基于电厂工况划分的双层聚类算法研究

以利用火电厂实时数据的工况划分为对象,提出了一种基于SOM网络模型和改进K-均值算法的双层聚类算法。海量数据通过SOM网络的压缩,神经元保持了与原始数据的相同结构;再利用优化了初始聚类中心和可自适应调整到最佳K值的改进K-均值聚类算法,将神经元进一步聚类。实现了在较短时间内合理划分电厂生产过程工况的目标。

Division of power plant operating conditions based on double clustering algorithm

This paper proposes a two-tier clustering algorithm means algorithm based on SOM network model enhancement K-means algorithm,which take advantage of real-time data of the operating mode is divided into objects.Massive data compression through the SOM network,neurons maintained the same structure with the original data;re-use of optimization of the initial cluster centers and adaptive to the best value improvement K-means clustering algorithm,the neuron further polyed class.Achieved the target of reasonable ...