基于MDN和VMD-TPA-LSTM的泵站主机组劣化趋势预测OA
Deterioration trend prediction of pump station host group based on MDN and VMD-TPA-LSTM
为提高泵站主机组的安全稳定运行能力,解析其运行状态,获取机组设备的健康状况,准确预测其未来发展趋势,提出一种基于混合密度网络(Mixture Density Networks,MDN)和融合变分模态分解(Variational Mode Decomposition,VMD)与基于时序模式和注意力机制的长短时记忆网络(Temporal Pattern Attention-Long Short-Term Memory Network,TPA-LSTM)的泵站主机组劣化趋势预测模型.模拟结果表明,此法能够准确地表达机组的劣化趋势并可有效提高其预测精度.
In order to improve the safe and stable operation capability of the pump station host group,analyze its operating status,obtain the health status of the unit equipment,and accurately predict its future development trend,a pump station host group degradation trend prediction model based on Mixed Density Networks(MDN),Variational Mode Decomposition(VMD),and Temporal Pattern Attention Long Short Term Memory Network(TPA-LSTM)based on temporal pattern and attention mechanism is proposed.The simulation results show that this method can accurately express the deterioration trend of the unit and effectively improve its prediction accuracy.
夏臣智;李英玉;吴子豪;李超顺;黃富佳;莫兆祥
南水北调(江苏)数智科技有限公司,江苏南京 210019||江苏省泵站工程技术研究中心,江苏南京 210019华中科技大学土木与水利工程学院,湖北武汉 430074
水利科学
趋势预测网络模型泵站机组MDNVMDTPA-LSTM
trend predictionnetwork modelpump station unitMDNVMDTPA-LSTM
《江苏水利》 2024 (009)
24-29 / 6
江苏省水利科技项目(2022001)
评论