论文部分内容阅读
一般工业测控现场的环境温度变化急剧,从而造成仪表输出值随环境温度的变化而变化,再加上气压、以及气体流量等因素,导致测量出误差,因此误差补偿问题一直是工业测控系统中的关键环节。利用神经网络来处理各种环境因素而产生的误差不失是一种低成本而可靠的方法,BP(Back-Propagation)算法主要的优点是简单、易于实现。本文将BP网络的学习算法通过专家系统自能校正学习参数。利用大量的样本数据训练构建了一种自适应的强收缩网络模型,实现了仪表高精度快速误差补偿。
General industrial monitoring and control scene temperature changes rapidly, resulting in changes in the instrument output value with the ambient temperature, coupled with pressure, and gas flow and other factors, resulting in the measurement of error, so the error compensation problem has been the industrial measurement and control system The key link. The use of neural networks to deal with various environmental factors and errors generated after all is a low-cost and reliable method, BP (Back-Propagation) algorithm has the main advantage of simple and easy to implement. In this paper, the learning algorithm of BP network can be self-tuning learning parameters through the expert system. Using a large number of sample data training to build an adaptive strong contraction network model, to achieve a high precision instrument rapid error compensation.