Date of Award
Doctor of Philosophy (PhD)
Elimination of aviation accidents is one of the primary goals of the Federal Aviation Administration (FAA) and the airline industry. A leading cause of aviation accidents is lack of oversight of various organizational issues, in particular, the organization's maintenance operation performance. The technologies used in the industry generate multiple risks, mostly from three domains: systems, hardware and people. Maintenance performance analyses identify the inherent risk in distributed, large-scale systems. Analysis of existing aviation maintenance data is a crucial step in meeting the aviation industry's need to improve aviation safety. Presently, we lack suitable tools to analyze large bodies of maintenance data. In this study, we generate models responsive to airline operation requirements using hierarchical logistic regression analysis based on historical auditing and surveillance data. These models helped to determine the organizational factors underlying aviation maintenance errors, ultimately helping airline personnel to manage the surveillance and auditing functions of aircraft maintenance. Three models were generated- one model each for an airline's technical audit, internal audit and surveillance work functions. These models were embedded in a web-based surveillance and auditing tool (WebSAT). Validation experiments were conducted to evaluate the utility of the model in WebSAT. Results indicated that there is significant improvement in vendor/ department performance prediction capabilities when the model is employed with WebSAT. The auditors and surveillance representatives' ability to understand the effect of a change in the level of a predictor on rejection rate improved significantly when the model was employed in WebSAT. The technical audit and surveillance managers' non-significant results indicate that the Audit Allocation and Surveillance Planning tools are not as useful for managers. It is important to improve the capabilities of the planning tools by employing more variables in the regression models including information on surveillance representatives and auditors.
Iyengar, Nikhil, "DEVELOPMENT OF PREDICTION MODELS TO MEASURE VENDOR PERFORMANCE IN SURVEILLANCE AND AUDITING OF AIRCRAFT MAINTENANCE" (2007). All Dissertations. 86.