基于栈式自编码BP神经网络预测水体亚硝态氮浓度模型
|
摘要
亚硝态氮对于水产养殖动物具有毒性,对于其含量的及时监控非常重要。基于光谱法和电极法设计的亚硝态氮传感器价格昂贵,难以大面积推广,因此急需研发一种能快速预测养殖水体亚硝态氮的模型。实验通过实验室构建的水质在线检测系统测定水体中温度、pH、溶解氧、氧化还原电位4个参数,同时用α-萘胺比色法测定水体中亚硝态氮的浓度,从4种参数中选取与亚硝态氮浓度相关的参数作为预测模型的关联变量。水质参数数据及亚硝态氮浓度数据分别经预处理后作为原始数据用于SAE神经网络的训练,训练方法采用无监督逐层贪婪训练法,用学习到的特征监督训练SAE-BP神经网络,利用反向传播算法(BP)优化模型。训练得到结构为4-5-4-3-1的SAE-BP神经网络模型,建立的神经网络模型对实验数据预测的拟合优度R2为0.95,预测结果的均方根误差RMSEP为0.099 71。研究表明,亚硝态氮预测模型可以较为精准地预测水体中亚硝态氮的浓度。本模型将为开发在线快速监测养殖水体亚硝态氮浓度提供新的思路。
Nitrite nitrogen is toxic to the aquatic animals. Monitoring the concentration of nitrite nitrogen is very critical for the culture of aquatic animals. Due to the high cost of the current commercial electrode sensor which is used to measure the concentration of nitrite nitrogen in water, this kind of sensor is very difficult to be popularized on a large scale. Therefore, it is an urgent need to develop another novel method to predict the concentration of nitrite nitrogen in water. In this paper, taking the advantage of the established online water monitoring system in our laboratory, the temperature, pH value, dissolved oxygen and oxidation-reduction potential were recorded from the water in tanks. Meanwhile, the actual concentration of nitrate nitrogen in water was measured using alphanaphthalene colorimetric method. The data after pretreatment were used as the original data to be used for SAE neural network training. Thereafter, unsupervised greed training method was applied. The learnt characteristics were used for the supervision and training of BP neural network. The model was optimized using the back propagation(BP) algorithm. The prediction model R2 of the nitrite nitrogen after training was 0.95, and root mean square error of the prediction(RMSEP) was 0.099 71, indicating that the model could accurately predict the concentrations of nitrate nitrogen in water. The established model will pave a new way for developing online system for monitoring the water nitrate nitrogen concentration in the future.
Nitrite nitrogen is toxic to the aquatic animals. Monitoring the concentration of nitrite nitrogen is very critical for the culture of aquatic animals. Due to the high cost of the current commercial electrode sensor which is used to measure the concentration of nitrite nitrogen in water, this kind of sensor is very difficult to be popularized on a large scale. Therefore, it is an urgent need to develop another novel method to predict the concentration of nitrite nitrogen in water. In this paper, taking the advantage of the established online water monitoring system in our laboratory, the temperature, pH value, dissolved oxygen and oxidation-reduction potential were recorded from the water in tanks. Meanwhile, the actual concentration of nitrate nitrogen in water was measured using alphanaphthalene colorimetric method. The data after pretreatment were used as the original data to be used for SAE neural network training. Thereafter, unsupervised greed training method was applied. The learnt characteristics were used for the supervision and training of BP neural network. The model was optimized using the back propagation(BP) algorithm. The prediction model R2 of the nitrite nitrogen after training was 0.95, and root mean square error of the prediction(RMSEP) was 0.099 71, indicating that the model could accurately predict the concentrations of nitrate nitrogen in water. The established model will pave a new way for developing online system for monitoring the water nitrate nitrogen concentration in the future.
引文
[1]Gowen R J,Bradbury N B.The ecological impact of salmonid farming in coastal waters:a review[J].Oceanography and Marine Biology Annual Review,1987,25:563-575.
[2]黄翔鹄,李长玲,郑莲,等.亚硝酸盐氮对凡纳滨对虾毒性和抗病相关因子影响[J].水生生物学报,2006,30(4):466-471.Huang X H,Li C L,Zheng L,et al.The toxicity of NO2-N on Litopenaeus vannamei and effects of NO2-N on factors relating to the anti-disease ability[J].Acta Hydrobiologica Sinica,2006,30(4):466-471(in Chinese).
[3]王鸿泰,胡德高.池塘中亚硝酸盐对草鱼种的毒害及防治[J].水产学报,1989,13(3):207-214.Wang H T,Hu D G.Toxicity of nitrite to grass carp(Ctenopharyngodon idellus)in ponds and its way of prevention[J].Journal of Fisheries of China,1989,13(3):207-214(in Chinese).
[4]张武肖,孙盛明,戈贤平,等.急性氨氮胁迫及毒后恢复对团头鲂幼鱼鳃、肝和肾组织结构的影响[J].水产学报,2015,39(2):233-244.Zhang W X,Sun S M,Ge X P,et al.Acute effects of ammonia exposure on histopathology of gill,liver and kidney in juvenile Megalobrama amblycephala and the post-exposure recovery[J].Journal of Fisheries of China,2015,39(2):233-244(in Chinese).
[5]陈家长,臧学磊,胡庚东,等.氨氮胁迫下罗非鱼(GIFTOreochromis niloticus)机体免疫力的变化及其对海豚链球菌易感性的影响[J].生态环境学报,2011,20(4):629-634.Chen J Z,Zang X L,Hu G D,et al.The immune response of GIFT Oreochromis niloticus and its susceptibility to Streptococcus iniae under stress in different ammonia[J].Ecology and Environmental Sciences,2011,20(4):629-634(in Chinese).
