File 1_papers 1 to 10

 

MDCCT 2012
6 – 7 February, 2012; Burdwan University

 

Miniaturization of a Rectangular Microstrip Antenna using Slots
and Defective Ground Structure

Mrinmoy Chakraborty1, Biswarup Rana2 and Ankita Mitra3
Dr. B.C.Roy Engineering College, Durgapur, West Bengal, India
Email: [email protected] [email protected] [email protected]

Abstract
In this paper a rectangular microstrip antenna with slots and defective ground structure (DGS) is
proposed. In the absence of DGS and slots the structure found to resonant at 3.21 GHz. When the DGS
and slots are introduced a frequency shift of 3.21 GHz to 1.69 GHz is observed. The main contribution of
this paper is the miniaturization of 72% which is very much encouraging.

Introduction

Microstrip antennas have been used widely in wireless communication due to their light weight, low profile, low
cost and ease of fabrication and excellent compatibility with planar integrated circuits and even non planar
surfaces. In recent years, as the demand of small systems have increased, small size antennas at low frequency
have drawn much interest from researchers [1]. Many kind of miniaturization techniques, such as using of
dielectric substrate of high permittivity [2], slot on the patch, DGS at the ground plane or a combination of them
proposed and applied to microstrip antennas. The other method to miniaturize the microstrip antenna is to
modify its geometry irises [3] or a folded structure [4], [5] based on perturbation effect [6]. In this paper DGS
and slots are used to miniaturize the rectangular microstrip antenna. The present works deals with design and
analysis of a compact microstrip for wireless application. The design incorporates defective ground structure
which is on the ground plane which disturbs shielded current distribution in ground plane [7], [8]. Initially the
antenna is designed for the resonant frequency of 3.21 GHz and then using DGS and slots the resonant
frequency is brought down to 1.69 GHz. So a size reduction of 72% is achieved.

Design Principles

The geometry of the proposed antenna is shown in Figure 1. The substrate Rogers RT/duriod 5880(tm) of
dielectric constant 2.2 and dielectric loss tangent of 0.0009 has been taken in this design. The antenna has been
designed and simulated with Ansoft Designer software. The length and the width of the rectangular patch are
40 mm and 30 mm. Slots as shown in the Figure 1 are introduced. The feed point is taken at (0,-6) in absence of
DGS at the ground plane. At this time the resonant frequency of the rectangular microstrip antenna is found to
be 3.21 GHz. The DGS as shown in Figure 2 is incorporated at the ground plane. Here D1 is taken as 7 mm.
Both L1 and L2 are taken as 22 mm. The feed point is taken at (2,-2.5) in presence of DGS at the ground plane.
At this time the resonant frequency of the structure with DGS is found to be 1.69 GHz.

Figure 1: Rectangular Patch Figure 2: DGS at the ground plane (right)

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Simulation results

Return Loss and Votage Standing Wave Ratio (VSWR): Return loss of the circular patch antenna with DGS
and without DGS is given in Figure 3. It is observed that return loss at 3.21 GHz is -30 dB in absence of DGS
at the ground plane and return loss at 1.69 GHz is -15 dB in presence of DGS at the ground plane. The VSWR
with DGS is shown in Figure 4.

0  50
40
-5  30
20
-10  VSWR 10
(dB-)1  5 
0
 11 1.50
S  -20 

-25  with DGS  1.75 2.00
with out DGS  Frequency(GHz)
-30  3.5 4.0
1.0  1.5  2.0 2.5  3.0 

Frequency(GHz)

Figure 3: RL of the Circular Patch with DGS and Figure 4: VSWR with DGS
without DGS

Radiation Pattern and other other Characteristics: The microstrip patch antenna radiates normal to its patch
surface. So the elevation pattern for φ= 0 and φ= 90 degrees are important for the measurement. Figure 5 below
shows the E-plane and H plane radiation pattern at 3.21 GHz. The maximum gain is obtained at resonant
frequency for the microstrip antenna without DGS at the ground plane is 7.8 dBi for both φ= 0 and φ= 90
degrees. Figure 6 below shows the E-plane and H plane radiation pattern at 1.69 GHz. The maximum gain is
obtained at resonant frequency for the microstrip antenna with DGS at the ground plane is 1.7 dBi for both φ= 0
and φ= 90 degrees.

Phi=0 deg Phi=90 deg

Figure 5: Radiation Pattern without DGS at 3.21 GHz Figure 6. Radiation Pattern with DGS at 1.69 GHz

 2

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Conclusion
Microstrip circular patch antenna with DGS is carried out in this work. A DGS in the ground plane found to give
a size reduction of about 72% and shift the resonant frequency 3.21 GHz to 1.69 GHz with 70 MHz bandwith
and -15 dB return loss facilitating the antenna to be used for wireless application
Acknowledgement
The authors like to acknowledge Dr. B. C. Roy Engineering College, West Bengal, India for providing
necessary support during this research work.
References

[1] A. K. Skrivernilk, Zurcher O. Staub and J. R. Mosig, “PCS antenna design: The challenge of miniatururization”, IEEE Antenna
Propagation Magazine, 43 (4), pp 12-27, August 2011M. Young, The Techincal Writers Handbook. Mill Valley, CA: University
Science, 1989.

[2] T. K. Lo and Y. Hwang, Microstrip antannas of very higy permittivity for personal communications, 1997 Asia Pacific Microwave
Conference, pp. 253-256

[3] J. S. Seo and J. M. Woo, “Miniaturizaation of microstrip antenna using iris,” Electron Lett., vol.40 no. 12, pp.718-719, Jun.2004
[4] K. L. Wong, Compact and Broadband Microstrip antennas. New York: Wiley-Interscience,2002, p.5
[5] H. M. Heo and J. M. Woo, “Miniaturization of microstrip antenna using folded structure,” in Proc. Int. Symp. Antennas Propag.,

Endai, Japan, Aug. 2004, pp. 985–988.
[6] R. F. Harrington, Time-Harmonic Electromagnetic Fields. Piscataway, NJ: IEEE Press, 2001
[7] A. K. Arya, M. V. Kartikeyan, A. Patnaik, “Efficiency enhancement of microstrip patch antennas with Defected Ground Structure,”

IEEE proc. Recent Advanced in Microwave theory and applications (MICROWAVE-08), pp.729–731 November 2008.
[8] F. Y. Zulkifli, E. T. Rahardjo and D. Hartanto,” Mutual Coupling Reduction using Dumbbell Defected Ground Structure for Multiband

Microstrip antenna array” Progress In Electromagnetics Research Letters, Vol. 13, 29-40, 2010
 

 3

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Design of a CPW-Fed Slot Antenna for Wireless Application

Mrinmoy Chakraborty1 and Biswarup Rana2
Dr. B. C. Roy Engineering College, Durgapur, West Bengal, India
Email: [email protected]; [email protected]

Abstract
In this paper a coplanar waveguide (CPW) fed slot antenna is proposed. The physical dimension of the
antenna is 60 mm × 80 mm × 2 mm. It is designed at centre frequency of 8.4 GHz with bandwidth of 1
GHz. This antenna can be used for wireless applications from 7.8 GHz to 8.8 GHz frequency band.
Introduction
Microstrip antennas are used for transmitting and receiving signals. Microstrip or printed antennas are very
popular and used widely in wireless communications, mobile communications, radar applications or any other
applications because of low profile, small size, light weight, ease of fabrication and excellent compatibility with
planar integrated circuits and even non planar surfaces. There are several feeding technique like coaxial probe
fed, microstrip line fed, edge fed, inset fed, CPW [1,2] fed. The CPW is the feeding which side-plane conductor
is ground and centre strip carries the signal. The advantage of CPW fed slot antenna is its wide band
characteristics. Hence CPW fed slot antenna is most effective and promising antenna for wideband wireless
application. Wideband characteristic in CPW fed antenna can be achieved by using different slot geometries
like bow-tie slot [3], wide rectangular slot [4], circular slot [5], and hexagonal slot [6]. The wide band
characteristic in CPW fed antenna can also be achieved by using coupling techniques like inductively and
capacitively coupled slots [7], dielectric resonant coupling [8] and other techniques such as using photonic band
gap [9]. In this paper a special shaped slot as shown in Figure 1 has been taken. The simulation software used
for this analysis is HFSS.
Design Principles
The geometry of the proposed antenna is shown in Figure 1. The substrate Arlon CuClad 217™ of dielectric
constant 2.17 and dielectric loss tangent of 0.0009 has been taken in this design. The antenna has been designed
and simulated with HFSS software. The details dimension of this antenna is shown in the Table 1. The resonant
frequency of this CPW-fed antenna is 8.4 GHz with band width of 1 GHz. This antenna can be used for wireless
applications from 7.8 GHz to 8.8 GHz frequency band.

Figure 1: Rectangular Patch

 4

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Parameters Value in mm
L 60
W 80
L1 15
W1 30
L2 13
W2 4
L3 4

Table 1: Detail antenna dimensions

Simulation results

Return Loss and VSWR: Return loss of the CPW-fed slot antenna is shown in Figure 2. It is observed from the
Figure 2 that return loss is -23 dB at 8.4 GHz. It is also seen that the band width of this proposed antenna is 1
GHz covering the frequency range 7.8 GHz to 8.8 GHz. The VSWR of the proposed antenna is shown in the
Figure 3. It is found from Figure 3 that VSWR is 1.2 at 8.4 GHz

0 18
-5 16
-10 14
-15 12
-20 10
8
6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 6
Frequency(GHz) 4
S11(dB) 2
VSW R 0

6 7 8 9 10
Frequency(GHz)

Figure 2: RL of the antenna Figure 3: VSWR of the antenna

Radiation Pattern: The microstrip patch antenna radiates normal to its patch surface. So the elevation pattern
for φ= 0 and φ= 90 degrees are important for the measurement. Figure 4 below shows the E-plane and H plane
radiation pattern at 8.4 GHz.

Figure 4: Radiation Pattern at 8.4 GHz
 5

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Conclusion

A CPW- fed slot antenna is carried out in this work. The CPW- fed antenna gives the bandwidth of 1 GHz at
the resoant frequency of 8.4 GHz facilitating the antenna to be used for wireless application.

