Development of an Assembly Consistency End-of-Line Inspection
System for Corn Harvester
Bingzhong Peng, Du Chen, Dong Sun, Shumao Wang
College of Engineering, China Agricultural University, Beijing 100083, China;
1
Beijing Key Laboratory of Optimized Design for Modern Agricultural Equipment, Beijing 100083, China
Keywords: Corn Harvester, Assembly Consistency, Vibration, Transmission.
Abstract: Nowadays, along with the application of intelligent equipment and new technology, automation degree of
manufacturing becomes more and more advanced. While corn harvester, as one of the most important
agricultural machineries, the assembly quality was still inspected by traditional method depending on workers’
experience in China. Transmission is the core part of corn harvester, which is also the vibration source of
whole machine. Detecting and keeping transmission assembly consistency can reduce the vibration degree
and increase the mean time to failure, and improve the performance of whole machine. In this paper, a
transmission assembly consistency inspection system was designed and a method for judging the assembly
consistency was proposed. The proposed approach included two parts, the vibration signals analysis, which
was processed based on six bands method to calculate the energy of six bands; and the consistency comparison,
which was used to judge the assembly consistency intuitively through the chart. In the last we chose the corn
harvester pulverizer as an analysis object. According to result, the third and the fifth are incongruous.
1 INTRODUCTION
Corn harvester, as one of the most complex
agricultural machinery, is mainly embodied in the
complexity of the transmission system. On the one
hand, the transmission system of corn harvester
consists of various transmission forms, including belt
drive, chain drive, gear transmission, cam drive and
so on. On the other hand, for the corn harvester, the
harvest process requires long time, and the
transmission system need to withstand interfere of
dust, straw, clods and other debris while working,
which increase the reliability and durability
requirements of transmission system. Therefore it is
necessary to inspect the assembly quality of corn
harvester to ensure its reliability. At present, the
inspection of assembly quality is still in a traditional
way, without automation. Inspecting with an
automated way can control assembly quality well.
At present, studies on assembly quality focus on
process quality control strategies and techniques.
Suzuki et al. (2001) aiming at the reliability of
assembly workshop, put forward a method (AREM)
of assembly reliability evaluation, in which, the
assembly failure rate was quantitatively studied by
designing factors and workshop factors. According to
this study, it can improve the reliability level of
assembly by controlling influence factors. Kurt et al.
(2000) used the AQM and DFA to evaluate the
assembly quality of the products, so as to improve the
assembly quality in design stage. Milberg and
Wisbacher (1992) analyzed noise of the assembled
products and assembly process, and then controlling
the assembly quality of products by using the noise
spectrum diagram. Su S. developed a marine diesel
engine assembly quality information management
system based on management system model, which
can extract diesel engine configuration information
and generate assembly quality inspection cards
suitable for specific diesel engines. Kong F.
established the prediction model of assembly quality
defect rate, it can be used as the monitoring means of
assembly quality, and can also be used as the basis for
product design and assembly process adjustment.
According to the references, we know that most of the
studies focused on design stage or components
assembly. However, there is no end of line assembly
quality evaluation method.
Specifically speaking, in order to ensure assembly
reliability of corn harvester and raise automation
degree of inspection, we proposed an end of line
assembly quality evaluation method. The rest of the
652
Peng, B., Chen, D., Sun, D. and Wang, S.
Development of an Assembly Consistency End-of-Line Inspection System for Corn Harvester.
In 3rd International Conference on Electromechanical Control Technology and Transportation (ICECTT 2018), pages 652-656
ISBN: 978-989-758-312-4
Copyright © 2018 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved

