Akademik Veri Yönetim Sistemi
Aircraft Engineering And Aerospace Technology, cilt.1, sa.94, ss.1-15, 2022 (SCI-Expanded)
Abstract
Purpose- Scheduling
of aircraft maintenance operations is a gap in the literature. Maintenance times
should be determined close to the real-life to schedule aircraft maintenance
operations effectively. The learning effect, which has been studied extensively
in the machine scheduling literature, has not been investigated on aircraft
maintenance times. In the literature, the production times under the learning
effect have been examined in numerous studies but for merely manufacturing and
assembly lines. A model for determining base and line maintenance times in
civil aviation under the learning effect has not been proposed yet. It is pretty
challenging to determine aircraft maintenance times due to the various aircraft
configurations, extended maintenance periods, different worker shifts and
workers with diverse experience and education levels. The purpose of this study
is to determine accurate aircraft maintenance times rigorously with a new model
which includes the group learning effect with the multi-products and shifts,
plateau effect, multi sub-operations and
labor firings/rotations.
Design/methodology/approach
– Aircraft
maintenance operations are carried out in shifts. Each maintenance operation
consists of many sub-operations that are performed by groups of workers. Thus, various models, e.g. Learning
Curve for Maintenance Line (MLC) Learning Curve for Maintenance Line with
Plateau factor (MPLC), Learning Curve for Maintenance Line with Group factor
(MGLC) were developed and employed in this study. The performance and
efficiency of our models were compared with the current models in the literature,
such as the Yelle Learning model (Yelle), Single Learning Curve model (SLC) and
Single Learning Curve with Plateau factor model (SLC-P). Estimations of all
these models were compared with actual aircraft maintenance times in terms of
Mean Absolute Deviation (MAD), Mean Absolute Percentage Error (MAPE) and Mean
Square of the Error (MSE) values. Seven years (2014-2020) maintenance data of
one of the top ten maintenance companies in civil aviation were analyzed for
the application and comparison of learning curve models.
Findings- The best
estimations in terms of MAD, MAPE and MSE values are respectively gathered by
MGLC, SLC-P, MPLC, MLC, SLC and YELLE models. We revealed that the models
(MGLC, SLC-P, MPLC), including the plateau factor, are more efficient in
estimating accurate aircraft maintenance times. Furthermore, MGLC always made
the closest estimations to the actual aircraft maintenance times. The results
show that our MGLC model is very accurate than all of the other models for all
sub-operations. The MGLC model is promising for the aviation industry in
determining aircraft maintenance times under the learning effect.
Originality/value- In this
study, learning curve models, considering groups of workers working in shifts,
have been developed and employed for the first time for estimating more
realistic maintenance times in aircraft maintenance. To the best of our
knowledge, the effect of group learning on maintenance times in aircraft
maintenance operations has not been studied. The novelty of our models are
their applicability for groups of workers with different education and
experience levels working in the same shift where they can learn in accordance
with their proportion of contribution to the work and learning continues
throughout shifts. The validity of the proposed models has been proved by
comparing actual aircraft maintenance data. In practice, the MGLC model could efficiently
be employed for aircraft maintenance planning, certifying staff performance
evaluations and maintenance trainings. Moreover, aircraft maintenance
activities can be scheduled under the learning effect and a more realistic
maintenance plan could be gathered in that way.