Machine learning Internship Projects by UCT

Deliverables for each Project:

Code: on Github

Project Report: also on Github, mentioning about company (UCT), Background of the project,
Problem statement relevance, Design, Implementation details, Result, your learnings.

Machine learning Internship project scope varies from current scenario, industrial demand, as well
as projects running in our company. So we divide them in below areas:

A. Agriculture
B. Predictive maintenance
C. Smart city
D. Industrial Manufacturing and Production

Project 1: Obsolete
Project 2: Obsolete
Project 3: Obsolete

A. Agriculture
1. Project 4: Prediction of Agriculture Crop Production in India

Context
Agricuture Production in India from 2001-2014

Content
This Dataset Describes the Agricuture Crops Cultivation/Production in india. This is
from https://data.gov.in/ fully Licensed

Acknowledgements
This Dataset can solves the problems of various crops Cultivation/production in india.

Columns
 Crop: string, crop name
Variety:string,crop subsidary name
state: string,Crops Cultivation/production Place
Quantity:Integer,no of Quintals/Hectars
production:Integer,no of years Production
Season:DateTime,medium(no of days),long(no of days)
Unit:String , Tons
Cost:Integer, cost of cutivation and Production
Recommended Zone:String ,place(State,Mandal,Village)

Inspiration
Across The Globe India Is The Second Largest Country having People more than 1.3 Billion.
Many People Are Dependent On The Agricuture And it is the Main Resource.
In Agricuturce Cultivation/Production Having More Problems.
I want to solve the Big problem in india and usefull to many more people

Data set Link:

https://drive.google.com/file/d/1zfqvs8-mAO6E0JpgvhBdueNx8Th03pUp/view?usp=sharing

2. Project 5: Crop and weed detection

Content
This dataset contains 1300 images of sesame crops and different types of weeds with each image
labels.
Each image is a 512 X 512 color image. Labels for images are in YOLO format.

Data Preparation
STEPS:
1. First of we have to collect dataset for it.For that we have to capture photos of weeds and crops. We
collected total 589 images
2. After collection of photos we have to clean the dataset. This step is very important because if any
bed photo is remain in dataset it causes worse effect in detection model. After cleaning we have 546
images.
3. Now time for image processing. Our photo size is 4000X3000 color which is very large and model
will take very long time for training so we convert all images to 512X512X3 size.
4. Now 546 image is not enough for training, so we have done some magic to convert 546 image into
1300 images. We used Data Augmentation technique to increase dataset.(Check it out keras
ImageDataGenerator on google)
5. This step is very tedious, Manual labeling of image data!! In this step we have to drow bounding
boxes on photos whether it weed or crop.

Problem
Weed is an unwanted thing in agriculture. Weed use the nutrients, water, land and many more things
that might have gone to crops. Which results in less production of the required crop. The farmer
often uses pesticides to remove weed which is also effective but some pesticides may stick with crop
and may causes problems for humans.

Aim
We aim to develop a system that only sprays pesticides on weed and not on the crop Which will
reduce the mixing problem with crops and also reduce the waste of pesticides.

Data set Link:

https://drive.google.com/file/d/1MNdDKYB0x0PEW7P71bE1Jx_uLllvORA0/view?usp=sharing

B. Predictive maintenance

Project 6: Predict the number of remaining operational cycles before

failure for Turbofan engine

Experimental Scenario
Data sets consists of multiple multivariate time series. Each data set is further divided into
training and test subsets. Each time series is from a different engine – i.e., the data can be
considered to be from a fleet of engines of the same type. Each engine starts with different
degrees of initial wear and manufacturing variation which is unknown to the user. This wear
and variation is considered normal, i.e., it is not considered a fault condition. There are three
operational settings that have a substantial effect on engine performance. These settings are
also included in the data. The data is contaminated with sensor noise.
The engine is operating normally at the start of each time series, and develops a fault at some
point during the series. In the training set, the fault grows in magnitude until system failure. In
the test set, the time series ends some time prior to system failure. The objective of the
competition is to predict the number of remaining operational cycles before failure in the test
set, i.e., the number of operational cycles after the last cycle that the engine will continue to
operate. Also provided a vector of true Remaining Useful Life (RUL) values for the test data.
The data are provided as a zip-compressed text file with 26 columns of numbers, separated by
spaces. Each row is a snapshot of data taken during a single operational cycle, each column is a
different variable. The columns correspond to:
1) unit number
2) time, in cycles
3) operational setting 1
4) operational setting 2
5) operational setting 3
6) sensor measurement 1
7) sensor measurement 2
...
26) sensor measurement 26

