云计算环境下智能电网短期负荷预测方法研究

智能电网短期负荷波动性大,传统预测方法无法解决波动性问题,预测结果不准确。为此,提出一种新的云计算环境下智能电网短期负荷预测方法。介绍了支持向量机理论,将一个含有所有某类样本在内的、由支持向量支撑的球面看作超球面,分析了分位数回归过程,将支持向量机和分位数结合在一起,构建支持向量-分位数回归预测模型。得到短期负荷概率密度函数,从而实现智能电网短期负荷预测。在进行实验时,完成对功率采样值和智能电网负荷属性的归一化处理,将其转换成0,1]区间内的数据。实验结果表明,所提方法预测精度和效率高、成本低。

Research on short term load forecasting method for smart grid under cloud computing environment

The short-term load fluctuation of smart grid is very large, and the traditional forecasting methods can not solve the problem of volatility, and the prediction results are inaccurate. To this end, proposes a new cloud computing environment of smart grid short-term load forecasting methods, introduces the theory of support vector machine, one containing all certain samples, supported by the support vector sphere as hypersphere, analyzed the quantile regression, support vector machine and quantile together, build quantile regression prediction model of support vector, get the short-term load probability density function, so as to realize the short-term load forecast of smart grid. During the experiment, the normalized sampling of the power sampling value and the smart grid load attributes is completed and converted into the data within the 0,1] interval. Experimental results show that the proposed method has high prediction accuracy and efficiency and low cost.