Navmar Applied Sciences Corporation and GCAS Incorporated have been tasked by the Navy with conducting data mining of Organizational, Intermediate and Depot (OID) level inspection and maintenance records to determine and predict the future costs of corrosion within the NAVAIR fleet. 

With the Navy air maintenance costs surpassing a billion dollars a year, identification and control over the affecting factors has a potential for high returns. Corrosion repairs could be scheduled more effectively so as to optimize reparability versus cost and improve the maintenance operations efficiency. Corrosion maintenance requirements at field and depot levels could be predicted for individual aircrafts. Manpower levels for corrosion control and maintenance could be matched more closely to needs.

Of paramount importance to these goals has been the determination of data quality in the inspection and maintenance databases. The records were assessed to determine their overall correctness. Bias and accuracy of the logged maintenance hours were evaluated. 

This analysis highlighted the inherent difficulty in field and depot data collection. Bias and clustering were observed at the O&I levels, where a small percentage of the maintenance actions accounted for most of the corrosion costs. This finding was evident over a variety of aircraft types and locations. The Depot-level records showed a clustering pattern as well, where only a couple of maintenance actions out of a dozen accounted for the majority of the costs. However, these data were easier to discriminate between usual costs of the great majority of maintenance actions and a limited number of outliers. This led to a good predictability of the overall Depot costs.

Specific recommendations were made for more accurate inspection and maintenance data recording and database improvements at all three levels, in order to facilitate future cost modeling and prediction activities.