1. Efficient Paired Design Treadwear Testing and Analysis Leighton Spadone +1 440 264 3638 Avrohom.Spadone@Gmail.Com

2. Objective Reduce treadwear-testing cost by reducing the number of test tires, vehicles, and the mileage needed to detect statistically significant treadwear improvements.

3. Paired Designs When two experimental treatments experience the same conditions, like experimental tires paired together by axle or when the same tire is tested under different conditions, like on road and on wheel, these are examples of Paired Designs.

4. Challenge Treadwear testing can be an expensive proposition because uncontrolled variation increases the noise in the test results and makes it difficult to detect significant wear differences. The Sea of Noise: driver, route, vehicle, tire position, inflation, load, road, speed, alignment, pavement, ambient temperature, and precipitation.

5. 12 Tires: Control A  4 Tires: A, B, C, D 4 Tires: A, H, K, L 4 Tires: A, E, F, G Block Variation: 4-Way Paired Design Compounds on each tire see the same test conditions.

6. Test Method Stop/Finish: 16K Miles, 25,750 Km Rotate Tires: 1K Miles Rotate Vehicles: 4K Miles 24 Tread Depth Measurements per Tire: –Four Compounds @ 6 Grooves –288 total data points

7. Sample Data: Remaining mm 12 Tires, 4 Each: ABEH, ACFK, ADGL Compounds, 6 Tire Grooves

8. Assumptions The wear differences measured are due to the compound formulation. –The test tires provided uniformity of compound, cure, and footprint shape. –The tread depth measurement system was accurate and repeatable. –Tire alignment, camber, toe-in toe-out, rims, etc. were uniform across vehicles. –The tire rotation schedules were followed.

9. Caveats This experiment test was stopped at 16K miles or 25,750 kilometers. Extrapolating these wear improvements to higher mileages is not reliable. This is a screening experiment designed to find promising tread compound candidates for further study. This analysis does not “prove” that any compound is better than another except in regards to the control compound.

10. Wear Process Analysis Is the wear process stable and predictable? Control Chart each Compound-Groove Shoulder Groove 1

11. Center Line Groove 3 Intermediate Groove 2

12. Intermediate Groove 5 Center Line Groove 4

13. Shoulder Groove 6 Control Charts show the wear process to be stable and predictable for all grooves and tires. All wear measurements are within the upper and lower red control limit lines.

14. Paired Comparisons are powerful The Paired methodology subtracts the wear measurements of the Control from each experimental compound by tire and tests to see if these differences are significantly greater than zero. This Single Mean type comparison has increased statistical power because it does not rely on pooling variance across the different experimental treatments.

15. Compare Each Compound-Groove 54 paired comparison’s in 4 tire groups. Significant wear improvements are highlighted. Prob>t < 0.05 or 20:1 Odds or Better CompoundCompound

16. Compare Compound–Center Grooves 9 paired compound comparison’s using n=16. Significant wear improvements are highlighted. Prob>t < 0.001 or 1000:1 Odds or better CompoundCompound

17. Mean Wear Improvements: G2-G5 This ranking does not imply any significant difference between experimental compounds.

18. Summary The wear process was stable and in-control for all compounds by tire and groove. Fast wearing shoulder grooves showed no significant compound wear improvements. Compounds B, C, D, H, K, & L show significant wear improvement vs. control in central grooves, G2-G5.

19. Conclusions Paired Design and associated analysis was able to detect statistically significant wear differences among 9 compounds by testing only 12 tires on 3 vehicles for 16K miles. The Paired Design approach provides a potential cost reduction for wear testing.