ANN Model Based Calculation of Tensile of Friction Surfaced Tool Steel

International Journal of Trend in Scientific Research and Development (IJTSRD)
Volume 3 Issue 6, October 2019 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470
@ IJTSRD | Unique Paper ID – IJTSRD29169 | Volume – 3 | Issue – 6 | September - October 2019 Page 494
ANN Model Based Calculation of
Tensile of Friction Surfaced Tool Steel
V. Pitchi Raju
Professor, Mechanical Engineering Department,
Indur Institute of Engineering & Technology, Hyderabad, Telangana, India
ABSTRACT
Friction surface treatment is well-established solid technology and is usedfor
deposition, abrasion and corrosion protectioncoatingsonrigid materials.This
novel process has wide range of industrial applications,particularlyinthefield
of reclamation and repair of damaged and worn engineering components. In
this paper, present the prediction of tensileoffrictionsurfacetreatedtool steel
using ANN for simulated results of friction surface treatment. Thisexperiment
was carried out to obtain tool steel coatings of low carbon steel parts by
changing input process parameters such as friction pressure,rotational speed
and welding speed. The simulation is performed by a 33-factor design that
takes into account the maximum and minimum limits of the experimental
work performed by the 23-factor design. Neural network structures, such as
the Feed Forward Neural Network (FFNN), were used to predict tensile tool
steel sediments caused by friction.
KEYWORDS: Friction surfacing, Artificial Neural Networks (ANN), Process
Parameters
How to cite this paper: V. Pitchi Raju
"ANN Model Based Calculation of Tensile
of Friction Surfaced
Tool Steel"Published
in International
Journal of Trend in
Scientific Research
and Development
(ijtsrd), ISSN: 2456-
6470, Volume-3 |
Issue-6, October 2019, pp.494-500, URL:
https://www.ijtsrd.com/papers/ijtsrd29
169.pdf
Copyright © 2019 by author(s) and
International Journal ofTrendinScientific
Research and Development Journal. This
is an Open Access
article distributed
under the terms of
the Creative Commons Attribution
License (CC BY 4.0)
(http://creativecommons.org/licenses/by
/4.0)
1. INTRODUCTION
Friction surfacing is an advanced technology that can
effectively deposit a metal on another metal. In this process,
the consumable rod is rotated and forced against the
substrate in the axial direction. A large quantity ofhotness is
produced due to the friction among the consumable rod and
the friction contact surface between the substrates, and the
contact end of the metal consumption rod is plasticizedafter
a certain period of time. The substrate is then horizontally
moved to a vertically consumable rod, so that a layer of
mechanical material is deposited on the substrate. Friction
surface treatment has been used for a varietyofhardsurface
metal coatings, such as mild steel or stainless steel coating
on the tool steel coating. In this process, the strong adhesion
between the coating and the substrate can only be achieved
by applying a high contact pressure, but this requires
expensive machinery [1,2]. Friction surface treatment has
significant advantages over conventional fusion welding
processes. This novel process correlates many process
parameters, which directly affect the quality of the deposit.
In this process, the obtained coating is fairly flat andregular,
and there is no conventional cross-sectional profile of the
invasive meniscus [3]. This process can be considered in
another key area that is damaged and damaged by the
reclamation and repair of engineering components [4]. A
number of industrial applications have been observed in
friction surface treatment and are mainly used to deposit
hard materials on the cutting edges of varioustoolsrequired
for the food processing,chemical andmedical industries.The
process can be widely used in tool steel, aluminum, stainless
steel and mild steel, copper-nickel alloy and other materials
[5-7]. This innovation process can be carried out in open air
[8], water [9] and inert gas [10]. In the process, the right
choice of process factors is critical to attaining the quality of
the coating. The axial force actingontheconsumablerod,the
rotational speed of the rod and the transverse velocityof the
substrate are the main process parameters affecting the
coating properties such as coating thickness, coating width
and adhesive strength. In order to achieve the desired
mechanical properties, it is necessary to understand the
correlation between mechanical properties and process
parameters. Okuyucu Kurt and Areaklioglu [11] obtained
correlation between mechanical properties and FSW
parameters using artificial neural networks (ANNs), whose
attempts focused on linking process parameters ratherthan
optimizing them. Now in the field of metal processing,
the use of artificial neural net works is also increasingly
important.
The focus of this study is on computer-aided ANN models to
predict the tensile and shear strength of tool steel M2
deposits formed by frictionsurfaces.Duetothelimitationsof
the experimental work, the simulation was carried out by
taking into account the maximum and minimum 33 factor
designs of the experimental work carried out by 23 factor
designs. The feed forward neural network (FFNN) was used
to predict the tensile and shear strength of the friction
surfacetoolsteelM2sediments.
IJTSRD29169

