工程实验设计课件Lecture 2-DOE.pdf

Experiments with Single Factor Introduction Compare two –conditions or -treatment experiments By hypothesis test Paired comparison design Analysis of variance One-way layout Introduction Two styles: 1) Different conditions or treatments; 2) Single factor Experiment with a levels Statistic model for a single factor experiment i 1,2 ⎧ y ij ui +εij ⎨ j 1,2,..., ni ⎩ Design of One or Two Sample Comparison Determine sample size n. δ= (µ - µ )/σ or δ= (µ - µ )/σ 1 0 2 1 Common practice : α= 0.05, β= 0. 10. Type I and II error in Hypothesis Test Example: Tension bond strength of portland cement mortar: to compare effects of modified formulation, as two levels of a formulation Compare two –conditions or -treatment Two sample t-Test Tests mean difference, variance unknown (equal) Suppose 2 2 2 , compare two treatment means σ σ σ 1 2 u , u in randomized design. 1 2 Test statistic y 1 −y 2 t 0 1 1 S + p n n 1 2 This /2 t Cement example: Tests mean difference, variance unknown (unequal) Tests