solar-intensity-forecasting


FINANCE

Adaptive Learning Hybrid Model for Solar Intensity Forecasting

Yu Wang, Member, IEEE, Yinxing Shen, Shiwen Mao, Senior Member, IEEE,
Guanqun Cao, and R. M. Nelms, Fellow, IEEE

Abstract—Energy management is indispensable in the smart grid, which integrates more renewable energy resources, such as solar and wind. Because of the intermittent power generation from these resources, precise power forecasting has become very crucial to achieve efficient energy management. In this paper, we propose a novel adaptive learning hybrid model (ALHM) for precise solar intensity forecasting based on meteorological data. We first present a time-varying multiple linear model (TMLM) to capture the linear and dynamic property of the data. We then construct simultaneous confidence bands (SCB) for variable selection. Next we apply the genetic algorithm back propagation neural network (GABP) to learn the nonlinear relationships in the data. We further propose ALHM by integrating TMLM, GABP and the adaptive learning online hybrid algorithm (ALOHA). The proposed ALHM captures the linear, temporal and nonlinear relationships in the data, and keeps improving the predicting performance adaptively online as more data collected. Simulation results show that ALHM outperforms several benchmarks in both short-term and long-term solar intensity forecasting. Index Terms—Solar intensity forecasting, online adaptive learning, local linear estimation, artificial neural network, genetic algorithm back propagation neural network.
I. INTRODUCTION
In recent years, Smart Grid (SG) has become an irreversible tendency in many countries all over the world. The advanced techniques from many fields, including industrial informatics, power electronics and automatic control make SG a sustainable power grid, which integrates more renewable energy sources, such as solar and wind [1], [2]. Because of the intermittency of renewable power generation, energy management is thus very important to improve the reliability, efficiency and utility of a SG [3]–[5]. It is mentioned in [5] that energy management efficiency can be greatly improved if the renewable energy generation can be predicted more accurately [6]–[8]. Thus, predicting renewable energy generation in the SG has attracted great interests [9], [10], mainly focusing on predicting solar power for their wide range of utilization. Solar power generation from solar panels are proportional to solar intensity, power generated per unit area. Therefore, predicted solar power can be acquired by predicting solar intensity, which is related to meteorological variables. Many recent works focus on the meteorological-data-based solar intensity forecasting problem by presenting different methods [2], [11].The work of [2] provides acceptable predicting results using SVM regression, and the author of [11] proposes the Hybrid Fuzzy Inference System algorithm (HyFIS) as solar intensity forecast mechanism But it lacks a deep analysis of the solar power generation and weather data. Learning techniques are also used to predict solar intensity, capturing the relationships between solar intensity and the meteorological variables. Artificial neural network (ANN) [12]– [14] is also a commonly used learning algorithm for complex

solar-intensity-forecasting
FREE DOWNLOAD


@ engpaper.com published paper

PUBLICATION PROCEDURE WITH US ENGPAPER.COM

ENGPAPER.COM PUBLISHED PAPERS


FREE IEEE PAPER AND PROJECT