| ||Format||Pages||Price|| |
|PDF (244K)||9||$25||  ADD TO CART|
This paper presents maximum likelihood methods and a computer program for fitting fatigue curves with nonconstant standard deviation to data with runouts. The methods yield estimates and confidence limits for curve coefficients and percentiles of the life distribution at any stress level. Also presented are (1) the likelihood ratio test for comparing the fit to data of different types of fatigue curves and (2) graphical analyses of residuals to check the assumed fatigue life distribution, to assess fit of a curve, and to spot peculiar data.
Statistician, General Electric Co. Corporate Research and Development, Schenectady, NY
Stock #: JTE10700J