EVALUATION AND OPTIMIZATION OF TENSILE STRENGTH

RESPONSES OF COPPER AT SUB ZERO TEMPERATURES

USING TAGUCHI ROBUST DESIGN AND MACHINE
LEARNING

PRESENTED BY
GUIDED BY DINESH VARMA (12241A0373)
DR. SWADESH KUMAR SINGH DEVARAJU ADHARAVENI (12241A0361)
RAVINDER MUPPIDI (13245A0314)
MADHU (13245A0319)
 CONTENTS TO BE COVERED

 Abstract
 Introduction
 Experiment
 Taguchi Method
 Machine Learning
 Experimental Data
 Results and Discussion
 Conclusion
 INTRODUCTION

Objectives

 To find the flow stress using Computer controlled UTM

 To predict the flow stress of alloy using Johnson Cook Model

 Resources to be used:

 CNC Milling Machine

 Vernier callipers

 Computer Controlled UTM

 Mat lab 2013

 MATERIAL

 Zirconium is a commercially available refractory metal with excellent properties like Composition
1. Corrosion resistance
2. Good mechanical properties
3.Very low thermal neutron cross section
4. Can manufactured by standard fabrication methods

Symbol Zr Zirconium Tin Chromium Iron

Discoverer Martin Heinrich

Discoverer In 1789
MATERIAL IN PERIODIC TABLE
 SPECIMEN AND ITS PREPARATION

Dimensions of the UTM specimen for tensile testing (in mm)

 EXPERIMENT
 The Zircalloy-4, with less oxygen content, of 1mm thick sheet was milled into
tensile test samples in rolling direction.
 The tensile tests were conducted on a computer controlled UTM, shown in Fig.,
which has a maximum load capacity of 100kN. The machine is equipped with a
control system to impose an exponential increase of the actuator speed to obtain
displacement . A contact type extensometer is used to measure the strains. The
resistance heating split furnace is used to heat the tensile test specimen up to
423K.
Strain Rates Used Temperature Used
0.001 s-1 298 K
0.005 s-1 348 K
Fig : Tensile test data range for JC model development
0.01 s-1 423 K
 JOHNSON COOK MODEL

 Development of Stresses in Johnson Cook Model depends on three independent phenomena

𝝈 = 𝑨 + 𝑩𝜺𝒏 𝟏 + 𝑪 𝐥𝐧 𝜺ሶ ∗ 𝟏 − 𝑻∗𝒎

𝑨 + 𝑩𝜺𝒏 - Strain Rate Hardening

𝟏 + 𝑪 𝐥𝐧 𝜺ሶ ∗ - Isotropic Hardening
𝟏 − 𝑻∗𝒎 - Thermal Hardening

where 𝜎 is the Von-mises flow stress A indicates the yield stress at reference temperature
and reference strain rate, B indicates the coefficient of strain hardening, 𝜺 indicates the true plastic strain,
𝒏 is strain hardening exponent, 𝑪 is the coefficient of strain rate hardening, 𝜺ሶ ∗ = 𝜺/
ሶ 𝜺ሶ 𝟎 , where 𝜺ሶ is true
strain rate, εሶ 0 is the reference strain rate, 𝑻∗ is the homologous temperature and the 𝒎 is the thermal
softening exponent.
 ISOTROPIC HARDENING

 For isotropic hardening, if you plastically

deform a solid, then unload it, then try
to reload it again, you will find that its
yield stress (or elastic limit) would have
increased compared to what it was in
the first cycle. This is isotropic
hardening.
 STRAIN RATE HARDENING

 The strain rate hardening is used to express the increase of flow stress with the increase of loading strain rate.

 Strain hardening is related to the increase in flow stress by increasing strain as a result of

(1) Interaction of the stress fields of the dislocations.

(2) Back-stress resulting from dislocation pile-ups at barriers which are sessile dislocations such as Lomer-Cottrell
barriers.

(3) Dislocation intersections which results in a small step or jog in the dislocation line and restricts its motion.