paper is organized as follows: Section 2 provided
method and materials in this work. Section 3
determined the proper sensor. In Section 4 the results
and discussion are presented and Section 5 gave some
conclusions.
2 METHOD AND MATERIALS
In the shop, the assembly quality of corn harvester is
inspected in an artificial way (Yi and Wang, 2009). To
make it automatic, we designed an inspection system,
which can acquire vibration signal and analyze
assembly quality with it. The inspection system
includes velocity sensor, NI-9234 AD input module,
NI cDAQ-9172 eight-slot USB chassis, and a signal
analysis software. The specific composition and flow
chart of measurement is shown in Figure 1.
Figure 1: The specific composition and flow chart of
measurement.
With the inspection system, to acquire vibration
signal, we designed a set of experimental scheme and
carried out in Tianjin Yongmeng Machinery Co. Ltd.
The whole detection flow chart mainly included five
steps, namely frequency pre-estimation, signal
acquisition and processing, band division, data
analysis, and results presentation. Type 4YZ4590 was
selected as experiment object. And selecting four key
vibration parts in a corn combine harvester, namely
header, engine, pulverizer and rear bracket to test.
Total 8 corn harvesters were tested. Each part tested 3
times. The sampling rate is set to 10 kHz. The
harvester worked both in idle rotational or maximum
rotational speed conditions. Figure 2 is the picture
taken while we were doing experiment.
Figure 2: Doing experiments.
3 DETERMINATION OF THE
PROPER SENSOR
In this paper, a preliminary experiment was carried
out to determinate the proper sensor between velocity
sensor and accelerometer, both using piezoelectric
principles. Important parameters of the two sensors
are shown in table 1.
Table 1: Main parameters of sensors.
Name
Performance
index
Technical
p
aramete
r
Type YSV201
Velocity sensor
range
response
frequency
sensitivit
y
185mm/s
4~1000Hz
27mV(mm/s)
Type
ULT2035V
Accelerometer
range
response
frequency
sensitivity
50g
1-12000Hz
98.24mV/g
All experiment materials comprised function
generator, power amplifier, vibration exciter, velocity
sensor, accelerometer, NI-9234 AD inputs module, NI
cDAQ-9172 eight-slot USB chassis, and a computer
with Signal Express software. Experiment installation
is shown in figure 3.
Figure 3: Experiment installation
In this experiment, according to the vibration
frequency range of corn harvester, which belongs to
low-frequency stage (Lin et al. 2015), it is better to
use velocity to reflect vibration features based on
vibration theoretical background. Therefore, velocity
peak-peak value are measured to make a performance
contrast. Data were acquired at the operational
condition of sinusoidal signal from frequency 4Hz to
44Hz. And to get velocity values, the data acquired
with accelerometer was processed in single
integration algorithm. In Figure 4, measured data are
shown.
Development of an Assembly Consistency End-of-Line Inspection System for Corn Harvester
653

Figure 4: Peak-peak value contrast histogram between
velocity sensor and accelerometer
From Figure 4, the two histograms are closed to
each other from frequency 4Hz to 14Hz, when the
frequency exceed 14Hz, they are alike in the tendency,
but the result tested by accelerometer fluctuate
tremendously. So the comparison leads to the
conclusion that the velocity sensor is better while the
frequency belongs to low-frequency stage. It is
correspond to what we know that the accelerometer is
suitable for frequency over 1000Hz. In a conclusion,
from the stability perspective, measured with the
velocity sensor is better.
4 RESULTS AND DISCUSSION
To judge assembly consistency, frequency band
energy is a most commonly used parameter. The total
energy generated by all the peaks in the band is
calculated by the following formula:
E=

∑
F
i
2
n
i=1

N
BF
(1)
Where the n is number of spectral lines in
frequency band, 

is the spectrum value, and 

is noise bandwidth of the selected window function.
According to the carried out experiment, corn
harvest pulverizer was chose as analysis object. In
order to analyze the test result, two kinds of data
processed methods were used. One is statistical
analysis. Total energy of corn harvester from
frequency 0 to 500Hz was calculated in level and
vertical directions. As we tested 3 times on each corn
harvester, then made an average of total energy. Table
2 showed the sum of averaged level and vertical total
energy. According to the data in table 2, first we verify
that it obey the normal distribution. Because the
sample is 8, we use the method of non-parametric test
with the statistical analysis software SPSS. Result is
shown in the following table 3:
Table 2: Total energy of frequency band.
Table 3: Single sample Kolmogorov-Smirnov test.
p
arameters energy
N8
Normal
parameter a, b
Mean value 537.4300
Standard
deviation
306.62526
Extreme
difference
A
b
solute value 0.335
Positive 0.335
Negative -0.210
Kolmogorov-Smirnov Z 0.947
Progressive significance(
d
ouble side) 0.332
a. the distribution of test is normal distribution.
b. calculated according to the data.
The progressive significance is 0.332>0.05, so it
obeys the normal distribution.
As it is verified obey the normal distribution, and
the total mean value, variance are unknown, we can
use T test and calculate the confidence interval of
consistency, of which the confidence is 95%. The
sample average, standard deviation and confidence
interval are calculated with the following formulas:
X

=
(x
1
+…+x
n
)
n
(2)
S=
∑
(x
i
-x)
2
n
i=1
n-1
(3)
(X

-t
α
2

(n-1)
S
√
n
,X

+t
α
2

(n-1)
S
√
n
) (4)
Where the n is sample, 

is sample average, S
is variance, and α is significance level.
The calculated result is X

