Assembly-specific database for predicting human reliability in assembly operations

• Autores: Christian Keen, Robert Refflinghaus
• Localización: Total Quality Management & Business Excellence, ISSN-e 1478-3371, Vol. 26, Nº. 9-10, 2015, págs. 1056-1070
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • The assembly takes up a special position in a company's order processing; as with regard to the company's service provision, it mostly comes at the end of the value chain. For that reason the assembly depicts a melting pot of many technical, organisational and qualitative errors. To mitigate technical failure, in the past versatile solutions were developed, but there are only little satisfactory solutions to consider human failure in the planning process of manual assembly operations. This paper presents a procedure that allows transferring the knowledge about human reliability gained in safety-critical industries to manual assembly operations. In this context, at the research institute of the authors, approaches have been developed to use the knowledge of the Expert System for Task Taxonomy – a method that originally was developed to evaluate error rates of control activities in safety-critical areas – for creating a new method that allows evaluating human reliabilities in manual assembly operations. The aim of this method is to be able to systematically predict human error probabilities for any manual work content of the industrial batch production. Considerable elements of the developed assembly-specific method are an own database containing assembly-specific standard terms and a calculation model which is adapted to low error rates of the manual assembly. First results at a heating manufacturer verified that the developed method enables reliable risk predictions for the field of manual assembly. This is particularly given by the creation of a new database, which is especially tailored to the application field of manual assembly and the development of assembly-specific calculation regulations. In the following steps, the comparison of the calculated error data and the actual error data of the industrial practice is extended, so that the method can be validated across all industries