变尺度混沌蚁群算法在NOx排放优化中的应用

为了控制燃煤电厂NOx排放，应用支持向量回归建立了大型四角切圆燃烧电站锅炉NOx 排放特性模型。利用大样本量的热态实炉NOx 排放试验数据对模型进行了训练和验证，结合NOx排放模型采用一种变尺度混沌蚁群算法对锅炉运行参数进行优化, 定量分析优化算法参数对优化结果的影响。计算结果表明，相对于BP神经网络，支持向量回归模型能更好地预测锅炉NOx排放；变尺度混沌蚁群算法能明显降低NOx排放，且具有较高的稳定性与鲁棒性，1.8 min的优化时间也便于在线应用；支持向量回归与变尺度蚁群混合算法能有效降低燃煤锅炉NOx排放，是锅炉NOx排放控制的有效工具。

Application of Scaleable Chaotic Ant Colony Algorithm in NOx Emissions Optimization

In order to control the nitric oxide emissions in coal-fired power plants, support vector regression (SVR) was employed to establish a mathematic model predicting the characteristics of nitric oxide emissions in large capacity corner-fired boilers. A large number of field test data from a full-scale operating boiler was used to train and validate the SVR model. Combining with NOx emission model, a scaleable chaotic ant colony optimization(SCACO) was used to optimize the operating parameters of the boiler. The inherent control parameters of the optimization algorithm were quantitatively analyzed in detail. The computational results show that SVR, compared to BPNN, predicts NOx emissions with better accuracy. ACO reduces NOx emissions significantly with high stability and robustness. One point eight minutes for optimization process is suitable for on line applications. The hybrid algorithm combining SVR and SCACO provides an efficient approach to reduce NOx emissions from the coal-fired utility boiler.