 Strain rate hardening is related to increase in flow stress by increasing strain rate as a result of high value of the
strain rate sensitivity (m), which is one of the major requirements to inhibit necking during deformation.
 THERMAL HARDENING

 In thermal hardening a hard layer is formed at the surface of plain carbon and low alloy steels of medium carbon
content (0.3 - 0.6%) by heating the steel till it reaches the austenistising temperature (about 850oC, depending
upon carbon content) then rapidly quenching it to form a martensitic structure.

 The depth of hardening can be controlled by the rate of heating, rapid heating for a 5 - 20 seconds with an
induction coil will limit the hardened zone to material adjacent to the surface. The more rapid the cooling the
greater will be the depth of hardening (up to the limit of the austenitised zone) - the thinner the section the
easier it is to cool rapidly. However the faster the quench the greater the danger of distortion or cracking.
Increasing alloying content also gives greater depth of hardening and will allow a less severe quench to be used.

 The depth of hardening is normally in the range of 1 - 5 mm.

DETERMINATION OF MATERIAL CONSTANTS IN JC MODEL

 Step 1: At reference strain rate and reference temperature.

 Step 2: At a fixed strain and reference temperature.
 Step 3: At a fixed strain and reference strain rate.
 Step 4: Optimization of the C and m values.
 Step 5: Predictability of the constitutive equation.
All the above equations are solved by using MATLAB.
Material constants for JC model

Material A B C n m
Constant
Value 283 774.2 0.0152 0.8499 0.8012
 EXPERIMENTAL DATA
Displacement Load
(mm) (KN) Displacement vs Load Data
0 0
0 0.01
Direction – Longitudinal Temperature - 298k
0.01 0.01
0.01 0.02 Strain Rate - 0.001
0.02 0.02
0.03 0.02 Load vs Displacement
0.03 0.03
0.03 0.02
0.03 0.03 2.5
0.04 0.03
0.05 0.03 2
0.05 0.04
1.5

Load(KN)
0.06 0.04
0.07 0.04
1
0.07 0.05
0.07 0.04 0.5
0.07 0.05
0.07 0.04 0
0 2 4 6 8 10 12 14
0.07 0.05 Displacement (mm)
0.08 0.05
 RESULTS AND DISCUSSION
600 (a) 0.01s-1
1
2
500
True Stress(MPa)

400

300

200

100

0
0 0.05 0.1 0.15 0.2 0.25 0.3
True Plastic Strain
Experimental- 298K 348K 423K
Predicted- 1: 298K 2: 348K 3: 423K

The equation for JC model is given by

𝝈 = 𝟐𝟖𝟑 + 𝟕𝟕𝟒. 𝟐 × 𝜺𝟎.𝟖𝟒𝟗𝟗 𝟏 + 𝟎. 𝟎𝟏𝟓𝟐 × 𝒍𝒏 𝜺ሶ ∗ 𝟏 − 𝑻∗𝟎.𝟖𝟎𝟏𝟐
 RESULTS AND DISCUSSION
600 R=0.7616

Predicted True Stress (Mpa)

Figures show the predicted values of flow stress for JC 500

model at strain rates of 0.01, 0.005 and 0.001s-1. However, 400

the R value for the model came out to be 0.7616(as shown 300

in Fig. ) and value of average absolute error was 11.95% and 200

standard deviation of 12.13. The predictive capability of the 100

JC model is low mainly due to the influence of the coupled 0

0 100 200 300 400 500 600
effect of strain and temperature, and of strain rate and Experimental True Stress (Mpa)
temperature. JC Model Linear (JC Model)

Correlation between experimental and predicted stress.

 CONCLUSION

 We have experimentally found the flow stress of the alloy at different

temperature and different strain rates
 We have developed the Johnson cook model for the alloy at different
temperature and different strain rates
 We have optimized the parameter of Johnson Cook model for the alloy
 We have predicted the flow stress using Johnson Cook model
Hence, it can be concluded that the flow behaviour of Zircalloy-4 with low oxygen content at elevated
temperature is not governed by three independent phenomena, viz., thermal softening, strain rate hardening and
isotropic strain hardening but a combination of them.
THANK YOU