针对燃煤电厂参与调峰负荷波动较大,出口SO_(2)浓度控制效果不佳的问题,建立了一种基于捕鱼优化算法(catch fish optimization algorithm,CFOA)优化融合神经网络的出口SO_(2)浓度预测模型。首先使用互信息算法筛选由机理分析得到的特征...针对燃煤电厂参与调峰负荷波动较大,出口SO_(2)浓度控制效果不佳的问题,建立了一种基于捕鱼优化算法(catch fish optimization algorithm,CFOA)优化融合神经网络的出口SO_(2)浓度预测模型。首先使用互信息算法筛选由机理分析得到的特征变量,并通过逐次变分模态分解对筛选后的辅助变量进行分解重构,保留相关性较大的重构分量作为输入变量。随后采用双向时间卷积网络、双向门控循环单元与多头自注意力机制构建融合神经网络模型,通过CFOA对模型超参数寻优以进一步提高精度。最后使用某660 MW燃煤电厂历史运行数据进行对比实验,实验结果表明,该模型在出口SO_(2)浓度剧烈波动的工况下仍能实现较好的预测效果。同多种模型对比,该模型具有更小的误差和更高的预测精度,体现出其在复杂变化环境中的鲁棒性和可靠性。展开更多
Sulfur dioxide(SO_(2)) and its derivatives have been recognized as harmful environmental pollutants.However,they are often produced during the processing of traditional Chinese medicines,potentially compromising the q...Sulfur dioxide(SO_(2)) and its derivatives have been recognized as harmful environmental pollutants.However,they are often produced during the processing of traditional Chinese medicines,potentially compromising the quality of these medicinal materials and contributing to various health issues.Due to a lack of effective monitoring and imaging tools,the physiological effects of excessive SO_(2) residues in traditional Chinese medicine remain unclear.Therefore,developing a rapid and effective tool for detecting SO_(2) is crucial for understanding its metabolic pathways and effects in vivo.In this study,we developed a near infrared(NIR) and ratiometric fluorescent probe,NIR-RS,which exhibits high sensitivity,selectivity,and rapid response for SO_(2) detection.Notably,NIR-RS accurately quantifies SO_(2) contents in Pinelliae rhizoma(P.rhizoma) samples,with recovery rates from 98.46 % to 102.40 %,and relative standard deviations(RSDs)< 5.0 %.For bioimaging applications,NIR-RS has low cytotoxicity and good mitochondrial-targeting ability,making it suitable for imaging exogenous and endogenous SO_(2) in mitochondria.Additionally,NIR-RS was successfully applied to image SO_(2) content of P.rhizoma samples within cells,revealing that high SO_(2) residue elevated mitochondria adenosine triphosphate(ATP) content,these findings reveal that P.rhizoma with excessive SO_(2) can affect the organism's growth mechanisms through alterations in ATP pathways.In vivo,SO_(2) was found to predominantly accumulate in the liver following gavage with P.rhizoma solution,with accumulation levels increasing in proportion to SO_(2) residue concentration.High SO_(2) concentrations in P.rhizoma can cause pulmonary fibrosis and gastric mucosal damage.This work provides a valuable tool for regulating SO_(2) content in P.rhizoma and may help researcher better understand the metabolism of SO_(2) derivatives and explore their physiological roles in biological systems.展开更多
文摘针对燃煤电厂参与调峰负荷波动较大,出口SO_(2)浓度控制效果不佳的问题,建立了一种基于捕鱼优化算法(catch fish optimization algorithm,CFOA)优化融合神经网络的出口SO_(2)浓度预测模型。首先使用互信息算法筛选由机理分析得到的特征变量,并通过逐次变分模态分解对筛选后的辅助变量进行分解重构,保留相关性较大的重构分量作为输入变量。随后采用双向时间卷积网络、双向门控循环单元与多头自注意力机制构建融合神经网络模型,通过CFOA对模型超参数寻优以进一步提高精度。最后使用某660 MW燃煤电厂历史运行数据进行对比实验,实验结果表明,该模型在出口SO_(2)浓度剧烈波动的工况下仍能实现较好的预测效果。同多种模型对比,该模型具有更小的误差和更高的预测精度,体现出其在复杂变化环境中的鲁棒性和可靠性。
基金supported by the Natural Science Foundation of Hubei Province (Nos.2023AFB376 and 2024AFD287)National Key Research and Development Program (No.2023YFC3503804)the National Natural Science Foundation of China (No.22077044)。
文摘Sulfur dioxide(SO_(2)) and its derivatives have been recognized as harmful environmental pollutants.However,they are often produced during the processing of traditional Chinese medicines,potentially compromising the quality of these medicinal materials and contributing to various health issues.Due to a lack of effective monitoring and imaging tools,the physiological effects of excessive SO_(2) residues in traditional Chinese medicine remain unclear.Therefore,developing a rapid and effective tool for detecting SO_(2) is crucial for understanding its metabolic pathways and effects in vivo.In this study,we developed a near infrared(NIR) and ratiometric fluorescent probe,NIR-RS,which exhibits high sensitivity,selectivity,and rapid response for SO_(2) detection.Notably,NIR-RS accurately quantifies SO_(2) contents in Pinelliae rhizoma(P.rhizoma) samples,with recovery rates from 98.46 % to 102.40 %,and relative standard deviations(RSDs)< 5.0 %.For bioimaging applications,NIR-RS has low cytotoxicity and good mitochondrial-targeting ability,making it suitable for imaging exogenous and endogenous SO_(2) in mitochondria.Additionally,NIR-RS was successfully applied to image SO_(2) content of P.rhizoma samples within cells,revealing that high SO_(2) residue elevated mitochondria adenosine triphosphate(ATP) content,these findings reveal that P.rhizoma with excessive SO_(2) can affect the organism's growth mechanisms through alterations in ATP pathways.In vivo,SO_(2) was found to predominantly accumulate in the liver following gavage with P.rhizoma solution,with accumulation levels increasing in proportion to SO_(2) residue concentration.High SO_(2) concentrations in P.rhizoma can cause pulmonary fibrosis and gastric mucosal damage.This work provides a valuable tool for regulating SO_(2) content in P.rhizoma and may help researcher better understand the metabolism of SO_(2) derivatives and explore their physiological roles in biological systems.
