Venkatappaiah, J. (1986, ASQC) Indian Statistical Institute, Hyderabad, India
A measure of human civilisation is the quality of products and services availed by the society; more civilised is the society, more demanding are the quality requirements. The efforts are towards improvement in quality - the universal equation for excellence, in all fields of human activity.The extensive use of statistical technology made Japanese Industry world leaders in quality. The word quality has acquired new significance with prime emphasis on product variability and influence on society - minimum variability from target and least adverse impact on society during the operating life of the product.The life requirements of some products are incorporated into quality specifications in the form of life tests. Since life tests cannot be done regularly or frequently, it is desirable to test the life of products under different design conditions during product development and process engineering stage to determine the optimum process conditions. In order to ensure good stability and enough length of life to the product, it is preferable to conduct broad based experiments by incorporating as many factors as possible.Dr. G. Taguchi, an eminent Japanese expert on Industrial Experimentation formulated versatile Orthogonal Array [O A] designs along with linear graphs for conducting experiments especially with large number of factors. The advent of O A designs have a tremendous boost to the application of experimental methods in Industry for quality improvement. New methods of analysis have also been evolved for extracting more information from experimental data. Minute Analysis  is one of such methods for analysing life data.In a plant manufacturing Telecommunication cables failure in a particular life test was high. The customer not only imposed a stiff penalty for the failed cables but also threatened to completely stop the orders unless the cable quality is improved.The plant technologists conducted experiments on a number of occassions and specified process conditions broadly. However, the efforts though indicated scope for improvement, did not lead to reproducible and sustained results.Experts in cable technology from a foreign collaborator, worked on this problem for a few years and reported that it would not be possible to meet the life test requirements with the equipment and technology of the plant. They opined that conformance to this rigorous life test is not necessary and recommended that the plant approach the customer for waiving the test.However, statistical investigation of the problem together with an L 18 OA experiment and Minute analysis of the data furnished vital clues for improvement in the life of cables. The experiment involved 8 factors, covering input materials and process parameters, at 2 or 3 levels; a few of the factors were nested within other factors. The Minute Analysis could give an insight with regard to the stability of a material and its application conditions during the period of test duration. This would not have come to light through usual methods of analysis.The parameter combinations that enable very high conformity to the life test were identified and these were adopted for regular manufacture after elaborate confirmatory trials.