Understanding and steering of metallurgical processes Dr. Sander Arnout – InsPyro Inspiration Day 4/12/2015
Understanding and steering of metallurgical processes
Dr. Sander Arnout – InsPyro Inspiration Day 4/12/2015
2
Merge into one operating framework
InsPyro’s vision on knowledge Two types of knowledge
Essential to run a process Control often depends on individual Changes by trial and error Mechanisms unclear Experience transfer is difficult
Essential to be in control Control depends on model Changes are based on physics Mechanisms are explicit(ly assumed) Transferrable
Experience: knowledge on how to run a process
Insight: understanding the science of a
process
3
Process development approach Stepwise process of increasing knowledge and experience:
1. Idea (opinion)2. Concept from literature or experience3. Process model to define expected working area4. Economic evaluation5. Lab or pilot scale experiments6. Validate process model, benchmarking7. Scale-up or adjustments
Innovation isn’t random but a structured approach, learning from failures Fact-based decision on the road ahead
4
Process improvement = development Lots of innovation happen on existing facilities
Increased energy efficiency Increase of input from secondary streams Increased complexity
Lots of potential in using existing processes optimally Nobody develops a process without a model, yet several processes are
run without an explicit model Use the same innovation approach:
Learn from process behaviour, history, trends, including mistakes and gut feeling Build on laws of physics to structure the chaos and avoid relying on opinions
5
Survey on data use in metallurgy Level of data management in the organization
Data collection is minimal
We collect lots of data but don't use it much
We use data for analysis but it requires a lot of effort
Data analysis is easy but it is difficult to draw conclusions
Data and analysis provide input for decisions with some effort
Data and analysis provide input for decisions in a structured and automated way
0% 10% 20% 30% 40% 50% 60%
6
Survey: data useWe lose time combining and cleaning data
YesMaybeNo
All kinds of data are stored in the same format or location
We have good tools for visualization
Some important data is only stored on paper
Some important data is not collected
NO
NO
YES
MAYBE
7
Survey: process There is a lot of fluctuation without a clear cause
YesMaybeNo
Process understanding is mainly in the brain of the people
We stick to known recipes to avoid problems
The process results can be predicted
The process is regarded as a black box
MAYBEYES
YESMAYBE
NOYES
Metallurgy & Business Intelligence ProOpt combines metallurgical insight with data management
8
9
ProOpt goal: increase value creation
World Class optimisation and control system for the process, melting and mining industry
Info.base: data information system Secure availability and quality of data when
you need it Reporting.base:
KPI’s, process and economical information available at your finger tips
Model.base: Process optimisation based on dynamic
modelling and statistical analysis – measure, monitor and optimise your process
Remote control room: Updated Experts available online
ProOptRemote Control
Room
ProOptModel.base
ProOptInfo.base
ProOptReporting.base
ProOptControl System
Expected impact of ProOpt system Engineers spend time on making improvements
– not on finding and checking the data Optimize feed mix to reduce fluctuation in
process and cost per produced unit Better understanding of process reduces
mistakes – makes complex plants manageable Wide insight in critical factors – also by
operators, management, purchasing Feed forward function reduces critical
happenings
Go beyond insight and optimise value
10
Value creation
Numbers Information Analysis Fact based decisions
Management
Purchasing
R&D and Engineering
Operation
11
ProOpt International: contact detailsLausanne office Leuven officeAvenue de Sevelin 6B Kapeldreef 60Lausanne 1007 3001 LeuvenSwitzerland Belgium
Dr. Sander [email protected]+32 16 298 491
12
This presentation was part of the seminar
Data Management and Fact-Based Decision Making in Metallurgical Operations
4th of December 2015 – Leuven, Belgium