BHM-Terrain-Paper

Code for the paper: A Bayesian hierarchical model to understand the effect of terrain on wind turbine power curves

Description:

This code repository contains the code to reproduce the tables and figures in the paper: A Bayesian hierarchical model to understand the effect of terrain on wind turbine power curves. This document explains the folder structure and provides instruction to execute the code.

Software:

The results are generated using shell scripts (bash) that wrap code from R software package version 3.6.3 or above and uses the following packages:

• DSWE
• mvtnorm
• FNN
• optparse
• truncdist
• latex2exp

Folder Structure:

The root folder contains the following sub-folders:

• src: A folder that contains the core code files.
• data: This folder contains all the relevant data files to reproduce the results in the paper. The code will automatically read the appropriate dataset for each bash script as long as the name and location of the files are not changed. See Datasets section for details about the files.
• artifacts: This folder will store all the model objects during execution and use them generate results.
• results: All the results from the paper will be stored in this folder.
• log: This folder will contain all the runtime log. If you get any error while running the code, please check the relevant logs in this folder.

Code Execution:

Each of the tables and figures have their own shell scripts. The following commands can be used to generate each of the results.

File	Command	# Cores	RAM	Runtime	Output files
Table1.sh	bash Table1.sh	1	2GB	5 minutes	Table1.txt
Figure2.sh	bash Figure2.sh	1	2GB	1 minute	Figure2.png
Figure3.sh	bash Figure3.sh	1	2GB	1 minute	Figure3.pdf
Figure4.sh	bash Figure4.sh	1	2GB	1 minute	Figure4.pdf
Figure5.sh	bash Figure5.sh	1	2GB	1 minute	Figure5.pdf
Figure7_Table3.sh	bash Figure7_Table3.sh	1	2GB	2 hours	Figure7a.png Figure7b.png Table3.txt
Figure8.sh	bash Figure8.sh	6	12GB	2.5 hours	Figure8a.pdf Figure8b.pdf Figure8c.pdf Figure8d.pdf Figure8e.pdf Figure8f.pdf
Table4_Figure9.sh	bash Table4_Figure9.sh	6	12GB	24 hours	Figure9a.pdf Figure9b.pdf Table4.txt

Note:

• Figure8.sh depends on the results obtained Figure7_Table3.sh. Please run Figure7_Table3.sh before running Figure8.sh.
• We provide a convenient wrapper Run_All.sh to execute all the code sequentially. The total runtime is approx 29 hours on Apple Silicon M1 chip.

Datasets:

The datasets used in this paper comprises of individual turbine operations data as well as terrain information data. The code will automatically read the appropriate dataset for each bash script. The code files assumes that the data is in data/ folder and have the same name as given below:

Files	Description
Turbine1_2017.csv to Turbine66_2017.csv	These files contain 10-min frequency wind power curve data for each of the 66 turbines used in the paper.
thinned_power.csv, thinned_wind_speed.csv, thinned_temperature.csv	These files contains thinned data using the thinning procedure described in the paper. Each column corresponds to a single turbine's data in each file.
weightedTerrainData.csv	This files contains the terrain variables values for each of the turbine weighted by the frequency of data falling in a wind direction sector as described in the paper.
location.csv	This file contains the location (latitude and longitude) of each turbine. The coordinates have been shifted by an unknown constant to preserve anonymity of the location.
terrain_level_agg.csv	This file contains the categorical level of the terrain for each turbine as described in the paper.