Wind speed data

Citation Author(s):
Zhang
Hengshan
Submitted by:
Zhang Hengshan
Last updated:
Sun, 12/10/2023 - 21:55
DOI:
10.21227/h85v-x637
License:
232 Views
Categories:
Keywords:
0
0 ratings - Please login to submit your rating.

Abstract 

1.    The release data includes the original data of wind turbine X30 at 2018 and 2019, which are files “2019_x30.csv” and “2018_x30.csv” in directory “Original data”.    For the turbine, the wind speed and direction are collected at each 30 seconds, then there are 2880 data at a day. In the experiment, we use the slide window with 120 data, which corresponding to an hours, and slide the window with 10 data step, which corresponding to 5 minutes. For the data in window, we select the data of final 10 minutes as the label. The data in a year for each turbine is decreased from 1051200 to 105120.

Instructions: 

1.    The release data includes the original data of wind turbine X30 at 2018 and 2019, which are files “2019_x30.csv” and “2018_x30.csv” in directory “Original data”.

2.    Preprocess:  For the turbine, the wind speed and direction are collected at each 30 seconds, then there are 2880 data at a day. In the experiment, we use the slide window with 120 data, which corresponding to an hours, and slide the window with 10 data step, which corresponding to 5 minutes. For the data in window, we select the data of final 10 minutes as the label. The data in a year for each turbine is decreased from 1051200 to 105120.

3.    The predicting results of individual learners are files “2019_x30.csv” and “2018_x30.csv” in directory “Prediction data of individual learners”.  The basis learners are “CNN and GRU are connected in serial (CG)”, “CNN and GRU are jointed in parallel (CGp)”, “ResNet and GRU are connected in serial (RG)”, “ResNet and GRU are jointed in parallel (RGp) ”, “Transformer(TRs)”.