Project phaseII Report

VISVESVARAYA TECHNOLOGICAL UNIVERSITY
Jnana Sangama, Belagavi-590018, Karnataka
PROJECT REPORT
(15ECP85)
ON
“PREDICTION OF REMAINING USEFUL LIFE OF AIRCRAFT
ENGINE”
Submitted in partial fulfillment of the requirements for the award of degree of
BACHELOR OF ENGINEERING
in
ELECTRONICS AND COMMUNICATION ENGINEERING
by
Karthik G Sarode 1BG16EC051
Mohit Kumar Pandey 1BG16EC064
Rajesh Kumar Jha 1BG16EC084
Chaitra Ilkal 1BG14EC021
Under the guidance of

Mr. Sudharshan
Assistant Professor,
Dept. of ECE, BNMIT
B.N.M. Institute of Technology

Approved by AICTE, Affiliated to VTU, Accredited as Grade A Institution by NAAC. All UG branches – CSE, ECE, EEE, ISE & Mech.E
Accredited by NBA for academic years 2018-19 to 2020-21 & valid upto 30.06.2021 Post box no. 7087, 27th cross, 12th Main,
Banashankari 2nd Stage, Bengaluru- 560070, INDIA Ph: 91-80- 26711780/81/82 Email: principal@bnmit.in, www.bnmit.org
Department of Electronics & Communication Engineering

2019 – 2020
B.N.M. Institute of Technology
Approved by AICTE, Affiliated to VTU, Accredited as Grade A Institution by NAAC. All UG branches – CSE, ECE, EEE, ISE &Mech.E Accredited by NBA for
academic years 2018-19 to 2020-21 & valid upto 30.06.2021 Post box no. 7087, 27th cross, 12th Main, Banashankari 2nd Stage, Bengaluru- 560070, INDIA
Ph: 91-80- 26711780/81/82 Email: principal@bnmit.in, www.bnmit.org
Department of Electronics and Communication Engineering
CERTIFICATE
Certified that the project (15ECP85) entitled “Prediction Of Remaining Useful Life Of Aircraft
Engine” carried out by Karthik G Sarode (1BG16EC051), Mohit Kumar Pandey (1BG16EC064),
Rajesh Kumar Jha (1BG16EC084), and Chaitra Ilkal (1BG14EC021) bona fide students of VIII
semester in partial fulfillment for the award of Bachelor of Engineering degree in Electronics and
Communication Engineering of the Visvesvaraya Technological University, Belagavi during the year 2019-
2020. It is certified that all corrections/suggestions indicated for internal assessment have been incorporated
in the report deposited in the department library. The project phase I report has been approved as it satisfies
the academic requirements in respect of Project work prescribed for the said degree.
Mr. Sudarshan Dr. P.A. Vijaya Dr. Krishnamurthy. G.N

Assistant Professor Professor and Head Principal
Dept. of ECE Dept. of ECE BNMIT
BNMIT, Bengaluru BNMIT, Bengaluru Bengaluru
ACKNOWLEDGMENT
On successful completion of our project, we would like to place on record our sincere
thanks and gratitude to the concerned people, whose suggestions and words of encouragement
has been valuable.
We express our heartfelt gratitude to BNM Institute of Technology, for giving us the
opportunity to pursue Degree of Electronics and Communication Engineering and helping us
to shape our career. We take this opportunity to thank Prof. T. J. Rama Murthy, Director,
BNMIT, Prof. Eishwar N Maanay, Dean, BNMIT and Dr. Krishnamurthy. G.N,
Principal, BNMIT for their support and encouragement to pursue this project. We would like
to thank Dr. P.A.Vijaya , Professor and Head, Dept. of Electronics and Communication
Engineering, for her support and encouragement.
We would like to thank our guide Mr. Sudarshan, Assistant Professor, Dept. of
Electronics and Communication Engineering, who has been the source of inspiration
throughout our project work and has provided us with useful information at every stage of our
project.
We would also like to thank our project coordinator Mrs. Anuradha J P Assistant
Professor, Dept. of Electronics and Communication Engineering, who is kind enough to extend
her help for our project whenever the need arose.
Finally, we are thankful to all the teaching and non-teaching staff of Department of
Electronics and Communication for their help in the successful completion of our project. Last
but not the least we would like to extend our sincere gratitude to our parents and all our friends
who gave us a constant source of inspiration.
Karthik G Sarode
Mohit Kumar Pandey
Rajesh Kumar Jha
Chaitra Ilkal
i
ABSTRACT
The technological development in recent years has caused an increase in the engineering
systems and their complexity, especially in the areas of electronics, nuclear energy,
automotive and aerospace. Although a good design is essential to obtain a high reliability,
the deterioration due to time, wear and working conditions, impacts directly on the
performance of the systems, which implies more efficient maintenance programs. As the
main power source for aircrafts, the reliability of an aero engine is critical for ensuring the
safety of aircrafts. Prognostics and health management (PHM) on an aero engine can not
only improve its safety, maintenance strategy and availability, but also reduce its operation
and maintenance costs.
Residual useful life (RUL) estimation is a key technology in the research of PHM.
According to monitored performance data from the engine’s different positions, how to
estimate RUL of an aircraft engine by utilizing these data is a challenge for ensuring the
engine integrity and safety.
In this project we aim to build a machine learning model to analyze the data form the sensors
of an aircraft engine and using Stochastic Gradient Descent (SGD) to predict the Remaining
Useful Life (RUL) of an aircraft engine.
ii
TABLE OF CONTENTS
Acknowledgment I
Abstract II
List Of Figures III
List of Tables IV
CHAPTER 1 Introduction 1
1.1Overview 1
1.2Motivation 3
1.3Existing Systems 4
CHAPTER 2 Literature survey 5
2.1New stochastic model to predict RUL 5
2.2RUL prediction using BDLM and Copulas 5
2.3Battery RUL prediction using SVM 6
CHAPTER 3 Methodology 7
3.1Experiment flow diagram 7
3.2SGD Algorithm 12
CHAPTER 4 Present work 15
4.1Flowchart of program 15
4.2System requirement 16
4.3Sensors used and threshold setting 25
CHAPTER 5 Results & Discussion 29
5.1Ordinary Least Squares or Linear Regression_____________________________________29
5.2Support Vector Regression____________________________________________________31
5.3Random forest______________________________________________________________32
5.4Neural network_____________________________________________________________33
CHAPTER 6 Conclusion & References_______________________________________________36
iii
III
List of Figures
Figure 3.1 Process flow diagram .................................................................................................. 07
Figure 3.2 Supervised vs Unsupervised learning .......................................................................... 10
Figure 3.3 Sample network ........................................................................................................... 11
Figure 3.4 Activation function to be used in algorithm ................................................................ 12
Figure 3.5 Sigmoid function and Rectifier function ..................................................................... 13
Figure 3.6 Path taken by Batch Gradient Descent ........................................................................ 15
Figure 3.7 Path taken by Stochastic Gradient Descent ................................................................. 15
Figure 4.1 Flowchart of the working of the algorithm.................................................................. 16
Figure 4.2 Data Visualization of various sensor outputs .............................................................. 24
Figure 4.3 Plot of all the sensor data............................................................................................. 25
Figure 5.1 Hypothesis Testing& OLS………………………………………………………….29
Figure 5.2 Chi Square Values…………………………………………………………………...30
Figure 5.3 SVR Result…………………………………………………………………………..31
Figure 5.4 Random Forest Result……………………………………………………………….32
Figure 5.5 Loss and Error……………………………………………………………………….33
Figure 5.6 Comparison between true and predicted RUL……………………………………….34
Figure 5.7 Final Result…………………………………………………………………………..35
IV
List of Tables
Table 4.1 Dataset description that would be used in program ................................................. 23
Table 4.2 Details of various sensor components and their units.............................................. 26
iv
Prediction of Remaining Useful Life of an Aircraft Engine
CHAPTER 1
INTRODUCTION
1.1 Overview:
As the technological development increases, the engineering systems are more complex, especially
in the areas of electronics, nuclear energy, automotive and aerospace. Although a good design is
essential to obtain a high reliability, the deterioration due to time, wear and working conditions,
impacts directly on the performance of the systems, which implies more efficient maintenance
programs. As the main power source for aircrafts, the reliability of an aero engine is critical for
ensuring the safety of aircrafts. Prognostics and health management (PHM) on an aero engine can
not only improve its safety, maintenance strategy and availability, but also reduce its operation
and maintenance costs. Residual useful life (RUL) estimation is a key technology in the research
of PHM. According to monitored performance data from the engine’s different positions, how to
estimate RUL of an aircraft engine by utilizing these data is a challenge for ensuring the engine
integrity and safety.
Through-life Engineering Services (TES) are essential in the manufacture and servicing of
complex engineering products. TES improves support services by providing prognosis of run-to-
failure and time-to-failure on-demand data for better decision making. The concept of Remaining
Useful Life (RUL) is utilised to predict life-span of components (of a service system) with the
purpose of minimising catastrophic failure events in both manufacturing and service sectors. The
purpose of this paper is to identify failure mechanisms and emphasise the failure events prediction
approaches that can effectively reduce uncertainties. It will demonstrate the classification of
techniques used in RUL prediction for optimisation of products’ future use based on current
products in-service with regards to predictability, availability and reliability. It presents a mapping
of degradation mechanisms against techniques for knowledge acquisition with the objective of
presenting to designers and manufacturers ways to improve the life-span of components.
Since the aircraft engine is a complex system, there are various monitored performance data from
different positions during its operation. How to estimate RUL of an aircraft engine by utilizing
these data has become the focus of most engine industries. Until now, approaches to predict system
Dept. Of ECE, BNMIT 1 2019-20

