You are on page 1of 10

Sensors & Transducers Journal, Vol. 129, Issue 6, June 2011, pp.

24-32

Sensors & Transducers


ISSN 1726-5479
© 2011 by IFSA
http://www.sensorsportal.com

Design, Development and Calibration of a Portable Capacitive


Moisture Meter
Mahmoud Soltani, Reza Alimardani
Department of Agricultural Machinery Engineering, Faculty of Agricultural Engineering &
Technology, University of Tehran, Karaj, Iran
E-mail: mahmoodsoltani39@yahoo.com

Received: 30 April 2011 /Accepted: 20 June 2011 /Published: 30 June 2011

Abstract: Dielectric properties of grains, fruits and vegetables are used in most fields of agriculture
such as microwave drying treatment and precision agriculture. In this paper a portable instrument was
designed and developed to predict moisture content of grains and seeds by measuring the dielectric
constant of sample. A high correlation was observed between moisture content and dielectric constant.
The best result was obtained at 1 MHz frequency. The system was calibrated for wheat, corn, millet,
lintel, pea and black eyed pea. Fast response, cheapness and reliability are the main characteristics of
this instrument. Copyright © 2011 IFSA.

Keywords: Capacitive sensor, Instrument, Moisture content.

1. Introduction
Today, the dielectric property of agricultural products is one of the subjects of researches and studies.
The dielectric properties are mainly used in microwave heating [12], prediction of fruit ripeness
[16, 13], estimation of the fruit and vegetable volume [9, 11], quality measurement of eggs [14] and
monitoring bees passing [8]. One of the most important applications of dielectric properties is
measuring the moisture content of grains and seeds. Many scientists studied the correlation between
moisture content and dielectric properties of grains and seeds. Casada and Armstrong (2008) measured
the moisture content of wheat with a fringing field capacitive sensor. They extracted the linear
calibration equations over a temperature range of 10 °C to 30 °C [6]. Das et al. (2010) used dielectric
constant of soybean at 2.45 GHz and 24 ˚C to predict its moisture content. They extracted a quadratic
equation between moisture content and dielectric constant of soybean [7]. Sacilik and Colak (2010)
determined the dielectric properties of corn seeds as functions of moisture content, bulk density and
frequency. They reported the moisture content was the most significant factor affecting the dielectric
24
Sensors & Transducers Journal, Vol. 129, Issue 6, June 2011, pp. 24-32

properties of corn seeds, also dielectric constant and loss factor increased with increasing moisture
content [15]. Moisture content is the most important property that influences the bread quality. Bread
undergoes several physicochemical, sensory and microbial changes during storage that results in a
rapid loss of freshness. These changes depend on the moisture content of bread [4]. Bhatt et al. (2008)
designed an instrument based on impedance spectroscopy technique to predict moisture content of
bread during storage. They obtained a linear relationship between measured impedance and moisture
content of bread [4]. These studies show the importance of moisture content measurements in
agriculture. Different methods of measuring moisture content were developed. One of these methods is
oven drying which is a destructive technique. Another technique is Nuclear Magnetic Resonance
(NMR) which is an expensive method, so a simple, rapid and nondestructive method is needed to
measure the moisture content of agricultural products. Capacitive method seems to be a good
technique for measuring moisture content of products. This technique has all attributes mentioned
above.

This study aims to design an instrument for predicting moisture content of grains and seeds. In this
research, the instrument was calibrated for wheat, corn, millet, lintel, pea and black eyed pea. The
system can be calibrated for other grains and seeds.

2. Materials and Methods

2.1. Sample Providing

The required quantity of grains (wheat, corn, millet, lintel, pea and black eyed pea) were provided and
cleaned to prepare samples at five levels of moisture content. At each level, about 100 cc of seeds were
provided. The initial moisture content of seeds was determined by oven method (level 3) (as reported
in Table 1). In order to reach a higher moisture level and to prohibit gemmating, the seeds must be
placed near the vapor of water, so the samples were exposed to saturated air in an isolated box at 30 ˚C
for 18 hours (level 4) and 36 hours (level 5), respectively. To reach a lower moisture content level, the
oven method was used at 60˚C for 24 hours (level 2) and 48 hours (level 1). After providing the
samples, they were kept in the refrigerator at 4 ˚C for 72 hours for equalization.

