首页> 外文会议>International conference on education technology and computer >Centrifugal Compressors Performance Forecast Model Based on Wavelet Neural Network
【24h】

Centrifugal Compressors Performance Forecast Model Based on Wavelet Neural Network

机译:基于小波神经网络的离心压缩机性能预测模型

获取原文
获取外文期刊封面目录资料

摘要

The traditional approach to forecast centrifugal compressors performance is always BP neural network, which lack of precision, converge slowly, and easily trapped into local optimal solutions.To get more accurate forecast on centrifugal compressors performance and discover hidden problems in advance, here, we have integrated immune algorithms, wavelet theory and neural network, and established centrifugal compressors performance forecast model based on WNN and immune algorithms.First, immune algorithms produce antibody group, get WNN coefficient for every antibody through iterative operation, then use back propagation algorithm to train WNN and approximate any nonlinear function.The simulation experiment his indicated that this forecast model can achieve accurate forecast and monitoring on centrifugal compressors performance. This forecast model has the merits of simple arithmetic, stable structure, high convergent speed and strong generalization ability;the accuracy of forecast can reach 99% which is 13% higher than traditional approach: it has certain practical value and theory value.
机译:传统的离心压缩机性能预测方法始终是BP神经网络,该方法缺乏精度,收敛速度较慢,容易陷入局部最优解。要获得更准确的离心压缩机性能预测并提前发现隐患,这里有结合免疫算法,小波理论和神经网络,建立了基于WNN和免疫算法的离心压缩机性能预测模型。首先,免疫算法产生抗体组,通过迭代运算获得每种抗体的WNN系数,然后使用反向传播算法训练WNN仿真实验表明,该预测模型可以对离心压缩机的性能进行准确的预测和监测。该预测模型具有算法简单,结构稳定,收敛速度快,泛化能力强等优点;预测精度可以达到99%,比传统方法高13%:具有一定的实用价值和理论价值。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号