Acknowledgement

The authors like to acknowledge Dr. B.C. Roy Engineering College, West Bengal, India for providing necessary
support during this research work.

References

[1] J. Y. Chiou, J. Y. Sze and K. L. Wong, “A broad-band CPW-fed striploaded square slot antenna,” IEEE Trans. Antennas and Propag.,
Vol. 51, No.4, April 2003, pp. 791-721.

[2] H. D. Chen, “Broadband CPW-fed square slot antenna with a widened tuning stub,” IEEE Trans. Antennas and Propag., Vol. 51, No 8,
Aug 2003, pp. 1982-1986.

[3] L. Marantis and Paul Brennan, “A CPW-Fed Bow-Tie Slot Antenna With Tuning Stub,” Proc. Of Loughborough Antennas &
Propagation Conference, Mar 2008, pp 389-393. 

[4] W. Q. Chen, G. F. Ding et.al, “Design and Simulation of Broadband Unidirectional CPW-Fed Rectangular Slot Antennas”, IEEE
International Symposium on Microwave, Antenna, Propagation and EMC Technologies for Wireless Communications, 2007, pp. 632-
635. 

[5] E. A. Soliman, S. Brebels, E. Beyne, and G. A. E. Vandenbosch,“CPW-fed cusp antenna,” Microwave Opt. Technol. Lett., Vol. 2,Aug.
1999, pp.288-290. 

[6] Kraisorn Sari-kha, Vech Vivek, and Prayoot Akkaraekthalin, “A Broadband CPW-fed Equilateral Hexagonal Slot Antenna,” Proc. Of
IEEE International Conference on Computer Systems and Information Technology, Sept. 2006, pp.783-786. 

[7] L. Giauffret, J. Laheurte, and A. Papiernik, “Study of various shapes of the coupling slot in CPW-fed Microstrip antennas,” IEEE
Trans.Antennas Propag., Vol. 45, April 1997, pp.642-647. 

[8] M.S. Al Salameh, Y.M.M Antar, and G. Seguin, “Coplanar waveguidefed slot coupled rectangular dielectric resonator antenna,” IEEE
Trans.Antennas Propag., Vol.50, No.10, Oct 2002, pp.1415-1419. 

[9] L.T. Wang, X.C. Lin, and J.S. Sun, “The broadband loop slot antenna with photonic bandgap,” Proc. Int. Conference on Antennas and
Propagation, Vol.2, 2003, pp.470-472. 

 6

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Earthquake Loss Estimation using Image Processing Techniques

K. Banerjee1, M. Iqbal2 and N. Satyam3
1Dr. B. C.Roy Engineering College, Durgapur, West Bengal

2VITS Engineering College, Jabalpur, Madhya Pradesh 
3Department of Civil, Earthquake Engg Research Centre, International Institute of Information Technology,

Gachibowli, Hyderabad, Andhra-Pradesh
Email: [email protected], [email protected] 

Abstract
In this paper the objective was development of an approach for effective earthquake loss estimation using
techniques of image processing coupled with pivot table application of Microsoft Excel. The study
incorporates two pre and post-earthquake imageries (QUICKBIRD images-61cm resolution), MATLAB
for soft computing, ERDAS for unsupervised image classification and MS Excel for result calculations.
The study involves Yushu earthquake in china on 14th April 2010. The results obtained shows that such a
method can be used with further precision for loss estimations of rail and road networks. Rightly with the
use of NDVI classification for a dense vegetation cover the results can be further made accurate.
Although some amount of error which detects vegetation as buildings or vice-versa may occur, yet the
results are overtly satisfactory. In the present paper, by building extraction in dense urban areas
estimation about the losses incurred in that locality was figured out, making use of the pre and post-
earthquake imageries. Buildings are seen as independent objects in the image. Hence primarily with a
building classification (unsupervised with 30 classes) followed by shadow detection and removal the
approximate buildings devastated by the earthquake were found out. Two methods have been undertaken
one using ERDAS and the other using MATLAB although better results have been obtained with the
prior software. At the end by using the MS Excel we calculate the percentage of loss for that area.

Introduction

The 2010 Yushu earthquake struck on April 14, 2010, and registered a magnitude of 6.9 or 7.1. It originated in
Yushu, Qinghai, China, at 7:49 am local time. According to the Xinhua News Agency, 1,144 people was
confirmed dead, 417 missing, and 11,744 injured of which 1,192 were severely injured. The epicenter was
located in Rima village, Upper Laxiu Township of Yushu County, in remote and rugged terrain, near the border
of Tibet Autonomous Region. The epicenter was about 30 km from Gyêgu town, the seat of Yushu County, and
about 240 km from Qamdo. The epicenter was in a sparsely populated area on the Tibetan plateau that is
regularly hit by earthquakes. At least 11 schools were destroyed in the earthquake Over 85% of buildings in
Gyegu, mostly of wood-earth construction, were destroyed, leaving hundreds trapped and thousands homeless.
Building retrieval from high resolution images is a difficult problem especially for earthquake devastated site.
The image is a projection of the real 3D object, therefore introducing deformation and multi-faces view of the
object; shadows appear more or less; occlusions are frequent in very dense areas. [1] Quite rightly mentioned in
the paper even very small and minor errors recorded while obtaining data can drastically change the result.
Therefore approaches relying on low level features detection to the proposed earthquake loss estimation would
inevitably fail. An earthquake loss which the present paper deals on is mainly loss of the man-made
architectures (buildings) with a 3D geometry. After an earthquake, obtaining information about the damaged
area is crucial for effective emergency management and allocation of limited resources [2]. True to the cause,
such a catastrophe demands immediate attention and augmentation of medical facilities mainly to avoid the
climbing causality lists. However without properly monitoring the most damaged civilian areas aids for affected
people cannot be rendered. Also to sustain the balance back in the economic framework within the country
accurate damage estimation must be done to avail contingent facilities and insurance coverage.

This study devised an approach for building-extraction in dense urban areas with estimation about the losses
incurred in that locality. Buildings are seen as independent objects in the image. Hence primarily a building
classification (unsupervised with 30 classes) which was analyzed mainly on the strength of spectral value
analysis, followed by shadow detection was used to estimate buildings devastated by the earthquake.
Manipulating the gray scale image so obtained the color /pseudo color was assigned to each of the classes to
find out the respective feature representation. Thereby the most appropriate among them were grouped one at a
time to visualize the extant and concentration of particular features like roads, pavements or buildings. A
shadow classification was done using the darker values (tends to value 0). Building classification was done
using classes mainly 24, 29, 30 and 26 using less darker value(tends towards 255). Among the two methods
undertaken one use ERDAS and the other use MATLAB although better results have been obtained with the
prior software. At the end by using the MS Excel the percentage of loss for that area was calculated. Rightly
with the use of NDVI classification for a dense vegetation cover the results can be further made accurate.

7 
 

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Although some amount of error which detects vegetation as buildings or vice-versa may occur, yet the results
are overtly satisfactory.

Methods

The main software used for this work is ERDAS due to its capability of implementing the ISODATA algorithm
for the unsupervised clustering required to find out a specific number of classes (feature vectors). This algorithm
is based on the k-means algorithm, and employs processes of eliminating, splitting, and clustering. The
advantages of the ISODATA are its self-organizing capability, its flexibility in eliminating clusters that are too
small, its ability to divide clusters that are too dissimilar, and its ability to merge clusters that are sufficiently
similar. Some disadvantages are: 1) multiple parameters must be given by the user, although they are not known
a priori; 2) a considerable amount of experimentation may be required to get reasonable values; 3) the clusters
are ball shaped as determined by the distance function; 4) the value determined for K depends on the parameters
given by the user and is not necessarily the best value; and 5) a cluster average is often not the best prototype for
a cluster [3]. Later on an analysis of the same earthquake hit area is done using MATLAB however it can be
seen that the results obtained are much vague and inadequate compared to the ISODATA clustering done using
ERDAS. The incoming sections present a brief explanation about the various images that was simulated and the
respective processes.

Results and discussions

Building and shadow classification using ArcGIS and ERDAS: The images used are of Yushu earthquake in
china on 14th April 2010 (pre-earthquake imagery- 6 November 2004 Figure 1, post-earthquake imagery -15th
April 2010 Figure 2, QUICK BIRD satellite0.61m resolution).

Figure 1: Pre earthquake image Figure 2: Post earthquake image

Unsupervised classification of Figure 1 and Figure 2 with 30 classes (more number of classes can be taken if the
software permits) was done using ERDAS. Raster property conversion toolbox was used to convert it into image
files with .img extension. These .img files were added to Arc Scene (ArcGIS 9.2). Therefore manipulating the
gray scale image so obtained respective color /pseudo color was assigned to each class. After respective feature
representation the most appropriate among them were grouped suitably e.g.: darker features assigned value 0.
Building classification was done using classes 24, 29, 30 and 26 (tends towards 255) (Fig3 Fig4). Finally the
building classified image and shadow classified image (Figure 5, Figure 6) was subtracted like (B-S) simply to
obtain the required results of just the building detection (Figure 7).The white patches in the classified image can
be accounted with certain features which is new to the post-earthquake image and no such features existed on
the pre-earthquake image. Since the sole aim was to detect the change the feature gets marked in white.

Figure 3: pre building Figure 4: post: Figure 5: pre shadow Figure 6: post shadow Figure 7: changes

classification building classification classification classification detected

Thus superimposing the two unsupervised classified image gave the changes in the infrastructure of the land

(new buildings constructed or old ones devastated by the earthquake).

Building and shadow classification using Matlab: the imread command was used to initially load the image
in Matlab.