=537.43,S=306.63, and
confidence interval is (281.58, 793.28). According to
the confidence interval, the third and fifth corn
harvests are incongruous. A reference index value K
can be calculated. K=793.28/281.58=2.82. If the K
value calculated less than 2.82, then the assembly is
consistent, otherwise, it is incongruous. K value
calculation formula:
K=
P
281.58
(5)
Where the P is the total power.
No 1 2 3 4
Total
power
290.43 305.61 1148.42 437.24
No567 8
Total
power
863.42 471.82 333.26 449.24
ICECTT 2018 - 3rd International Conference on Electromechanical Control Technology and Transportation
654

Six bands spectrum is an effective way to analyze
vibration signals by dividing the frequency domain
signals into six frequency bands. On each band, it
reflects correspond faults. Usually the first frequency
band is less than 1 times working frequency, second,
third, fourth and fifth frequency band contain 1times,
2 times, 3 times and 4 times, sixth band contains 4 to
12 frequency doubling (James, 1990). As the engine
speed of tested corn harvesters is 2400rpm, so for the
corn harvest pulverizer, its major energy focuses on
the frequency from 0 to 500Hz, we divide it into 6
frequency band, first is from 0 to 35Hz, second is 35
to 75, and the increment is 40 from the second to the
fourth, the last frequency band is from 195 to 500Hz.
On each band, energy was calculated and made an
average with 3 times. Figure 5 and figure 6 show the
frequency band of 8 corn harvesters in level and
vertical directions, respectively.
Figure 5: The energy of line chart in level direction.
Figure 6: The energy of line chart in vertical direction.
From the line chart, it is easy to discern the third
corn harvest and the fifth is incongruous with the
others both in level and vertical direction. According
to the knowledge of mechanical vibration fault
diagnosis, we know rotary machine exist rotor
unbalance if it is abnormal in working frequency, and
if it is unusual in 2 times working frequency indicates
rotor misalignment. The pulverizer is a kind of
rotating machine, from figure 4 and figure 5, we can
judge the third one with fault of unbalance and the
fifth one exist misalignment.
Statistical analysis and six bands can judge
assembly congruous of corn harvester both well. The
statistical analysis makes a judgment quantitatively,
while six bands judges more intuitionistic, which can
also reflect specific fault.
5 CONCLUSIONS
In this work, it demonstrates an assembly consistency
inspection system and a method of consistency
judgment about corn harvest transmission system.
The result using vibration signals corroborate its
efficiency. The use of frequency bands energy help to
separate the energy of different corn harvests in
different bands, which make the corn harvest, whose
assembly is incongruous, remarkable and easy to
distinguish form the chart. It gives us an intuitionistic
judgment.
This method proves to be extremely useful in
cases where the vibration signals are the most
important information for the complicated corn
harvests. And it make the inspection of consistency
more scientific, efficient and automatic, which can
guarantee the whole performance of corn harvests
well before they start working.
ACKNOWLEDGMENTS
This work was financially supported by National
Science & Technology Pillar Program during the
thirteenth Five-year Plan Period (2017YFD0700204).
The corresponding author is Du Chen.
REFERENCES
Suzuki, T. et al. 2001. Assembly Reliability Evaluation
Method. International Symposium on Assembly and
Task Planning Soft Research Park, Fukuoka, Japan May
28-29, 2001.
Kurt, A. et al. 2000. Assembly quality method: a tool in aid
of product strategy, design, and process improvements.
Design Engineering Technical Conferences Baltimore,
MD September 10-13, 2000.
Milberg, J. & Wisbacher, J. 1992. Acoustic Test Procedures
—A Powerful Method for Quality Assurance and
Process Monitoring in Assembly [J] CIRP Annals -
Manufacturing Technology. 41(1):25-28
Development of an Assembly Consistency End-of-Line Inspection System for Corn Harvester
655

Yi, K. & WANG, M. 2009. Analysis and countermeasures
for Chinese agricultural machinery product quality [J].
Chinese Agricultural Mechanization, (5): 07~10
Lin, J. et al. 2015. Status of Maize Harvesting Machinery
and Its Development Trend in China [J]. Agricultural
Science & Technology and Equipment(6)
James, E, Berry. 1990. How to Specify Machinery
Vibration Spectral Alarm Bands. Sound and Vibration.
ICECTT 2018 - 3rd International Conference on Electromechanical Control Technology and Transportation
656