Qualification of second-source raw materials in the presence of multiple batch effects

A system comprising at least one electronic processor that executes instructions to perform operations comprising:

• selecting a final linear mixed model for a quantifiable property of at least one part based on an empirical data set obtained by measuring the quantifiable property in each of a plurality of samples of the at least one part, wherein the at least one part is produced by a candidate supplier using a bulk manufacturing process, and wherein the selecting comprises, for at least one ordering of variables corresponding to a plurality of batch effects comprising batch effects produced by the bulk manufacturing process, including in a linear mixed model variables from the at least one ordering of variables that significantly affect the quantifiable property, and wherein the selecting further comprises, for a plurality of permutations of variables corresponding to the plurality of batch effects, iteratively incorporating in a respective linear mixed model variables from a respective permutation of variables that significantly affect the quantifiable property, whereby a plurality of candidate models are produced, wherein the final model is selected among the plurality of candidate models;
• fitting the final linear mixed model to the empirical data set to obtain a fitted linear mixed model;
• extracting a mean and a standard deviation of the quantifiable property for the at least one part from the fitted linear mixed model;
• determining an acceptance parameter from the mean and the standard deviation, wherein the acceptance parameter comprises at least one of a process capability index or a tolerance interval bound;
• and
• qualifying the candidate supplier based at least in part on comparing the acceptance parameter to a threshold, wherein the at least one part is incorporated into an aircraft manufacturing process.