Statistical tolerance analysis for non-normal or correlated normal component characteristics

The paper presents a simple approximation technique for statistical tolerance analysis, namely, the allocation of component tolerances based on a known assembly tolerance. The technique utilizes a discretized, multivariate kernel density estimate and a simple transformation to approximate the probability distribution of the overall assembly characteristic. The data-driven approach is suitable for real-world settings in which components are randomly selected from their respective manufacturing processes to form mechanical assemblies. Demonstrated is the numerical approach in two dimensions for two distinct cases: first, when component characteristics are non-normal, independent random variables, and second, when they are highly correlated, normal random variables. The results are promising in initial test problems.