2.2 Instrumentation

An instrument based on capacitive technique was designed and developed to measure dielectric
constant of grain at various moisture contents. The block diagram of the designed instrument is shown
in Fig. 1. The instrument consists of a signal conditioning circuit, a 10-bit Microcontroller
(ATmega 32) interfaced with a 16×2 LCD display and sinusoidal function generator (XR2206).
Function generator produces an AC current with variable magnitude and frequency. The produced sine
signal is fed to capacitive sensor and the output signal from the sensor is sent to the signal conditioning
circuit. The final output voltage is measured by ADC and the capacitance and dielectric constant of
sample is computed by a microcontroller and the results are displayed on an LCD. Specific software
was developed by C − language for calculations.

Fig. 2 shows the system's electrical circuit. A sine wave in the range of 1 kHz to 1 MHz was used to
extract the best calibration equation. The main components of sinusoidal signal generator circuit are
IC-XR2206, resistors R1 and R3. R1 at pin 7 provides the desired frequency tuning and R3 at pin 3
adjusts the amplitude of signal. The output voltage from the sensor is converted to DC current by a
diode bridge, and then the A/D unit of ATmega 32 measures the voltage. The system was calibrated by
standard capacitors and the relation between measured voltage and capacitance had previously been
extracted.
25
Sensors & Transducers Journal, Vol. 129, Issue 6, June 2011, pp. 24-32

Fig. 1. Block diagram of instrument for measuring the dielectric constant of seeds.

2.3. Cylindrical Capacitive Sensor

Fig. 3 shows the capacitive sensor that was used in this research. The material for electrodes material
was selected from aluminum. To avoid any occurrence of conduction, two polyethylene plates were
used in constructing the sensor. Each electrode was covered by a polyethylene layer with 1 mm
thickness.

2.4. Dielectric Calculation

The capacitance of a cylindrical capacitor can be calculated by Eq.1.

2  r  0 h
C
b (1)
Ln( )
a

where εr is the dielectric constant of material, ε0 is the permittivity of air (8.85×10-12 F/m), h is the
height of material, b and a are the radius of the outer and inner concentric cylinders.

It can be seen that each side of the polyethylene intermediary in Fig. 3 is in contact with the electrode
and seed, so polyethylene layer became the series capacitance to the measuring system. The equivalent
circuit diagram is shown in Fig. 4. In the diagram, CP1 and CP2 are the polyethylene capacitance, Cm is
the measured capacitance, and Ceq is the equivalent capacitance of the sample (CS) and the air gap
(Cair) that existed among the seeds in the container, so the seeds and air make parallel capacitors. To
measure the dielectric constant of polyethylene, a rectangular parallel plate capacitor with
polyethylene dielectric material was constructed and its dielectric constant was calculated.

26
Sensors & Transducers Journal, Vol. 129, Issue 6, June 2011, pp. 24-32

Fig. 2. Electronic circuit diagram of the device for banana ripeness measurement.

27
Sensors & Transducers Journal, Vol. 129, Issue 6, June 2011, pp. 24-32

Fig. 3. Cylindrical capacitive sensor filled with seed. (a) Section cut. (b) Top view.

Fig. 4. The equivalent circuit of capacitive sensor.

Ceq can be calculated by Eq. 2.

1
C eq 
1 1 1 (2)
 
Cm C p1 C p 2

The ratio of air gap volume to total volume of filled capacitor is defined as the porosity (P) of seed, so
the height of air gap in capacitor (hair) is P×h and the height of sample (hs) is (1-P)×h. therefore:

28
Sensors & Transducers Journal, Vol. 129, Issue 6, June 2011, pp. 24-32

2 π ε0 Ph
C air 
b (3)
Ln( )
a

C s  C eq  C air (4)

b
C s Ln( )
a (5)
s 
2   0 ( 1  P )h

The porosity of each grain had been calculated by previous researchers whose findings were used in
this study [1-3, 5, 10, 17].

2.5. Experiments

Dielectric measurement of seeds was carried out at 5 levels of moisture content at 1 kHz, 10 kHz,
100 kHz, 500 kHz and 1 MHz frequencies. After electrical experiments, the moisture content of each
sample was measured using oven method. The moisture content (%MCdb) is calculated on dry basis by
Eq.6. All measurements were performed in a laboratory with an average room temperature of 25 ˚C.

ww ( w  wd )
% MC db   100  i  100 (6)
wd wd

where wi is the initial weight of sample, ww is the weight of sample water and wd is the weight of dried
sample.

