基于L-M算法优化BP神经网络的飞灰含碳量测量

Measurement System for Unburned Carbon in Fly Ash Based on L-M Algorithm Optimized BP Network

针对目前用于测量飞灰含碳量的烟道微波飞灰测碳仪没有充分考虑烟气密度影响的问题,设计了一种微波幅度和相位组合测试系统用于补偿烟气密度变化。通过对飞灰含碳量影响因素的机理分析以及灰色关联法对其相关性的计算,以微波功率衰减、微波相移量、燃烧器摆角、一次风总风压、二次风挡板开度作为模型输入参数,构造了基于L-M算法优化BP神经网络的数据融合技术的飞灰含碳量测量系统。仿真实验结果表明该系统对飞灰含碳量的测量较烟道式飞灰测碳仪的测量结果更加真实和准确。

Microwave-based monitors of unburned carbon content in fly ash is widely used nowadays, but the effect of smoke density is not fully considered. A microwave amplitude and phase composition test system is designed to compensate for the change of smoke density. The fly ash carbon content measurement system which is based on L-M algorithm to optimize the BP neural network data fusion technology is constructed, using the microwave power attenuation, microwave phase shift amount, burner tilt, primary air pressure, secondary air baffle opening angle as input. The simulation results show that the improved measurement system is more authentic and accurate.