基于BP神经网络的循环流化床锅炉密相区高度检测的模拟研究

STUDY ON SIMULATION OF HEIGHT DETECTION FOR DENSE PHASE ZONE ON CFB BASED ON BP NEURAL NETWORK

基于BP神经网络算法对循环流化床(CFB)锅炉密相区高度检测进行了模拟研究。结果表明,当系统内气体介质光学厚度为38时,该算法的计算值与真实值非常接近;随着光学厚度的增加,计算误差也逐渐增大,当光学厚度达到75时,计算值与真实值相差较大,这是由于光学探测器(CCD)的检测精度降低所致。

Study on simulation of height detection for dense phase on circulating fluidized bed （CFB） based on BP neural network has been carried out,the results show that the values calculated by using the said algorithm was very close to the true values when the optical thickness of gas medium in the system was less than 38,along with increasing the optical thickness,calculation error also gradually goes up;when the optical thickness reached to 75 ,the calculated values were greatly deviated from the true values, this is due to decreasing in detection precision of used optical detector.