Machine Predictive Maintenance for Industrial Machines

Feynn Labs Internship Project1

I started this project at feynn labs repo and I create this separate repo to continue this project.

About

Predictive maintenance (PdM) is a maintenance strategy that uses data analysis to predict when a machine or equipment is likely to fail. This allows for preventive maintenance to be performed before the failure occurs, avoiding costly downtime and repairs. By combining PdM with ML, it is possible to develop models that can accurately predict machine failures days, weeks, or even months in advance. This allows organizations to schedule preventive maintenance at the most opportune time, minimizing downtime and maximizing uptime.

Download Repository

git clone --depth 1 https://github.com/Adhiban1/Machine-Predictive-Maintenance.git

Change directory

cd Machine-Predictive-Maintenance

Create virtual environment

python -m venv .venv

Activate virtual environment

For windows

.venv/Scripts/activate 

For linux

source .venv/bin/activate

Install requirements

pip install -r requirements.txt

Train

Run train.py this will train the model and save the models into models folder

python myapp/modules/train.py

Test

Run test.py to verify saved models work well

python myapp/modules/test.py

App

Run web app

python manage.py runserver 5000

Setup

You can run above commands one by one or you can use setup/setup.sh for linux or setup/setup.bat for windows. You should modify the code in setup.sh or setup.bat file for your cases. And then run this one file this will download repository, change directory, creating virtual environment, activating virtual environment, upgrading pip, installing requirements, training the model, testing the model, opening Flask App.

Once you run setup file, then don't run again to open flask app. Just run run/run.bat for windows or .run/run.sh for linux or you can activate .venv manually and run app.py.

Download setup file and modify it according to your needs and run this.

Windows

D:\test>dir
setup.bat

D:\test>setup.bat
"Git Cloning..."
Cloning into 'feynn-labs'...
...
...

Linux

adhiban@DESKTOP:/test$ ls
setup.sh
adhiban@DESKTOP:/test$ bash setup.sh
Git Cloning...
Cloning into 'feynn-labs'...
...
...

Dark mode:

Light mode:

Input Data

>>> df = pd.read_csv('predictive_maintenance.csv')
>>> df.head()
   UDI Product ID Type  Air temperature [K]  ...  Torque [Nm]  Tool wear [min]  Target  Failure Type
0    1     M14860    M                298.1  ...         42.8                0       0    No Failure
1    2     L47181    L                298.2  ...         46.3                3       0    No Failure
2    3     L47182    L                298.1  ...         49.4                5       0    No Failure
3    4     L47183    L                298.2  ...         39.5                7       0    No Failure
4    5     L47184    L                298.2  ...         40.0                9       0    No Failure

[5 rows x 10 columns]
>>> df.columns
Index(['UDI', 'Product ID', 'Type', 'Air temperature [K]',
       'Process temperature [K]', 'Rotational speed [rpm]', 'Torque [Nm]',
       'Tool wear [min]', 'Target', 'Failure Type'],
      dtype='object')

Output Data

>>> output = pd.read_csv('output.csv')
>>> output.head()
   UDI Product ID Type  ...  Failure Type  is_failure  failure_type
0    1     M14860    M  ...    No Failure           0    No Failure
1    2     L47181    L  ...    No Failure           0    No Failure
2    3     L47182    L  ...    No Failure           0    No Failure
3    4     L47183    L  ...    No Failure           0    No Failure
4    5     L47184    L  ...    No Failure           0    No Failure

[5 rows x 12 columns]
>>> output.columns
Index(['UDI', 'Product ID', 'Type', 'Air temperature [K]',
       'Process temperature [K]', 'Rotational speed [rpm]', 'Torque [Nm]',
       'Tool wear [min]', 'Target', 'Failure Type', 'is_failure',
       'failure_type'],
      dtype='object')

Here failure_type, is_failure columns are added in output.csv

I give the same training dataset to the Flask app to check whether it predicts correct or not. Thats why Target, Failure Type and other columns are displayed in the terminal output.

Docker

Docker pull

docker pull adhiban/machine-predictive-maintenance

Docker run

docker run -d -p 5000:5000 adhiban/machine-predictive-maintenance

Open it in browser http://127.0.0.1:5000/

[!NOTE] This Django project is in Debug=True mode