Preprocessing: preprocessing for the reference background image is done to subtract any new image from it so
that we can compute the changes that have taken place in between the two images. Here pre earthquake image

8 
 

MDCCT 2012
6 – 7 February, 2012; Burdwan University

acts as the reference while the post-earthquake image acts as the current image. im2double (I); im2double (I1);
where I=pre-earthquake (Figure 1), I1=post-earthquake (Figure 2) was used to convert the image to double for
accurate calculation. Now the range of values are [0..1] the image was thus normalized in R and G in RGB
image like misc = (X(:,:,1)+X(:,:,2)+X(:,:,3)); where X defines respective image arrays in the R, G, and B
bands. After aggregating R, G and B, if the value of misc amounts to zero should be replaced with 0.001 to
avoid mathematical errors of infinite value like to overcome “divide by zero error” ( misc (misc==0) = 0.0001).
Normalizing R factor is mandatory for keeping mathematical and statistical calculations simple. This was done
using: X(:,:,1) = X(:,:,1)./misc. similarly G band image array was normalized too. To get the difference of two
images (post & pre images) imabsdiff (X, X1); (Figure 8) was helpful where X1 is the double type conversion of
the Post-Earthquake image. It subtracts each element in array X1 from the corresponding element in array X and
returns the absolute difference in the corresponding element of the output array A providing an approximate
judgment about any important change that has occurred over the terrain of interest. Threshold of 0.31 is applied
to the array A to get the shadow free image. The logic was taken as: (A<0.31) = 0, (A>=0.31) = 1. Now an
image which is shadow free was obtained (Figure 9). Now we get an image Res which is shadow free, but we
need to reconstruct it. So Morphological operation Dilation is applied on A. Reconstruction was done using
imdilate (A, ones (9)) (Figure 10). We need exact boundary of the objects and their shadows. To do so we
applied background subtraction on gray scale image. rgb2gray is used to convert RGB image to gray scale and
imfill with ‘holes’ is used to fill out the holes in a bounded area of an image (Figure 11).

Reconstruction and Object Classification: To get the reconstructed boundary of the object without shadow
region, point wise multiplication of objects and its shadow is multiplied with the dilated object region (Figure
12). We see that the result gradually degrades therefore dilation operation to sort out the damaged areas is error
prone thus we proceed on further calculation on the ERDAS results.

Figure 8 Figure 9 Figure 10 Figure 11 Figure 12

Normalized Differential Vegetation Index (NDVI): Here Landsat image of area under study maybe used
based on vegetation analysis requirements. Primarily Landsat consists of 7 bands of which for the NDVI
classification the value of R and NIR bands that is bands 4, 3 and 2. Import CIR Bands from a BIL Image File,
Enhance the CIR Composite with a decorrelation Stretch, Construct an NIR-Red Spectral Scatter Plot, Compute
Vegetation Index via MATLAB Array Arithmetic, Locate Vegetation- Threshold the NDVI Image, Link
Spectral and Spatial Content.

Use of Microsoft Excel on ERDAS classified images to obtain an estimated area devastated with respect to
pre and post earthquake images

The pivot table options were used to classify the row labels as buildings and shadows from the ERDAS
classified pre-earthquake image. The attributes to which are Sum of histogram values and sum of resolution. In
the second post-earthquake image the row labels like shadows, buildings and blank (which takes in all other
attributes apart from these two e.g.: roads, pavements.) This model was developed in IIIT Hyderabad, India as a
summer project workout. All the results along with the Matlab and ERDAS compilations were shown pertaining
to each step. Calculation of the net result was done with the pivot chart implementation to make a pre
earthquake and post-earthquake analysis as shown in the images below and find out the area under supervision.
Net area under supervision was the total histogram value*resolution=744038*3721cm^2=27.68 hectors. This is
mention ably the pre-earthquake value under analysis taking into account the building and shadows. Where
3721= 612 cm sq. as per the resolution of the QUICKBIRD imagery taken for analysis. Whereas the post value
under analysis taking into account the building and shadow attributes amounts to: (92560+84021) * 3721cm^2.
The estimation was done as (pre value-post value)/pre value *100 = [3721(744038-(92560+84021))]/
(3721*744038) =.800*100=80%.

9 
 

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Pre earthquake analysis Post-earthquake analysis

Conclusion

Loss estimation using Image processing techniques is considered a quite economical process which continues to
yield satisfactory results. The objective of this study is to investigate the use of Image processing along with MS
Excel to make an effort of even better economy and results without having to use other sophisticated software.
The aim was to develop a model that would enable designers to design practical values of the system
components with even better results owing to finer analysis. In order to corroborate the practical aspects of the
paper the results obtained were shown from both the Matlab and ERDAS. Also the complexities associated with
the analysis were minimized with careful use of mathematical tools and certain approximations were done to get
the result in whole number. The working was noted after the simulation experiments was compared to the real
time data collected and it yielded quite a good result. The close proximity between these two data serve as
Irrefutable proof that an implementation of the above mentioned model in the practical domain is very much
feasible.

Future Scope

As will be seen in this study image processing forms a very integral part of the approach that had been designed.
Determination of the pre earthquake buildings shadows and there acceptability in the real time domain can serve
as an important premise for future research work. The paper described here focuses its attention towards the
modeling of a single and simple method that permits only one way communication for pre and post-earthquake
loss estimates using satellite imagery. A much more challenging and interesting approach can be to design a
similar model for a two way communication network where a NDVI classification can be involved for which
Landsat imageries are required. An NDVI classification to obtain the vegetation (values as low as .3-.9) strength
in the area yields better perception to minimize errors can be used. Finally some errors are encountered within
the classification like some parts of open land and roads getting selected along with the buildings but it can be
avoided by increasing the no of classes taken at the time of classification. Further use of better algorithms can be
efficient in paying attention to details like road or rail network destructions. Thus instead of just estimation of
the losses a probable estimation even before earthquake takes place can be demonstrated and effective measures
taken on its behalf.

References

[1] Liu Wei, Veronique Prinet, “Building Detection from High-resolution Satellite Image Using Probability Model”, International
geosciences and remote sensing and symposium IGARSS, July 2005 (Seoul), Korea

[2] E. Sumer, M. Turker, “Building Damage Detection from Post-Earthquake Aerial Imagery using Building Grey-Value and Gradient
Orientation Analyses”, 0-7803-8977-8/05/$Z0.00 02005 IEEE.

[3] Carl G. Looney “Chapter 5. Fuzzy Clustering and Merging”.
[4] M. J. Sabin, “Convergence and Consistency of Fuzzy c-means /ISODATA Algorithms”, IEEE Trans. Pattern Anal. Machine Intell.,

September 1987.
[5] Kannika Komwong, Ramphing Simking, Panu Nuangjumnong and Sanwit Iabchoon. “First Phase of Suitable Model for 2d Building

Extraction in Thailand: Bangpli District, Samut Prakan Province”, PN-244
[6] J. M. Bardsley, Marylesa Wilde, Chris Gotschalk, M. S. Lorang. MATLAB Software for Supervised Classification in Remote Sensing

and Image Processing. Journal of Statistical Software, Volume VV, Issue II.

1 0
 

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Automated Detection of Blood Vessels from
Normal Retinal Image

Tapas Chakrabarti1 and Siladitya Sen2
Department of ECE, Heritage Institute of Technology, Kolkata 700 017

Email: [email protected]; [email protected]

Abstract
The information of human retinal blood vessels can be used in grading severity of diseases or as part of
the process for automated diagnosis of diseases with ocular manifestation. The blood vessels of retina can
have measurable abnormalities in diameter, color and tortuosity, e.g. Hypertension may result in focal
constriction of retinal arteries, neovascularization due to diabetics, dilated tortuous veins due to central
retinal vein occlusion, etc. The human eye or the retina of eye has a unique characteristic that, in this part
of the blood vessels can be observed directly. The eye specialist generally dilate the pupil, making wider,
to pass more light and using ophthalmoscope the retinal image been observed for diagnosis. The retinal
image resolution is high as it consists of a complex vascular network; hence manually it is a tiresome job.
The retinal image generally used to capture with the fundus camera directly or by fluorescein
angiography. The extraction of blood vessels from retinal image is a challenging task in medical analysis
and diagnosis. Moreover the fluorescein angiographic retinal image capturing is a highly super special
technique with an operative task, specialist doctor is required for this type of job and risk and side affect
are also there. In this paper we tried to develop a method of automated detection of blood vessels, of
human retina, using existing image processing tools in DSP, on directly captured fundus image. Especially
in rural area where the specialized doctors are not available, the normal fundus image can be captured by
a ordinary trained technician and after developing the image, it can be transmitted to a specialized doctor
using existing communication system. In this system the rural patient can take specialized treatment from
their rural area.

Introduction

The extraction of blood vessels from retinal images is an important and critical task in medical science. In this
paper we studied the retinal images using template matching method. In this template matching method we
applied Prewitt Operator, Sobel Operator, and Canny Edge detector for edge detection. We applied the Kirsch
matched template to extract the blood vessels.

Introduction to Digital Image Processing

An image may be defined as a two dimensional function, f(x,y), where x and y are spatial coordinates. The
function f(x,y) is the accumulation of energy from the object’s radiant energy distribution. When x,y and the
intensity values of f are all finite, discrete quantities, the image is called digital image. The digital image
processing involves the steps of preprocessing & edge detection, segmentation, normalization, feature
extraction.

Preprocessing: The preprocessing is used to eliminate the spurious noises which may insert during digitization.
Original images are first smoothed by means of filters, to reduce the effect of these spurious noises. In this work
we did all the works in Mat lab platform. We first convert both the retinal image i.e. the normal digital fundus
image and the image captured after using fluorescein dye in the blood of the patient, in to grayscale image, i.e.
we convert it in to two dimensional images.

Edge Detection: Edge defines the boundaries between regions in an image, which helps with segmentation and
object recognition [1, 2]. They can show where shadows fall in an image or any other distinct change in the
intensity of an image. Edge detection is a fundamental of low level image processing and good edges are
necessary for higher level processing. The edge detection is the boundary between two regions, with relatively
distinct gray level properties. In a continuous image, a sharp intensity transition between neighboring pixels is
considered as edge [3].

The techniques used for extracting information from an image are known as image analysis techniques. As an
image is composed of edges and shades of gray, corresponds to high frequency and low frequency. Separation
(Filtering) of high frequency information means Edge Detection. In Matlab there are two types of edge detection
operators as (i) Gradient Operators [2] and compass operators [4]. These operators are called Masks.

