The Right Blend

The primary objective in purchasing optimization software is to produce a blended product for sale at the highest possible profit. Several programs are currently on the market that will serve this purpose, but choosing among the various options can be a confusing task. Scrap processors should insist on certain capabilities in a product.

Unless a company only makes one product and will never make more than one product, processors will need a program that can make multiple chemistries at the same time. For instance, if a blend of alloy "A" is needed today, while a blend of alloy "B" will be needed tomorrow, a processor should be able to shadow alloy "B" while making today’s blend. With an inventory having a large selection of alloys, it is impossible to know whether an item will be more profitable in alloy "A" or alloy "B." If both are blended at the same time, and only alloy "A" is posted against inventory, a processor will have assured himself that he will not only make the most profit possible today, but also tomorrow. In the event that more items are received into the inventory before it is necessary to actually make alloy "B," there very well could be an increase in profit but there will never be a reduction. It is not uncommon in systems that make one alloy at a time to make alloy "A" and then, because certain items were used up in that blend, not be able to make alloy "B." This situation will result in lower overall profits and, worse, being late on a delivery.

The number of elements that a program should consider with maximums and minimums is almost always under-specified. The need to add more elements will be one of the first issues that arise as a company starts to use its new software. When considering various programs it is important to determine how easy it is to add more elements. Some software will require a re-write by the vendor. The ideal package should allow the operator to simply add the new element to the inventory and blending specifications.

In blending scrap for furnace-direct sales, the form of the scrap is an important property. Top-line blending software should allow the operator to control the material used in a blend by its physical form. For example, a customer may specify that the blends delivered to it cannot be more that 15 percent turnings and 5 percent grindings. These limits should be easy to meet if the software used allows blending by physical form.

In the scrap metal industry, the need to cut a large solid could result in a significant loss in light of material loss and the cost of cutting. Ideally, a blending program should be capable of "mixed integer programming," which, simply put, means that items in the inventory can be designated for utilization only as a whole. In essence, use all of it or none of it. Without this capability, a blend program may call for 100 pounds of a 300-pound piece of material.

Another important consideration in complex multi-alloy inventories is whether to use an item in a blend or to sell it as itself. If overall profitability is the goal, the ability to make this determination should be part of any comprehensive blending system.

There is a difference between least cost and maximum profit. With many of the big melters paying based on meltdown chemistry, there will be many times when the least-cost blend is not the most profitable blend. A comprehensive blending program should be able to solve both problems. The program should be able to determine profit based on being paid on total weight or being paid on the weight of specific elements delivered.

Scrap processors will invariably ask two questions while making blends: 1) How much weight can I make of this blend? 2) How much weight should I make (sell) of this blend?

A good blending system should be able to answer both of these questions. Both questions require the blending program find a solution with a non-specific final weight. In both cases, the minimum weight must be set at "0" and the maximum weight set at some large number that exceeds the total weight of the inventory. In the first case, the weight of the blend will be used as the function to maximize (the blend will most likely not be profitable). In the second case, the blend should be made to maximize profit. The result will be the weight of that blend that makes the most profit (any more weight, and the profit will start to decrease). The decision on how much weight to make will be the operator’s, but effective software ensures that it will be an informed decision.

All scrap processors have made good buys and bad buys. A good buy will disappear quickly in a dynamic blending system. A bad buy may stay with you forever. Your blending program should allow you to force the use of a specific item through a constrained minimum weight or coerce the use or non-use of an item through a false price for blending calculation only (results should be reported accurately).

Programmers have built many features into their blending software. These are just a few that experience has shown to be necessary to extract the most money from an inventory. Other capabilities could be considered important, but they fall under the topic of "modeling," which is perhaps the most beneficial aspect of computer blending and a topic for another article.

The increased profits that will be realized with computer blending are difficult to quantify. I can only speak from experience, and that experience has been that managers that see the results of an optimized blend compared to what they normally produce demand the implementation of blending software.