Professional Documents
Culture Documents
Prediction of Seed Distribution in Rectangular Vibrationg Tray Using Grey Model and Artificial Neural Network 2019
Prediction of Seed Distribution in Rectangular Vibrationg Tray Using Grey Model and Artificial Neural Network 2019
Prediction of Seed Distribution in Rectangular Vibrationg Tray Using Grey Model and Artificial Neural Network 2019
ScienceDirect
Research Paper
Zhan Zhao a,b,*, Mingzhi Jin a, Chunjie Tian a, Simon X. Yang b,**
a
Institute of Agriculture Engineering, Jiangsu University, Zhenjiang, 212013, China
b
School of Engineering, University of Guelph, Guelph, N1G 2W1, Canada
article info
To maintain good continuous working performance in a vacuum plate seeder, it is impor-
Article history: tant to monitor the distribution of seeds in real time and automatically adjust vibration
Received 27 June 2018 parameters accordingly. Seed motion in a rectangular vibrating tray with vibration varying
Received in revised form with time and interference by direction angle was simulated using the discrete element
30 August 2018 method (DEM). A plane model P was used to describe the variation of seed layer thickness.
Accepted 19 September 2018 Four square areas on the bottom of the tray were divided symmetrically near the four
corners to measure seed layer thickness, and a monitoring plane model Pm was established.
DEM simulation results showed that the models Pm and P had the similar change rules,
Keywords: although there were some differences in fitting parameters. There was obvious time delay in
Seed distribution the change of Pm compared with P. Therefore, a grey system model (GM) was adopted to
Prediction method predict the change of Pm, and two back-propagation (BP) neural networks which take GM
Vibrating tray prediction results as input parameters were developed respectively. Then, according to the
Discrete element method BP neural networks outputs, a prediction plane model Pp was proposed to predict the seed
Grey model distribution. Experiments were carried out on a test-rig to validate these predictions. The
Neural network seed distribution plane P was measured manually, the monitoring plane Pm was established
using seed layer thickness measurement results and the prediction plane Pp was established
using the GM and BP neural networks. The results indicated that the proposed method had
good precision and stability and provides the basis for the design of an automatic control
system for the vibrating tray to promote a uniform seed distribution.
© 2018 IAgrE. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Institute of Agriculture Engineering, Jiangsu University, Zhenjiang, 212013, China. Fax: þ86 (0)511 88797338.
** Corresponding author. Fax: þ1 519 8244120 (ext. 52437).
E-mail addresses: zhaozhan@ujs.edu.cn (Z. Zhao), syang@uoguelph.ca (S.X. Yang).
https://doi.org/10.1016/j.biosystemseng.2018.09.017
1537-5110/© 2018 IAgrE. Published by Elsevier Ltd. All rights reserved.
b i o s y s t e m s e n g i n e e r i n g 1 7 5 ( 2 0 1 8 ) 1 9 4 e2 0 5 195
reciprocating vibration was analysed using DEM, and a According to the structure of the vacuum plate seeder, the
monitoring method for seed mass in a small area was pro- size of rectangular tray model was 700 mm 420 mm and the
posed (Zhao, Wu, Yin, & Tang, 2015). In this study seed motion material was aluminium alloy # 7075. Taking the centre of the
in a rectangular vibrating tray is simulated, under vibration tray as the origin, an inertial coordinate X0 ¼ [X Y Z] is estab-
with interference in direction angle. The objective is to lished, shown in Fig. 1. In ideal conditions, the tray is vertically
establish a seed distribution model according to seed distri- vibrated along the Z axis. But, some uncertain interference
bution state, and to propose a corresponding monitoring factors such as the precision of installation and the horizontal
method. Then, according to the monitoring results, an posture of the frame are inevitable in practical operations.
improved prediction method which combines grey system This usually leads to a small time varying change in vibration
and artificial network was developed and the prediction direction. Here, the angle between the vibration direction
model optimised based on the error analysis and validated by vector of the tray nv and the Z axis is defined as gv, and the
experiment to evaluate its prediction accuracy. angles between the projection of nv and the X, Y axes are
defined as av, bv.
