Analysis of thermal fatigue failure using IoT

Analysis of Thermal Fatigue Failure in Material
by using the Temperature Prediction Polynomial
Regression Algorithm Employing
Bolt Wi-Fi Module
By
Manupratap Singh Parmar
Department of Mechanical Engineering,
Shrama Sadhana Bombay Trust's College Of Engineering and Technology,
Jalgaon, Maharashtra.

CONTENTS
ABSTRACT
INTRODUCTION
FATIGUE FAILURE INSPECTION METHODS
EFFECT OF TEMPERATURE VARIATION ON FATIGUE LIFE
PRACTICAL SET UP AND ANALYSIS OF THE ACCURACY OF PREDICTION BY BOLT
WORKING
CONCLUSIONS
REFERENCES

The failure caused by fatigue has a majority of share in the industries. The recent
trends and studies have shown that 85 to 90 % components fail not due to static
loading but due to cyclic loadings and fluctuating conditions around.
ABSTRACT
The most common and prominent being the thermal fatigue failure. Due to uneven
temperature conditions the components under cyclic loading tend to show a decreased
fatigue life as compared to the ideal conditions thus their probability of failure is quite
high and sometimes dangerous too.
This project presents an approach to do the online temperature analysis of the
conditions and predict the further failure conditions by analysing the thermal
stresses caused and preventing the potential sudden failure of the component by
using polynomial regression algorithm to predict the temperature conditions based
on the initial conditions and correlating it to the magnitude of thermal stress
induced and analysing the failure possibility of the component.

In the modern structures the static loading failure are comparatively less as compared to
the failure caused by cyclic loadings..
INTRODUCTION
S-N curve for Aluminium and low Carbon Steel

At high temperatures, the limiting factor in design is usually static strength, but resistance to fatigue
is an important consideration in engine design, particularly when static and alternating stresses are
combined.
Thermal stresses arise in materials when they are heated or cooled. Thermal stresses effect the
operation of facilities, both because of the large components subject to stress and because they are
effected by the way in which the plant is operated.
This idea also presents a supporting towards the prevention from thermal fatigue by using the
polynomial regression algorithm to predict the temperature conditions based on the initial
conditions and correlating it to the magnitude of thermal stress induced and analysing the failure
possibility of the component.
To use composite structures to their full potential, design strain levels will have to rise and a
partial growth criteria needs to be adopted; if this is to happen, an accurate fatigue lifing
methodology needs to be established

FATIGUE FAILURE INSPECTION METHODS
Visual Inspection
is the oldest and most basic method
of inspection . It is the process of
looking over a piece of equipment using
the naked eye to look for flaws
Liquid Penetrant Examination
The technique works via the principle of “capillary action,”
a process where a liquid flows into a narrow space
without help from gravity. Because it is one of the easiest
and least expensive NDE techniques to perform,
.

EFFECT OF TEMPERATURE VARIATION ON FATIGUE LIFE
S-N curve for DIN 35 NiCrMoV 12 5 steel at
three different temperatures
Coefficient of thermal expansion of nickel base
superalloy single crystal, nickel aluminide, and
NiCoCrAlY

Influence Of Temperature on Fatigue Crack Propagation
We know that according to theory,
σT = E (α ΔT)
ΔT = Tf - Ti where Tf > Ti or Tf < Ti
Where σT is the induced thermal stress, E is the Young’s Modulus,
(α ΔT) is the thermal Strain.
Α being the coefficient of thermal expansion and Ti is the steady temperature of
performance but subjected to temperature fluctuation of magnitude ΔT.

Building on this work, the BOLT Wi-Fi module was employed to predict the future temperature limits by using
the polynomial regression algorithm using a temperature sensor for a sample and it was subjected to different
temperature limits over time and the corresponding stress induced was recorded .
Thus this was in the initial phase, which served as a database for the machine learning algorithm to predict the
future fluctuation in temperature to indicate any sudden increase in thermal stress over a particular period of
time by using the graphical depiction of real time operation and accordingly the safety measures may be taken if
any parameter crosses the set limits.

It’s a IoT platform which enables us to control
the things through internet .Connect the
sensors, actuators etc. to bolt, write a short code
and it’s good to go. It collects, monitors and
visualise the data through the sensors
embedded
LM35 (Temperature sensor)
Schematic set of a shaft of suitable material fixed from ends
but free to rotate
PRACTICAL SET UP AND ANALYSIS OF
THE ACCURACY OF PREDICTION BY
BOLT

Schematic diagram of the practical setup to analyse the effect of temperature
variation and prediction and the thermal stresses induced

Hardware Connections
 For the LM35 connections, the
LM35 has 3 pins namely the VCC,
Output and Gnd.
 The VCC pin of the LM35 connects
to 5v of the Bolt Wi-Fi module.
 Output pin of the LM35 connects to
A0 of the Bolt Wi-Fi module and
Gnd pin of the LM35 connects to
the GND
 After the connections are done
power the Bolt Wi-Fi Module to
laptop via the USB cable.