  11

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Grayscale image with Grayscale image without
Fluorescin Angiogram Angiogram

Methodology

In this work we (i) Collect the different retinal images with fluorescein dye through angiogram and without dye
i.e. normal digital (color) image. (ii) We applied different gradient operator as Prewitt Operator, Sobel Operator,
Canny Operator, in a same image, for edge detection. We observed that the result with Canny Operator is better
than other.

(iii). as the canny operator gives the better result we applied it on the grayscale image of retina with angiogram.
(iv). Then we changes the threshold value of canny operator applied in the normal digital image and choose the
value which brings the image in to nearer to the image with fluorescein dye.

  12

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Extraction of Blood Vessels using Matched Filter

Matched filter is one of the template matching algorithms that are used in the detection of the blood vessels in
retinal images and other application as well. It is based on the spatial properties of the object to be recognized.
The idea of the matched filter is introduced by taking a number of samples for a cross section of retinal blood
vessels, the gray level profile of these samples is then approximated by Gaussian shaped curve. Matched filter
designing is based on a number of properties for blood vessels.
i) Vessels can be approximated as anti parallel segments.
ii) Vessels have lower reflectance than other retinal surfaces, so they appear darker relative to the background.
iii) Vessels size may decrease when moving away from the optic disk, the width of a retina vessel may lie

within the range of 2-10 pixels.
iv) The intensity profile varies by a small amount from vessel to vessel.
v) The intensity profile has a Gaussian shape.

Kirsch template matching

A new method is there for color images. In this algorithm another edge detector called Kirsch Operator is used
for detection of blood vessels in colored digital fundus retinal images. The Kirsch Templates are shown in Table
1 [5]:

h(1) = , h(2) = , h(3) = , h(4) =

h(5)= , h(6)= , h(7)= , h(8)=

Table 1: -Kirsch Template

The Kirsch Operator is one of first derivative operator used for edge detection. For detecting edges the operator
is rotated in all eight directions like East (h1), West (h2), Northeast (h3), Southwest (h4), North (h5), South
(h6), Northwest (h7), Southwest (h8).

Algorithm for Retinal Blood Vessels Detection Using Matched Filter

Input: A Retinal Fundus image (Digital) without angiograph.
Output: Blood vessel extracted image.

Step-1: First collected retinal fundus images are preprocessed i.e. brightness and contrast is enhanced.
Step-2: It is converted into gray monochrome image.
Step-3: The image is filtered through Adaptive median filter for-

i) To remove impulse noises
ii) Smoothing of other noise
iii) Reduce distortion, like excessive thinning or thickening of object boundaries.
Step-4: To detect vessels on all possible orientations, the kernel must be rotated to all possible vessel
orientations and the maximum response from the filter bank is registered. Matched filter is rotated by
an amount of 150 which results a filter bank of 12 kernels. Matched kernels are convoluted to get
fundus image and at each pixel only the maximum of their responses is retained.
Step-5: The convoluted output is threshold using entropic thresholding to get the retinal vasculature
structure.
Step-6: The image is filtered by adaptive median filter to remove isolated pixel to some extent.
Step-7: Finally the image is filtered by a Length filter to remove remaining isolated pixels and to get
complete retinal vasculature structure.

  13

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Extracted Retinal Blood Vessels: The digital fundus image-1, has been developed using different image
processing tools, increasing the signal information, the blood vessels are extracted. In this process we use the
Canny Operator, on converted gray scale image. We observed that the edge of blood vessels has been detected.
Now we use the Kirsch Matched Template based filter, on that image, and we found the blood vessels of that
image. To get the blood vessels more proper noise free, we applied Gaussian Filter and Median filter. The
extracted blood vessels are become more prominent and noise free.

Conclusion
From the above experimental works, we may conclude that the retinal fundus normal digital images can be
developed with the different techniques, e.g. contrast of image can be enhanced and through median filter,
Gaussian filter and the blood vessels can be extracted by matched filter, in a developed form. The improved
image bears more signals then the normal digital image, and the improved retinal images, can be used to detect
the diseases, without doing angiography.
References

[1] S. Chaudhury, S. Chatterjee, N. Katz, M. Nelson, and M. Goldbaum, “Detection of blood vessels in retinal images using two
dimensional matched filters”, IEEE Trans. Medical imaging, vol.8 No.3, September 1989.

[2] Ehsan Nadernejad, “Edge Detection Techniques: Evaluations and comparisons”, Applied Mathematical Sciences, Vol. 2, 2008, no. 31,
1507 – 1520.

[3] John Canny, “A Computational Approach to Edge Detection”, IEEE Transactions on pattern analysis and machine intelligence, Vol.
PAMI-8, No.6, November 1986.

[4] Madhuri A. Joshi, “Digital Image Processing An Algorithmic Approach”.
[5] http:\\ www.kirsch-operator-wikipedia, the free encyclopedia.

  14

MDCCT 2012
6 – 7 February, 2012; Burdwan University

A Slotted Microstrip Patch Antenna for WLAN Applications

Biswarup Rana1, Chandan Kumar Ghosh2, Subhajit Sinha3, Sabyasachi Guha4 and Susanta Kumar Parui5
1,2,3 Dr. B.C.Roy Engineering College, Durgapur, West Bengal, India

4Institute of Science & Technology, Chandrakona Town, Paschim Medinipur, West Bengal, India

5Bengal Engineering and Science University, Shibpur, West Bengal, India
Email: [email protected]

Abstract
In this paper, a microstrip patch antenna has been described that operates from 4.86 GHz to 6.91 GHz
frequency band covering WLAN band with bandwidth 2GHz. Two slots are incorporate which is the
responsible for the generation of second resonant frequency. This second resonant frequency gives the
wideband characteristic.

Introduction

Wireless local area networks (WLAN) are being used extensively. The IEEE 802.11b and 802.11g standard use
2.4-GHz ISM band which is license-free. Bluetooth, microwave, cordless phone and other devices also use the
same frequency band. Hence, there will be interference with ISM band. The 802.11a standard uses the 5-GHz
band which is cleaner to support high-speed WLAN. However, the segment of frequency band used varies from
one region of the world to another. In the U.S., the 802.11a system may use the 5.15–5.35 GHz band and 5.725–
5.825 GHz band. Some countries allow the operation in the 5.47–5.825 GHz band. A traveler with 802.11a
transceiver that can cover the frequency range from 5.15 GHz to 5.825 GHz will be able to gain access to a local
WLAN network in different parts of the world [1]. Microstrip patch antenna is the most promising and effective
choice for WLAN because of its low-cost, very lightweight, easy to fabrication, ease of integration with
microwave circuits. But it suffers from narrow bandwidth. The bandwidth of the microstrip antenna can be
increased by several ways. Using the air substrate the band width can be increased. [2]. Bandwidth of microstrip
antenna can also be increased by incorporating the slot in the patch, slot in the ground plate, increasing the
substrate thickness, etc. This paper describes the design, analysis and simulation of patch antenna for recent
wireless communication system. The antenna is designed, optimized and analyzed with IE3D electromagnetic
simulator. Two slots are incorporated as shown in the Figureure1 that provides the wideband characteristics of
the proposed antenna. By the adjustment of the length of slots (Figure 1) second resonant frequency is achieved.
This second resonant frequency in conjugation with first one provides the necessary wideband characteristics.
The schematic diagram of the proposed wide band antenna is illustrated in the Figure 1 and the dimension of the
patch and slots are shown in the Table 1.

Design Principles

The geometry of the proposed antenna is described in the Figure 1. A FR4 substrate has been used with ground
20x34 sq mm. FR4 substrate has the thickness of 1.6mm, relative permittivity 4.5 mm and loss tangent 0.02.
The patch has the length 10 mm and width 15.8 mm. A 50 ohm feed with width 2.8 mm was used to excite the
microstrip antenna. Two slots of each length 0.5 mm and width 5.5 mm are incorporated in the patch as shown
in Figure1. The patch has one resonant frequency. Two slots are incorporated which disturb the surface current.
As a result of it a local inductive effect is introduced that is responsible for the excitation of a second resonant
mode.

Parameters Dimensions

L 10.0 mm

W 15.8 mm

L1 12.0 mm

W1 5.5 mm

W2 6.0 mm

S1 0.5 mm

S2 0.5 mm

S3 2.8 mm

Table 1: Antenna dimensions

  15

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Figure 1: Geometry of the proposed antenna

Simulation results

The microstrip patch antenna has been designed with Zeland make IE3D simulator which works on principle of
FIT. The return loss of the proposed antenna is shown in the Figure 2. It is seen from the Figure that antenna can
operate from 4.86 GHz to 6.91 GHz with -10 dB characteristic bandwidth of 2 GHz. The VSWR is shown in the
Figure 3. The radiation patter at 6.2 GHz is shown in the Figure 4.

0 8
7
-5 6 W1=5.5 mm
5 W1=5.0 mm
-10 4 W1=4.5 mm
3 W1=4.0 mm
-15 2
1
-20S 11(d B ) 0
VSW R
-25 5
W1=5.5 mm 67 8
Frequency(GHz)
-30 W1=5.0 mm 7 8

-35 W1=4.5 mm
W1=4.0 mm

-40
456

Frequency(GHz)

Figure 2: RL of the antenna Figure 3: VSWR of the antenna

  16

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Figure 4: Radiation Pattern at 6.2 GHz

 

Conclusion

The bandwidth can be tuned by varying the length (W1) or width (S1, S2) or both of the slots on proposed

antenna. In this investigation, we have achieved a B. W. of 2.0 GHz at center frequency of 5.25 GHz. This
antenna may be found suitable for WLAN application. 
 

Acknowledgement

The authors like to acknowledge Bengal Engineering and Science University, Shibpur, India for providing
necessary support during this research work.

References

[1] B. K. Ang and B. K. Chung, “A wideband e-shaped microstrip patch antenna for 5–6GHz wireless communications” Progress In
Electromagnetic Research, PIER 75, 397–407, 2007.

[2] M. Abbaspour and H. R. Hassani, “Wideband star-shaped microstrip patch antenna”, Progress In Electromagnetic Research Letters,
Vol. 1, 61-68, 2008.