lifetime can be broadly categorized into three types: physics-based models, data-driven approaches
and hybrid approaches. Generally, a physics-based model utilizes the failure physical model of the
system/component to estimate its RUL, which is usually based on the system/component’s physics
of failure or physics of dynamics deeply. It can usually obtain reasonable and accurate predictions
of RUL based on physical models with limited historical data. However, it is usually different or
too expensive to apply a physics-based model to a complex system. Besides, this approach has
shown significant limitations due to the assumptions and simplifications of the adopted models.
The data-driven approach utilizes the monitored operational data relating to system health for RUL
estimation, which is preferred when the system’s failure physics is complicated or unavailable but
systems’ degradation procedure and degradation data are available. Note from that the data-driven
approach provides accurate RUL predictions for a complex system, which can be applied quickly
and cheaply compared to the physics-based model. Furthermore, recent development of sensor
technology and simulation capabilities enables us to continuously monitor the healthy situation of
a complex system and obtain the related large amount of performance index data. Prognosis Health
Management (PHM) is a term introduced in the field of medicine as the prediction of the
development and outcome of a disease.
Derived from the same concept, several prognosis methods have been developed for machinery
maintenance during the last 14 years this is a field of research and application that uses data of the
current condition of the system to predict when and how the system is more prone to fail during
its useful life. This basically refers to the prediction of the remaining useful life (RUL), which is
an important concept for decision making in preventive maintenance. Techniques based on
machine learning are appropriate when there is not enough information to understand the
behaviour of a system, or when a system is complex enough to develop precise physics-based
models. The objective is to use the data from the measurements of important variables that account
for the condition of the system. For example, pressure, temperature, speed, vibration, electric
current, among others, and from them to create a model that correlates the behaviour of these
parameters with the level of degradation of the system with the objective of predicting its RUL.

1.2 MOTIVATION:
Every year the flight industry incurs huge direct (i.e. plane loss, after damage repair, life and
personal loss, legal costs) and indirect costs (i.e. customer dissatisfaction, reputation damage) due
to a failure in any of the equipment’s currently deployed over the aircraft.
It is to the interest of the flight companies to look for a solution how they can predict the time
when any of the mechanical parts would need a maintenance. which leads to the concept of
predictive maintenance.
List of accident due to engine failure:-
1)Indian Airlines Flight 171 crashed at Bombay following an in-flight fire caused by an
uncontained engine failure on 12 October 1976, killing all 95 on board.
2)Air India Flight 855 crashed off the coast of Bandra, Bombay, (now Mumbai) on 1 January 1978
when the captain became spatially disoriented after the failure of one of the flight instruments in
the cockpit. All 213 aboard were killed.
3)Baikal Airlines Flight 130: a starter of engine No. 2 on a Tu-154 heading from Irkutsk to
Domodedovo, Moscow in 1994, failed to stop after engine startup and continued to operate at over
40,000 rpm with open bleed valves from engines, which caused an uncontained failure of the
starter. A detached turbine disk damaged fuel and oil supply lines (which caused fire) and hydraulic
lines. The fire-extinguishing system failed to stop the fire, and the plane diverted back to Irkutsk.
However, due to loss of hydraulic pressure the crew lost control of the plane, which subsequently
crashed into a dairy farm killing all 124 on board and one on the ground.
So,these are few accident where people lost their lives due to failure of aircraft engines.
So,taking these accident as an issue and in order to predict the remaining useful life and ensure
that all the parts of the engine are properly working in order to avoid loss of property and
money.We come up with an software model which will ensure that all the parts are working
properly and are functional with the help of sensors ,placed near the corresponding parts to ensure
that the parts are properly functional and showing correct values.So, in this way,Depending on the

