Trending of equipment inoperability for commercial aircraft

Abstract

Abstract The Safety Performance Analysis System (SPAS) is a Web-based analytical tool, or rather a database system, that is being fielded by the Federal Aviation Administration (FAA). It is intended to provide the FAA’s Aviation Safety Inspectors (ASIs) with the means to evaluate and control appropriate surveillance levels for aircraft operators. The SPAS enables qualitative trend analysis to be performed at the national, regional, and district levels. This will permit the system to be more sensitive to particular areas of concern for a given region. For example, inspectors particularly desire a tool to help them recognize problems which are occurring in the fleet. Especially desirable is a tool that could do so automatically by searching for trends and alerting the relevant inspectors. The Service Difficulty Reporting (SDR) system, one of the SPAS databases, provides FAA inspectors with information related to aircraft equipment inoperability, such as in-service difficulties, malfunctions, and defects. The SDR data provide information for planning, directing, controlling, and evaluating certain assigned safety and maintenance programs. This paper presents previous SDR trending results for DC-9 operators with heterogeneous fleets, but now uses data from two specific operators with homogenous fleets consisting entirely of 737 aircraft. This new research is an extension of the previous SDR forecasting research, but now provides more meaningful information, as the DC-9 aircraft data were composed of numerous operators with mixed fleets and differing operating and maintenance policies. Multiple regression and neural network models are the principal two forecasting methods examined. A population modeling concept, or data grouping strategy, appears to be an effective technique for trending SDRs for operators of either heterogeneous and homogeneous fleets. The forecasting methods presented in this paper offer technical enhancements for SDR trending compared to the current qualitative method of visual observation of graphical plots.