Chris Swartz, John MacGregor
McMasters Chemical Engineering Department is a leader in the
field of process control, and operates the highly successful McMaster Advanced Control
Consortium (MACC). This group has developed empirical modelling techniques which can be
applied to large process databases containing flawed information. The resulting models can
help with monitoring and control, troubleshooting, or product development. MACCs
activities have already led to tangible benefits for steel processes, and the Chair focused on steel related applications of these concepts has fostered
interdisciplinary problem solving. We currently have active work on control and optimization systems for
electric arc furnace steelmaking. There
is potential for a wide variety of work in this field:
- Multivariate analysis to optimize discontinuous multistage
processes. Steel operations generate a wealth of data on process conditions and product
attributes. Database analysis techniques can quickly pinpoint the variables responsible
for quality excursions, identify operating windows and even guide product development.
- Feedback control methods can be applied to complex processes where
uncertainty in the data would defeat conventional approaches. An extension of this is the
"soft sensors" idea where a property may be inferred from other data in the
absence of instruments. These techniques have been used to predict and prevent slab caster
breakouts at Dofasco.
- Process scheduling is another area with major cost implications for
steel operations. Schedules based on historical data have improved the throughput
Dofascos batch annealing shop.
- Multivariate methods are ideal for examining the output of
sophisticated sensors. An obvious use in steel production is to interpret surface
inspection data where multispectral imaging is emerging to overcome the drawbacks of
- Control and Optimization of EAFs is underway.