摘要
面向旋转机组工作状态的趋势预测 ,结合组合预测模型和人工神经网络预测模型的各自优点 ,提出了一种变权重神经网络组合预测模型的新方法。该方法是利用变权重神经网络的非线性映射能力 ,依据各单项预测模型对预测结果的不同影响 ,动态确定组合预测模型的权系数。在大型旋转机组工业现场应用中 ,采用这种模型进行预测提高了预测精度 ,获得较好的预测效果。
In the trend predicting to working conditions of rotary sets, integrating the virtues of the Combined Predicting Model and the Artificial Neural Network Predicting Model, a new predicting model named Combined Predicting Model based on neural network of variable weighting modulus is put forward to forecast the working conditions. The core idea of the new method is to determine the weighting modulus of the Combined Predicting Model dynamically, utilizing non-linear mapping ability of the Artificial Neural Network of Variable Weighting Modulus and the different influence of the single predicting model on the predicting result. Using this new method, the predicting precision is improved and more satisfactory predicting results are obtained to large rotary sets in the industrial applications.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2003年第4期332-335,共4页
China Mechanical Engineering
基金
国家自然科学基金资助项目 (597750 0 2
关键词
旋转机组
趋势预测
组合预测模型
变权重神经网络
rotary sets trend predicting combined predicting model neural network of variable weighting modulus
