Study of forecasting solar irradiance based on neural networks combined with wavelet analysis

Forecast of solar irradiance plays an important role in the load forecast for air-conditioning systems. It is also one of the preconditions of optimal operation of air-conditioning systems in achieving the goals of energy efficient use and energy management of the systems. For the sake of higher accuracy in the forecast of solar irradiance, the artificial neural networks are used as the basis of methodology, combining wavelet analysis. In this paper, the data sequence of solar irradiance as samples is mapped into several time-frequency domains using wavelet transformation, and a recursive BP network is established for each domain. The solar irradiance can be forecasted with the algebraic SLIM of the components, which were forecasted by the established networks, of all the time-frequency domains. A discount coefficient method is adopted in updating the weights and thresholds of the networks so as to make the closest forecasts playing more important roles. Oil the basis of the principle of combination of artificial neural networks and wavelet analysis, a model is completed for forecasting solar irradiance. Based on historical day-by-day records of solar irradiance in Shanghai as samples an example is presented with the forecasted total solar irradiance. The results of the example indicate that the method makes the forecasts much more accurate than the forecasts using the artificial neural networks without combination with wavelet analysis.