Expert System for Evaluating Human Reliability in Manual Assembly Operations

Manual assembly tasks constitute a key element to create stable and efficient production processes. For reaching a high quality level, production processes are necessary that can be carried out both economically and reliably. In practice, creating robust and efficient assembly processes often fails because the occurrence of errors caused by persons acting in the production system cannot be quantified prospectively. Consequently, frequently occurring human errors and process steps prone to errors are often only detected in the ongoing operation. However, then system adjustments can only be realized with high temporal and financial expenditure. For that reason, at the institute of the authors, a procedure has been developed to use the knowledge about evaluating human reliability gained in safety critical areas (e.g. aerospace, nuclear power engineering) for creating a computer aided expert system that allows predicting human error probabilities for manual assembly operations. This paper illustrates how human error probabilities in manual assembly processes can be determined and linked to pre-determined motion times of the assembly process. Afterwards, those error probabilities in common with needed knowledge can be depicted in an Excel-based software tool, so that reliable time and risk analyses of manual work tasks can be drawn up with less expenditure.