FMODE Statement

FMODE keyword=variable (’value1’’valuen’) < / options> ;

Use the FMODE statement with data that have failures attributable to multiple causes (failure modes). You can analyze data by either keeping, eliminating, or combining specific failure modes with the FMODE statement. Use this statement with the KEEP or ELIMINATE keyword in conjunction with the ANALYZE, MODEL, PROBPLOT, or RELATIONPLOT statement. Use this statement with the COMBINE keyword with the ANALYZE or PROBPLOT statement. You can place an FMODE statement anywhere after the PROC RELIABILITY statement and before the RUN statement.

If you specify the keyword KEEP, the life distribution for only the identified failure modes is estimated, with all other failure modes treated as right-censored data. If you specify the keyword ELIMINATE, the life distribution that results if the failure modes identified are completely eliminated is estimated. The keyword ELIMINATE causes the failure modes identified to be treated as right-censored data and causes a single life distribution to be estimated for the remaining data. If you specify the keyword COMBINE, the data are analyzed with all the specified failure modes combined acting. See the section Weibull Probability Plot for Two Combined Failure Modes for an example of a Weibull plot of data with two combined failure modes. The failure mode for an observation in the input data set is identified by the value of variable, where variable is any numeric or character variable in the input data set. You must identify a failure mode for each observation that is not right-censored. You specify failure modes to keep, eliminate, or combine by listing variable values (value1valuen) in parentheses after the failure mode variable name. The list of variable values must have entries separated by blanks or commas. You can specify the following options after the slash (/). These options will affect the analysis only when you use the COMBINE keyword.

Table 16.10: FMODE Statement Options




Specifies a LEGEND statement for individual mode fit lines.


Suppresses legend for individual mode fit lines.


Plots individual failure distribution lines on probability plot.