提出一种基于人工神经网络的锅炉燃烧智能化系统，学习锅炉运行参数与煤汽比间复杂的非线性映射关系，基于学到的映射关系动态调节锅炉运行参数，有效提高锅炉热效率，降低司炉工的劳动强度。该系统安装在35 t锅炉上，进行为期6个月的实验，实验结果表明，系统能够提高锅炉热效率,提高吨煤产汽量5.42%，节煤5.58%（平均231.31 t/月），降低司炉工的劳动强度95%。

An artificial neural network based intelligent boiler combustion system is proposed to enhance the thermal efficiency of boiler and reduce the labor intensity of boiler workers.The system firstly learns the complex nonlinear mapping between boiler operating parameters and coalsteam ratio,then dynamically adjusts boiler operating parameters according to the learned nonlinear mapping.The system was deployed on our company's 35ton boilers for a sixmonth experiment.The experimental results show that the proposed system improves the thermal efficiency of boiler,increases the steam production per ton of coal by 5.42%,saves coal bt 5.58% (231.31 tons per month on average) and reduces the labor intensity of boiler workers by 95.0%.

TK32

国家自然科学基金（51475264）；湖北省自然科学基金 （2016CFB565）；水电机械设备设计与维护湖北省重点实验室（三峡大学）开放基金（2016KJX03）； 宜昌市科技局项目(A14-302-a03)