基于优化型K-means聚类算法的锅炉热效率研究

Research on Boiler Thermal Efficiency Based on Optimized K-means Clustering Algorithm

针对K-means聚类算法存在初始聚类中心影响聚类精度的问题,提出采用生物地理学算法优化K-means聚类中心,使其能提高聚类算法的准确率。在基准数据集中对本算法进行实验,其结果表明改进算法具有良好的性能。其次,采用改进的K-means聚类算法对不同工况下的锅炉燃烧工艺参数进行聚类,并挖掘出每一类中热效率最高时的燃烧工艺参数作为最佳工艺参数,使锅炉在最佳工艺参数下进行燃烧,达到提高热效率的目的。为了验证最佳工艺参数的有效性,采用贝叶斯最小二乘支持向量机辨识锅炉热效率模型,结果显示热效率明显提高,说明经过优化型K-means聚类算法挖掘的最佳工艺参数是有效的。

For the K-means clustering algorithm,there is a problem that the initial clustering center affects the clustering accuracy.Therefore,the biogeography-based optimization is proposed to optimize the K-means clustering center so that the accuracy of the clustering algorithm can be improved.Biogeography-based K-means clustering algorithm is tested in benchmark dataset and its results show that the improved algorithm has good performance.Secondly,the improved K-means algorithm is adopted to cluster the boiler combustion process parameters in different cases and mine the combustion optimal process parameters with the highest thermal efficiency as the optimal process parameters in every classification.Keep the boiler runs under the optimal process parameters to achieve the purpose of improving thermal efficiency.To verify the validity of the optimal process parameters,the Bayesian least squares support vector machine is applied to determine the boiler thermal efficiency model.The result shows that the thermal efficiency is improved obviously and indicates that the optimal process parameters are effective.