Jerrod Stutzman created a centralized spatial system called the Spatial Reasoning System (SRS) to store and display Devon Energy's SCADA data. The SRS uses OpenGeo Suite with POSTGIS and GeoServer to store and serve spatial data fast. Apache SOLR is used for search. FME and FME Server are used to synchronize data between different databases. This overcomes challenges with inconsistent location data from SCADA and integrates data from multiple sources for consistent viewing on desktop and mobile apps. Examples shown use FME to transform Excel, Oracle, and seismic data into the standardized POSTGIS spatial database.
2. Project Goal
CONNECT. TRANSFORM. AUTOMATE.
! Create a centralized spatial system for storing
and displaying SCADA (supervisory control and
data acquisition) data
! View data on map via desktop and mobile
! Google-like search and performance
3. Challenges
! Need to see consistent data on desktop web app
and custom iPhone app
! Location information from SCADA database is
inconsistent and unreliable – need to join with
reliable data for mapping
! Need to show on map along with other Devon
assets (wells, pipelines, leases, etc)
! ArcSDE is authoritative spatial database
! Won’t be used if it isn’t fast!
4. Solution
! Spatial Reasoning System (SRS)
! OpenGeo Suite (POSTGIS data served via
GeoServer using GeoWebCache)
! Apache SOLR search platform
! Internally developed web and mobile apps
! FME & FME Server for data creation and
synchronization
! Custom code for certain integrations
5. Why OpenSource
! PERFORMANCE
! Biggest pain point historically is our well dataset.
Over 4 million features + 3 geometries (surface
location, bottom location, directional path)
! POSTGIS + GeoServer render and cache this
dataset quite fast
! POSTGIS can store all 3 geometries on one row!
! Reduced database complexity
! Low cost
6. ! FME is used within the SRS
to create and sync data for
wells, pipelines, leases,
seismic, facilities, etc
! FME Server handles
scheduling and failure
reporting (email)
SRS Integrations Diagram
8. Selected Workbench Example 1
EXCEL Spreadsheet to POSTGIS
! SCADA Radio Tower data is maintained in Excel
! Updated quarterly. Full dataset update
! GPS Location and Range need to be shown
! Need to store 1 to 3 geometries for each tower
11. Selected Workbench Example 2
Non-Spatial Oracle to POSTGIS
! Non-spatial well production data from Oracle DB
is synchronized with POSTGIS
! Workflow:
Retrieve last
run time from
POSTGIS
Query
ORACLE for
changes
Query
POSTGIS for
matches
Compare to
determine
INSERT/UPDATE
Determine
DELETES
Update
POSTGIS
DB triggers
update
SOLR
14. Selected Workbench Example 3
Seismic Data from Oracle DB
! Seismic data can be represented spatially as lines
or polygons
! This is stored in the seismic DB as a coordinate list
of vertices (numbers only, non-spatial DB)
! This data needs to be ‘spatialized’ and stored in a
spatial database – synced nightly (incremental
updates)
! Complexity Reduction – store all data (2 different
geometry types) in same table
17. Future
! Ability to do an “UPSERT” DB transaction would
greatly simplify some of the logic required to
determine DB transaction!
! ARCSDE30 Writer actually had this capability, if it
did not find the record to update, it would insert
! Need to revise old workbenches to use best
practices and gain efficiencies
! Scaling SRS to handle more data and users
18. Thank You!
! Questions?
! For more information:
! Jerrod Stutzman
! Jerrod.Stutzman@dvn.com
! Devon Energy Corporation
CONNECT. TRANSFORM. AUTOMATE.