基于区间二型模糊逻辑的短期风速预测研究  

FORECASTING STUDY OF SHORT-TERM WIND SPEED BASED ON INTERVAL TYPE-2 FUZZY LOGIC METHOD

作　　者：郑高[1] 肖建[1] 王婧[2] 周文卫[1] 

机构地区：[1]西南交通大学电气工程学院,成都610031 [2]吉林大学珠海学院计算机科学与技术系,珠海519041

出　　处：《太阳能学报》2011年第12期1792-1797,共6页Acta Energiae Solaris Sinica

基　　金：国家自然科学基金(60674057);教育部博士点基金(20060613003)

摘　　要：针对风速具有强随机性而难以准确预测的问题,引入二型模糊逻辑以减小预测误差。基于反向传播-奇异值分解混合迭代算法,建立了区间二型模糊模型用于风速的短期时间序列预测,首先通过反向传播算法调节模型输入及规则前、后件集合的参数,然后利用奇异值分解法优选模型规则。同时,基于反向传播算法建立了一型模糊模型和区间二型模糊模型,将其预测结果作为检验性能的基准。仿真结果表明,基于混合迭代算法的区间二型模糊模型能有效精简系统规则并提高预测精度,具有一定的实用价值。In order to reduce the with random characteristic. An time series based on the singular forecast error, the type-2 fuzzy logic method was introduced to forecast wind speed interval type-2 fuzzy logic model was proposed to forecast short-term wind speed value decompose and back-propagation （SVD-BP） hybrid iterative arithmetic, the BP arithmetic was used to tune input, antecedent and consequent parameters, the SVD method was used to choose reasonable rules. For testing the performance, both a type-1 fuzzy logic model and an interval type-2 fuzzy logic model using only BP arithmetic were designed as comparable benchmarks. The simulation results show that the interval type-2 fuzzy logic model using the SVD-BP hybrid iterative arithmetic has the best performance and practicability because it has the least rule base and the minimal forecast error.

关 键 词：风速预测 时间序列预测 二型模糊集合 区间二型模糊逻辑 反向传播-奇异值分解 

分 类 号：N94[自然科学总论—系统科学]