【摘要】 由于背压设定值基本为运行人员经验设置,致使机组经济性能差。通过现场数据计算出机组70%～100%负荷下部分关键工况点,采用BP神经网络对该数据联想记忆,近似得到该负荷段下所有背压设定值模型。同时采用粒子群算法对该模型进行优化,以提高BP神经网络模型的泛化能力和模型精度。将该模型应用到某一典型工况下的背压控制中,通过仿真得出平均每个风机功耗降低13.6 kW,而背压与改进前相差不多,此时机组净功率增加326.4 kW,经济效率得到显著提升。该方法对今后现场背压优化控制也具有一定借鉴意义。更多还原

【Abstract】 Because the setting value of back pressure is basically set by the operator experience,the economic performance of the unit is poor. This paper calculates some key operating points of the unit under 70% ～100% load by field data,associates and memorizes the data by BP neural network,and approximates all back pressure setting models under the load section. At the same time,the model is optimized by particle swarm optimization to improve the generalization ability and model accuracy of BP neural network model. The model is applied to the back pressure control under a typical working condition. The average power consumption of each fan is reduced by 13.6 k W by simulation,while the back pressure is similar to that before improvement. At this time,the net power of the unit increases by 326.4 k W,and the economic efficiency is significantly improved. This method can also be used for reference in the field of back pressure optimization control.更多还原