摘要
为保证旋转机械注水泵机组安全、稳定运行,应采用合适的预测模型对其状态评定参数进行预测。提出基于均值函数新息加权的遗传算法优化神经网络预测模型(MWGANN模型),用此模型能够优化神经网络结构参数,并可利用时间序列数据新旧程度的不同提高预测的精度和实时性。工业现场采集大型旋转注水泵机组振动烈度时间序列数据,应用MWGANN模型和基于人工经验设定神经网络结构参数的模型分别对其进行预测并比较,结果表明MWGANN模型在预测精度、预测实时性方面取得了较好的效果。
To ensure safe and stable operation of the rotary mechanical injection pump unit, status assessment parameters should be predicted by using appropriate prediction model. A genetic algorithm optimization neural network prediction model is presented based on mean function new information-weighted theory （MWGANN prediction model ）. MWGANN prediction model can optimize the neural network structure parameters, at the same time improve the prediction accuracy and real-time by using the recency difference of time series data. Large rotating injection pump unit vibration intensity time series are collected in the industrial site. MWGANN model and artificial experience parameters neural network model are applied to predict trends and compare the results. The results show that MWGANN model achieved good results in prediction accuracy and real-time prediction.
出处
《北京信息科技大学学报（自然科学版）》
2012年第1期1-4,共4页
Journal of Beijing Information Science and Technology University
基金
国家自然科学基金项目(50975020)
北京市人才强教深化计划项目(PHR20090518)
北京市引进国外技术重点项目(B201101010)
关键词
趋势预测
均值函数
新息加权
神经网络
prediction model mean function new information-weighted neural network
