One of the aims of this case study is to show how to
build monitoring and diagnostic applications on top of
the RapidBase platform. RapidBase is a relational
database management system with temporal extensions and a
versatile trigger system for reacting to database
updates. RapidBase databases are located in main memory
instead of the disk which makes database operations
faster than typical with disk based database systems.
Another aim has been to experiment with different feature
extractors to be able to detect different fault
situations.
Our experiences indicate that at least in well-known
situations visual inspection may be used to find
diskriminatory features capable of diskriminating between
different machine states. The Matlab programming
environment provides a large set of mathematical and
signal processing functions that may be used to
experiment with various feature extractors. Matlab
compiler provides a bridge between modeling and
operational usage by converting the Matlab functions into
C++ code that may be embedded into one's own system
architecture. We have demonstrated this by using the
converted Matlab functionality in the RapidBase triggers
to create user defined functionality.
The performance of the pilot system has also been studied
and it appears that with the tested standard laptop
computer at most 130 records (with 7 floating point
values) of raw measurement data could be processed per
second by the test application and the suggested analysis
algorithms could still be run in real time mode as
database trigger applications.
| Original language | English |
|---|
| Place of Publication | Espoo |
|---|
| Publisher | VTT Technical Research Centre of Finland |
|---|
| Number of pages | 15 |
|---|
| Publication status | Published - 2004 |
|---|
| MoE publication type | D4 Published development or research report or study |
|---|
| Series | VTT Information Technology. Research Report |
|---|
| Number | TTE1-2004-52 |
|---|