Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 3
3 1.3 Ruggedness Experiments are determined from theoretical considerations or previous experience with the test. For this study, information obtained 1.3.1 Background from testing completed in Project 9-19 and in Phase II of The purpose of ruggedness testing is to improve a test Project 9-29 was used to select the factors and their levels. The method by determining which controllable testing conditions selection of the factors and their levels are discussed in detail most influence the results, and establishing limits for their later in this Chapter. Test data are collected for specific com- control. A ruggedness evaluation should always precede an binations of the factors and their levels as outlined in Table 1. interlaboratory study for a test method. The purpose of an This table uses the nomenclature from ASTM D1067. The interlaboratory study is to establish the precision of a test seven factors are designated by letters A through G. Capital method. It involves testing of multiple materials in multiple letters indicate high levels for the factors while lower case letters laboratories, and requires a significant commitment of time indicate low levels. Thus, determination 1 will be made with and resources. If critical testing conditions are not first iden- factors A, B, and E at low levels and factors C, D, F, and G at tified and controlled through a ruggedness evaluation, then high levels. With replication, the experiment requires 16 tests, an interlaboratory study will likely yield poor precision for two for each of the specific combinations indicated in Table 1. the test method. Perhaps more important than a finding of The order of the tests should be randomized within each poor precision, is the fact that data from the interlaboratory replication of the experiment. study is not generally useful for determining how to improve Analysis of the resulting data is straightforward as described the precision of the test. This was the unfortunate finding of in ASTM D1067. It involves determining effects for each of an interlaboratory study that was recently completed for the the factors included in the partial factorial design, and an dynamic modulus test (3). This study identified high vari- estimate of the variance of a single measurement. An F-test or ability in dynamic modulus data obtained from several labo- linear regression can then be used to assess the statistical signif- ratories, but was not able to establish reasons for the high icance of the factor effects relative to the variance of a single variability or to identify procedural changes that would result measurement. in more acceptable testing error. By systematically varying The major considerations in the design of a ruggedness test testing conditions and quantifying their effect on the meas- are (1) selection of the factors and their levels, (2) selection of ured data, a ruggedness evaluation is able to identify important a range of materials or test conditions for the evaluation, and sources of testing error and help establish limits to reduce (3) selection of an appropriate number of laboratories to par- testing error to a tolerable level. ticipate in the ruggedness testing. The experimental design in Since ruggedness testing is a critical part of the development Table 1 uses seven factors at two levels. This design is consid- of a test method, efficient statistical designs have been devel- ered appropriate for the proposed ruggedness testing for the oped and standardized for ruggedness tests. ASTM E1169, simple performance tests. ASTM D1067 recommends using Standard Guide for Conducting Ruggedness Tests, describes the three to five materials covering the expected range of proper- partial factorial Plackett-Burnam designs most often used in ties to be measured in the test. The results from each material ruggedness testing. These designs are very efficient for simul- are analyzed separately; therefore, 16 measurements are needed taneously evaluating the effect of changes in a number of for each material included in the evaluation. ASTM E1169 operating conditions when there is no interaction between and ASTM D1067 differ on the number of laboratories to be the operating conditions being evaluated. Inherent to this type included in the ruggedness testing. ASTM E1169 recommends of statistical design is the assumption that the effect of each of the operating conditions on the test result is independent. Table 1. Experimental design Therefore, the observed effect resulting from simultaneous for a two level, seven factor variation of several operating conditions is simply the sum of ruggedness test. the individual effects. Since ruggedness testing is concerned Determination Number with the evaluation of the effect of changes in testing condi- tions and not necessarily the form of the effect, each testing Factor 1 2 3 4 5 6 7 8 condition is usually evaluated at only two levels. Replication A a A A a A A A A should be included in the design when an estimate of the vari- B b B B B b B B B ance of a single measurement is not known. C C C C c C C C c ASTM D1067, Standard Practice of Conducting a Ruggedness D D D D d d D D D or Screening Program for Test Methods for Construction Mate- E e E E E E E E e rials, describes the two-level, seven-factor design with repli- F F F F F F F f F cation recommended for ruggedness testing for construction G G G G G g G G g materials tests. The factors to be evaluated and their two levels