values given by the sensors we can predict the useful life of aircraft engines and save lives and
money.
1.3 Existing system:
Because they are key components of aircraft, improving the safety, reliability and economy of
engines is crucial. To ensure flight safety and reduce the cost of maintenance during aircraft engine
operation, a prognostics and health management system that focuses on fault diagnosis, health
assessment, and life prediction is introduced to solve the problems. Predicting the remaining useful
life (RUL) is the most important information for making decisions about aircraft engine operation
and maintenance, and it relies largely on the selection of performance degradation features. The
choice of such features is highly significant, but there are some weaknesses in the current algorithm
for RUL prediction, notably, the inability to obtain tendencies from the data. Especially with
aircraft engines, extracting useful degradation features from multisensor data with complex
correlations is a key technical problem that has hindered the implementation of degradation
assessment. To solve these problems, deep learning has been proposed in recent years to exploit
multiple layers of nonlinear information processing for unsupervised self-learning of
features.Deep learning approach to predict the RUL of an aircraft engine based on the sensor data
and SGD algorithm. The data obtained from the sensors are given to the multilayer self-learning
and hence finding the RUL using SGD algorithm. However, the hyperparameters of the deep
learning, which significantly impact the feature extraction and prediction performance, are
determined based on expert experience in most cases.

CHAPTER 2
Literature Survey
2.1 A New Method Based on Stochastic Process Models for Machine Remaining Useful Life
Prediction
Remaining useful life (RUL) prediction is a key process in condition-based maintenance for
machines. It contributes to reducing risks and maintenance costs and increasing the
maintainability, availability, reliability, and productivity of machines. This paper proposes a new
method based on stochastic process models for machine RUL prediction. First, a new stochastic
process model is constructed considering the multiple variability sources of machine stochastic
degradation processes simultaneously. Then the Kalman particle filtering algorithm is used to
estimate the system states and predict the RUL. The effectiveness of the method is demonstrated
using simulated degradation processes and accelerated degradation tests of rolling element
bearings. Through comparisons with other methods, the proposed method presents its superiority
in describing the stochastic degradation processes and predicting the machine RUL. The limitation
is that the model parameters need to be initialized using large numbers of available measurements
that are hard to be acquired in real applications.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 65,
NO. 12, DECEMBER 2016
2.2 Remaining Useful Life Prediction for a Machine with Multiple Dependent Features Based
on Bayesian Dynamic Linear Model and Copulas
Degradation modeling and remaining useful life (RUL) prediction for products with multiple
degradation features are hot topics in the prognostic and health management. The key to this
problem is to describe the dependence among multiple degradation features effectively. In this
paper, a multivariate degradation modeling approach based on the Bayesian dynamic linear model
(BDLM) is proposed to calculate the RULs of degradation features, and the Copula function is
employed to capture the dependence among RUL distributions. A combined BDLM is used to
establish the multivariate degradation model, which includes two typical BDLMs, namely, the
linear growth model and seasonal factors model. After the model parameters get calibrated by the
maximum likelihood estimation, the model can predict the degradation process of features. Once

the failure thresholds are given, the probability density function and cumulative distribution
function (CDF) of RUL for each degradation feature can be obtained. Since these RUL
distributions are not independent of each other, the Copula function is adopted herein to couple
the CDFs. Finally, some practical testing data of a microwave component, which has two
degradation features, are utilized to validate our proposed method. This paper provides a new idea
for the multivariate degradation modeling and RUL prediction. The current study just selects two
typical BDLMs, linear growth model and seasonal factors model. In the future, more types of
BDLM should be investigated.
Date added to IEEE Access, August 4, 2017,
2.3 Battery remaining useful life prediction algorithm based on support vector regression
and unscented particle filter
Batteries are used in many areas, such as the electronic, aeronautics, astronautics, automobile and
energy, etc. However, there are many explosion and fire accidents caused by the battery aging, so
accurately estimating its remaining useful life (RUL) is very critical. In this paper, an improved
method is proposed by using support vector regression-unscented particle filter (SVR-UPF), which
increases the accuracy of the RUL prediction results. Firstly, an exponential model is adopted to
approximately express the degeneration of battery capacity. Secondly, a novel SVR-UPF method
is presented to solve the degeneracy phenomenon of the UPF algorithm, and then it is applied to
predict the battery RUL. Finally, some experiments and comparisons have been done to validate
the improved SVR-UPF prediction method. The results show that the proposed method is better
than the standard particle filter (PF) prediction method and the standard unscented particle filter
(UPF) prediction method. Future works include changing the regression algorithm in re-sampling
process and finding a better degeneration model to imitate the evolution of capacity.Date Added
to IEEE Xplore: 15 August 2016

CHAPTER 3
Methodology
3.1 Experiment flow diagram
This project uses the regression, classification and neural network approach of the machine
learning to predict the RUL for an engine and also compare the performance of the different
models. Keras sequential model has been used with SGD at the different learning rate (i.e. rate =
0.001, 0.002 etc.) to train the system.
Figure 3.1 Process flow diagram

Data collection:
We collect the data from legit sources , in this project we collect the data from NASA's Turbofan
Engine Degradation Simulation, wherein they simulated run to failure events from 100 different
engines. The data package comes with 4 datasets each for training and testing the model and one
set containing RUL values.
Data analysis is a process of inspecting, transforming and modeling data with the goal of
discovering useful information, informing conclusions and supporting decision-making. Data