[3] C. Ying, G. Li, and Y. Zhang, “An LTCC planar ultra-wideband antenna," Microwave and Optical Technology Letters, Vol. 42, 220-
222, 2004.

  17

MDCCT 2012
6 – 7 February, 2012; Burdwan University

A CPW-Fed Microstrip Antenna with Four Slots
for Wireless Application

Biswarup Rana1, Subhajit Sinha2, Chandan Kumar Ghosh3 and Susanta Kumar Parui4
1,2,3Dr. B.C.Roy Engineering College, Durgapur, West Bengal, India

4Bengal Engineering and Science University, Shibpur, West Bengal, India
Email: [email protected]

Abstract
In this paper a CPW-Fed slot microstrip antenna is proposed as shown in the Fig. 1. This antenna is
suitable for wireless applications covering frequency 5.05 GHz to 6.7 GHz. It has bandwidth of 1.65 GHz
with centre frequency of 5.8 GHz. Four slots are introduced on the patch which gives the wide band
characteristics.

Introduction

As wireless communication applications require more and more bandwidth, the research in area of design of
compact broadband antenna for WLAN applications is increasing exponentially. For these modern
communication systems, size reduction of antenna with wide bandwidth and multiband operation are becoming
important design considerations for realistic applications. Many authors have presented antenna designs suitable
for WLAN operates in 2.4 GHz band and also in (5–6 GHz) [1-5]. Because of the flourishing demand in
wireless communication system applications, microstrip patch antennas have attracted much interest due to their
low profile, light weight, ease of fabrication and compatibility with printed circuits. Therefore CPW fed planar
slot antennas are most promising design for wideband wireless applications. However, they also have some
drawbacks, ranging from narrow bandwidth to low gain. In this paper, a narrative design of a broadband planar
antenna with CPW-feed technology, consisting of a rectangular shape patch embedded with four plus sign slots,
is offered for WLAN applications. By adjusting the size of the spacing gap S and the ground plan with feed line,
the overall performance of the proposed antenna can be improved. The simulated results show that the proposed
antenna has good impedance bandwidth and radiation characteristics in the operating bands which cover the
required band width of the WLAN bands.

Design Principles

Figure 1 shows the geometry of the proposed CPW-Fed antenna. A FR4 substrate of thickness 1.588 mm and
relative permittivity 2.2 dielectric loss tangent of 0.001 has been taken. A 50-ohm feed-line of width Wf = 4mm
was used to excite the antenna. The FR4 is a fire electrical grade dielectric made with epoxy material reinforced
with a woven fibreglass material. FR4 means “flame retardant” and type 4 indicates woven glass reinforced
epoxy resin. The proposed antenna has a single layer metallic structure on one side of FR4 substrate layer
whereas the other side is without any metallization. On each side of the CPW feed-line two equal finite ground
planes are placed symmetrically. The basis of the proposed antenna structure is a rectangular patch, which has
dimensions of length L and width W, and connected at the end of the CPW feed-line. . The four slots are
symmetrical of their length. The optimized geometric parameters of the proposed antenna are: length of
rectangular patch L= 20 mm, width of rectangular patch W= 20mm, ground plane length Lg= 10 mm, ground
plane width Wg = 10 mm, feed-line width Wf = 4 mm, slot length L1= 6 mm, slot length L2 = 6 mm, slot length
L3= 6 mm, slot length L4 = 6 mm, slot width Ws , spacing between ground plane and feed length d = 1 mm and
spacing between rectangular patch and ground plane S = 3.5 mm. The antenna performance is analyzed using
method of moment based Zeland IE3D simulator.

  18

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Figure 1: Geometry of the proposed antenna

Simulation results

The microstrip patch antenna has been designed with Zeland make IE3D simulator which works on principle of
FIT. The return loss of the proposed antenna is shown in the Figure 2. It is seen from the Figure that antenna can
operate from 5.05 GHz to 6.7 GHz with bandwidth of 1.65 GHz. The return loss at 5.8 GHz is -23 dB. The
VSWR is shown in the Figure 3. The VSWR at 5.8 GHz is 1.12. The radiation pattern is shown in the Figure 4.

0 25
-5
-10 20
-15
-20 15
-25
10
3456789
Frequency(GHz) 5

Figure 2: RL of the antenna 0
S (dB) 456789
11 Frequency(GHz)

VSW R Figure 3: VSWR of the antenna

  19

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Figure 4: Radiation Pattern at 5.8 GHz
 

Conclusion

A CPW- fed slot antenna is carried out in this work. The bandwidth can be tuned by varying the length or width

or both of them of the slots on proposed antenna. In this investigation, we have achieved a band width of 1.65
GHz at resonant frequency of 5.8 GHz. This antenna may be found suitable for WLAN application.  

Acknowledgement

The authors like to acknowledge Bengal Engineering and Science University, Shibpur, India for providing
necessary support during this research work.

References

[1] T. H. Kim and D. C. Park, “CPW-fed compact monopole antenna for dual-band WLAN applications,” Electronics Letters, Vol. 41,
No. 6, March 2005.

[2] Liu Wen-Chung, “Broadband Dual-Frequency CPW-Fed Antenna with a Cross-Shaped Feeding Line for WLAN Application,”
Microwave and Optical Technology Letters, vol. 49, pp. 1739-1744, 2007.

[3] Y. F. Ruan, Y. X. Guo, K. W. Khoo and X. Q. Shi, “Compact wideband antenna for wireless communications,” IET Microwave
Antennas Propag., Vol. 1, (3), pp. 556-560, 2007.

[4] Nan Chang and Jing-Hae Jiang, “Meandered T-Shaped Monopole Antenna,” IEEE Transactions on Antennas and Propagation, Vol.
57, No. 12, December, 2009.

[5] M.T. Islam, “Broadband E-H Shaped Microstrip Patch Antenna for Wireless Systems,” Progress in Electromagnetics Research, PIER
98, 163-173, 2009.

  20

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Design and Analysis of a Compact Circular Microstrip Antenna
Using Defective Ground Structure

Mrinmoy Chakraborty1 and Biswarup Rana2
Dr. B. C. Roy Engineering College, Durgapur, West Bengal, India
Email: [email protected]; [email protected]

Abstract
In this paper a circular microstrip antenna with defective ground structure (DGS) is proposed. In the
absence of DGS the structure found to resonant at 8.74 GHz. When the DGS is introduced a frequency
shift of 8.74 GHz to 2.86 GHz is observed. The main contribution of this paper is the miniaturization of
87% which is very much encouraging.

Introduction

Microstrip antennas have been used widely in wireless communication due to their light weight, low profile, low
cost and ease of fabrication and excellent compatibility with planar integrated circuits and even non planar
surfaces. In recent years, as the demand of small systems have increased, small size antennas at low frequency
have drawn much interest from researchers [1]. Many kind of miniaturization techniques, such as using of
dielectric substrate of high permittivity [2], slot on the patch, DGS at the ground plane or a combination of them
proposed and applied to microstrip antennas. The other method to miniaturize the microstrip antenna is to
modify its geometry irises [3] or a folded structure [4], [5] based on perturbation effect [6]. In this paper DGS is
used to miniaturize the circular microstrip antenna. The present works deals with design and analysis of a
compact microstrip for wireless application. The design incorporate dumble shaped defective ground structure
which is on the ground plane which disturbs shielded current distribution in ground plane [7], [8]. Initially the
antenna is designed for the resonant frequency of 8.74 GHz and then using DGS the resonant frequency is
brought down to 2.86 GHz. So a size reduction of 87% is achieved.

Design Principles

The geometry of the proposed antenna is shown in Figure 1. The substrate Rogers RT/duriod 5880(tm) of
dielectric constant 2.2 and dielectric loss tangent of 0.0009 has been taken in this design. The antenna has been
designed and simulated with Ansoft Designer software. The diameter of the circular patch is 13 mm. The feed
point is taken at (-2,-0.8) in absence of DGS at the ground plane. At this time the resonant frequency of the
circular microstrip antenna is found to be 8.74 GHz. The dumble shaped DGS as shown in Figure 2 is
incorporated at the ground plane. Here D1 is taken as 10 mm. Both L1 and L2 are taken as 30 mm. S is taken as
1.5 mm. The feed point is taken at (-1.9,-1.9) in presence of DGS at the ground plane. At this time the resonant
frequency of the structure with DGS is found to be 2.86 GHz.

Figure 1: Circular Patch Figure 2: DGS at the ground plane
  21

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Simulation results

Return Loss and VSWR: Return loss of the circular patch antenna with DGS and without DGS is given in
Figure 3. It is observed that return loss at 8.74 GHz is -43 dB in absence of DGS at the ground plane and return
loss at 2.86 GHz is -20 dB in presence of DGS at the ground plane. The VSWR with DGS is shown in Figure 6
and Figure 7 respectively.

0S11(dB) 20
-5 VSWR 18
-10 16
-15 With DGS 9 10 14 2.75 3.00
-20 With out DGS 12 Frequency(GHz)
-25 10
-30 345678 8
-35 Frequency(GHz) 6
-40 4
2
2 0

2.50

Figure 3: RL of the Circular Patch with DGS and Figure 4: VSWR. With DGS
without DGS

Radiation Pattern and other other Characteristics: The microstrip patch antenna radiates normal to its patch
surface. So the elevation pattern for φ= 0 and φ= 90 degrees are important for the measurement. Figure 5 below
shows the E-plane and H plane radiation pattern at 8.74 GHz. The maximum gain is obtained at resonant
frequency for the microstrip antenna without DGS at the ground plane is 6.9 dBi for both φ= 0 and φ= 90
degrees. Figure 6 below shows the E-plane and H plane radiation pattern at 2.86 GHz. The maximum gain is
obtained at resonant frequency for the microstrip antenna with DGS at the ground plane is 1.8 dBi for both φ= 0
and φ= 90 degrees.

Phi=0 deg Phi=90 deg

Figure 5: Radiation Pattern without DGS at 8.74 GHz Figure 6: Radiation Pattern with DGS at 2.86 GHz

  22

MDCCT 2012
6 – 7 February, 2012; Burdwan University

Conclusion

Microstrip circular patch antenna with DGS is carried out in this work. A dumble shaped DGS in the ground
plane found to give a size reduction of about 87% and shift the resonant frequency 8.74 GHz to 2.86 GHz with
50 MHz bandwith and -20 dB return loss facilitating the antenna to be used for wireless application.

