Data integration including spectroscopic Data: - Vision and benefit of an overdue issue

Apr
10
2018

Analytica Conference > GDCh / Room 3 > From Big Data to Smart Data

10. Apr 2018
11:00-11:30 h  |  ICM - Internationales Congress Center München Room 3

Additional informations in detail view:
o Attachments
o Presscontact
o Alternative Events

Details

Subjects: 

Analytic

|

Digital Transformation / Smart Laboratory

Speaker: 

Dr. Matthias Odenweller 

(EvonikTechnology & Infrastructure GmbH)

Type:
Lecture
Speech: English

Within the chemical industry, spectroscopy is a widely used tool. In most cases, however the implementation into the process automation is via a dedicated and isolated system, that operates the spectrometer, calculates the predictions and communicates these predictions to the control System. Often this communication is realized via analogue signals and Bus systems in more sophisticated applications. The actual spectral data, however are not transmitted. They are saved as isolated files on a more or less isolated computer. Any aggregation or integration has to be done separately, in most cases manually. Because of this isolation only a very small part of the collected data actually can be used. Using only the build predictions makes the spectrometer a mere sensor far behind its actual potential.
The talk will focus on the opportunities that arise, if these spectral data are collected automatically and stored as spectral data in a database. Within this Database the information can be combined with Laboratory results, process parameters, SAP and further specific Information from the spectroscopic system. Envisioning such a setup, multiple benefits become apparent:
From the spectroscopy point of view, the external data, for example necessary reference values to enable the modelling of a prediction model, can be assigned easily or even automatically to the individual spectra. The required resources to build the model can be significantly reduced and errors by merging the data can be avoided.
Additional utilization of self-learning algorithms even give way to the vision of a completely automated calibration. Furthermore, similar applications can be combined and aggregated to have a suitable model for new applications based on similar existing applications.
On the other hand, also the process itself can benefit from spectral data directly accessible. Data mining could be applied using normal process data as well as the rich spectral data to predict properties of the process. Instead of predicting mere concentrations, patterns can be correlated or new indicators for an optimized process can be derived.

Speaker
Dr. Matthias Odenweller, EvonikTechnology & Infrastructure GmbH Show Profile

Location
Outdoor areas
EntranceNorth-East
EntranceNorth-West
EntranceNorth
ConferenceCenter Nord
ICM -InternationalesCongressCenterMünchen
B0
C1
C2
C3
C4
C5
C6
B1
B2
B3
B4
B5
B6
EntranceWest
Atrium
A1
A2
A3
A4
A5
A6
EntranceEast
EntranceNorth-West
EntranceNorth
Outdoor areas
EntranceNorth-East
ConferenceCenter Nord
ICM
B0
C1
C2
C3
C4
C5
C6
B1
B2
B3
B4
B5
B6
EntranceWest
Atrium
EntranceEast
A1
A2
A3
A4
A5
A6
Apr
10
2018

Analytica Conference > GDCh / Room 3 > From Big Data to Smart Data

Machine Learning zur Prozessanalytik am Beispiel hyperspektraler Bildgebung

10. Apr 2018
10:30-11:00 h  |  ICM - Internationales Congress Center München Room 3
Speaker: Dr. Jan Makowski (LuxFlux GmbH)

Additional informations in detail view:
o Attachments
o Presscontact
o Alternative Events

Details
Apr
10
2018

Analytica Conference > GDCh / Room 3 > From Big Data to Smart Data

Cloud Computing – The next generation of material scanning solutions

10. Apr 2018
11:30-12:00 h  |  ICM - Internationales Congress Center München Room 3
Speaker: Prof. Dr. Dominik Rabus (Spectral Engines)

Additional informations in detail view:
o Attachments
o Presscontact
o Alternative Events

Details
Apr
10
2018

Analytica Conference > GDCh / Room 3 > From Big Data to Smart Data

Big Data Approaches for Process Sensors - First Experiences and Challenges

10. Apr 2018
10:00-10:30 h  |  ICM - Internationales Congress Center München Room 3
Speaker: Prof. Dr. Christian Decker (Hochschule Reutlingen)

Additional informations in detail view:
o Attachments
o Presscontact
o Alternative Events

Details