analysis has multiple facets and approaches, encompassing diverse techniques under a variety of
names, and is used in different business, science, and social science domains. In today's business
world, data analysis plays a role in making decisions more scientific and helping businesses
operate more effectively.
Data cleaning is the process of detecting and correcting (or removing) corrupt or
inaccurate records from a record set, table, or database and refers to identifying incomplete,
incorrect, inaccurate or irrelevant parts of the data and then replacing, modifying, or deleting
the dirty or coarse data. Data cleansing may be performed interactively with data wrangling tools,
or as batch processing through scripting.
Data visualization is the graphic representation of data. It involves producing images that
communicate relationships among the represented data to viewers of the images. This
communication is achieved through the use of a systematic mapping between graphic marks and
data values in the creation of the visualization. This mapping establishes how data values will be
represented visually, determining how and to what extent a property of a graphic mark, such as
size or color, will change to reflect change in the value of a datum.
Feature engineering is the process of using domain knowledge of the data to create features that
make machine learning algorithms work. Feature engineering is fundamental to the application of
machine learning, and is both difficult and expensive. The need for manual feature engineering
can be obviated by automated feature learning.
Data splitting is a process where we split the data in two catogories for training and testing,
generally 80% of the data is used to train the model and remaining 20% is used to check the
behaviour of the model.
Machine learning (ML) is the scientific study of algorithms and statistical models that computer
systems use to perform a specific task without using explicit instructions, relying on patterns
and inference instead. It is seen as a subset of artificial intelligence. Machine learning algorithms

build a mathematical model based on sample data, known as "training data", in order to make
predictions or decisions without being explicitly programmed to perform the task.
Supervised learning algorithms build a mathematical model of a set of data that contains both the
inputs and the desired outputs. The data is known as training data, and consists of a set of training
examples. Each training example has one or more inputs and a desired output, also known as a
supervisory signal. In the mathematical model, each training example is represented by an array or
vector, sometimes called a feature vector, and the training data is represented by a matrix.
Unsupervised learning algorithms take a set of data that contains only inputs, and find structure
in the data, like grouping or clustering of data points. The algorithms therefore learn from test data
that has not been labeled, classified or categorized. Instead of responding to feedback,
unsupervised learning algorithms identify commonalities in the data and react based on the
presence or absence of such commonalities in each new piece of data. A central application of
unsupervised learning is in the field of density estimation in statistics, though unsupervised
learning encompasses other domains involving summarizing and explaining data features.

Fig 3.2 Supervised v/s Unsupervised learning
A neural network is a network or circuit of neurons, or in a modern sense, an artificial neural

network, composed of artificial neurons or nodes. Thus a neural network is either a biological
neural network, made up of real biological neurons, or an artificial neural network, for
solving artificial intelligence (AI) problems. The connections of the biological neuron are modeled
as weights. A positive weight reflects an excitatory connection, while negative values mean
inhibitory connections. All inputs are modified by a weight and summed. This activity is referred
as a linear combination. Finally, an activation function controls the amplitude of the output. For
example, an acceptable range of output is usually between 0 and 1, or it could be −1 and 1.
In machine learning, support-vector machines (SVMs, also support-vector networks[1])
are supervised learning models with associated learning algorithms that analyze data used
for classification and regression analysis. Given a set of training examples, each marked as
belonging to one or the other of two categories, an SVM training algorithm builds a model that
assigns new examples to one category or the other, making it a non-probabilistic binary linear

classifier (although methods such as Platt scaling exist to use SVM in a probabilistic classification
setting). An SVM model is a representation of the examples as points in space, mapped so that the
examples of the separate categories are divided by a clear gap that is as wide as possible. New
examples are then mapped into that same space and predicted to belong to a category based on the
side of the gap on which they fall.
The Model Comparison tool compares the performance of one or more different predictive models
based on the use of a validation, or test data set. It generates a report, a table of basic error
measurements, and a table of prediction results of each model. The tool supports all binary
classification, where the target variable has only two levels, such as "Yes" and "No", multinomial
classification, where the target variable has more than two levels, and regression (continuous target
variable) models.
Fig 3.3 Sample Network

In artificial neural method we check the model performance for the different activation function
(i.e. linear, rectifier, sigmodal etc.) in the below figure we easily see how activation function is
used in the network, without activation function neural network is just a linear regression model.

Here some example of the activation function:
Fig 3.4 activation functions to be used in the algorithm
In the context of artificial neural networks, the rectifier is an activation function defined as the
positive part of its argument: where x is the input to a neuron. This is also known as a ramp function
and is analogous to half-wave rectification in electrical engineering
The state of the art of non-linearity is to use rectified linear units (ReLU) instead of sigmoid
function in deep neural network. What are the advantages? I know that training a network
when ReLU is used would be faster, and it is more biological inspired.
The sigmoid function is a activation function in terms of underlying gate structured in co-
relation to Neurons firing, in Neural Networks. The derivative, also acts to be a activation
function in terms of handling Neuron activation in terms of NN's. The differential between the
two is activation degree and interplay.
The main reason why we use sigmoid function is because it exists between (0 to 1). Therefore, it
is especially used for models where we have to predict the probability as an output.Since
probability of anything exists only between the range of 0 and 1, sigmoid is the right choice.

Fig 3.5 Sigmoid function & rectifier function
3.2 Stochastic Gradient Descent (SGD)
Before talking about Stochastic Gradient Descent (SGD), let’s first understand what is Gradient
Descent? Gradient Descent is a very popular optimization technique in Machine Learning and
Deep Learning and it can be used with most, if not all, of the learning algorithms. A gradient is
basically the slope of a function; the degree of change of a parameter with the amount of change
in another parameter. Mathematically, it can be described as the partial derivatives of a set of
parameters with respect to its inputs. The more the gradient, the steeper the slope. Gradient
Descent is a convex function.
Gradient Descent can be described as an iterative method which is used to find the values of the
parameters of a function that minimizes the cost function as much as possible. The parameters
are initially defined a particular value and from that, Gradient Descent is run in an iterative
fashion to find the optimal values of the parameters, using calculus, to find the minimum
possible value of the given cost function.
Types of Gradient Descent:
Typically, there are three types of Gradient Descent:

1. Batch Gradient Descent

2. Stochastic Gradient Descent
3. Mini-batch Gradient Descent
The word ‘stochastic’means a system or a process that is linked with a random probability. Hence,
in Stochastic Gradient Descent, a few samples are selected randomly instead of the whole data set
for each iteration. In Gradient Descent, there is a term called “batch” which denotes the total
number of samples from a dataset that is used for calculating the gradient for each iteration. In
typical Gradient Descent optimization, like Batch Gradient Descent, the batch is taken to be the
whole dataset. Although, using the whole dataset is really useful for getting to the minima in a less
noisy or less random manner, but the problem arises when our datasets get really huge.
Suppose, you have a million samples in your dataset, so if you use a typical Gradient Descent
optimization technique, you will have to use all of the one million samples for completing one
iteration while performing the Gradient Descent, and it has to be done for every iteration until the
minima is reached. Hence, it becomes computationally very expensive to perform.
This problem is solved by Stochastic Gradient Descent. In SGD, it uses only a single sample, i.e.,
a batch size of one, to perform each iteration. The sample is randomly shuffled and selected for
performing the iteration.
SGD algorithm:
So, in SGD, we find out the gradient of the cost function of a single example at each iteration
instead of the sum of the gradient of the cost function of all the examples.
In SGD, since only one sample from the dataset is chosen at random for each iteration, the path
taken by the algorithm to reach the minima is usually noisier than your typical Gradient Descent
algorithm. But that doesn’t matter all that much because the path taken by the algorithm does not
matter, as long as we reach the minima and with significantly shorter training time.