Acknowledgement

The authors like to acknowledge Dr. B.C. Roy Engineering College, West Bengal, India for providing necessary
support during this research work.

References

[1] A. K. Skrivernilk, Zurcher O. Staub and J. R. Mosig, “PCS antenna design: The challenge of miniatururization”, IEEE Antenna
Propagation Magazine, 43 (4), pp 12-27, August 2011M. Young, The Techincal Writers Handbook, Mill Valley, CA: University
Science, 1989.

[2] T. K. Lo and Y. Hwang, “Microstrip antannas of very higy permittivity for personal communications”, Asia Pacific Microwave
Conference, pp. 253-256, 1997

[3] J. S. Seo and J. M. Woo, “Miniaturizaation of microstrip antenna using iris,” Electron Lett., vol.40 no. 12, pp.718-719, Jun.2004
[4] K. L. Wong, Compact and Broadband Microstrip antennas. New York: Wiley-Interscience,2002, p.5
[5] H. M. Heo and J. M.Woo, “Miniaturization of microstrip antenna using folded structure,” in Proc. Int. Symp. Antennas Propag., Endai,

Japan, Aug. 2004, pp. 985–988.
[6] R. F. Harrington, Time-Harmonic Electromagnetic Fields, Piscataway, NJ: IEEE Press, 2001
[7] A. K. Arya, M. V. Kartikeyan, A. Patnaik, “Efficiency enhancement of microstrip patch antennas with Defected Ground Structure”,

IEEE proc. Recent Advanced in Microwave theory and applications (MICROWAVE-08), pp.729–731 November 2008.
[8] F. Y. Zulkifli, E. T. Rahardjo, and D. Hartanto,” Mutual Coupling Reduction using Dumbbell Defected Ground Structure for

Multiband Microstrip antenna array”, Progress In Electromagnetics Research Letters, Vol. 13, 29-40, 2010

 

  23

MDCCT 2012
6 - 7 February, 2012; Burdwan University

Characteristics Studies of a
Metal-Semiconductor-Metal (MSM) Photodiode

D. Chowdhury1, S. Bhunia1 and T. K. Barik2
1Department of Electronics and Communication Engineering, 2School of Applied Sciences and Humanities

Haldia Institute of Technology, West Bengal, India
E-mail: [email protected], [email protected], [email protected]

Abstract
Metal-Semiconductor-Metal (MSM) structure based on silicon (Si) and aluminum (Al) as metal has been
fabricated. The voltage-current (V-I) characteristic of the MSM structure has been studied theoretically
and experimentally under different strength of illumination. The variation of knee voltages of the
structure with illumination is also been studied. The experimental results excellently matched with the
theoretical variation.

Introduction

Metal-Semiconductor-Metal (MSM) devices have emerged as important devices in recent years in view of their
major applications as photo detectors and also as Barrier Injection Transmit Time (BARITT) diodes. In
optoelectronic integrated circuit (OEIC), MSM devices are used as detectors of the photo-receiver system which
possess advantages over photo-detectors (like PIN, APD etc.) for easier planar integration. MSM photodiodes
can be used as high speed detectors for photonic application. Because of their simple structure few processing
steps are required and those can be easily integrated with other electronic devices. The integrated schottky finger
spacing is in the sub micrometer range. Therefore it can be predicted theoretically that MSM devices will be
capable of detecting signals of frequency as high as 100 GHz. Communication technologies has undergone rapid
development due to the invention of the optical fibers, modulators, modulator-controller coplanar strip line
transmission, tapered transmission etc. It is well known that the optical signal has very high frequency and very
short wavelength. The favorable characteristics of a photodiode are low capacitance (high bandwidth), short
transit time (high bandwidth), high responsivity (high sensitivity) and low dark current (low short noise). A lot
of research work has been carried out in above mentioned areas and experimental reports on the fabrication of
the photo receiver using MSM devices and HBT or HEMT as a pre-amplifier on the same chip are available in
the published literature [1-5].

A metal-semiconductor-metal (MSM) In0.53Ga0.47As photodiode using a transparent cadmium tin oxide (CTO)
layer for the interdigitated electrodes was investigated. CTO has greater transmission properties at the

wavelengths appropriate for In0.53Ga0.47As photodiodes over conventional transparent conductors such as indium
tin oxide (ITO) [6]. The thick-film sensors, prototypes and the array’s components, were fabricated on the basis

of commercial sensor platforms

[7]. A two-dimensional self-

consistent time-dependent

simulation technique has been

used to investigate electron–hole

transport processes in the active

region of metal–semiconductor–

metal photodiode structures

(MSM -PD) and to analyze their

high-speed response at different

energy levels of the optical

illumination. Charge Figure1: MSM structures

accumulation and screening of

the dark electric field at high

optical excitation levels greatly modify the drift conditions of the photogenerated electrons and holes in the

active region of the MSM –PD [8]. The use of high-temperature (HT)-AlN interlayer in Al0.3Ga0.7N/GaN epi
layer and its effects on the grown structures were explored. The Al0.3Ga0.7N-based inter digitated MSM photo
detector has been successfully fabricated and characterized [9]. Transparent conducting oxides thin layers, due

to their optical and electrical properties, can be used as transparent electrodes in various optoelectronic devices

[10]. Mg0.20Zn0.80O films have been prepared on quartz substrates by using radio frequency (RF) magnetron
sputtering, and metal–semiconductor–metal (MSM) structured photo detectors have been fabricated on the

Mg0.20Zn0.80O films employing inter digital Au as metal contacts [11]. In this article the basic characteristics of
an MSM photodiode has been studied under dark and illuminated conditions.

24

MDCCT 2012
6 - 7 February, 2012; Burdwan University

Theoretical Analysis

For a simple illuminated PN junction diode, the incident light generates electron-hole pairs which will move in
opposite directions for a time τ (recombination time) before recombining. The applied voltage appears almost
entirely across the space charge region-1, the strong internal field of the space charge region will rapidly sweep
the carriers within a diffusion length of the space charge region (Region-2), and they may also diffuse into the
region and be swept across by the field. For the photo generated current

Iph = qPinc η/hν * ΓG, Where, ΓG = photo current gain, Pinc= incident optical power, η = external quantum
efficiency.The transit times of carriers through the device are given by-

ttre = W/ve = W2/µeV for electrons and for holes ttrh = W/vh = W2/µhV, where ve and vh are saturation velocities of
electrons and holes respectively.

Before recombination the optical gain is given by

ΓG= Where W= the length of the semiconductor.

When the applied voltage V is small the drift velocity of the photo-generated carrier is small. Therefore the
probability of the recombination of EHP is high; this accounts for the loss in the number of the carriers reaching
at the terminal. The average time tetr/2, thtr/2>τ, where τ is the recombination lifetime. Hence the optical gain is
given by:

ΓG = [µe(V/w)τ+µh(V/w)τ]/W = {(µe+µh)/µe} τ/ttre

Where µe and µh are the electron and hole mobility. V is the applied bias.
For the case of two rectifying (blocking) contacts the reinjection of both types of carrier is prevented. When
Voltage (V) is small tetr> τ and tthr> τ, hence ΓG< 1. For larger bias values tetr> τ and tthr> τ, the electron path is
restricted to an average of w/2 because there is no reinjection. Hence ΓG= (1/2+τ/thtr) <1. Finally for large V,
tetr> τ and thtr< τ. Because there is no possibility of reinjection, the path lengths of both carriers are restricted and
ΓG=1.

It is observed that the optical gain increases almost linearly and then exponentially followed by saturation value
1 with the increase in applied voltage. At lower applied voltage the drift velocity of carriers is so small that loss
due to re- combination predominates over the gain due to photo generated carriers and therefore there is no net
gain. Optical gain starts when electron hole pair (EHP) generation rate predominates over the recombination of
EHP. The drift velocity increases with applied voltage which results in increase in optical gain. Figure- shows
the variation of optical gain for different optical generation gain (gop/gth), with applied voltage, where gop is the
optical generation rate and gth is the thermal generation rate. Experimentally it can be observed that the knee
voltage decreases almost exponentially throughout the range of normalized optical generation gain (gop/gth) used
in our calculation.

Figure 2: EHP generation process in simple Figure 3: Variation of optical gain with
p-n junction biasing voltage for different normalized

25 optical generation rate

MDCCT 2012
6 - 7 February, 2012; Burdwan University

Experimental Results

A Metal-Semiconductor-Metal (MSM) structure with aluminum (Al) as a metal and n-silicon as the base
semiconductor has been fabricated shown in Figure-4 and the corresponding voltage-current characteristics have
been measured under dark and different illuminated conditions shown in Figure 5.

It has been observed that the dark current due to thermally generated carrier is very small compared to the photo
generated current. The total current and photo generated current under different illuminating conditions
increases with applied voltage shown in Figures 6 and 7.

(a) (b) Figure 5: V-I characteristics of MSM photodiode
Figure 4: (a) Wafer (b) Fabricated Device under dark and illuminated conditions dark and
illuminated conditions

Figure 6: Total current under different Figure 7: Photo generated current under
illuminating conditions different illumination level

26

MDCCT 2012
6 - 7 February, 2012; Burdwan University

The knee voltage (voltage at where the current achieves the saturation value) decreases with increase in
illumination level shown in Figure-8. The knee voltage is 7.02volt for .05 Sun, 6.78volt for 0.1 Sun, 6.75volt for
0.25 Sun, 6.36volt for 0.4 Sun. It has also been observed that for a fixed bias voltage the photo generated current
increases with increase in illumination level.

Conclusion

In this article the voltage – current characteristics of a MSM photo diode has been studied under dark and
different illumination level. It has been found that the current gradually increases with illumination level as
incident photon rate is higher for higher illuminating level which results increase in electron-hole pairs in the
active region. But the saturation value of the current is same for different illuminating level which is actually

Figure 8: Knee voltage variation with illumination level
determined by the maximum carrier concentration depending on the doping concentration. The knee voltage is
gradually decreasing with increase in illuminating level due higher EHP generation rate for higher level of
illumination.