For rice nursery seeding operations, many experimental
2. Materials results have indicated that the optimum vibration frequency f
and amplitude A of the tray are near 11 Hz and 4 mm, and the
In this section, under the excitation of vibrating tray, the average seed layer thickness h0 in the tray should be main-
motion of seeds with different layer thickness was simulated tained in a range of 15e25 mm (Chen, Gong, Li, Li, & Xu, 2015).
using DEM. A plane model was used to describe the distribu- Therefore, DEM simulations are carried out with f of 11 Hz and
tion of seeds in the tray. According to the central coordinates A of 4 mm, h0 is selected as 15, 20 and 25 mm respectively. The
and the corresponding seed layer thickness of all regions, simulation process was divided into two stages. In the initial
plane fitting models of seed distribution were established 10 s, the tray was sinusoidal vibrated in vertical direction.
using multiple regression method. The variation of seed dis- Then, a continuous random time varying interference angle gt
tribution under the vibration with time varying interference in was applied to the vibration direction nv from 10 to 70 s.
direction angle was discussed. Variations of the vibration direction in DEM simulations are
given in Table 2.
2.1. DEM simulations
2.2. Seed distribution model
In the DEM, Newton's second law of motion and Euler's dy-
namics equation are usually used to describe the translational The DEM simulation of seeds motion in the tray is shown in
and rotational movements of particulate assemblies. Inter- Fig. 2. In order to analyse the seed distribution state, the tray
action force models are established in the framework of was divided into 10 6 small square regions.
classical mechanics (Horabik & Molenda, 2016). In order to Under the excitation of vibration, the rice seeds in the tray
represent the various shapes of agricultural seeds, a multi- are always in discrete motion state. Thus, the relationship
sphere approach method was adopted, in which a certain between number of seeds and the stable seed layer thickness
number of spheres were “glued” together (Radvilaite, Ramı́res- when the tray is static was firstly established. The stable seed
Go mez, & Kac ianauskas, 2016; Wang, Yu, Lv, Wang, & Fu, layer thickness in each region was then calculated by count-
2017). A commercial three-dimensional DEM code (EDEM® ing the number of seeds. From DEM simulation of parameter #
2.5, EDEM, Edinburgh, UK) was used. According to the physical 2 (see Table 2), the calculated seed layer thickness in all re-
properties of rice seeds, a triaxial ellipsoidal particle model gions at time t of 10, 30 and 50 s is shown in Fig. 3. When the
was established using multiple spheres. Its semi-axes were tray was vibrated vertically, the seeds were always in vibra-
2.85, 1.55 and 1.35 mm, and the mass of single seed was tional motion at their original locations and seed flow in the
26.5 103 g. Contact forces were calculated using the Hertz- tray could be ignored. As shown in Fig. 3 (a), seed distribution
Mindlin model. Table 1 summarises the values of the material in the tray can be well fitted using a horizontal plane model.
properties required in the DEM. When the tray is vibrated with interference direction angle gt,
seeds will flow within the tray. The distribution presented
Fig. 3 e Fitting results of seeds distribution using a plane model: (a) t ¼ 10 s, (b) t ¼ 30 s, (c) t ¼ 50 s.
198 b i o s y s t e m s e n g i n e e r i n g 1 7 5 ( 2 0 1 8 ) 1 9 4 e2 0 5
simulation data showed that the root mean squared error monitoring plane Pm, an improved prediction method incor-
(RMSE) was less than 2.5 mm and the coefficient of determi- porating a grey system model and back-propagation (BP)
nation was >0.9. neural network is proposed.