WORKING
ML Polynomial Regression
The main steps involved in
Polynomial Regression are given
below:
o Data Pre-processing
o Build a Linear Regression model and
fit it to the dataset
o Build a Polynomial Regression model
and fit it to the dataset
o Visualize the result for Linear
Regression and Polynomial
Regression model.
o Predicting the output.

Applying the Polynomial Regression algorithm for the
temperature monitoring and induced thermal stress
calculation.
The temperature conditions in the system are varied and the variations in a pictorial format are
observed on the respective output device.
The machine learning requires some of the initial data to predict the further conditions so keeping
that in mind, some reading of temperature were influenced by the temperature control unit for the
system and the results can be seen clearly .
After a specific time period the temperature data is pushed to the cloud and through google chart
library the graph can be plotted toward data visualization.

Initial readings to serve as a feed to the ML
algorithm to assist prediction.
Practical Results
Real time temperature monitoring with specific time
duration.

Output for the predicted temperatures and a
comparative analysis of actual and predicted
σT = E (α ΔT)
ΔT = Tf - Ti where
Tf > Ti or Tf < Ti

CONCLUSIONS
Thermal fatigue failure is quite a big issue for the components subjected to cyclic loading along with
fluctuating temperature conditions. Thus there was an alarming need for the online temperature
monitoring system, which can analyse and predict the nature and magnitude of thermal stress induced.
The real time thermal stress induced can be calculated and by using a suitable software a separate
visualization can also be formulated by setting the threshold stress limits as it will show the variation of
stresses induced due to different temperature conditions
Thus will also serve as an indicator to schedule the inspection and maintenance of the mechanical
components by analysing the number of times the induced stress crossed the safety level and in the long run
such installations can also improve the working life of the components by a significant amount
Also it will save the components which are directly connected to the component being monitored as if it will fail
it will also affect the performance of the other components thus it will be very beneficial as far as the assembly
is taken into account.

REFERENCES
[1]. Deformation and Fracture Mechanics of Engineering
Materials -Hertzberg, Richard W. – John Wiley & Sons
1996.
[2]. Yield Point Phenomena in Metals and Alloys – E. O.
Hall –Plenum Press New York 1970.
[3]. Materials Science and Engineering, an Introduction
3rd ed. -Callister, William D. Jr. - New York: John Wiley &
Sons, Inc., 1994.
[4]. Mechanics of Materials 2nd ed. - Beer, Ferdinand P.,
and E.Russell Johnston, Jr. - New York: McGraw-Hill, Inc.
1992.
[5] Wood, M.I., 1989. The mechanical properties of
coatings and coated systems. Mater. Sci. Eng. A121, 633–
643
[6] EFFECT OF TEMPERATURE ON FATIGUE
PROPERTIES OF DIN 35 NiCrMoV 12 5 STEEL
A THESIS SUBMITTED TO THE GRADUATE SCHOOL OF
NATURAL AND APPLIED SCIENCES OF THE MIDDLE
EAST TECHNICAL UNIVERSITY BY ORKUN UMUR ÖNEM
July 2003
[7] F .C. Campbell, “Fatigue,” in Elements of Metallurgy and Engineering
Alloys, F.C. Campbell, Ed. ASM International, 2008, pp. 243–264
[8] Fatigue Of Metals – Forrest, Peter George – Owford, New York,
Pergamon Press, 1962.
[9] Fatigue At Elevated Temperatures - J. Wareing, B. Tomkins, and G.
Sumner – A.E. Carden, A.J. McEvily, and C.H. Wells (ed.) – ASTM Special
Technical Publication 520, 1972
[10] Pint, B.A., Haynes, J.A., More, K.L., Wright, I.G., Layens, C., 2000.
Compositional effects on aluminide oxidation performance: objectives
for improved bond coats. In: Pollack, T.M., et al. (Eds.), Superalloy 2000;
Pt aluminide data from Cheng, J., Jordan, E.H., Barber, B., Gell, M., 1998.
Thermal/residual stress in thermal barrier coating system. Acta Mater.
46, 5839–5850.
[11] Taplak, Hamdi & Uzmay, Ibrahim & YILDIRIM, Sahin. (2006). An
artificial neural network application to fault detection of a rotor bearing
system. Industrial Lubrication and Tribology. 58. 32-44.
10.1108/00368790610640082.
[12]https://microcontrollerslab.com/lm35-temperature-sensor-
pinout-interfacing-with-arduino-features/
[13] www.boltiot.com , Inventrom Private Limited, India
[14]https://www.javatpoint.com/machine-learning-polynomial-
regression

Analysis of thermal fatigue failure using IoT

More Related Content

What's hot (18)

Similar to Analysis of thermal fatigue failure using IoT (20)

Recently uploaded (20)

Analysis of thermal fatigue failure using IoT