中文摘要：

高精度的风电功率预测对于电力系统的安全经济运行具有重要意义。基于大量风电功率历史数据，结合相关性分析和K近邻算法，提出一种新的多输出模型的风电功率超短期多步预测方法。以东北地区2个风电场实测风电功率数据为例进行分析计算，使用国家能源局提供的风电功率实时预测评价指标对两种多步预测方式进行评价。结果表明该方法预测精度高，方法简单，具有一定的工程实用价值。

英文摘要：

Accurate wind power prediction is important for power system planning and operation. Based on extensive wind power historical data, a new short-term prediction method of multiple output model is proposed by combing correlation analysis and K-Nearest Neighbor algorithm. Taken field measurement data from two wind farms in Northeast region as example, the two multi-step prediction methods are evaluated by using index defined by National Energy Board. The results show high precision and simplicity of proposed method.