Data Set: FD001

Train trajectories: 100
Test trajectories: 100
Conditions: ONE (Sea Level)
Fault Modes: ONE (HPC Degradation)

Data Set: FD002

Train trajectories: 260
Test trajectories: 259
Conditions: SIX
Fault Modes: ONE (HPC Degradation)

Data Set: FD003

Train trajectories: 100
Test trajectories: 100
Conditions: ONE (Sea Level)
Fault Modes: TWO (HPC Degradation, Fan Degradation)

Data Set: FD004

Train trajectories: 248
Test trajectories: 249
Conditions: SIX
Fault Modes: TWO (HPC Degradation, Fan Degradation)

Data set Link:

https://drive.google.com/file/d/1dgWM0KKOnoN9kVObbA-GahsgXPJBCT4c/view?usp=sharing

Project 7: Predict life time of a bearing in manufacturing industry

Data Structure
Three (3) data sets are included in the data packet (IMS-Rexnord Bearing Data.zip). Each data set
describes a test-to-failure experiment. Each data set consists of individual files that are 1-second
vibration signal snapshots recorded at specific intervals. Each file consists of 20,480 points with the
sampling rate set at 20 kHz. The file name indicates when the data was collected. Each record (row) in
the data file is a data point. Data collection was facilitated by NI DAQ Card 6062E. Larger intervals of
time stamps (showed in file names) indicate resumption of the experiment in the next working day.

Set No. 1:
Recording Duration: October 22, 2003 12:06:24 to November 25, 2003 23:39:56
No. of Files: 2,156
No. of Channels: 8
Channel Arrangement: Bearing 1 – Ch 1&2; Bearing 2 – Ch 3&4;
Bearing 3 – Ch 5&6; Bearing 4 – Ch 7&8.
File Recording Interval: Every 10 minutes (except the first 43 files were taken every 5 minutes)
File Format: ASCII
Description: At the end of the test-to-failure experiment, inner race defect occurred in bearing 3 and
roller element defect in bearing 4.

Set No. 2:
Recording Duration: February 12, 2004 10:32:39 to February 19, 2004 06:22:39
No. of Files: 984
No. of Channels: 4
Channel Arrangement: Bearing 1 – Ch 1; Bearing2 – Ch 2; Bearing3 – Ch3; Bearing 4 – Ch 4.
File Recording Interval: Every 10 minutes
File Format: ASCII
Description: At the end of the test-to-failure experiment, outer race failure occurred in bearing 1.

Set No. 3 :
Recording Duration: March 4, 2004 09:27:46 to April 4, 2004 19:01:57
No. of Files: 4,448
No. of Channels: 4
Channel Arrangement: Bearing1 – Ch 1; Bearing2 – Ch 2; Bearing3 – Ch3; Bearing4 – Ch4;
File Recording Interval: Every 10 minutes
File Format: ASCII
Description: At the end of the test-to-failure experiment, outer race failure occurred in bearing 3

Data set Link:

https://drive.google.com/file/d/12rV9AhpqbMivYCu4WVM7DhXpO1aO98k_/view?usp=sharing

Project 8: Predictive maintenance of Gearbox using vibration sensors

Predictive maintenance allows manufacturers to lower maintenance costs, extend equipment
life, reduce downtime and improve production quality by addressing problems before they cause
equipment failures.
Gearbox Fault Diagnosis Data set include the vibration dataset recorded by using
SpectraQuest’s Gearbox Fault Diagnostics Simulator. Dataset has been recorded with
the help of 4 vibration sensors placed in four different direction. Data set has been
recorded under variation of load from '0' to '90' percent. Data set has been recorded in
two different scenario:
1) Healthy condition and
2) Broken Tooth Condition

