Modify Manual Observation System Data to Automatic Observing System Data on Wind Speed Using Bp Neural Network

This paper uses back propagation (BP) neural network to modify manual observing system data to automatic observing system data in wind speed data. First, we prepare six factors which are wind speed data six hours ahead to one hour ahead as input data. The tests indict when these factors selected, the modify results are better than others. The training data used for the neural network is the data from 1 Jan 2005 to 31 Dec 2007. The data from 1 Jan 2004 to 31 Dec 2004 are employed to test the model. The results show that the proposed method can produce satisfactory results in modifying manual observing system data to automatic observing system data. After modifying the manual data using BP neural network, the modified time series achieves -0.05m/s in Mean Bias Error compared to automatic observing system data. For wind speed, it could be a promising candidate for modifying manual observing system data to automatic observing system data.