References:

[1] V. Hurm, W. Benz, M. Bronner, G. Kaufel, K. Koholer, M. Ludwig, E.Olander, B. Rayner and J.Rosenweigh, “20Gb/S fully
integrated MSM/HEMT photoreceiver”, Electronics Lett., 32, pp.683-685,1996.

[2] Z. Lao, V. Hurn, W.Bronner, A. Hulsmann, T. Jakobus, N. Schechtweg amd A. Thied, “ 20 Gb/s 14 Ghz trans impedance long-wave
length MSM-HEMT photoreceiver OEIC, “IEEE photon. Technol. Lett, 10, pp. 710-712, 1998.

[3] V. Hurm, J. Rsenberg, M. Ludwig, W. Benz, M. Berroth, A. Huelsmann, G. Kaufel, K. Kochler, B. Rayner and J. Schneider, “8 GHz
bandwidth monolithic integrated optoelectronic receiver using MSM photodiode and AlGaAs-GaAs HEMT”, Electronics Lett.,27,
pp.734-735,1991.

[4] S. Chandrashekhar, L. M Lunardi, A. H. Gnauck, D. Ritter, R.A. Hamm, M.B. Panish and G.J. Qua, “A 10 Gb/S OEIC photoreceiver
using InP/InGaAs HBT”, Electronics Lett., 28, pp.466-468,1992.

[5] L. M. Lunardi, S. Chandrashekhar, A. H. Gnauck and C. A. Burrns, “20 Gb/S monolithic p-I-n/HBT photoreceiver module for
1.55µm applications”, IEEE photonics Technol. Lett, 7,pp.1201-1203,1995.

[6] S. V. Averine, R. Sachot, “Transit-time considerations in metal-semiconductor-metal photodiode under high illumination conditions”,
Solid-State Electronics, vol.44, pp.1627-1634, 2000.

[7] K. Aliberti, M. Wraback, M. Stead, P. Newman, H. Shen, “Measurements of InGaAs metal-semiconductor-metal photodetectors
under high illumination conditions”, Appl.Phys. Lett., Vol.80, No 16, pp.2848-2850, 2002.

[8] S. V. Averine and R. Sachot , Solid-State Electronics, Volume 44, Issue 9,Pages 1627-1634 , 1 September 2000.
[9] In-Seok Seo, In-Hwan Lee, Yong-Jo Park and Cheul-Ro Lee,Journal of Crystal Growth, Volume 252, Issues 1-3, Pages 51-57 May

2003.
[10] E. Budianu, M. Purica, F. Iacomi, C. Baban, P. Prepelita and E. Manea, Thin Solid Films, Volume 516, Issue 7, Pages 1629-1633,15

February 2008
[11] Dayong Jiang, Xiyan Zhang, Quansheng Liu, Zhaohui Bai, Liping Lu, Xiaochun Wang, Xiaoyun Mi, Nengli Wang and Dezhen Shen

Materials Science and Engineering: B, Volume 175, Issue 1, Pages-41-43.15 November 2010.

27

MDCCT 2012
6-7 February, 2012; Burdwan University

Effect of Junction Temperature and Current Density on

Admittance Profile of Si DDR IMPATT Designed at 94 GHz

Arpan Deyasi1 and Swapan Bhattacharyya2
1Department of Electronics & Communication Engineering, RCCIIT, West Bengal, INDIA
2Department of Computer Science & Engineering, Asansol Engineering College, West Bengal, INDIA

Email: [email protected], [email protected]

Abstract
Extensive numerical analysis is carried out to estimate the effect of junction temperature and bias current
density on electric field profile, admittance characteristics and conversion efficiency based on a double
iterative computer technique developed on simultaneous numerical solution of Poisson’s equation, carrier
diffusion equation, continuity equation and Runge-Kutta method in addition with the effect of mobile
space charge subject to appropriate boundary conditions for electric field and normalized current density
at depletion layer edges for Si DDR IMPATT structure designed at 94 GHz. Dependence of optimum
frequency on junction temperature and bias current density is also computed, and degradation of
performance on these controlling parameters are studied. Simulation results will be useful to analyze Si
IMPATT’s performance in different mm and sub-mm wave regions, and hence design or proper heat
sink.

Introduction

Si-based IMPATT diode is now in the focus of semiconductor microwave device family due to its capability of
generating r.f power at microwave and millimeterwave frequency bands. The SDR (p+nn+) avalanche transit
time device operating at CW mode are the simplest structures from economical and commercial point of view,
when designed at atmospheric window frequencies. Shockley etc. first proposed [1] semiconductor transit-time
diode oscillator, which was later theoretically modeled by Read [2] through incorporation of impact ionization
and avalanche multiplication phenomena; and Lee [3] suggested first working SDR structure. Theoretical results
provided by eminent researchers established the importance of d.c field and carrier profiles to predict the
behavior of SDR and DDR IMPATT [4-5]. Xizeng [6] compared theoretically power output and efficiency
between single and double drift devices including the effect of doping profile and current density.

DDR devices attracted a greater interest of researchers for providing higher output power and efficiency in high
frequency range at room temperature [7]. But in mm-wave range, with increase of bias current density, problem
of higher junction temperature come to the focus [8]-[9] along with mobile space-charge effect and parasitic
series resistance degradation. Roy etc. [10-11] proposed double iterative method for simultaneous solution of
Poisson’s equation and continuity equation which also includes the effect of mobile space-charges for
numerically calculating performance of different IMPATT structures. Later, Mukherjee et. al. [12-15] analyzed
small-signal parameters for DDR IMPATT’s for different materials and operating at different frequency regions.
The present paper deals with the effect of junction temperature and bias current density on dc and ac properties
of symmetric DDR Si-IMPATT structure, which was numerically evaluated by double iterative technique
involving simultaneous solution of Poisson’s equation with continuity equation satisfying appropriate boundary
conditions. Based on dc analysis, admittance profile is obtained by Runge-Kutta method. Distortion of electric
field at higher current density is established with study on degradation of dc-to-rf conversion efficiency at
different junction temperatures. Optimum frequency profile is also calculated as a function of current and
junction temperature. Mobile space charge effect is also incorporated where saturated carrier velocities are
assumed to be independent of field throughout the space-charge layer.

Mathematical Modeling

Theoretical investigation of DDR IMPATT structure begins with double iterative technique where computation

is initialized from field maximum near the metallurgical junction. D.C field and current profiles for the diode
that involves simultaneous solution of Poisson’s equation, current density equations and continuity equations

incorporating mobile space charge effecting the depletion layer; where iteration is carried out over magnitude of

field maximum and its location in the depletion layer. Avalanche multiplication and carrier current flow across
the junction of a p+pnn+ diode is given by-

 E(z)  q  NA  p(z)  n( z )] (1)
x  [ND

Continuity equations for electron and hole current are given by

28

MDCCT 2012
6-7 February, 2012; Burdwan University

 (n, p)  1 J n, p  g Un,p (2)
t q z

In this analysis, we excluded the contribution of Un,p because the transit time of carriers in the depletion layer of
the device is several orders of magnitude lower than the recombination time. If the diffusion component of

current caused by carrier concentration gradient in the space charge layer is considered, then expressions of Jn

and Jp are given by-

J n,p  q(n, p)vn, p  qDn , p  (n, p) (3)
z

Diffusion current has been considered as a perturbation term over the major drift current in avalanche region.

Using perturbation technique, the expressions for hole and electron concentration in the space charge layer

(considering both diffusion and drift) can be written as-

(n, p)(z)   J n, p    Dn , p  J n, p    D2   2 J n,p    D3  3 J n, p   .. (4)
qvn, p  qvn, p 2 z  n, p  z 2   n, p  z 3 

qvn, p 3 qvn, p 4

Hence net mobile space charge density due to both drift and diffusion component is obtained as-

(z)  p(z)  n(z)   Jp    Jn    Dp  J p    Dn  J n  
 vp  vn vp2 z vn 2 z 

 Dp2   2 J p    Dn 2  2Jn    Dp3   3 J p    Dn 3  3Jn   .. 
vp3  z 2  vn3 z 2 vp4  z 3  vn 4 z 3
(5)

Thus reverse-bias condition can be represented as-

 1  Jp   Jn  
( q vp  vn 
 N D  NA )   

(z)   (zi )  q z  Dp  (6)
 zi vp2 .......
 1   J p   Dn  J n  
 q z  vn 2 z  


Field boundary conditions are given by-

 (z1 )   (z2 )  0 (7)

where z1 and z2 define n+ and p+ edges of depletion layer. Similarly boundary conditions for normalized current

density J(z) are-

J (z1 )  2 M p  1 (8.1)

and

J(z2) 1 2 (8.2)
Mn

where M n, p  J 0 J ns.ps are computed considering J p  J n  0.95J 0 , extension of avalanche zone has been

defined as the distance between two points (z1, z2) on either side of avalanche center.

Computation starts from the position of field maximum (z0) near metallurgical junction and then proceeds
towards edges of depletion layer. Double iterations are then carried out over field magnitude (E0) and position of
field maximum (z0) in depletion region to satisfy boundary conditions for field and current density. Once the
values of maximum electric field (Em) and it’s position (z0) for which both the boundary conditions are
simultaneously satisfied are computed at a given current density, this computer program provides more accurate
profiles of electric field and carrier current. Simultaneous satisfaction of boundary conditions at depletion layer
edges gives appropriate solution and dc field and current profiles are obtained.

Small-signal analysis of DDR IMPATT is carried out based on Gummel-Blue approach once dc field and
current profiles are obtained. The edges of depletion layer of the diode, obtained from dc analysis, considered as
starting and end points of small-signal analysis, which provides diode impedance by the double-iterative scheme
incorporating modified Runge-Kutta method. The diode negative resistance and reactance are computed over
active space charge layer as-

29

MDCCT 2012
6-7 February, 2012; Burdwan University

z2 (9.1)

 Z R    Rdx
z1

and

z2 (9.2)

 Z X    Xdx
z1

Diode admittance YD is given by-

YD  Z R 1   (G  jB) (10)
 jZ X

The negative conductance (G), susceptance (B) and quality factor (Q) of the device can be calculated as-

  G   ZR (11.1)
ZR2  ZX2

and

 B  Z X (11.2)
ZR2  ZX 2

It may be noted that –G and B are normalized to the area of the diode.

