Vol.:(0123456789) 1 3
International Journal of Aeronautical and Space Sciences
https://doi.org/10.1007/s42405-019-00204-2
ORIGINAL PAPER
Estimation of Maximum Strains and Loads in Aircraft Landing Using
Artifcial Neural Network
Seon Ho Jeong
1
· Kyu Beom Lee
1
· Ji Hoon Ham
1
· Jeong Ho Kim
1
· Jin Yeon Cho
1
Received: 17 January 2019 / Revised: 17 June 2019 / Accepted: 29 July 2019
© The Korean Society for Aeronautical & Space Sciences 2019
Abstract
Hard landings account for a large proportion of aircraft accidents and are generally judged by the intuition of pilots. For this
reason, there are frequent false judgments that lead to unnecessary and costly ground inspections. False judgments can be
reduced signifcantly if detailed load information is available, such as strains or loads on critical areas of aircraft structures.
This study used an artifcial neural network (ANN) to develop a numerical model that can estimate the maximum strains
for areas of interest and landing loads from basic fight parameters. The results can be used to provide the required detailed
load information. An efcient and accurate landing simulation model was constructed and used to build reliable datasets
for training. Basic fight parameters from immediately after touchdown were used as input data for training, and the cor-
responding maximum values of strains and landing loads were obtained from the landing simulation model as target data
for training. This information was used to train the ANNs with the Levenberg–Marquardt backpropagation algorithm. A
performance evaluation using test data confrmed that the trained ANNs can successfully estimate strains and landing loads
with sufcient accuracy.
Keywords Structural health monitoring · Hard landing · Artifcial neural network · Supervised learning
1 Introduction
Most aircraft accidents occur during takeof and landing.
Figure 1 shows the fatal accidents and onboard fatali-
ties from 2007 to 2016, and about half of fatal accidents
occurred during fnal approach and landing [1]. Even though
the fnal approach and landing account for only about 4% of
the total fight time, accidents occur most frequently at these
times. Therefore, monitoring aircraft at the landing stage is
very important.
Hard landings are one of the most frequent accidents that
occur during landings. Hard-landing accidents had the larg-
est proportion of accidents (54 of 385 cases) among large
Western-built commercial jet airplanes from 1993 to 2002
[2]. Hard landing should be monitored as one of the key
parameters of structural integrity, as shown in Fig. 2 [3].
Hard landings occur when landing with a relatively large
sink rate and load compared to a normal landing. The main
causes are incorrect landing fares, exceeding the vertical
rate of descent on the fnal approach, insufcient crew coor-
dination, and special meteorological conditions [4].
Although most hard landings do not lead to serious
human injury, omitting necessary maintenance due to mis-
judgment may eventually lead to fatal accidents in succes-
sive fights. Therefore, pilots who generally report hard-
landings tend to judge them very conservatively and submit
hard-landing reports for any suspicious landing situations
because of safety concerns. When a report is submitted, the
aircraft is checked for damage. It usually takes more than
1.5 h for the visual inspection as a frst stage and more than
8 h for non-destructive examination as a second stage [5].
However, it has been reported that more than 85% of air-
craft are not damaged after the examination, which leads
to unnecessary expense and wasted time [6]. Therefore, a
hard-landing determination method based on accurate data
and objective criteria is required to minimize the unneces-
sary execution of expensive and time-consuming precision
inspections of aircraft.
In this regard, Young et al. [6] examined the efect of
using flight parameters and design modeling tools on
determining hard-landing decisions. They estimated that
* Jin Yeon Cho
cjy@inha.ac.kr
1
Department of Aerospace Engineering, Inha University, 100
Inha-ro, Michuhol-Gu, Incheon 22212, Republic of Korea