Search

Artificial neural network method was used to forecast the wind speed, and two wind speed forecasting models were built based on BP and RBF neural network methods. 24 hours continuous wind speed forecast was conducted for a single wind turbine in wind farm. The results show that the models built are reasonable and have high prediction accuracy. By comparing the two kinds of wind speed forecasting models, BP neural network forecasting model has higher prediction accuracy than RBF neural network forecasting model in wind speed, but it demands much more training time.

The rice husk from China has been non-isothermally pyrolysed on thermogravimetric analyzer(TGA). The analyses were performed at different heating rates (20, 40, 60, 80, 100°C/min) up to 900°C with nitrogen as purge gas. The weight loss curve showed that the main pyrolysis of rice husk took place in the range of 200~500°C. On the basis of experiment data, a pyrolysis kinetic model was proposed. The kinetic parameters of activation energy(E) and frequency factor(A) were obtained by the Direct Arrhenius Plot Method. There was no clear relationship between activation energy and heating rate.