The seed distribution plane model P can be expressed by
3.1. Monitoring method
P : Ax þ By þ Cz þ D ¼ 0 (1)
where the normal vector of the plane np ¼ [A, B, C], the angle A seed layer thickness monitoring method and device for
between the np and the Z axis is defined as gp, and the angles measuring seeds impact force in a small area has been
between the projection of np and the X, Y axes are defined as developed (Zhao et al., 2015). Under vibration excitation, seeds
ap and bp. The fitting plane model P can be determined by gp impact with the tray and generate a periodic force signal. In
and ap because there is ap þ bp ¼ 90 . DEM simulation data was order to obtain the stable impact force, a second-order But-
sampled with frequency of 1 Hz. The seed distribution plane terworth low pass filtering calculation was used. The linear
models at different times were established according to vari- relationship between the seed layer thickness and the stable
ations of the seed layer thicknesses. The calculated variations impact force was established.
of gp and ap at different times are given in Fig. 4. In practice, it is impossible to monitor the seed layer
thickness in many regions. For the plane distribution model,
four 60 mm 60 mm square areas on the bottom of the tray
3. Methods were delineated, shown in Fig. 1. The distances between the
centres of each area and the corresponding vertex in X-axis
Four square areas on the base of the tray were divided to and Y-axis directions were 120 and 95 mm. With the simula-
measure the seed layer thickness. According to the mea- tion parameters given in Table 2 and the critical frequency of
surement results, a monitoring plane model Pm of seed dis- low pass filter of 1 Hz, the received variations of seed layer
tribution was established. By analysing the difference thickness h1, h2, h3 and h4 in four monitoring areas are shown
between the measured seed distribution plane P and the in Fig. 5. The measured seed distribution basically followed
Fig. 4 e Comparation of the actual distribution plane P and the monitoring plane Pm: (a) Parameter # 1, (b) Parameter # 2,
(c) Parameter # 3.
b i o s y s t e m s e n g i n e e r i n g 1 7 5 ( 2 0 1 8 ) 1 9 4 e2 0 5 199
Fig. 5 e Variations of seeds layer thickness in four monitoring areas: (a) Parameter # 1, (b) Parameter # 2, (c) Parameter # 3.
the plane model. Therefore, a plane model was used to fit the 3.2. Grey prediction model
seed distribution. According to the central coordinates of four
areas [xm1, ym1], [xm2, ym2], [xm3, ym3] and [xm4, ym4], a seed To improve the real time performance of the measurement,
distribution monitoring plane model Pm was established using predicting the change of monitoring plane Pm was required.
multiple regression, which was expressed as With signal processing this can be calculated accurately.
However, seed motion under vibration excitation with inter-
Pm : Am x þ Bm y þ Cm z þ Dm ¼ 0 (2)
ference from direction angle is complex and almost impos-
The normal vector of the plane is nm ¼ [Am, Bm, Cm]. The sible to establish using a defined mathematical model.
angle between the nm and the Z axis is defined as gm, and the Therefore, the variation of seed distribution can be regarded
angles between the projection of nm and the X, Y axes are as a grey system.
defined as am and bm. The multiple regression results showed Grey theory demonstrates the optimal ability of perform-
that the root mean squared error (RMSE) was generally less ing fitting predictions using small data sets and limited in-
than 2 mm. The variations of gm and am at different times are formation (Bezuglov & Comert, 2016; Kayacan, Ulutas, &
given in Fig. 4. Kaynak, 2010; Yin, 2013). It employs a data generation
The variation of measured seed distribution plane P was in method, rather than a statistical regulation, to obtain a more
real time because it was established by counting the number regular generation sequence from the initial data. Grey pre-
of seeds. The monitoring plane Pm was established by dictions have been utilised in agriculture, industry, economics
measuring the seed layer thickness, and it was inevitable that and to study natural phenomenon. Although various types of
the signal processing caused some time delay. So, it can be grey models have been developed, most of the research are
seen from Fig. 4 that the variations of gm and am both have still focused on the first order with one variable grey model
obvious time delays relative to gp and ap, but the delay period GM(1,1) in prediction applications because the GM(1,1) has
was generally <3 s. The variation of gm basically reflected the good computational efficiency and also is the foundation of
change rule of gp although there were some differences in the other grey prediction models.
