Bayesian hierarchical model with MCMC for optical cable flame retardancy tests

We report a model and method for calculating the probability of an optical fiber cable passing a flame retardancy test (FRT). The method uses a Bayesian approach that accounts for variations in the experimental conditions between tests. We show that adding a hierarchy to the empirical model shrinks the posterior distributions and discuss the applicability of the hierarchical model to various test environments. The model uses the average smoke parameter as the predictor and can be extended to other test responses such as the peak smoke, flame spread, or toxicity.