Merge branch 'master' of github.com:PML-UCF/pinn_wind_bearing

README.md

@@ -1,25 +1,23 @@
-<h1>Wind Turbine Main Bearing Fatigue Life Estimation with Physics-informed Neural Networks</h1>
+# Wind Turbine Main Bearing Fatigue Life Estimation with Physics-informed Neural Networks
 
-<p>This repository is provided for replication of results presented in the article:</p>
-<p><i>Y.A. Yucesan and F.A.C. Viana, "Wind turbine main bearing fatigue life estimation with physics-informed neural networks," Proceedings of the Annual Conference of the PHM Society 2019</i></p>
+This repository is provided for replication of results presented in the article:
 
-<p>This repository includes two versions of the code, named as: basic and advanced.</p>
+*Y.A. Yucesan and F.A.C. Viana, "Wind turbine main bearing fatigue life estimation with physics-informed neural networks," Proceedings of the Annual Conference of the PHM Society 2019*
 
-<p>Basic code has two main files:
-<ul>
-  <li>run01 file trains the RNN using pretrained MLP model with fixed initial weights.</li>
-  <li>run02 file predicts the fatigue damage accumulation of the wind turbine main bearing for 6 months.</li>
-</ul></p>
+This repository includes two sets of the code.
 
-<p>Advanced code has four main files:
-<ul>
-  <li>run01 file generates a random plane approximation for MLP training.</li>
-  <li>run02 file trains the MLP with randomly generated initial weights.</li>
-  <li>run03 file trains the RNN using trained MLP model.</li>
-  <li>run04 file predicts the fatigue damage accumulation of the wind turbine main bearing for 30 years.</li>
-</ul></p>
+**Basic:**
+- run01 file trains the RNN using pretrained MLP model with fixed initial weights.
+- run02 file predicts the fatigue damage accumulation of the wind turbine main bearing for 6 months.
 
-<p>Please refer to following source for required data:</p>
-<p><i>Yucesan, Yigit, 2019, "Wind Turbine Main Bearing Fatigue Life Prediction with PINN", <a href="url">https://doi.org/10.7910/DVN/ENNXLZ</a>, Harvard Dataverse, V1, UNF:6:o3b2Pkuz0uIgkQ57jEKGOA== [fileUNF]</i></p>
+**Advanced:**
+- run01 file generates a random plane approximation for MLP training.
+- run02 file trains the MLP with randomly generated initial weights.
+- run03 file trains the RNN using trained MLP model.
+- run04 file predicts the fatigue damage accumulation of the wind turbine main bearing for 30 years.
 
-<p>Download the data and <strong>extract folders inside wind_bearing_dataset to the directory where this repository is cloned.</strong></p>
+In order to run the codes, you will also need to:
+1. Install the PINN python package: https://github.com/PML-UCF/pinn
+2. Download the data:
+Yucesan, Yigit, 2019, "Wind Turbine Main Bearing Fatigue Life Prediction with PINN" (https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/ENNXLZ), Harvard Dataverse, V1, UNF:6:o3b2Pkuz0uIgkQ57jEKGOA== [fileUNF]
+and extract folders inside wind_bearing_dataset to the directory where this repository is cloned.