Paired Design Treadwear Analysis

 

Abstract

Efficient Paired Design Treadwear Analysis

Introduction:

Presenting an approach that reduces treadwear testing cost: in time, vehicles, and tires; and improves the power of detecting compound treadwear improvements.

Background:

Treadwear testing has always been an expensive proposition due to the uncontrolled variables that increase the uncertainty in test results, thus making it difficult to detect statistically significant wear differences between compounds without increasing the number of test tires and vehicles.

Summary:

This example is a Passenger Tire Paired Design that evaluates the treadwear vs. a control for nine experimental tread compounds using 12 tires and 3 vehicles. The experimental design assumptions, analysis methodology, caveats, and conclusions are reviewed.

The retreaded tire experiment uses 4 different tread compounds on the same tire. The 4-Way-Treads have different experimental compounds for each ¼ of the tread circumference. One of these 4 compounds is always the control.

The Paired methodology subtracts the wear measurements of the Control on each tire from each Experimental Compound within each groove and then tests to see if these differences in wear are statistically greater than zero.

Benefits:

Paired Designs blocks-out the noise or variation due to tire position, vehicle, driver, route, load, inflation, speed, and road and weather conditions during the treadwear testing, thereby increasing the experiments power to detect wear differences between compounds.

The Paired Design provides meaningful compound comparisons by improving the statistical power of discovering and quantifying a real treadwear differences using fewer test tires.