Durability Tests Can Hone Competitive Edge

One of the more daring challenges American auto makers give their cars is called the “Pike’s Peak test.” In this test, a car is driven on a road down the Colorado mountain in high gear while the driver keeps a foot on the brake pedal to maintain a safe speed. The test, used by General Motors, Ford and Chrysler, is designed to give engineers an idea of how brakes will perform in extreme situations.

Although auto makers have traveled to Pike’s Peak to test their brakes for several years, the manufacturers are also looking for more sophisticated and comprehensive tests to check the durability of almost all their parts and systems. Computer simulations, for example, can tell--long before actual hardware is built--whether brake systems can pass the Pike’s Peak test. The implications of these tests for improved durability are not limited to auto manufacturers but stretch across a wide variety of industries. As airlines face the challenge of maintaining fleets of older aircraft, for example, the issue of product durability becomes of critical importance. On the ground, companies making high-tech, computerized commercial food equipment know that for restaurants and fast-food outlets, sales are directly tied to the ability of this equipment to reliably deliver menu items on demand.

The reason improved testing methods are needed for cars is simple: State and federal regulations and market forces are driving auto makers and their suppliers to make their products more durable. Beginning in 1993, for example, California and eight Northeastern states are expected to impose stricter standards on emission systems, including warranties of 10 years or 100,000 miles, whichever comes first. The federal government is expected to adopt similar standards by the mid-1990s.

Some car makers have already extended warranties on certain systems to seven years/70,000 miles, and they expect their suppliers to make parts that can also meet these standards.

But as manufacturers and regulators extend warranties, they are learning that many of the parts now in use simply weren’t designed to last that long. How long can they last? Many auto makers and parts manufacturers don’t know because they have no way of accurately measuring the durability of their products--especially new products for which actual operating information is limited or non-existent. But if frequent failures begin to occur during the longer warranty periods, car makers and suppliers could face massive losses. For the entire industry, this is uncharted territory.

In general, durability is difficult to estimate from laboratory, test-track, field or warranty data taken at early stages in a part’s life. This difficulty arises from the uncertainty and variation in the operating environment, including road and weather conditions, the manufacturing process and the properties of the various materials.

(Many auto makers today advertise “high quality” based on consumer surveys that measure the frequency of problems reported by owners. But the survey results cited in most of these ads only cover the first 90 days of ownership, a rather short-term index of quality and not an index of durability.)

More information on long-term performance, as well as better methods for analyzing and interpreting this data, is needed to accurately predict durability.

But the recent trend toward extending warranties for cars and the shrinking time available for development have put heavy pressure on auto makers and suppliers to come up with reliable testing methods that can catch any problems early in the engineering phase. In addition to computer simulations, many companies are increasing their testing of parts taken from used vehicles, employing accelerated testing and using a technique known as statistical design of experiments.

Accelerated testing, for example, is an economical way of estimating the durability of a part in much less time than its normal service life. A road simulator can put as much wear on an automobile in three weeks in the laboratory as would 100,000 miles of driving on the road. Specific accelerated tests, for instance, can measure the durability of pistons and seals in shock absorbers. Each automotive system is unique, however, and the challenge for engineers is to determine the exact relationship between accelerated test results and durability under actual driving conditions.

Statistical design of experiments is a technique for planning experimental programs so they yield credible information in the most efficient way possible. It pinpoints sources of inconsistencies in design and manufacturing, establishes useful performance limits and compares alternative product designs and production processes. This technique helps isolate the factors that have the greatest impact on durability. One supplier used this method to find the best thickness for chrome plating on the shock absorber rods.

Finally, computer simulation, which can replicate the results of the Pike’s Peak test, can also perform a variety of “what if” experiments more quickly, cheaply and safely than they could be done in the laboratory or in the field. Computer models can also simulate combinations of operating conditions and designs that would be too expensive and time-consuming to evaluate with experiments. In addition to brake tests, computer simulations have been used to eliminate thermal fatigue in aluminum alloy engine heads and blocks and to test the control valves for active suspension systems soon to be installed in some of the more exotic cars.

These and other tests will help manufacturers and suppliers estimate durability and warranty costs, design better parts, shorten the product development cycle--and make them more competitive.