Fig 3.6 Path taken by Batch Gradient Descent Fig 3.7 Path taken by Stochastic Gradient
Descent
One thing to be noted is that, as SGD is generally noisier than typical Gradient Descent, it usually
took a higher number of iterations to reach the minima, because of its randomness in its descent.
Even though it requires a higher number of iterations to reach the minima than typical Gradient
Descent, it is still computationally much less expensive than typical Gradient Descent. Hence, in
most scenarios, SGD is preferred over Batch Gradient Descent for optimizing a learning algorithm.

CHAPTER 4
PRESENT WORK
4.1 Flow Chart of the program
Fig 4.1 Flowchart of the working of the algorithm

4.2 System requirements

Software:
• Anaconda
• jupyter notebook
Module & packages used
• Sklearn
• Seaborn
• Pandas
• Numpy
• Keras
Anaconda
Anaconda is a free and open-source distribution of the Python and R programming languages
for scientific computing (data science, machine learning applications, large-scale data
processing, predictive analytics, etc.), that aims to simplify package management and deployment.
Package versions are managed by the package management system conda The Anaconda
distribution is used by over 15 million users and includes more than 1500 popular data-science
packages suitable for Windows, Linux, and MacOS. Anaconda distribution comes with more than
1,500 packages as well as the conda package and virtual environment manager. It also includes a
GUI, Anaconda Navigator[7], as a graphical alternative to the command line interface (CLI).
The big difference between conda and the pip package manager is in how package dependencies
are managed, which is a significant challenge for Python data science and the reason conda exists.
Open source packages can be individually installed from the Anaconda repository, Anaconda
Cloud (anaconda.org), or your own private repository or mirror, using the conda
install command. Anaconda Inc compiles and builds all the packages in the Anaconda
repository itself, and provides binaries for Windows 32/64 bit, Linux 64 bit and MacOS 64-bit.
Anything available on PyPI may be installed into a conda environment using pip, and conda will
keep track of what it has installed itself and what pip has installed.

Anaconda Navigator
Anaconda Navigator is a desktop graphical user interface (GUI) included in Anaconda distribution
that allows users to launch applications and manage conda packages, environments and channels
without using command-line commands. Navigator can search for packages on Anaconda Cloud
or in a local Anaconda Repository, install them in an environment, run the packages and update
them. It is available for Windows, macOS and Linux.
The following applications are available by default in Navigator:
• JupyterLab
• Jupyter Notebook
• QtConsole
• Spyder
• Glueviz
• Orange
• Rstudio
• Visual Studio Code
Jupyter Notebook
Jupyter Notebook (formerly IPython Notebooks) is a web-based interactive computational

environment for creating Jupyter notebook documents. The "notebook" term can colloquially
make reference to many different entities, mainly the Jupyter web application, Jupyter Python web
server, or Jupyter document format depending on context. A Jupyter Notebook document is
a JSON document, following a versioned schema, and containing an ordered list of input/output
cells which can contain code, text (using Markdown), mathematics, plots and rich media, usually
ending with the ".ipynb" extension.
A Jupyter Notebook can be converted to a number of open standard output formats

(HTML, presentation slides, LaTeX, PDF, ReStructuredText, Markdown, Python) through
"Download As" in the web interface, via the nbconvert library or "jupyter nbconvert" command

line interface in a shell. To simplify visualisation of Jupyter notebook documents on the web, the
nbconvert library is provided as a service through NbViewer which can take a URL to any publicly
available notebook document, convert it to HTML on the fly and display it to the user.
Scikit-learn
Scikit-learn (formerly scikits.learn and also known as sklearn) is a free software machine
learning library for the Python programming language. It features
various classification, regression and clustering algorithms including support vector
machines, random forests, gradient boosting, k-means and DBSCAN, and is designed to
interoperate with the Python numerical and scientific libraries NumPy and SciPy. Scikit-learn is
largely written in Python, and uses numpy extensively for high-performance linear algebra and
array operations. Furthermore, some core algorithms are written in Cython to improve
performance. Support vector machines are implemented by a Cython wrapper around LIBSVM;
logistic regression and linear support vector machines by a similar wrapper around LIBLINEAR.
In such cases, extending these methods with Python may not be possible.
Scikit-learn integrates will with many other Python libraries, such as matplotlib and plotly for
plotting, numpy for array vectorization, pandas dataframes, scipy, and many more.
NumPy
NumPy is a library for the Python programming language, adding support for large, multi-
dimensional arrays and matrices, along with a large collection of high-
level mathematical functions to operate on these arrays. The ancestor of NumPy, Numeric, was
originally created by Jim Hugunin with contributions from several other developers. In
2005, Travis Oliphant created NumPy by incorporating features of the competing Numarray into
Numeric, with extensive modifications. NumPy is open-source software and has many
contributors.
Using NumPy in Python gives functionality comparable to MATLAB since they are both
interpreted, and they both allow the user to write fast programs as long as most operations work

on arrays or matrices instead of scalars. In comparison, MATLAB boasts a large number of

additional toolboxes, notably Simulink, whereas NumPy is intrinsically integrated with Python, a
more modern and complete programming language. Moreover, complementary Python packages
are available; SciPy is a library that adds more MATLAB-like functionality and Matplotlib is a
plotting package that provides MATLAB-like plotting functionality. Internally, both MATLAB
and NumPy rely on BLAS and LAPACK for efficient linear algebra computations.
Pandas
Pandas is a popular Python library for data analysis. It is not directly related to Machine Learning.
As we know that the dataset must be prepared before training. In this case, Pandas comes handy
as it was developed specifically for data extraction and preparation. It provides high-level data
structures and wide variety tools for data analysis. It provides many inbuilt methods for groping,
combining and filtering data.
Pandas is mainly used for machine learning in form of data frames. Pandas allow importing data
of various file formats such as csv, excel etc. Pandas allows various data manipulation operations
such as group by, join, merge, melt, concatenation as well as data cleaning features such as filling,
replacing or imputing null values.
Matplotlib
Matpoltlib is a very popular Python library for data visualization. Like Pandas, it is not directly
related to Machine Learning. It particularly comes in handy when a programmer wants to visualize
the patterns in the data. It is a 2D plotting library used for creating 2D graphs and plots. A module
named pyplot makes it easy for programmers for plotting as it provides features to control line
styles, font properties, formatting axes, etc. It provides various kinds of graphs and plots for data
visualization, viz., histogram, error charts, bar chats, etc
Matplotlib 2.0.x supports Python versions 2.7 through 3.6. Python3 support started with Matplotlib
1.2. Matplotlib 1.4 is the last version to support Python 2.6. Matplotlib has pledged to not support
Python 2 past 2020 by signing the Python 3 Statement.

