Abstract
A typical gas turbine part that is fitted into an aero engine goes through many manufacturing processes spread across multiple supply chain organizations. Despite high precision standards in comparison to other industry segments, geometrically complex parts in gas turbine do end up deviating from design geometries. These parts go through visual inspection at every stage in addition to dimensional measurement of critical characteristic features. Operator involvement is un-avoidable at present and despite being process driven, operator level variances are not ignorable in manufacturing operations. With Industry 4.0 being adopted by many industries, and 5.0 emerging, user involvement is minimized, and a large amount of data is generated and captured throughout the process of manufacturing, testing, verification, assessment and sentencing of parts for use in aerospace industry. As data is becoming increasingly centralized by adoption of cloud technologies, it becomes more efficient to troubleshoot problems through mathematical modelling — the brain behind Artificial Intelligence (AI). Structural Equation Modelling (SEM) has been in use for long in management sciences and decision making involving ill-defined factors and that have strong interdependencies.
This article presents a methodology of leveraging SEM technique in identifying correlation of qualitative and quantitative factors with part deviation assessment. A framework for converting qualitative inspection information to quantitative information is presented to leverage SEM. The implementation is explained using a representative data for a turbomachinery airfoil to demonstrate the approach and considerations. Few use cases of SEM in the industry are also presented.
