Introduction to Statistical Quality Control (5th edn)
- Montgomery DC
− The traditional control charts are based on the assumption that the measurement data relating to the average of each subgroup of the parameters under observation of a specific quality characteristic of product, process, system or service, are reported regardless of time stability of its own parameters. This condition is generally not verified in real cases, where we should take into account the lifecycle of measuring. In the case of monitoring of quality characteristic of a product for example, a parameter that is not in fact enjoy the property of time stability is its reliability, which is time-variant and it’s decreasing over time in force of degradation to which the product is subject. The traditional approach involves the construction of control charts with stationary limits, ignoring the possibility of an "adjustment" of those limits. This conceptual innovation determines the possibility of having an instrument to check, in a way as “tracking trajectory ", modulated on possible changes of the measuring. These variations may, for example, be due to degradation of the detained characteristics. The modelling for the scenario described is based on the implementation of reliability function, as a known source of the variability of control limits, as pre-determinable through the lifecycle of the product analyzed. The results achieved through the proposed approach can improve the sensitivity of the monitoring instrument through the modulation of control limits, which thus acquires a beneficial discriminatory capacity, able to separate the effects of random variability arising from the common causes, from those owned to deterministic variability related to predictable changes of the measuring. Numerical results, regarding an application to a manufacturing national company operating in the automotive sector, through fielded units taken in differing operating periods and tested, are reported.