| Greetings,
Below are listed some of the Steel projects done here.
Far from being an uncommon domain, steel manufacturing is one of the most
common domains for us to work in.
The person you should definitely talk to is Ram Josyula (ESIS::JOSYULA).
He is the Steel King Of AI. He is a KE in the US Fellowship Program who has
specialized in Steel projects over the last few years, travels all
over the US and Asia consulting on Steel AI projects, and is an affable
knowledgeable guy.
A US Steel Company
Tools: OPS5, FORTRAN on MicroVAX
Project: Blast Furnace Operations Advisor
Purpose: Provide all Blast Furnace Operators at Steel
Works with expert advice in adjusting and controlling the iron
making process. Provide consistent recommendations for adjusting
process parameters in order to produce a consistent, quality product.
A Korean Steel Company
Tools: Nexpert, Fortran
Project: Blast furnace trouble/inefficiency diagnosis
Purpose: Diagnostic system to isolate causes and suggest solutions to blast
furnace inefficiencies. Prototype is standalone, but finished system will be
realtime linked to data feeds from blast furnace.
Reason: Not dealing with this topic might result in inferior product (steel)
being produced or unnecessary loss of time in operation or danger to blast
furnace operators. Experts are scarce and busy in this domain.
A Canadian Steel Company
Project: Hot Metal Allocation
Purpose: To assist with the decision-making involved in delivering
hot metal from the blast furnacaes to the melt shops.
An aid to optimization of hot metal usage, maximizing yield
with minimal energy losses.
Tools/Languages: OPS5 / Fortran
A Canadian Steel Company
Project: Set Up Assistant for a Five Stand Cold Reduction Mill.
Purpose: To provide set up parameters to minimize wastage.
Reason: Currently set up is done manually by operators of varying
skills. These operators rely on past experience and intuition
to achieve good set up. Once a good initial set up is achieved
an automatic gauge control mechanism takes over.
During the initial set up phase the operator adjusts the mill
while it is running. This produces waste material during the
adjustment phase. The expert system will provide the less
skilled operators with the expertise of the few senior rolling
experts to minimize adjustment time.
Tools/Languages: Nexpert, C
A US Steel Company
Project: Diagnostics and Preventive Maintenance Expert System for
Continuous Caster Water Systems.
Purpose: To reduce caster down-time and reduce maintenance cost.
Reason: To apply AI Technology to
on-line process monitoring, systems diagnostics and the
prediction of system and component failure.
Tools/Language: Gensym G2
A Canadian Steel Company
Project: Assistant Billet Scheduler
Purpose: Schedule billets and lowest-cost steel substitutes in the most
efficient and optimal ways based on chemistry, grade and
priorities. Also, captures the expertise of the scheduling
expert.
Reason: Shorten the scheduling time and increase the accuracy of
present scheduling process.
Tools/Language: OPS5
Dikk
|