摘要
针对测量不确定评定数学模型无法通过确定的数学关系进行表达的问题,将BP神经网络算法引入测量不确定评定,通过将产生不同确定度的分量的影响因素作为神经网络的输入,合成不确定度和扩展不确定度作为神经网络输出,建立神经网络的不确定评定数学模型。以游标卡尺测量结果不确定度为研究对象,运用BP神经网络不确定评定数学模型,进行不确定评定。仿真结果表明,BP神经网络的不确定评定结果较好、精确度较高,接近于线性预测。
In view of the problem that measurement uncertainty evaluation mathematical model could not be expressed by the certain mathematical relationship,the BP neural network algorithm is introduced into the measurement of uncertainty evaluation,the influence factors of component that produce different uncertainty are taken as the inputs of neural network,and the synthetic uncertainty and expanding uncertainty are taken as the neural network outputs,with which the uncertainty evaluation mathematical model of neural network is established. Taking the uncertainty of measurement results with vernier caliper as research object,the mathematical model of uncertainty evaluation of BP neural network is applied to evaluate uncertainty.The simulation results show that the uncertainty evaluation results of BP neural network are good,high precision,close to the linear prediction.
出处
《四川理工学院学报（自然科学版）》
CAS
2015年第4期22-26,共5页
Journal of Sichuan University of Science & Engineering(Natural Science Edition)
关键词
测量不确定度
BP神经网络
数学模型
合成不确定度
扩展不确定度
measurement of uncertainty
BP neural network
mathematical model
synthesis of uncertainty
expanded uncertainty
