Failure Time Prediction
This example trains a model to predict the failure time of a system. The failure time allows to schedule maintenance to repair or replace system before its downtime.
The example is based on the research of Sampaio. The respective paper is published at MDPI: Prediction of Motor Failure Time Using An Artificial Neural Network. The Accelerometer Dataset (00611) is found at the UCI Machine Learning Repository https://archive.ics.uci.edu/ml/datasets/Accelerometer.
Downloaded Data
Download the Dataset and save the respective csv-file at data/download.csv of the example's root directory.
The tabular dataset consists of time series data in the following format:
| wconfid | pctid | x | y | z |
|---|---|---|---|---|
| 1 | 20 | 21.004 | 0.090 | -0.12 |
| 1 | 20 | 1.004 | -0.043 | -0.125 |
| ... | ... | ... | ... | ... |
Preprocess to Generate Labels
The labels of an estimated failure time are generated using formulas presented in the paper.
The preprocessing.ipynb notebook is used to reproduce the steps and generate a labeled dataset in the following format:
| x-Amplitude | x-Frequency | y-Amplitude | y-Frequency | z-Amplitude | z-Frequency | Growth-rate | Estimated-Failure-Time |
|---|---|---|---|---|---|---|---|
| 0.27 | 0.28 | 0.03 | 0.28 | 0.06 | 0.28 | 0.05 | 1093.65 |
| ... | ... | ... | ... | ... | ... | ... | ... |
The resulting file is stored at data/processed_data.csv.
Run this example
To run this example first make sure you followed the installation instructions to setup python and just.
Afterwards you can either use just or run the examples from source.
Just
The easiest way to run this example is using just.
Follow the installation guide to clone flowcean but stop before installing it or any of its dependencies.
Now you can run the example using
just examples-failure_time_prediction
This command will take care of installing any required dependencies in a separate environment. After a short moment you should see the learning results and the achieved metric values.
From source
Follow the installation guide to install flowcean and it's dependencies from source.
Afterwards you can navigate to the examples folder and run the examples.
cd examples/failure_time_prediction
python run.py