Modeling and Understanding Curtailment in Orkney
This is the repo for my Msc Thesis project, revolving around modeling and understanding generator curtailment in the Active Network Management system of the Orkney power grid.
See http://curtailment.net for a graph tool to explore the data.
See http://forecast.curtailment.net for a curtailment forecast based on the models we have developed.
- Datasets
- The main dataset used for modeling is found in
datasets/Dataset-01-12-2018-to-01-03-2019.csv - A clean version is found in
datasets/Dataset-01-12-2018-to-01-03-2019-cleaned.csv - Eday data found in
datasets/eday-winter-2018-2019.csv - The dataset used for evaluating predictive models is found in
datasets/predictive_dataset_april_2019.csv
- The main dataset used for modeling is found in
- Scripts
scrape.pycontains the data collection and storing.prepros.pycontains data preprocessing, alogn with database querying and data cleaning.descriptive.pycontains the rule-based descriptive models and evaluation methods.model.pycontains methods for training, evaluating, saving and loading ANN-based models.plotter.pycontains methods for generating visualizations and plots.prepros-eday-data.py, a simple script the parses the eday CSV-files and creates a dataframe.eday-power-curve.pycontains the code for constructing the power curves for the ERE wind turbine.metoffice.pycontains the methods for evaluating the predictive models.
To run the files, you need to make a config.py file with the following variables and place it in the main folder:
reader_user = "xxxx" # username for DB user with reading rights
reader_pw = "xxxxxx" # password for DB user with reading rights
writer_user = "xxxxx" # username for DB user with writing rights
writer_pw = "xxxxxx" # password for DB user with writing rights
API_KEY = "xxxxxxxx" # API Key for OpenWeatherMap
MET_API_KEY = "xxxxxxxxxxxxxxxxx" # API Key for UK Met office
SERVER = "xxx.xxx.xxx.xxx:xxxxx" # Server address and port for MongoDB
DATA_PATH = "./datasets/" # Or whereever you wanna put it.