Abstract

A systematic approach for the design and implementation of quality control systems has been developed: a model in which a large number of variables that interact stochastically in a quality control system are studied using the principles behind the technique known as Design of Experiments. The model consists of four phases. In the first phase, an ISO 9000 quality control system is designed using current practices: experience and common sense. The second phase is a screening phase, where each inspection point is reclassified as relevant or non-relevant by measuring its influence on the final quality of the product. Those points classified as relevant are analyzed in detail in a third phase, the scrutinizing phase, in which combinations of factors and levels are determined by using principles from Design of Experiments (DOE) such as orthogonal arrays and from Taguchi signal-to-noise ratio. Factors, such as frequency of calibration, training of personnel and instruments and levels such as every month/every two months, in-house training/formal training and vernier/micrometer are determined based on experience and research of the manufacturing process and available inspection techniques. A series of experiments are carried out to obtain the best combination of factors and levels that confirm that the product is exceeding customers' expectations at a minimal cost. The different effects of each of these combinations on the final quality is measured. With the final phase, data analysis and a confirmatory run, inspection and quality control efforts are then directed to those areas where they are truly required. The use of Design of Experiment ensures that the selected combination of factors and levels reduces the variability of the results of the implemented ISO 9000 quality control system, hence there is confidence that defective pieces are detected prior to shipment to the customer.