Microsoft Excel 2007 was used to analyze data and determine the regression models between the
studied attributes.

3. Results and Discussion


The moisture content (%db) of specimens at 5 levels is presented in Table 1.

Table 1. The provided moisture content for each grain.

Black-eyed
Level Wheat Millet Pea Corn Lentil
pea
1 5.76 5.44 4.34 5.3 6.97 5.47
2 8.48 8.98 6.15 7.78 8.35 7.81
3 11.36 13.64 9.73 12.11 13.13 10.82
4 17.31 16.87 17.17 19.68 16.14 16.55
5 20.36 18.04 19.35 20 17.88 23.14

Relation between εs and moisture content was investigated at mentioned frequencies. The best equation
that fitted to data was found as a quadratic function. The experimental data were fitted to a quadratic
model as:
29
Sensors & Transducers Journal, Vol. 129, Issue 6, June 2011, pp. 24-32

% MC db  a  2S  b  S  c (7)

The parameters of the model (a, b, and c) and R2 are given in Table 2. Results show that at all
frequencies, there is a high correlation between εs and moisture content.

Table 2. Results of the quadratic model of moisture content (%db) as a function of dielectric constant fitted by
regression analysis to the experimental data.

Frequency Wheat Millet Pea Black-eyed pea Corn Lentil


a 0.097 a -0.02 a -0.61 a -0.21 a 0.096 a -0.01
b 3.59 b 1.79 b 9.65 b 4.65 b -0.23 b 2.67
1 kHz
c -6.73 c -0.15 c -19.1 c -7.35 c 5.41 c -3
2 2 2 2 2 2
R 0.981 R 0.977 R 0.994 R 0.989 R 0.997 R 0.989
a -0.26 a -.305 a -1.82 a -0.35 a -0.06 a 0.316
b 6.07 b 6.12 b 21.34 b 7.39 b 2.96 b 0.834
10 kHz
c -14.1 c -13.03 c -43.8 c -14.74 c -7.42 c 0.41
2 2 2 2 2 2
R 0.969 R 0.978 R 0.963 R 0.99 R 0.993 R 0.971
a -0.61 a -0.57 a -1.33 a -0.8 a -0.38 a 0.318
b 10.9 b 9.7 b 16.98 b 12.1 b 8.44 b 2.28
100 kHz
c -28.3 c 23.1 c -35.6 c -25.46 c -28.6 c -4.99
2 2 2 2 2 2
R 0.988 R 0.987 R 0.973 R 0.996 R 0.993 R 0.979
a -0.62 a -1 a -2.1 a -1 a 0.18 a 0.004
b 11.77 b 14.63 b 24.4 b 14.1 b 1.75 b 5.64
500 kHz
c 31.12 c 35.03 c -52.1 c -29.16 c -7.58 c -12.8
R2 0.999 R2 0.998 R2 0.988 R2 0.998 R2 0.988 R2 0.996
a -0.48 a -0.74 a -1.65 a -1.08 a 0.82 a -0.38
b 10.5 b 11.76 b 19.82 b 14.9 b -6.24 b 9.39
1 MHz
c -28.3 c -28.06 c -40.2 c -30.8 c 15.81 c -20.6
R2 0.999 R2 0.997 R2 0.994 R2 0.999 R2 0.989 R2 0.992

To select the best frequency for system calibration, the average value and standard deviation (SD) of
coefficient of determination (R2) were calculated for each frequency (Table 3).

Table 3. The average value and standard deviation (SD) of R2 for each frequency.

Frequency Average SD
1 kHz 0.988 0.0076
10 kHz 0.979 0.014
100 kHz 0.986 0.0086
500 kHz 0.9945 0.0051
1 MHz 0.995 0.0042

30
Sensors & Transducers Journal, Vol. 129, Issue 6, June 2011, pp. 24-32

Results show that 1 MHz signal has the highest value of R2 and the lowest value of SD. So this 1 MHz
sinusoidal wave was selected for system calibration.

The results were in agreement with the similar reported papers for corn. Sacilik and Colak (2010)
proposed the second and third-order polynomial equations to describe the existing relationship
between dielectric properties and moisture content of corn seed [15]. They obtained the best result at
1 MHz with R2 of 0.998. Similar result was reported by Das et al. (2010) for soybean [7].