value. Smaller gp indicated that the seed distribution was Assuming the obtained time data sequence with n samples
uniform. With the increase of gp, it was shown that ap and am (time point) and named as original numerical sequences X(0) is
have stable and consistent change rules. given by
200 b i o s y s t e m s e n g i n e e r i n g 1 7 5 ( 2 0 1 8 ) 1 9 4 e2 0 5
0:5 xð1Þ ð2Þ þ xð1Þ ð1Þ 1 xð0Þ ð2Þ hj ¼ g vji xi ; i ¼ 0; 1; ,,,; N; j ¼ 1; 2; ,,,; L (10)
6 0:5 xð1Þ ð3Þ þ xð1Þ ð2Þ 17 6 ð0Þ 7
7; Y ¼ 6 x ð3Þ 7
i¼0
B¼6
4
/ /5 4 / 5
where vij is the connection weighs between neuron i in input
0:5 xð1Þ ðnÞ þ xð1Þ ðn 1Þ 1 xð0Þ ðnÞ
layer and the neuron j in hidden layer, xi is the ith input var-
By solving Eq. (6), the prediction formula can be presented iable, N is the number of input neurons and L is the number of
as neurons in hidden layer. The activation function of the output
neurons is a linear function and the outputs Ok are given by
b ðnþp1Þ b
bð1Þ
x m ðn þ pÞ ¼
bð1Þ
x m ð0Þ e þ (8)
a a X
l
Ok ¼ vjk hj ; k ¼ 1; 2 (11)
Hence, the desired prediction output at n step and p pre- j¼0
Fig. 6 e GM prediction results of gm and am: (a) Parameter # 1, (b) Parameter # 2, (c) Parameter # 3.
It is important to determine the number of neurons in increasing the number of hidden neurons (Zhang, Yang,
hidden layer L for achieving the desired precision and good Mittal, & Yi, 2002). The initial learning rate h0 was 0.005. To
generalisation performance. Based on the recommended avoid oscillation or divergence and improve training speed,
calculation method, an L value of 15 was determined by the learning rate h was select as an exponential decay func-
repeated trials because the accuracy did not improve while tion hðtÞ ¼ h0 expt=t .
Fig. 7 e Three-layer BP neural network: (a) prediction of ap, (b) prediction of gp.
202 b i o s y s t e m s e n g i n e e r i n g 1 7 5 ( 2 0 1 8 ) 1 9 4 e2 0 5
Fig. 8 e BP model prediction results of gp and ap: (a) Parameter # 1, (b) Parameter # 2, (c) Parameter # 3.
Finally, using the output xBP, yBP and gBP, a prediction plane according to the error analysis with different BP model input
model Pp was established to describe the actual seed distri- parameter combinations. Finally, experiments are carried out
bution plane model P, which can be expressed by on the test-rig to verify the feasibility and accuracy of the
proposed prediction method.
Pp : singBP cosaBP x þ singBP sinaBP y þ cosgBP z þ DP ¼ 0 (16)
4.1. Prediction results
4. Results and discussion From three DEM simulations 70% data and the corresponding
GM prediction results were selected for training and the
In this section, the prediction results are calculated using the
remaining 30% data was used for testing the BP models. In
DEM simulation data. Then, the prediction model is optimised
order to analyse the influence of GM outputs on the prediction
network, the error of prediction model Pp was effectively Altikat, S., Celik, A., & Gozubuyuk, Z. (2013). Effects of various no-
reduced. till seeders and stubble conditions on sowing performance
With the average seed layer thickness h0 of 15, 20 and and seed emergence of common vetch. Soil and Tillage
Research, 126, 72e77.
25 mm, experiments were repeated for 5 times for each
Atici, U. (2011). Prediction of the strength of mineral admixture
parameter. When the gp was greater than 1 , the prediction concrete using multivariable regression analysis and an
results indicated that the difference between ap and aBP was artificial neural network. Expert Systems with Applications, 38,
less than 7.5 and the difference between gp and gBP was less 9609e9618.
than 0.25 . Bezuglov, A., & Comert, G. (2016). Short-term freeway traffic
parameter prediction: Application of grey system theory
models. Expert Systems with Applications, 62, 284e292.