Seaborn
Seaborn is a library for making statistical graphics in Python. It is built on top of matplotlib and
closely integrated with pandas data structures.
Here is some of the functionality that seaborn offers:
• A dataset-oriented API for examining relationships between multiple variables

• Specialized support for using categorical variables to show observations or aggregate
statistics
• Options for visualizing univariate or bivariate distributions and for comparing them
between subsets of data
• Automatic estimation and plotting of linear regression models for different
kinds dependent variables
• Convenient views onto the overall structure of complex datasets
• High-level abstractions for structuring multi-plot grids that let you easily
build complex visualizations
• Concise control over matplotlib figure styling with several built-in themes
• Tools for choosing color palettes that faithfully reveal patterns in your data
Seaborn aims to make visualization a central part of exploring and understanding data. Its
dataset-oriented plotting functions operate on dataframes and arrays containing whole datasets
and internally perform the necessary semantic mapping and statistical aggregation to produce
informative plots.
Keras
Keras is an open-source neural-network library written in Python. It is capable of running on top
of TensorFlow, Microsoft Cognitive Toolkit, R, Theano, or PlaidML. Designed to enable fast
experimentation with deep neural networks, it focuses on being user-friendly, modular, and
extensible. It was developed as part of the research effort of project ONEIROS (Open-ended
Neuro-Electronic Intelligent Robot Operating System) and its primary author and maintainer is

François Chollet, a Google engineer. Chollet also is the author of the XCeption deep neural
network model.
Keras contains numerous implementations of commonly used neural-network building blocks such
as layers, objectives, activation functions, optimizers, and a host of tools to make working with
image and text data easier to simplify the coding necessary for writing Deep Neural Network code.
The code is hosted on GitHub, and community support forums include the GitHub issues page,
and a Slack channel. In addition to standard neural networks, Keras has support
for convolutional and recurrent neural networks. It supports other common utility layers like
dropout, batch normalization, and pooling.
Keras allows users to productize deep models on smartphones (iOS and Android), on the web, or
on the Java Virtual Machine. It also allows use of distributed training of deep-learning models on
clusters of Graphics Processing Units (GPU) and Tensor processing units (TPU) principally in
conjunction with CUDA.
Data Analyzing
Data sets consists of multiple multivariate time series, there are 26 attributes of the engine which
are unit id, time in cycle, operational setting and sensors. There are 3 operational setting and 21
sensors in the dataset. We drop the sensors which measurements were constant and Nan. There are
certain technical issues with the dataset.
• The training data does not include RUL as a target variable.
• The only provided RUL is for the last cycle of each engine.
• A lot of the sensor measurements were constants.
• In other words, the training data is unlabeled, the testing data is partially labelled.

Table 4.1 Dataset description that would be used in program
Index Data Fields Type Description
1 Unit_id integer Identifier range[1,260]
2 time_in_cycle integer Time in cycles
3 op_setting1 float Sensors Noise
4 Op_setting2 float Sensors Noise
5 Op_setting3 float Sensors Noise
6 Sensor_01 float Total temperature at fan Inlet
7 Sensor_02 float Total temperature at LPC Inlet
… …. …. ….
26 Sensor_21 float HPT coolant bleed
Data Visualization
By visualizing we quickly combining different sources of data, improving productivity,
accelerating time value, eliminating latency, maintaining data warehouse, and reducing the need
for multiple copies of data as well as less hardware. In this process we focus on creating virtual
structures for big data systems the benefit from big data virtualization is because it enables them
to use all the data assets, they collect to achieve various goals and objectives.
Engine unit_id = 1
Visual analysis of the engine of unit_id 1. In this figure we visualize the necessary details of
sensors of the aircraft engine.

Plot is always a good idea. In this way we can have an impressive and general overview of the data
at our disposal. At the end of the majority of the series we can observe a divergent behavior, which
announces a future failure.
Fig 4.2 Data visualization of the various sensor outputs

In the fig 4.3 we plot the data of engine having unit id 1, when we visualize the plot, we easily
see that some sensors having constant value and some sensors did not varies thereby we remove
that sensors and select remaining.

Fig 4.3 plot of all the sensor data

4.3 Sensors used in the prediction of RUL of the aircraft engine and threshold
setting
Table 4.2 Details of vaious sensor components and their units
Sensor No. Symbol Description Units

1 T2 Total temperature at fan inlet °R
2 T24 Total temperature at LPC outlet °R
3 T30 Total temperature at HPC outlet °R
4 T50 Total temperature at LPT outlet °R
5 P2 Pressure at fan inlet psia
6 P15 Total pressure in bypass-duct psia
7 P30 Total pressure at HPC outlet psia
8 Nf Physical fan speed rpm
9 Nc Physical core speed rpm
10 Epr Engine Pressure ratio −
11 Ps30 Static pressure at HPC outlet psia
12 Phi Ratio of fuel flow to Ps30 pps/psi
13 NRf Corrected fan speed rpm
14 fNR Corrected core speed rpm
15 BPRht Bypass ratio −
16 carBN Burner fuel–air ratio −
17 BleedP Bleed enthalpy −
18 f-dmd Demanded fan speed rpm
19 CNfR-dmd Demanded corrected fan speed rpm
20 W31 HPT coolant bleed lbm/s
21 W32 LPT coolant bleed lbm/s