Results & Discussion

Spatial field distribution for DDR structure gives the position of metallurgical junction. With increasing junction
temperature, electric field increases for specified current density, profile becomes perturbed with higher current
density. This is evident from Figure 1. Also keeping current density constant, peak of the electric field increases
linearly over junction temperature, as shown in Figure 2. These profiles are important for computation of
admittance of the device.

Admittance profile, obtained from small-signal analysis, is dependent on both bias current density and junction
temperature. It is observed from Figure 3 that negative conductance increases with current density at constant
junction temperature, and peak of the profile has more negative conductivity with increase of current density.
Keeping current density constant, negative conductance is highest at lower temperature and it decreases with
temperature. This is plotted in Figure 4.

30

MDCCT 2012
6-7 February, 2012; Burdwan University

Increasing junction temperature affects conversion efficiency and it reduces with higher temperature, as evident
from Figure 5. It decreases with increasing current density, and ultimately becomes equal to the efficiency of
SDR device designed with same doping concentration. Optimum frequency, i.e., the frequency at which
negative conductance is a maximum at a particular current density, is a function of both temperature and current
density. It increases with current density and decreases with temperature. It is shown in Figure 6.

References

[1] W. Shockley, Bell Syst. Tech. J., 33, 799, 1954.
[2] W. T. Read, Bell Syst. Tech. J., 37, 401, 1958.
[3] C. A. Lee, R. L. Batdorf, W. Wiegman and G. Kaminsky, Applied. Physics Letters, 6, 89, 1965.
[4] D. L. Scharfetter, D. J. Bartelink, H. K. Gummel, R. L. Johnston, IEEE Trans. Elec. Dev, ED-15, 691, 1968.
[5] J. Pribetich, M. Chive, E.Constant and A. Farrayre, J.Appl. Phys., 49, 5584, 1978.
[6] F. Xizeng, S. Wenmiao, Journal of Electronics, 4, 266, 1987.
[7] T. E. Siedel, R. E. Davies, D. E. Iglesias, Proceedings of the IEEE, 59, 1222, 1971.
[8] M. Mukherjee, P. Tripathi, S. P. Pati, Archives of Applied Science Research, 2, 42, 2010.
[9] M. Mukherjee, J. Mukherjee, J .P. Banerjee, S.K.Roy, ICMMT-2008.
[10] S. K. Roy, M. Sridharan, R. Ghosh, B.B.Pal, NASECODE-I, 266, 1979.
[11] S. K. Roy, J. P. Banerjee, S.P.Pati, NASECODE-IV, 494, 1985.
[12] M. Mukherjee, J. P.Banerjee, Int. J. Adv. Sci. Tech., 19, 1, 2010.
[13] M. Mukherjee, S. Banerjee, J. P.Banerjee, CODEC 2009.
[14] M. Mukherjee, S. K. Roy, Current Applied Physics, 10, 646, 2009.
[15] M. Mukherjee, S. Banerjee, J. P.Banerjee, Terahertz Science and Technology, Vol.3, 97, 2010.

31

MDCCT 2012
6-7 February, 2012; Burdwan University

Mini Cement Plant using PLC

Alik Saha1, Sushil Kundu2 and Apurba Ghosh3
Department of Applied Electronics & Instrumentation Engineering,
University Institute of Technology, Burdwan University, Burdwan – 713 104, West Bengal, India
Email: [email protected], [email protected], [email protected]

Abstract
This paper analyses the simulation study of PLC based mini cement plant system by developing ladder
diagrams using the addressing modes of Allen Bradley PLC. This paper presents comprehensive
simulations of the performance of the different electrical actuators such as stepper motor, solenoid valve
etc., the position sensor and the conveyer belt. Program execution status can also be monitored by the
PLC. It scans memory, inputs and outputs in a deterministic manner. The user can access the logic,
counting and memory functions easily by using this adopted simulation strategy with good performance
and more reliability.
Introduction
Industrial automation is the use of control systems to control industrial machinery and processes, reducing the
need for human intervention. If we compare a job being done by human and by automation, the physical part of
the job is replaced by use of a machine, whereas the mental capabilities of the human are replaced with the
automation. PLC [1 – 5] is the abbreviated form of Programmable Logic Controller. Historically a PLC forte
was in discrete control of manufacturing processes. Most of the inputs and outputs for discrete control are
binary, meaning they have only two states i.e. on and off, like a switch. Little processing power is needed for
computing on/off signals and hence PLC tended to be very fast and are used in machine tool and other
industries. PLCs started becoming popular in the 1970s they were often called "relay replacers", since the logic
for on/off type operations was done with relays arranged in a digital logic format. Manufacturers of the most
popular PLCs include Allen-Bradley, Siemens, Modicon and a bunch of others. The design is for the production
of large volume high quality dry powdered cement with PLC used as its controller, which implements ladder
diagram logic to monitor the output signals from the different input sensors[4].

Figure 1: Block diagram of mini cement plant.

  32

MDCCT 2012
6-7 February, 2012; Burdwan University

System description

The gypsum and clinker are fed in a subsequent manner depending on the time adjusted on the timer into a
dump hoper and then it is collected into a crusher as shown in the Fig. 1. After crushing for some time, the
material is sent to bucket elevator-2 used for clinker via conveyer belt-2 using the motor-2. When gypsum is
crushing over, it is fed to bucket elevator-1 through conveyer belt-2 and conveyer belt-1 with the help of both
the motors M-2 and M-3. Further the crushed products of clinker are again sent to the bucket elevator-3 using
the conveyer belts 1 and 2 with the help of motors M-2 and M-4. The crushed gypsum is stored in the tank
(GYP) using motor, M-5. The crushed clinker is deposited in the tanks, CLH-1, CLH-2, CLH-3 and CLH-4 by
running the motors M-6 and M-7. Then the gypsum and the clinker from the respective storage tanks are fed to
the millfed bucket elevator operating the motors, M-8, M-9, M-10, M-11, M-12 and M-13. The materials are
then sent to the cement mill using motor, M-14. Once the production of cement [6] is complete, the finished
product is transferred using bucket elevators, motors (M-15& M-16) and solenoid valve (V-1) to large storage
silos (SILO-1 & SILO-2) in the shipping department. The final product is then transferred for packing and
loading using the motors, M-17, M-18, M-19, M-20, M-21 and M-22.

Explanation of ladder diagram

The logic implemented in PLCs is based on the three basic logic functions (AND, OR, NOT). These functions
are used either singly or in combinations to form instructions that will determine if a device is to be switched
ON or OFF.

The complete ladder diagram can be thought of as being formed by individual circuits, each circuit having one
output. Each of these circuits is known as a rung. Therefore, a rung [3] is the contact symbology required to
control an output in the PLC. A complete PLC ladder diagram program [2] then consists of several rungs, each
controlling an output interface which is connected to an output field device.

Here sensor (I:1/0) is implemented as start switch normally open (NO). If I:1/0 gets high value then the motor-
1(M1) of crusher starts crushing for some time which is run by timer T4:0 as shown in Fig. 2. Motor-2(M2),
motor-3(M3), motor-4(M4) are made on by using timer T4:0, T4:1 and T4:2 respectively having addresses of
O:2/1,O:2/2,O:2/3 respectively. The output address of O:2/4, O:2/5 and O:2/6 are used for running of motor-
5(M5), motor-6(M6) and motor-7(M7) respectively with the help of timer T4:3. Then the motors, M8, M9,
M10, M11, M12, and M13 are connected in the different addresses with the help of timer T4:4. When T4:5/TT
goes high then the motor -14(M14) connected in the address of O:2/13 will start. The motor-15(M15) would be
running when T4:6/TT goes high. The solenoid valve (valve-1) of cement mill is made on when T4:7/TT high
and T4:8/DN low. If T4:8/TT gets high then the M16 starts operating. M17 and M18 are on when T4:9/TT gets
high. When T4:9/DN, I:1/4 or O:4/3 are high and C5:0/DN (here C5:0 is implemented as counter) , T4:10/DN
are low then the motor-19(M19) connected in the address of O:4/3 and the motor-21(M21) connected in the
address of O:4/4 are on. If T4:9/DN high, I:1/5 or O:4/5 high, C5:1/DN low,T4:10/DN low are connected in
cascade then the M20 and M22 runs. Here the NO switch (I:1/1) is used for the system stop. Once system stop
gets high then overall system will shutdown.

  33

MDCCT 2012
6-7 February, 2012; Burdwan University

Figure 2: Ladder diagram.
  34

MDCCT 2012
6-7 February, 2012; Burdwan University

Conclusion

This Paper has discussed an application of PLC for the mini cement plant system. The results of simulation
study were presented as far as the steps of process progress. All of our personal sketches and drawings of the
setup and prospected design of the mini cement plant were saved for reference for the actual design process. A
simple decision logic scheme on receiving the input signals was used to process the outputs of the mini cement
plant system to indicate the status of the process.

The study also indicates the condition of the motor of the conveyer belt and the status of the solenoid valve.
Further work is requested to indicate the use of on-line parameter identification to improve the performance of
the decision logic.

Hence, it can be concluded that the system designed and developed set up works satisfactorily and can be used
for demonstrating an application of PLC.
References

[1] Gray Dunning, Introduction to Programmable Logic Controllers, Delmar Thomson Learning, 1998.
[2] Programmable Logic Control : Principles & Applications, NIIT, PHI, 2004.
[3] John W. Webb, Ronald A. Reis, Programmable Logic Controllers, principles & Applications, 5th Edition, PHI; 2003.
[4] Madhuchanda Mitra, Samarjit Sen Gupta, Programmable logic controllers & industrial automation.
[5] L. A. Bryan, E. A. Bryan, Programmable Controllers Theory and Implementation, Second Edition, 1997.
[6] J. P. Saxena, Refractory Engineering and Kiln Maintenance in Cement Plants, 2003

  35