Boac, J. M., Ambrose, R. P. K., Casada, M. E., Maghirang, R. G., &
5. Conclusions
Maier, D. E. (2014). Applications of discrete element method in
modeling of grain postharvest operations. Food Engineering
DEM simulation results have shown that the seed distribution Reviews, 6(4), 128e149.
in a vibrating tray can be fitted using a plane model P. By Chen, J., Gong, Z. Q., Li, Y. M., Li, J. H., & Xu, Y. (2015).
monitoring the seed layer thickness in four square areas Experimental study on nursing seedlings of super rice
located on the bottom of the tray, a monitoring plane model precision seeder device. Transactions of the Chinese Society for
Agricultural Machinery, 46(1), 73e78.
Pm of seed distribution was established. Influenced by the
Craessaerts, G., Saeys, W., Missotten, B., & De Baerdemaeker, J.
signal processing of seeds thickness measurement, the
(2010). Identification of the cleaning process on combine
change of monitoring plane Pm was obviously time delayed harvesters, Part II: A fuzzy model for prediction of the sieve
with respect to the measured seed distribution plane P. losses. Biosystems Engineering, 106(2), 97e102.
The comprehensive application of GM and BP was pro- Ding, Y. C., Liao, Q. X., & Huang, H. D. (2013). The seeding
posed to predict the angle gp between the normal vector np of migration trajectory extraction and analysis of pneumatic
plane model P and the Z axis, and the angle ap between the precision metering device for rapeseed. Journal of Food
Agriculture and Environment, 11(1), 477e482.
projection of np and the X axis respectively.
Fanigliulo, R., & Pochi, D. (2011). Air-flow distribution efficiency of
Results for seed distribution prediction using DEM simu- a precision drill used in the sowing of different graded seeds.
lation showed that optimal precision was received while using Journal of Agricultural Science and Technology A, B1, 655e662.
the GM outputs with prediction step of 0, 1 and 2 as the BP Gaikwad, B. B., & Sirohi, N. P. S. (2008). Design of a low-cost
neural networks input parameters. When gp increased from 1 pneumatic seeder for nursery plug trays. Biosystems
to 2 and 3 , the corresponding difference between ap and aBP Engineering, 99(10), 322e329.
gradually decreased from 22.8 to 9.3 and 3.2 , and the dif- Gong, Z. Q., Chen, J., Li, Y. M., & Li, J. H. (2014). Seed force in
airflow field of vacuum tray precision seeder device during
ference between gp and gBP could be controlled below 0.3 .
suction process of seeds. Transactions of the Chinese Society for
Repeated experiments were carried out on vacuum plate Agricultural Machinery, 45(6), 92e97.
seeder test-rig. When gp was >1 , the obtained difference be- He, R. Y., Luo, H. Y., Li, Y. T., Wang, X. H., & Zhang, L. (2008).
tween ap and aBP was less than 7.5 and the difference be- Comparison and analysis of different rice planting methods in
tween gp and gBP was less than 0.25 . China. Transactions of the Chinese Society of Agricultural
The paper thus provides an effective way for prediction of Engineering, 24(1), 167e171.
Horabik, J., & Molenda, M. (2016). Parameters and contact models
seed distribution in a rectangular vibrating tray. This is
for DEM simulations of agricultural granular materials: A
important for automatic adjusting the vibration parameters to
review. Biosystems Engineering, 147, 206e225.
promote the uniform distribution of seeds in the tray, and to €
Karayel, D., Barut, Z. B., & Ozmerzi, A. (2004). Mathematical
improve the continuous working performance of a vacuum modelling of vacuum pressure on a precision seeder.
plate seeder. Biosystems Engineering, 87(4), 437e444.
Kayacan, E., Ulutas, B., & Kaynak, O. (2010). Grey system theory-
based models in time series prediction. Expert Systems with
Applications, 37(2), 1784e1789.
