As published in Recycling Today
March 2000

Managing Variation in the new Millennium

By J. Edward Norris

Managing variation in the new millennium does not have to be as difficult as in the last millennium. The technological explosion of the recent past has produced a deluge of PC software programs aimed at managing variation and reducing raw material costs for steelmakers.

Computer-based SPC programs, while simple, can be a highly effective tool for identifying, tracking and understanding variation in the raw materials stream. A wealth of information can be obtained from careful examination of the data collected. A few SPC software solutions are available as Microsoft Excel add-ons.

There are other choices for statistical-software programs designed for after-the-fact raw material analysis. A multiple linear regression program will extend the level of understanding even further than SPC.

With this software, steelmakers can regress melt data and accurately back-calculate the chemistry of each scrap melted. Furthermore, variables like kWh and power-on times can be regressed for each scrap melted.

Melt managers can now ask questions that are more informed: Which material most negatively impacts our power-on time? Which shredded scrap requires the least kWh to melt? Raw material purchases are now driven by fact, not opinion.

Armed with a fundamental understanding of normal material variation and regressed chemical value for raw materials melted, steelmakers may now opt for one of the linear optimization models available today.

These programs typically offer a wide assortment of features coupled with a least-cost charge algorithm. Linear optimization models offer a quick return on investment -- and who isn't looking for that? Successful implementation of computer-based linear optimization systems will create the long-term, bottom line savings sought by steel producers.

So far we have discussed software designed to deal with normal process variation. One other element that must be examined is random causal variation. Software programs cannot reduce exposure to a 400-lb. mill roll bearing buried in a railcar of P & S. Beyond physical inspection or random quality audits there aren't many choices available to protect the operation from this type of event.

Software systems discussed to this point have one thing in common. They lack a feedback loop to the supplier. The scrap producer/supplier is actually the starting point of the steelmaking process.

It is essential that information collected concerning the suppliers' process be shared with the supplier. By sharing information, the scrap consumer and supplier can successfully manage their shared variation

J. Edward Norris is president of Ferrous Solutions, Inc., a Canton, Ohio-based supplier of custom-blended scrap materials.