Data set Link:

https://drive.google.com/file/d/1nNNnjMPntlo5X0t_cif7cmlJhikCyWyP/view?usp=sharing

C. Smart city
Project 9: Forecasting of Smart city traffic patterns
We are working with the government to transform various cities into a smart city. The vision is to
convert it into a digital and intelligent city to improve the efficiency of services for the citizens. One of
the problems faced by the government is traffic. You are a data scientist working to manage the
traffic of the city better and to provide input on infrastructure planning for the future.
The government wants to implement a robust traffic system for the city by being prepared for traffic
peaks. They want to understand the traffic patterns of the four junctions of the city. Traffic patterns
on holidays, as well as on various other occasions during the year, differ from normal working days.
This is important to take into account for your forecasting.

Data set Link:

https://drive.google.com/file/d/1y61cDyuO9Zrp1fSchWcAmCxk0B6SMx7X/view?usp=sharing
 Industrial Manufacturing and Production
Project 10: Quality Prediction in a Mining Process
Explore real industrial data and help manufacturing plants to be more efficient

Context
It is not always easy to find databases from real world manufacturing plants, specially mining plants.
This database comes from one of the most important parts of a mining process: a flotation plant.
The main goal is to use this data to predict how much impurity is in the ore concentrate. As this
impurity is measured every hour, if we can predict how much silica (impurity) is in the ore
concentrate, we can help the engineers, giving them early information to take actions (empowering!).
Hence, they will be able to take corrective actions in advance (reduce impurity, if it is the case) and
also help the environment (reducing the amount of ore that goes to tailings as you reduce silica in
the ore concentrate).

Content
The first column shows time and date range (from march of 2017 until september of 2017). Some
columns were sampled every 20 second. Others were sampled on a hourly base.
The second and third columns are quality measures of the iron ore pulp right before it is fed into the
flotation plant. Column 4 until column 8 are the most important variables that impact in the ore
quality in the end of the process. From column 9 until column 22, we can see process data (level
and air flow inside the flotation columns, which also impact in ore quality. The last two columns are
the final iron ore pulp quality measurement from the lab.
Target is to predict the last column, which is the % of silica in the iron ore concentrate.

Expected submission
 Is it possible to predict % Silica Concentrate every minute?
 How many steps (hours) ahead can we predict % Silica in Concentrate? This would help engineers
to act in predictive and optimized way, mitigating the % of iron that could have gone to tailings.
 Is it possible to predict % Silica in Concentrate without using % Iron Concentrate column (as they
are highly correlated)?

Dataset
This dataset is about a flotation plant which is a process used to concentrate the iron ore. This
process is very common in a mining plant.

Data set Link:

https://drive.google.com/file/d/1N80d8eTDAf1JMQXGQbHDAUaMGRyA8QG3/view?usp=sharing
 Project 11: Multi-stage continuous-flow manufacturing process
Real process data to predict factory output

Context
This data was taken from an actual production run spanning several hours. The goal is to predict
certain properties of the line's output from the various input data. The line is a high-speed,
continuous manufacturing process with parallel and series stages.

Expected submission
We are always looking for the best predictive modeling approaches to use in real time production
environments. Models are employed for several use cases such as development of real time
process controllers (use the models in simulation environments) and anomaly detection (compare
model predictions to actual outputs in real time).

Dataset
The data comes from a continuous flow manufacturing process with multiple stages. Sample rates
are 1 Hz.

 In the first stage, Machines 1, 2, and 3 operate in parallel, and feed their outputs into a step that
combines the flows.
 Output from the combiner is measured in 15 locations. These measurements are the primary
measurements to predict.
 Next, the output flows into a second stage, where Machines 4 and 5 process in series.
 Measurements are made again in the same 15 locations. These are the secondary measurements to
predict.
Data set Link:
https://drive.google.com/file/d/1yvZzslpbWw2mpCVF5QqueSkNrNHmtvDE/view?usp=share_link