Reduce Cost by Reducing Weight?

from the Perspective of Competitive Advantage

by
Edwin B. Dean

----------------------------------------------

[NASA Logo] When I first began to address the question "Why does it cost how much," I asked an engineer how he was going to reduce cost in the system he was designing. The answer, of course, was "I am going to reduce the weight." Alas, the weight based cost estimating relationship (CER) has done us a great disservice. The graph of a weight based CER is a straight line with the natural logarithm of cost (ln(cost)) as the y axis and the natural logarithm of weight as the x axis (ln(weight)). The line is at an angle a1 with respect to the x-axis. The graph corresponds to the equation

ln(cost) = a0 + a1 ln(weight).

Looking at the graph, it is immediately obvious that the way to reduce cost is to reduce weight. Wrong!

What is wrong with this graph is that the second, and most important variable, is implicit in the graph and is not perceived by the engineer. The constant a0 = cmplx is the value of ln(cost) at 1 unit of weight. If a1 is assumed to be the same for all systems then cmplx is the complexity of the given system being measured and exp(cmplx) is the cost of the first unit of weight. The equation should be viewed as

ln(cost) = cmplx + a1 ln(weight)

where the second variable, cmplx, has a value associated with the category being developed.

For example, a structure made of composites has a higher complexity than the same structure made of aluminum, given the same part count. A typical way for the engineer to reduce weight would be to use composites instead of aluminum. The result is that the equation for the composite structure is different from the equation for the aluminum structure. This is graphically represented by two parallel lines which cross the one weight unit axis at different places. The line for the composite structure is much higher than for the aluminum structure. Given the same part count, it is usually the case that the cost of the reduced weight composite structure will be greater than the original aluminum structure because of the height differential between the two lines. Since complexity is a function of part count, the lines may be brought together more closely by reducing the part count in the composite structure. This is usually possible. It still may not be possible, however, to end up with a composite structure which costs less than the aluminum structure.

The weight based CER, as are many other common perceptions of cost, is quite misleading unless fully understood from a cost perspective and from an engineering perspective.

Another case: The aerospace industry usually assumes that the minimization of weight is equivalent to the minimization of cost. I submit that this is not true. The rationale is simple as illustrated by the following oversimplistic example.

Allocate the cost and weight of an aircraft to the three subsystems: avionics, engines, and the rest of the aircraft. Assume that the cost/unitweight of the rest of the aircraft is 1, the cost/unitweight of the engines is 10, and the cost/unitweight of the avionics is 25. These numbers are for use in this example only but are somewhat representative of the cost range in reality. To reduce structural weight by ten and increase avionics weight by 1 to contol the more flimsey structure adds cost because of the relative cost/unitweight between the two. If the weight were equally divided between the three subsystems, then there is no cost incentive to reduce structural weight. Cost could better be reduced by beefing up the less expensive structure and removing the more expensive avionic structural stability controls. Note further than the weight added to contain fuel is relatively inexpensive in terms of aircraft cost. However, it has additional effects on total flight cost through aircraft performance.

The first message is that the type of weight is totally important when considering cost because of the complexity differentials. Weight should be viewed as a nonlinear interpolator for complexity and, hence, cost. The second message is that cost must be modeled appropriately and that the model used instead using of existing paradigms associated with cost. If not, then the cost answer will probably be wrong. If aircraft economics were the major performance criteria, as it usually is today, then wrong answers could be disastrous.

----------------------------------------------

Table of Contents | Cost Technologies | Use

----------------------------------------------

Originated on 940905 | Improved on 951124
Author Ed Dean | Curator Al Motley