Improving the Accuracy of Forecasting the Labour Intensity of Aircraft Heavy Maintenance

One of the critical tasks of aircraft heavy/base maintenance check planning is to predict the additional production output for defects rectification and calculation of turnaround time. At the same time, the additional output should be distinguished by different specializations of the personnel involved in the maintenance (e.g., mechanics, avionics, structure etc.). The current approach implies the calculation of labour intensity based on the history of similar checks, considering the age of the aircraft and its total accumulated flight hours. The classical approach defines the number of additional man-hours by ratio to routine man-hours (final work scope includes routine man-hours and additional man-hours output resulting from results of routine inspections). Ratio increases with the age of the aircraft. However, modern maintenance programs involve a variety in the filling of maintenance checks. The content of checks may vary depending on the intensity of aircraft operation. Also, operation in different climatic conditions differently affects the technical condition of the aircraft and, accordingly, the number of resources spent on repair as part of maintenance. The present article proposes mathematical models for predicting additional production output. The models are based on machine learning algorithms using a decision tree as a non-parametric supervised learning method used for both classification and regression tasks. The models are developed based on expert analysis of the causes affecting additional labour. Proposed models have been tested in a real maintenance, repair, and overhaul organization (MRO). The results obtained are described. The article presents an additional set of functions for the algorithm. Examples of increasing prediction accuracy by adding information about aircraft operation history to the decision tree are presented in detail. The geography of airplane flight areas and configuration as parameters for the mathematical model are considered.

• Record URL:
  https://doi.org/10.1007/978-3-030-96196-1
• Availability:
  • Find a library where document is available. Order URL: http://worldcat.org/isbn/9783030961954
• Supplemental Notes:
  • © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2022.
• Corporate Authors:

  Springer International Publishing

  ,    
• Authors:
  • Tyncherov, Timur
  • Rozkova, Liubov
• Conference:
  • 21st International Multi-Conference Reliability and Statistics in Transportation and Communication (RelStat 2021)
  • Location: Riga , Latvia
  • Date: 2021-10-14 to 2021-10-15
• Publication Date: 2022-2

Language

• English

Media Info

• Media Type: Web
• Features: References;
• Pagination: pp 223-236
• Monograph Title: Reliability and Statistics in Transportation and Communication: Selected Papers from the 21st International Multidisciplinary Conference on Reliability and Statistics in Transportation and Communication, RelStat2021, 14-15 October 2021, Riga, Latvia
• Serial:
  • Lecture Notes in Networks and Systems
  • Volume: 410
  • Publisher: Springer Cham
  • ISSN: 2367-3370
  • Serial URL: https://www.springer.com/series/15179

Subject/Index Terms

• TRT Terms: Aircraft; Decision trees; Labor; Maintenance personnel; Predictive models; Vehicle maintenance
• Subject Areas: Aviation; Maintenance and Preservation; Planning and Forecasting; Safety and Human Factors;

Filing Info

• Accession Number: 01890800
• Record Type: Publication
• ISBN: 9783030961954
• Files: TRIS
• Created Date: Aug 23 2023 3:24PM