International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
Artificial neural networks (ANNs) can be used in various
fields of engineering applications, by using the input data to
obtain the required information, to overcome the
shortcomings of traditional methods [12]. The prediction of
the friction surface response is carried out by the
mathematical modeling of ANN,whichrepresentsthetensile
and shear strength of the input parameters.
The structure of the feedforward neural network (FFNN)
consisting of three layers consists of a concealment and
output layer and an arbitrary activation functionisa general
approximator [12]. The architecture of the FFNN network
model is shown in Figure 1.
Fig1. Architecture of FFNN network model
Figure 1 depicts the network model of input neurons, hiding neurons and output neuronal structures. Input layers include
network input process parameters suchasfrictionpressure,velocity,andweldingspeed.Hiddenlayersincludeneuronsknown
to map points in the input area to coordinates in the output area. The output area is called the transfer function of the
activation function processing input layer. In this case, the hyperbolic tangent function is selected as the activation function
because it tests the minimum mean square error between the other functions,suchasGaussianandlogarithmicfunctions[12].
2. EXPERIMENTAL WORK
The main process constraints such as friction, rotational speedandweldingpressureareselectedastheprocessconstraintsfor
the investigational procedure of numerous manageable process parameters which affect the tensile and shear strength of the
friction surface tool steel M2. For the experimental work, the range of the friction pressure (X1) was set to 105 kN, the
mechanical speed (X2) was (100-300 rpm), and thesubstratetraversespeed(X3)was(40-60mm/min).Themain parameters
with 23 factor designs were selected and the tool steel M2 was deposited on the mild steel. The sedimentsobtainedfromthese
eight treatments are shown in Table 1. After each test, a preliminarytest wasconductedinthe workshoptodeterminethebond
strength of the low carbon steel tool steel deposits.
Table.1.The process parameters used and tool steel deposits over low carbon steel obtained in the experimental
work.
The tensile strength was determined experimentally by applying a tangential force in the contact area by tensile test method
and shear strength. The values obtained are listedinTable1.Tensilestrengthisofparamountimportancefordesigningvarious
engineering components such as containers, pressure vessels,turbine blades,helicopter bladesandpumps.Samplesfortensile
strength tests have square-sized deposits and have round holes from the other side of the sample. The friction surface of the
tool steel deposits is separated from the low carbon steel substrate by the influence of the tension applied bytheindenter.The

tensile strength of the sample is then calculated by the ratio of the applied tensile load to the bearing area of the sample. The
shear strength test is of the utmost importance for the designpointofview, whichisdetermined byapplyingtangential loadsto
the deposit. The shear strength of the specimen is calculated by dividing the applied load by the shear area. Tensile strength
test before and after the test sample in Table 2.
Table2. Specimens before and after testing
3. EFFECT OF PROCESS PARAMETERS ON BOND STRENGTH OF THE DEPOSIT
3.1. Determination of Regression Equations
Construct the variance analysis (ANOVA) table to check the importance of all theprocessparametersforthetensile strengthto
determine the regression equation. The regression equation for the response to tensile strength, after eliminating less
important terms, can be rewritten as y = 103.4 + 15.125X1 + 21.875X2 + 6.875X3-18.87 X1X3 + 21.875X2X3
Similarly, the regression equation for the shear strength after eliminating the least significant term can be rewritten as
y = 43.75 + 4.5X1 + 9.75x2 + 4.25X3-9X1X3 + 11.25X 2X3
3.2. Prediction of Tensile and Shear Strength by using Artificial Neural Network (ANN)
The MATLAB R2012a version of the neural network toolbox is used to develop artificial neural networks (ANN) models for
predicting the tensile and shear strength of frictional surface sediments. The input layer consists of three processparameters,
namely, friction pressure, speed and welding speed, the output layer represents the tensile strength and shear strength.
Initially enter the input data into the neural network, and then simulatetoachievetheoutput.Whencreatinga neural network,
the velocity constants and the maximum number of neurons are changed to achievedifferentresults.Thisisdone byusingtrial
and error methods. The experimental parameters of the artificial neural network (ANN) model are shown in Table 3.1.
Table3.1 Experimental plan for selecting process parameters
The effect of process parameters such as friction pressure, rotational speed and welding speed on meanoftensilepotencyand
mean of shear potency are indicated in the figures 3.3, 3.4, 3.5, 3.6, 3.7 and 3.8