Units of the sensor outputs

°R, rankine temperature scale,
psia, pounds per square inch absolute,
rpm, revolutions per minute,
pps, pulse per second;
psi, pounds per square inch,
lbm/s, pound mass per second.
Threshold values of the Sensor values and RUL for optimal working
The Threshold values for the optimal working of the components is obtained by running the
component until the end of the RUL, and by performing the statistical analysis of the obtained data
we can obtain various Parameters like mean, standard deviations, variance etc.
Examples: Sensor 2 indicates Total Temperature at LPC outlet. By analysing the Data obtained
Turbofan Engine Degradation Simulation, wherein they used the simulated run to failure events
from 100 different engines, we see that the mean value of the Sensor 2 is around 579.67 oR.,
with a standard deviation of 37.28 oR. So we can set the optimal working range of the sensor
between 579.67 +37.28 oR and 579.67-37.28 oR. A suitable tolerance would be provided for the
practical working of the component. Any value beyond this range would indicate faulty working
of the component and would raise and alert even though the RUL is in a safe range.
Similarly, Sensor 15 indicates the Bypass Ratio value with a mean value of 9.32965 with a standard
Deviation of 0.74934. SO the optimal working range is between 9.32965 ± 0.74934.
The RUL values can be seen depending on the outputs of various sensors. The RUL prediction is
a continuous process and thus we can get the RUL values at regular intervals. Even though the
Maximum life is the RULmax-1, nearly 75% of RUL is considered safe. The remaining 25% of the
can be considered as reserve RUL, and may not be recommended for a lengthy usage. The
threshold values are specific to the kind of turbo engine subject to experiment. For different kinds
of engine, the algorithm works well only with a slight modification in the parameters.

CHAPTER 5
RESULTS & DISCUSSION
In this section we discuss about the results that we obtained by applying the different machine
learning model such as random forest and neural network and also discuss about the different
methods for feature selection like chi2, Hypothesis testing.
5.1 Ordinary Least Squares or Linear Regression
Fig 5.1 Hypothesis Testing & OLS

OLS is the method used to find the simple linear regression of a set of data. In statistics, ordinary
least squares are a type of linear least squares method for estimating the unknown parameters in a
linear regression model. The general linear statistical model can be described in matrix notation as
𝑦 = 𝑚𝑥 + 𝑐 .
Hypothesis testing is an act in statistics whereby an analyst tests an assumption regarding a
population parameter. The methodology employed by the analyst depends on the nature of the data
used and the reason for the analysis. Hypothesis testing is used to assess the plausibility of a
hypothesis by using sample data A statistical hypothesis is a hypothesis that is testable on the basis
of observing a process that is modeled via a set of random variables.
Fig 5.1 showing the result performed by ordinary least squares and hypothesis testing, in the above
section of the figure R-Squared and Adj. R-Squared showing the performance of the regression
model, 4th column of middle section showing the T-test and 5th column shows the comparison
between P-value and T-value.
Fig 5.2 Chi Square Values

A chi-squared test, also written as χ² test, is a statistical hypothesis test that is valid to perform
when the test statistic is chi-squared distributed under the null hypothesis. Before training the
model we first know about the importance of the feature such as which feature is more useful.
There is different feature selection technique which is used for the feature selection like select k-
best hypothesis testing and chi-square test. Fig 5.2 showing the chi2 test scores of the different
features used in this project, the scores are in descending order the variables which is more
important at top and so on.
5.2 Support Vector Regression
Support Vector Machine - Regression (SVR) Support Vector Machine can also be used as a
regression method, maintaining all the main features that characterize the algorithm (maximal
margin). The Support Vector Regression (SVR) uses the same principles as the SVM for
classification, with only a few minor differences also In machine learning, support-vector
machines are supervised learning models with associated learning algorithms that analyze data
used for classification and regression analysis.
Fig 5.3 SVR Result

Before describing the plot first, we see value of R-square which is almost 1 that implies that the
accuracy of the model is 100% and it is not good for the model, now we observe the fig 5.3 in
which the x(Actual value of X_test i.e y_test) and y(predicted value of X_test) variables are fitted
linearly and this is the perfect case of Overfitting(high bias and a low variance). So, we did not
use this model for further prediction or operation.
5.3 Random Forest
Random Forest is a flexible, easy to use machine learning algorithm that produces great results
most of the time with minimum time spent on hyper-parameter tuning. It has gained popularity
due to its simplicity and the fact that it can be used for both classification and regression tasks. It
also increases predictive power of the algorithm and also helps prevent overfitting.
Fig 5.4 Random Forest Result
Fig 5.4 shows the relation between the x (Actual value of X_test i.e y_test) and y (predicted value
of X_test). The R-squared value is .55 and mean square value is 2123.71. In Regression model the
accuracy of the model is determined by their R-square value greater the R-square value higher is
the accuracy and lower the mean square error better is the result. By viewing the result of random
forest, we did not satisfy because the accuracy is not good as we want.

5.4 Neural Network
A neural network is a series of algorithms that endeavors to recognize underlying relationships in

a set of data through a process that mimics the way the human brain operates, it is used for solving
many business problems such as sales forecasting, customer research, data validation, and risk
management. For example, at Statsbot we apply neural networks for time-series predictions,
anomaly detection in data, and natural language understanding.
The sequential model (also known as the Keras model) is a theory that describes cooperativity of
protein subunits. It postulates that a protein's conformation changes with each binding of a ligand,
thus sequentially changing its affinity for the ligand at neighboring binding sites. In this sequential
model we use the different number of epochs (200, 300 and 2000) and batch size (100, 200 and
900) and checks the performance of the model. In the model two activation function Sigmoidal
and Rectifier function is used, the learning rate is 0.001 and the optimizer is Stochastic Gradient
Descent (SGD).
Fig 5.5 Loss and Error

In statistics, mean absolute error is a measure of errors between paired observations expressing
the same phenomenon. Examples of Y versus X include comparisons of predicted versus observed,

subsequent time versus initial time, and one technique of measurement versus an alternative
technique of measurement. In statistics, the mean squared error (MSE) or mean squared
deviation (MSD) of an estimator (of a procedure for estimating an unobserved quantity) measures
the average of the squares of the errors—that is, the average squared difference between the
estimated values and the actual value.
Fig 5.5 showing the Error (mean absolute error) and Loss (mean square error) at each iterations or
epoch. Viewing of both figures we observe that with increment of number of epochs the mean
absolute error and mean square error is reduces this implies that the losses and error is very low in
this model which increases the performance of the model.
Fig 5.6 Comparison between true and predicted RUL
Fig 5.6 showing the difference between predicted and true remaining useful life, here the x-axis
(i.e unit_id of the engines) is common for both True and predicted RUL. By viewing this plot, we
say that the performance of the model is good and it is observed that the accuracy of the model is
very good if the remaining useful life of any engine exceeds the value of 200 life cycles.

