在天然气净化过程中,强腐蚀性气体会导致反应容器出现不同程度的损耗。尽管传统腐蚀探针可以提供腐蚀预警,但其存在一定的滞后性。为解决这一问题,提出了基于小波变换-长短期记忆(Wavelet Transform-Long Short Term Memory, WT-LSTM)...在天然气净化过程中,强腐蚀性气体会导致反应容器出现不同程度的损耗。尽管传统腐蚀探针可以提供腐蚀预警,但其存在一定的滞后性。为解决这一问题,提出了基于小波变换-长短期记忆(Wavelet Transform-Long Short Term Memory, WT-LSTM)网络的腐蚀数据趋势预测方法。对原始数据采用db10小波进行分解,获取各细节分量后,利用LSTM网络对各分量进行预测并重构信号。实验结果显示,WT-LSTM模型的均方根误差(Root Mean Square Error, RMSE)为0.002 562,低于仅使用LSTM模型的RMSE值0.003 178,表明WT-LSTM网络在趋势预测上更加精准。基于WT-LSTM网络的预测方案能够有效跟踪腐蚀数据的变化,尤其在数据突变时效果显著,从而增强了腐蚀探针的在线监测能力,实现了对腐蚀情况的预测和预警,确保了天然气净化过程的安稳运行。展开更多
混凝土在开裂瞬间由于裂纹表面电荷的振荡产生电磁信号。针对电磁信号容易受环境噪声的影响,信噪比低的问题,提出一种混凝土开裂电磁信号的去噪方法。文中给出一种新的复合判定指数作为灰狼算法(grey wolf algorithm,GWO)的优化目标,确...混凝土在开裂瞬间由于裂纹表面电荷的振荡产生电磁信号。针对电磁信号容易受环境噪声的影响,信噪比低的问题,提出一种混凝土开裂电磁信号的去噪方法。文中给出一种新的复合判定指数作为灰狼算法(grey wolf algorithm,GWO)的优化目标,确定变分模态分解(variational mode decomposition,VMD)的最优模态分量个数和惩罚因子。对原始电磁信号进行分解,通过峭度原则选取适合的模态分量。结合小波去噪(wavelet transform,WT)对信号进行重构,得到去噪后的电磁信号。仿真及实验结果表明:与传统电磁信号去噪方法相比,文中方法处理信号的信噪比更高,均方差更低,波形相似度更高,对非平稳混凝土开裂电磁信号的适应能力更强。展开更多
This paper discusses an experimental investigation into the fluidity of AZ91D-1 wt.%Ca O magnesium melt via induction for thin-section investment casting.Plaster molds with thin spiral cavities(0.5 to 1.5 mm square se...This paper discusses an experimental investigation into the fluidity of AZ91D-1 wt.%Ca O magnesium melt via induction for thin-section investment casting.Plaster molds with thin spiral cavities(0.5 to 1.5 mm square sections)were designed and manufactured to assess the impact of casting conditions on filling length,as magnesium alloys cause severe melting and melt-mold exothermic reactions,making investment casting challenging.Combinations of traditional Mg-mold reaction mitigation techniques,such as applying a protective mold coating(Yttria)and vacuum,were examined to determine their role in the filling process.The results suggest that when induction is employed to melt reactive alloys,these methods are not always beneficial,as initially thought.Particularly at higher melt temperatures,the combination of Yttria-coated molds with low-pressure vacuum induction significantly reduce fluidity:vacuum induced melt levitation which promotes oxidation with the residual atmosphere;and Yttria-coating cracking due to thermal stress during the mold fabrication slows filling and promotes significant melt-mold reaction.This study shows that best results to investment cast thin-sections are obtained by avoiding both vacuum and protective coatings,providing a viable route for the precision manufacturing of stent biomedical devices.展开更多
文摘在天然气净化过程中,强腐蚀性气体会导致反应容器出现不同程度的损耗。尽管传统腐蚀探针可以提供腐蚀预警,但其存在一定的滞后性。为解决这一问题,提出了基于小波变换-长短期记忆(Wavelet Transform-Long Short Term Memory, WT-LSTM)网络的腐蚀数据趋势预测方法。对原始数据采用db10小波进行分解,获取各细节分量后,利用LSTM网络对各分量进行预测并重构信号。实验结果显示,WT-LSTM模型的均方根误差(Root Mean Square Error, RMSE)为0.002 562,低于仅使用LSTM模型的RMSE值0.003 178,表明WT-LSTM网络在趋势预测上更加精准。基于WT-LSTM网络的预测方案能够有效跟踪腐蚀数据的变化,尤其在数据突变时效果显著,从而增强了腐蚀探针的在线监测能力,实现了对腐蚀情况的预测和预警,确保了天然气净化过程的安稳运行。
基金financed by National Funds through the Portuguese funding agency,FCT–Funda??o para a Ciência e a Tecnologia,within the strategic projects UIDB/04436/2020,UIDB/00481/2020 and LA/P/0063/2020(DOI 10.54499/LA/P/0063/2020)。
文摘This paper discusses an experimental investigation into the fluidity of AZ91D-1 wt.%Ca O magnesium melt via induction for thin-section investment casting.Plaster molds with thin spiral cavities(0.5 to 1.5 mm square sections)were designed and manufactured to assess the impact of casting conditions on filling length,as magnesium alloys cause severe melting and melt-mold exothermic reactions,making investment casting challenging.Combinations of traditional Mg-mold reaction mitigation techniques,such as applying a protective mold coating(Yttria)and vacuum,were examined to determine their role in the filling process.The results suggest that when induction is employed to melt reactive alloys,these methods are not always beneficial,as initially thought.Particularly at higher melt temperatures,the combination of Yttria-coated molds with low-pressure vacuum induction significantly reduce fluidity:vacuum induced melt levitation which promotes oxidation with the residual atmosphere;and Yttria-coating cracking due to thermal stress during the mold fabrication slows filling and promotes significant melt-mold reaction.This study shows that best results to investment cast thin-sections are obtained by avoiding both vacuum and protective coatings,providing a viable route for the precision manufacturing of stent biomedical devices.
