at:8080/most-rest-1.0.0-SNAPSHOT/rest/dp/occ6/periodicdata?from=2011-05-04T14:40:00&to=2011-05-06T10:00:00&period=3600&mode=1

mode 2:

http://demo-most.bpi.tuwien.ac.at:8080/most-rest-1.0.0-SNAPSHOT/rest/dp/occ6/periodicdata?from=2011-05-04T14:40:00&to=2011-05-06T10:00:00&period=3600&mode=2

quality = 1/count (number of measurements in this period)

value = is calculated differently based on requested mode

See figure 3 to 7 in the attached paper for details about modes.

If you want to know where the demo data comes from, here is a link to a web interface implemented with GWT:

http://demo-most.bpi.tuwien.ac.at:8080/most/

The building model used in the "3d-viewer" has currently about 90MByte. You need a fast internet connection and some tome if you want to try it.

We worked on "on demand" loading of object (like zooming into google maps) together with the bimserver.org people, but haven't integrated it yet.

Best Regards,

Robert Zach

-- 
Department of Building Physics and Building Ecology 
Vienna University of Technology 
Karlsplatz 13 
A-1040 Vienna 
Austria 

Univ.-Ass. Mag. Dipl.Ing.(FH) Dr.techn. Robert Zach 
mob. phone.: +43 664 2314337 
phone: +43 1 58801 27038 
fax: +43 1 58801 27093 
email: robert.zach@tuwien.ac.at 
web: http://www.bpi.tuwien.ac.at 

Attachment: MOST.pdf
Description: Adobe PDF document