Fig 3.3: Variation of mean of tensile strength at different friction pressures
Fig 3.4: Variation of mean of tensile strength at different rotational speeds
Fig 3.5: Variation of mean of tensile strength at different welding speeds

Fig 3.6: Variation of mean of shear strength at different friction pressures
Fig 3.7: Variation of mean of shear strength at different rotational speeds
Fig 3.8: Variation of mean of shear strength with welding speeds

4. RESULTS AND DISCUSSIONS
From the regression equation, the results show that the tensile strength is proportional to the combined effect of friction
pressure, rotational speed, speed and welding speed.Fromthepredictionofartificial neural network (ANN)model,theaverage
value of tensile strength and shear strength increases with the increase of friction pressure, rotational speed and welding
speed. Therefore, ANN can be used to determine the effect of process parameters on bond strength. Figure 4.1 depicts the
predicted values of the tensile strength of the sediments using FFNN and the experimental values. Figure 4.2 shows the
predicted Vs experimental values for shear strength.
Fig.4.1. Depicts the variation of predicted and experimental values of tensile strength of the deposit by using FFNN
Fig 4.2: Predicted Vs Experimental Values for Shear Strength
CONCLUSIONS
By changing the input process parameters, 23 factordesigns
were used to perform experiments on a friction surfacing
machine. Due to the limitations of the experimental work,
taking into account the maximum and minimum
experimental work, carried out a 3 3 times the design of the
simulation. The tensile of the samples were measured using
a universal testing machine. It can be seen from the
regression equation that the tensile strengthproportional to
the friction pressure, the rotational speed and the welding
speed. The tensile and shear strength of the tool steel M2
deposit produced by the friction surface treatment is
predicted by a feedforward neural network (FFNN) using
artificial neural networks (ANN). The results show that the
predicted values are closely related to the experimental
values. Thus, ANN technology is the most effective method
for predicting the tensile and shear strength in friction
surface treatment and can also be tested in many other
surface modification processes. Therefore, ANN is an
alternative to validating experimental values.
REFERENCES
[1] P.Lambrineas and al.1990 Institute of Engineering,
Australian Tribology Conference. edited by D. Scott,
Institute of Engineering Australian National
Conference, 90(14) 23

[2] B M Jenkins and E D Doyle 1987 Proceedings of the
International Tribology Conference Australian
National conference, 87(18)
[3] VSugandhi and V. Ravi Kumar 2012 Optimization of
Friction Surfacing Process Paramters for AA1100
Aluminum Alloy Coating with Mild Steel Substrate
using Response Surface Methodology (RSM)
Technique. Modern Applied Science, published by
Canadian Center of Science and Education, 6(2) 69-73
[4] Yamashita Y and Fujita K 2002 Newly developed
repairs on welded area of LWR stainless steel by
friction surfacing. Journal of Nuclear Science
Technology, 105-12.
[5] Puli R and Ram GDJ 2012 Microstructures and
Properties of Friction Surfaced Coatings in AISI 440C
Stainless Steel. Surf Coat Tech,207(310)
[6] Hanke S and al. 2011 Cavitation Erosion of NiAl –
Bronze Layers Generated by Friction surfacing .Wear,
273(32)
[7] Govardhan D and al. 2012 Characterization of
Austenitic Stainless Steel Friction surfaced Deposit
Over Low Carbon Stee. 36(206)
[8] Tokisue H and Katoh K 2005 Structures and
mechanical properties of multilayer friction surfaced
aluminum alloys. Report of the research Institute of
Industrial Technology Nihon University,78
[9] Li JQ and Shinoda T 2000 Underwater friction
surfacing. Surface Engineering, 16(1) 31-35
[10] Chandrasekaran M and Batchelor AW 1997 Study of
the interfacial phenomena during friction surfacing of
aluminum with steel. Journal of Material Sciences,(32)
6055-6062
[11] H Okuyucu and al. 2007 Artificial neural network
application to the friction-stir welding of aluminum
plates .Materials & Design , 28 (1), 78-84
[12] K. Brahma Raju and al.2012 Prediction of Tensile
Strength of Friction Stir Welded Joints Using Artificial
Neural Networks. International Journal of Engineering
Research & Technology (IJERT), 1(9)

ANN Model Based Calculation of Tensile of Friction Surfaced Tool Steel

More Related Content

What's hot (17)

Similar to ANN Model Based Calculation of Tensile of Friction Surfaced Tool Steel (20)

More from ijtsrd (20)

Recently uploaded (20)

ANN Model Based Calculation of Tensile of Friction Surfaced Tool Steel