Intelligent Component Failure Forecasting

Stay Ahead of Breakdowns

About The Dataset And The Model

Dataset

The LSTM model has been built and run seperately and not integrated within the application. The results of the model were interpreted by knowledge from the domain. The given dataset contains two files,

• The dataset itself
• Threshold values for the dataset The dataset has been preprocessed and split into 5 subsequent datasets namely
  • Articulated_Truck
  • Asphalt_Paver
  • Backhoe_Loader
  • Dozer
  • Excavator Each subsequent dataset has been further normalized. The normalized datasets can be found under /static/normal.xlsx

The threshold file contains the threshold limit for each component and parameter individually, which are mapped and learned in the model.

LSTM Model

LSTM models are particularly useful for predicting machine failures.

Time-Series Nature of Machine Data: The data from heavy machinery typically involves time-series data, where the sequence and timing of events are crucial. LSTM models excel at capturing temporal dependencies, meaning they can understand how previous states influence future states.

Long-Term Dependencies: Machinery failure can be influenced by patterns and events that occur over long periods.

Non-linear and Complex Relationships: The relationship between different parameters and the likelihood of failure can be complex and non-linear.

Detecting Anomalies: LSTMs can be trained to recognize normal operating patterns and detect deviations from these patterns, which can indicate potential failures.

Noise Handling: LSTM models can handle noisy data better than traditional methods, which is often the case with sensor data from heavy machinery.

Run Locally

Clone the project

  git clone https://github.com/lohithgsk/intelligent-component-failure-forecasting.git

Go to the project directory

  cd my-project

Install dependencies

  pip install -r requirements.txt

Start the server

  flask run

Feedback

If you have any feedback, please reach out to us.