Acknowledgements Kruggel-Emden, H., Rickelt, S., Wirtz, S., & Scherer, V. (2008). A
study on the validity of the multi-sphere discrete element
This work is supported by the National Natural Science method. Powder Technology, 188(2), 153e165.
Foundation of China (No. 51775246), the Science Foundation of Liu, C. L., & Song, J. N. (2004). Influence of seed tray vibration on
work performance of suction-vibration type precision seed
Anhui Province (No. 1808085QE172) and a Project Funded by
device. Journal of China Agricultural University, 9(2), 12e14.
the Priority Academic Program Development of Jiangsu Higher Movahedi, E., Rrzvani, M., & Hemmat, A. (2016). Design,
Education Institutions (PADP). development and evaluation of a pneumatic seeder for
automatic planting of seeds in cellular trays. Journal of
Agricultural Machinery, 4(1), 65e72.
references €
Onal, I., Degirmencioglu, A., & Yazgi, A. (2012). An evaluation of
seed spacing accuracy of a vacuum type precision metering
unit based on theoretical considerations and experiments.
Abdolahzare, Z., Asoodar, M. A., Kazemi, N., Rahnama, M., & Turkish Journal of Agriculture and Forestry, 36, 133e144.
Radvilaite, U., Ramı́res-Go & Kac
mez, A., ianauskas, R. (2016).
Mehdizadeh, S. A. (2016). Optimization of the most important
operational parameters of a pneumatic seeder using real-time Determining the shape of agricultural materials using
monitoring for cucumber and watermelon seeds. Journal of spherical harmonics. Computers and Electronics in Agriculture,
Agricultural Machinery, 6(1), 35e48. 128, 160e171.
b i o s y s t e m s e n g i n e e r i n g 1 7 5 ( 2 0 1 8 ) 1 9 4 e2 0 5 205
Tijskens, E., Ramon, H., & De Baerdemaeker, J. (2003). Discrete Yazgi, A., & Degirmencioglu, A. (2014). Measurement of seed
element modeling for process simulation in agriculture. spacing uniformity performance of a precision metering unit
Journal of Sound and Vibration, 266, 493e514. as function of the number of holes on vacuum plate.
Topakci, M., Karayel, D., Canakci, M., Furat, S., & Uzun, B. (2011). Measurement, 56(6), 128e135.
Sesame hill dropping performance of a vacuum seeder for Yi, S. J., Liu, Y. F., Wang, C., & Tao, G. X. (2014). Experimental study
different tillage practices. Applied Engineering in Agriculture, on the performance of bowl-tray rice precision seeder.
27(2), 203e209. International Journal of Agricultural and Biological Engineering,
Wang, X. M., Yu, J. Q., Lv, F. Y., Wang, Y., & Fu, H. (2017). A multi- 7(1), 17e25.
sphere based modelling method for maize grain assemblies. Yin, M. S. (2013). Fifteen years of grey system theory research: A
Applied Power Technology, 28, 584e595. historical review and bibliometric analysis. Expert Systems with
Xing, H., Wang, Z. M., Luo, X. W., Cao, X. M., Liu, C. B., & Zang, Y. Applications, 40, 2767e2775.
(2017). General structure design and field experiment of Zhang, Q., Yang, S. X., Mittal, G. S., & Yi, S. J. (2002). Prediction of
pneumatic rice direct-seeder. International Journal of performance indices and optimal parameters of rough rice drying
Agricultural and Biological Engineering, 10(9), 31e42. with neural networks. Biosystems Engineering, 83(3), 281e290.
Yang, Z. Y., Li, N., Ma, J., Sun, Y. J., & Xu, H. (2014). High-yielding Zhao, Z., Wu, Y. F., Yin, J. J., & Tang, Z. (2015). Monitoring method
traits of heavy panicle varieties under triangle planting of rice seeds mass in vibrating tray for vacuum-panel
geometry: A new plant spatial configuration for hybrid rice in precision seeder. Computers and Electronics in Agriculture, 114,
China. Field Crops Research, 168, 135e147. 25e31.