True Predicted Difference

RUL RUL
259 [270.0] [11.0]
258 [205.0] [-53.0]
257 [203.0] [-54.0]
256 [190.0] [-66.0]
255 [190.0] [-65.0]
254 [198.0] [-56.0]
253 [186.0] [-67.0]
252 [227.0] [-25.0]
251 [195.0] [-56.0]
250 [192.0] [-58.0]
Fig 5.7 Final Result
Fig showing the result of our model here we store the data into the csv or excel file in which there
are 3 column first column is the True value of the RUL 2nd column is the Predicted value of the
RUL and 3rd column is the difference between the Predicted value and True value.
By comparing all the models that we used in this project we confidently say that neural network
model’s performance is better than the other models.

CHAPTER 6
CONCLUSION
In this work, RUL prognosis model for turbofan is proposed, using the Turbofan Engine
Degradation Simulation Data Set by NASA. Although the use of dynamic networks such as LSTM
and RNN is preferred for problems with time series, these models are more expensive
computationally in embedded system applications. Thus, in this work the use of MLP is adopted
due to their low computational cost. So, in this way we with the help of data values given by
different sensors that are discussed above we give these data to the algorithm i.e. (SGD) and hence
the RUL of the engine is found. So, in this way the RUL prediction of an aircraft engine can not
only improve its safety, maintenance, and availability, but also reduce its operation and
maintenance costs.
This also enables the airline companies to reduce the losses occurred from the regular maintenance
and other indirect effects due to the unpredictable breakdown of engines.
This work has some limitations, which present potential areas for further research. Arrival and
international flights information were not used to predict delays due to a lack of reliable relevant
data. Further research can be then conducted to enhance the accuracy of the proposed model by
using both the domestic and international flights information.

References
[1]. Saxena, Abhinav, et al. "Damage propagation modeling for aircraft engine run-to-failure
simulation." 2008 international conference on prognostics and health management. IEEE, 2008.
[2]. Lei, Yaguo, Naipeng Li, and Jing Lin. "A new method based on stochastic process models
for machine remaining useful life prediction." IEEE Transactions on Instrumentation and
Measurement 65.12 (2016): 2671-2684.
[3]. Zhang, Heng, et al. "A Remaining Useful Life Prediction Framework for Multi-sensor
System." 2019 IEEE 19th International Conference on Software Quality, Reliability and Security
Companion (QRS-C). IEEE, 2019.
[4]. Xia, Min, et al. "A two-stage approach for the remaining useful life prediction of bearings
using deep neural networks." IEEE Transactions on Industrial Informatics 15.6 (2018): 3703-
3711.
[5]. Ramasso, Emmanuel, and Abhinav Saxena. "Review and analysis of algorithmic approaches
developed for prognostics on CMAPSS dataset." 2014.
[6]. Vieira, Darli Rodrigues, and Paula Lavorato Loures. "Maintenance, repair and overhaul
(MRO) fundamentals and strategies: An aeronautical industry overview." International Journal
of Computer Applications 135.12 (2016): 21-29.

Project phaseII Report

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Project phaseII Report

Uploaded by

Copyright:

Available Formats

VISVESVARAYA TECHNOLOGICAL UNIVERSITY

Jnana Sangama, Belagavi-590018, Karnataka

Under the guidance of

B.N.M. Institute of Technology

Department of Electronics & Communication Engineering

Department of Electronics and Communication Engineering

Mr. Sudarshan Dr. P.A. Vijaya Dr. Krishnamurthy. G.N

Dept. Of ECE, BNMIT 1 2019-20

Dept. Of ECE, BNMIT 2 2019-20

Dept. Of ECE, BNMIT 3 2019-20

Dept. Of ECE, BNMIT 4 2019-20

Dept. Of ECE, BNMIT 5 2019-20

Dept. Of ECE, BNMIT 6 2019-20

Figure 3.1 Process flow diagram

Dept. Of ECE, BNMIT 7 2019-20

Dept. Of ECE, BNMIT 8 2019-20

Dept. Of ECE, BNMIT 9 2019-20

Fig 3.2 Supervised v/s Unsupervised learning

A neural network is a network or circuit of neurons, or in a modern sense, an artificial neural

Dept. Of ECE, BNMIT 10 2019-20

Fig 3.3 Sample Network

Dept. Of ECE, BNMIT 11 2019-20

Here some example of the activation function:

Fig 3.4 activation functions to be used in the algorithm

Dept. Of ECE, BNMIT 12 2019-20

Fig 3.5 Sigmoid function & rectifier function

3.2 Stochastic Gradient Descent (SGD)

Types of Gradient Descent:

Typically, there are three types of Gradient Descent:

Dept. Of ECE, BNMIT 13 2019-20

1. Batch Gradient Descent

Dept. Of ECE, BNMIT 14 2019-20

Dept. Of ECE, BNMIT 15 2019-20

Fig 4.1 Flowchart of the working of the algorithm

Dept. Of ECE, BNMIT 16 2019-20

4.2 System requirements

Dept. Of ECE, BNMIT 17 2019-20

The following applications are available by default in Navigator:

Jupyter Notebook (formerly IPython Notebooks) is a web-based interactive computational

A Jupyter Notebook can be converted to a number of open standard output formats

Dept. Of ECE, BNMIT 18 2019-20

Dept. Of ECE, BNMIT 19 2019-20

on arrays or matrices instead of scalars. In comparison, MATLAB boasts a large number of

Dept. Of ECE, BNMIT 20 2019-20

Here is some of the functionality that seaborn offers:

• A dataset-oriented API for examining relationships between multiple variables

Dept. Of ECE, BNMIT 21 2019-20

Dept. Of ECE, BNMIT 22 2019-20

Table 4.1 Dataset description that would be used in program

Index Data Fields Type Description

1 Unit_id integer Identifier range[1,260]

2 time_in_cycle integer Time in cycles

3 op_setting1 float Sensors Noise

4 Op_setting2 float Sensors Noise

5 Op_setting3 float Sensors Noise

6 Sensor_01 float Total temperature at fan Inlet

7 Sensor_02 float Total temperature at LPC Inlet

26 Sensor_21 float HPT coolant bleed

Dept. Of ECE, BNMIT 23 2019-20

Fig 4.2 Data visualization of the various sensor outputs

Dept. Of ECE, BNMIT 24 2019-20

Fig 4.3 plot of all the sensor data