/Smart-Industrial-Predictive-Solutions

Smart-Industrial-Predictive-Solutions Feynn Labs Internship project

Primary LanguageHTML

Smart Industrial Predictive Solutions - Feynn Labs Internship 3rd project

Python Machine Learning Data Analysis Pandas NumPy scikit-learn Jupyter Notebook Windows Terminal Shell Script Flask Docker

Introduction

This is Feynn Labs Internship's 3rd project. Here, we have to choose one best project of all our 1st projects.

Then need to do:

  • Data collection
  • Data analysis
  • Machine learning
  • Creating business model
  • Creating finantial equation with the help of data analysis and machine learning.
  • If possible, we have to create app.

We selected Machine Predictive Maintenance for Industrial Machines project

Web App: Link1 (render.com) and Link2 (GCP)

About

Smart Industrial Predictive Solutions is a forward-looking project with a big goal: to make industries smarter and more efficient. It does this by using fancy technology like data analysis and machine learning to predict when machines might break down. Imagine having a crystal ball that tells you when your car is going to have engine trouble before it even happens. That's what this project aims to do for factories and big machines in industries. By doing this, it helps companies save money, work more safely, and keep everything running smoothly. It's like having a super-smart maintenance team that never misses a beat!

Docker

Docker image

Team members

Name GitHub link
Adhiban Siddarth Me Team Lead GitHub Link
Karakavalasa venkata pranay GitHub Link
Malay Vyas GitHub Link
Shreyash Banduji Chacharkar GitHub Link
Yash Mayur GitHub Link

Workflow

  • Getting details of 1st project of each member
  • Selecting one project from our 1st project
  • Or selecting new best project topic other than our projects
  • Role selection (Business Specialists, Data Scientists and Software Developer)
  • Building business model, creating financial equation, collecting data, doing data analysis & Machine Learning
  • Developing web or desktop or mobile app (optional)

Team members' 1st projects

Name Project Title Link
Adhiban Siddarth Machine Predictive Maintenance for Industrial Machines Github Link
Karakavalasa venkata pranay Personal Medical Assistant PDF Link
Malay Vyas University Chances Predictor PDF Link
Shreyash Banduji Chacharkar Clothing Recommendation System Based on Body Data PDF Link
Yash Mayur College Recommender GitHub Link

Project Evaluation Matrix for Feasibility, Viability, and Monetization

  • Feasibility: Product/Service can be developed in short term future. (2-3 years)
  • Viability: Product/Service should be relevant or able to survive in long term future. (20-30 years)
  • Monetization: Product/Service should be monetizable directly. (indirectly monetizable Product/Service should be dropped for this Project)

Project-Assessment.md

Project Title Feasibility (2-3 years) Viability (20-30 years) Monetization
Machine Predictive Maintenance for Industrial Machines Feasible Viable Direct monetization through maintenance services
Personal Medical Assistant Less feasible Potential viability May require complex monetization strategies
University Chances Predictor Feasible Viable Subscription services or partnerships
Clothing Recommendation System Based on Body Data Feasible Viable Monetization through affiliate marketing or personalized shopping experiences
College Recommender Feasible Viable Subscription services or partnerships

Among these, we selected Machine Predictive Maintenance for Industrial Machines and we renamed it as Smart Industrial Predictive Solutions for our 3rd project.

Roles

Role Person(s)
Business Specialists (2) Karakavalasa venkata pranay & Shreyash Banduji Chacharkar
Data Scientists (2) Malay Vyas & Yash Mayur
Software Developer (1) Adhiban Siddarth

Business Model

Real time Web Based monitoring and predictive maintenances app:

prototype of real time web based predictive maintenance services

flowchart LR
A[Data collection\nthrough\nIOT based\nsensor]
B[Cloud\nstorage]
C[Data processing\nData checkup]
D[Analytics \n& Prediction]
E[Web based\napplication]
A-->B-->C-->D-->E
Loading

Business Model is prepared by Karakavalasa venkata pranay and Shreyash Banduji Chacharkar

ML Model ➡️ Web app

Yash Mayur's Machine Learning Model - Github repo

Malay Vyas's Machine Learning Model - Github repo

flowchart LR
subgraph Data Scientist
A[(Dataset)] --Data Analysis-->B[Clean Dataset]
B --ML--> C[ML Model] --> D[Pickle file]
C --> E[ML Function]
end
subgraph Software Developer
E --> F[Web app]
D --> F
end
Loading

ML Function gets csv file of machines' data and returns failear dataframe of all machines.

def ml_function(csv_path, model_path):
    # loading pre-trained model from model_path
    # pandas dataframe(df) from csv_path
    # df cleaning & preprocessing...
    ...
    ...
    # failear_df prediction from df using the ML model
    return failear_df

Web app architecture

flowchart LR
A[Machines]
B[Front-end]
C[ML Function] 
E[ML-Model.pkl]
G[IOT]
H[Manual]

G --Machines data--> B

subgraph Web app
B --Machines data--> C
C--Failear Report--> B
E --> C
end

B --Failear report--> G

subgraph Industry

A --Data--> G
G --Repair--> A
H
end

A <--> H
H --database--> B
B --Failear Report--> H
Loading

Upload csv format

Type Air temperature [K] Process temperature [K] Rotational speed [rpm] Torque [Nm] Tool wear [min] Power
L 276 316 1244 15 192 29575
H 333 302 1925 77 32 12799
M 286 311 2678 80 50 15481
... ... ... ... ... ... ...
H 338 301 2214 34 245 90977

You can include extra columns for machine names or number but the above columns in the table must be there in your upload csv.

The report csv file has only failear machines data, for easily getting which machine is failear, another column called index is added in report csv.

Web App Deployment

We deployed web app in GCP and render.com. Shreyash Banduji Chacharkar deployed web app in GCP Github repo. I deployed web app in render.com using docker image - adhiban/smart-industrial-predictive-solutions:alpha

Web App: Link1 (render.com) and Link2 (GCP)

Project video

flask-app.mp4

Gratitude for Exceptional Teamwork: A Journey with My Team

I wanted to express my heartfelt appreciation for the incredible support and expertise my team members have provided throughout this project. My team members' skills in data analysis and machine learning are truly exceptional, and I've had the opportunity to learn a tremendous amount from each member. This project has undoubtedly been the highlight of my team experiences, and I consider myself fortunate to have had the privilege of working with such an outstanding group. Although I've led teams for college projects in my native language, Tamil, I initially had some reservations about leading a team in English. However, working with my team members has been an absolute pleasure. My team members' effective communication and seamless collaboration have made this journey a joy, and I am genuinely grateful to have my team members as my partners in this endeavor. Thank you all so much for your hard work and dedication!