4. Conclusion
To predict moisture content of grains and seeds, an instrument was designed, developed and calibrated.
Relation between dielectric constant and moisture content was extracted and quadratic trend line was
fitted to data. Results of regression treatment show that the dielectric constant highly correlates with
moisture content. This instrument applied a nondestructive method to predict moisture content of
grain. Designed capacitive moisture meter is a low cost, portable, rapid and reliable system.

References
[1]. Amin, M. N., Hossain, M. A., Roy, K. C., Effects of moisture content on some physical properties of lentil
seeds, Journal of Food Engineering, Vol. 65, 2004, pp. 83–87.
[2]. Ayman, H., Amer, E., Mohamed, M. A., Moustafa, H., Abdul Rahman, O. A. Moisture dependent physical
and mechanical properties of chickpea seeds, International Journal of Agricultural & Biolologycal
Engineering, Vol. 3, 4, 2010, pp. 80-93.
[3]. Baryeh, E. A., Physical properties of millet, Journal of Food Engineering, Vol. 51, 2002, pp. 39-46.
[4]. Bhatt, C. M., Nagaragu, J., Instrumentation to estimate the moisture content in bread using electrical
impedance spectroscopy, Sensors & Transducers, Vol. 97, 10, 2008, pp. 45-54.
[5]. Bulent, C., Yalc, I., Cengiz, O., Physical properties of sweet corn seed (Zea mays saccharata Sturt.),
Journal of Food Engineering, Vol. 74, 2006, pp. 523–528.
[6]. Casada, M, E., Armstrong, P. E., Wheat moisture measurement with a fringing field capacitive sensor, An
ASABE Meeting Presentation, 2008, Paper Pumber. 085207.
[7]. Das, P., Ahmad, Z., Singh, H. N., Microwave density –independent permittivity functions as soybean
seeds’moisture calibrators: A new approach, Ashutosh Prasad Agric Eng Int: CIGR Journal, Vol. 12, 3,
2010, pp. 107-114.
[8]. Campbell, M. J., Dahn, D. C., Ryan, D. A. J., Capacitance-based sensor for monitoring bees passing
through a tunnel, Measurement Science and Technology, Vol. 16, 2005, pp. 2503–2510.
[9]. Jarimopas, B., Nunak, T. Nunak, N., Electronic device for measuring volume of selected fruit and
vegetables, Postharvest Biology and Technology, Vol. 35, 2005, pp. 25–31.
[10].Karimi, M., Kheiralipour, M., Tabatabaeefar, A., Khoubakht, G. M., Naderi, M., Heidarbeigi, K., The
effect of moisture content on physical properties of wheat, Pakistan Journal of Nutrition, Vol. 8, 1, 2009,
pp. 90-95.
[11].Kato, K., Electrical density sorting and estimation of soluble solids content of watermelon, Journal of
Agricultural Engineering Research, Vol. 67, 1997, pp. 161 – 170.
[12].Nelson, S. O., Bartley Jr, P. G. Measuring frequency-and temperature-dependent dielectric properties of
food materials, Transactions of the ASAE, Vol. 43, 6, 2000, pp. 1733–1736.
[13].Nelson, S. O, Wen-Chuan, G., Samir, T., Stanley, J. K., Dielectric spectroscopy of watermelons for quality
sensing, Measurement Science and Technology, Vol. 18, 2007, pp. 1887–1892.
[14].Ragni, L., Gradari, P., Berardinelli, A., Giunchi, A., Guarnieri, A., Predicting quality parameters of shell
eggs using a simple technique based on the dielectric properties, Biosystems Engineering, Vol. 94, 2, 2006,
pp. 255–262.
[15].Sacilik, K., Colak, A., Determination of dielectric properties of corn seeds from 1 to 100 MHz, Powder
Technology, Vol. 203, 2010, pp. 365–370.
[16].Soltani, M., Alimardani, R., Omid, M., Prediction of banana quality during ripening stage using
capacitance sensing system, Australian Journal of Crop Science, Vol. 4, 6, 2010, pp. 443-447.
31
Sensors & Transducers Journal, Vol. 129, Issue 6, June 2011, pp. 24-32

[17].Unal, H., Isik, E., Alposy, H. C., Some physical and mechanical properties of black-eyed pea, Pakistan
Journal of Biological Sciences, Vol. 9, 9, 2006, pp. 1799-1806.

___________________

2011 Copyright ©, International Frequency Sensor Association (IFSA). All rights reserved.
(http://www.sensorsportal.com)

32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

You might also like