Statistical tools applied for the reduction of the defect rate of coffee degassing valves

Coffee is a very common beverage exported all over the world: just after roasting, coffee beans are packed in plastic or paper bags, which then experience long transfers with long storage times. Fresh roasted coffee emits large amounts of CO2 for several weeks. This gas must be gradually released, to prevent package over-inflation and to preserve aroma, moreover beans must be protected from oxygen coming from outside. Therefore, one-way degassing valves are applied to each package: their correct functionality is strictly related to the interference coupling between their bodies and covers and to the correct assembly of the other involved parts. This work takes inspiration from an industrial problem: a company that assembles valve components, supplied by different manufacturers, observed a high level of defect rate, affecting its valve production. An integrated approach, consisting in the adoption of quality charts, in an experimental campaign for the dimensional analysis of the mating parts and in the statistical processing of the data, was necessary to tackle the question. In particular, a simple statistical tool was made available to predict the defect rate and to individuate the best strategy for its reduction. The outcome was that requiring a strict protocol, regarding the combinations of parts from different manufacturers for assembly, would have been almost ineffective. Conversely, this study led to the individuation of the weak point in the manufacturing process of the mating components and to the suggestion of a slight improvement to be performed, with the final result of a significant (one order of magnitude) decrease of the defect rate.