[6]韩英,张辉,王琨.亚硝态氮对鲤鱼种血液SOD及GSH-Px的影响[J].淡水渔业,2007,37(1):66-68.Han Y,Zhang H,Wang K.Activity of SOD and GSH-Px in the blood of fingerling Cyprinus carpio under different nitrite concentration[J].Freshwater Fisheries,2007,37(1):66-68(in Chinese).
[7]Dunier M,Siwicki A K,Dema?l A.Effects of organophosphorus insecticides:effects of trichlorfon and dichlorvos on the immune response of carp(Cyprinus carpio):Ⅲ.In vitro effects on lymphocyte proliferation and phagocytosis and in vivo effects on humoral response[J].Ecotoxicology and Environmental Safety,1991,22(1):78-87.
[8]洪美玲,陈立侨,孙新谨,等.亚硝酸盐急性胁迫对中华绒螯蟹幼体相关免疫指标和应激蛋白(HSP70)表达的影响[J].应用与环境生物学报,2011,17(5):688-693.Hong M L,Chen L Q,Sun X J,et al.Effects of acute nitrite exposure on immunity indicators and HSP70expression in Chinese mitten-hand crab(Eriocheir sinen- sis)[J].Chinese Journal of Applied&Environmental Biology,2011,17(5):688-693(in Chinese).
[9]周鲜娇,邱德全.亚硝酸氮和副溶血弧菌对凡纳滨对虾部分免疫指标的影响[J].水生态学杂志,2009,2(1):49-53.Zhou X J,Qiu D Q.Effects of nitrite nitrogen and Vibrio parahaemolyticu on some immunity indicators of Litopenaeus vannamei[J].Journal of Hydroecology,2009,2(1):49-53(in Chinese).
[10]邓嫔,刘威,李小明,等.pH、ORP监控在亚硝酸型生物脱氮过程中的应用[J].环境科学与技术,2007,30(3):97-99.Deng P,Liu W,Li X M,et al.Application of pH,ORPon-line monitor in shortcut biological nitrogen removal process[J].Environmental Science&Technology,2007,30(3):97-99(in Chinese).
[11]Pagacova P,Blstakova A,Drtil M,et al.Continually measured orp and Ph signal for control of nitrogen removal[M]//Hlavinek P,Bonacci Q,Marsalek J,et al.Dangerous Pollutants(Xenobiotics)in Urban Water Cycle.Dordrecht:Springer,2008:287-296.
[12]Bastian T,Brondum J.Do traditional measures of water quality in swimming pools and spas correspond with beneficial oxidation reduction potential?[J].Public Health Reports,2009,124(2):255-261.
[13]Hanaki K,Wantawin C,Ohgaki S.Nitrification at low levels of dissolved oxygen with and without organic loading in a suspended-growth reactor[J].Water Research,1990,24(3):297-302.
[14]Hellinga C,Schellen A A J C,Mulder J W,et al.The SHARON process:an innovative method for nitrogen removal from ammonium-rich waste water[J].Water Science and Technology,1998,37(9):135-142.
[15]杨琴,谢淑云.BP神经网络在洞庭湖氨氮浓度预测中的应用[J].水资源与水工程学报,2006,17(1):65-70.Yang Q,Xie S Y.Application of BP neural network into predicting NH3-N concentration of Dongting Lake[J].Journal of Water Resources&Water Engineering,2006,17(1):65-70(in Chinese).
[16]乔俊飞,马士杰,许进超.基于递归RBF神经网络的出水氨氮预测研究[J].计算机与应用化学,2017,34(2):145-151.Qiao J F,Ma S J,Xu J C.The water ammonia nitrogen prediction research based on a recursive RBF neural network[J].Computers and Applied Chemistry,2017,34 (2):145-151(in Chinese).
[17]冯红利,刘秀红,杨庆,等.低溶解氧下氨氧化过程神经网络预测控制模型[J].中国环境科学,2017,37(1):139-145.Feng H L,Liu X H,Yang Q,et al.Neural network prediction and control model for ammonia oxidizing process under low DO concentration[J].China Environmental Science,2017,37(1):139-145(in Chinese).
[18]管秋,王万良,徐新黎,等.基于神经网络的污水处理指标软测量研究[J].环境污染与防治,2006,28(2):156-158.Guan Q,Wang W L,Xu X L,et al.Soft-sensing method development based on neural network for wastewater treatment[J].Environmental Pollution and Control,2006,28(2):156-158(in Chinese).
[19]汪翔,何吉祥,佘磊,等.基于NAR神经网络对养殖水体亚硝酸盐预测模型的研究[J].渔业现代化,2015,42(4):30-34.Wang X,He J X,She L,et al.Establishment of aquaculture water nitrite prediction model based on NARneural network[J].Fishery Modernization,2015,42(4):30-34(in Chinese).
[20]林志贵,姚芳琴,冯林强,等.结合自适应遗传算法与弹性BP神经网络的亚硝酸盐预测模型[J].天津工业大学学报,2015,34(3):67-72.Lin Z G,Yao F Q,Feng L Q,et al.Nitrite prediction model based on adaptive genetic algorithm and elastic BP neural network[J].Journal of Tianjin Polytechnic University,2015,34(3):67-72(in Chinese).
[21]杨磊.水中亚硝胺类消毒副产物生成规律及其前质去除方法研究[D].哈尔滨:哈尔滨工业大学,2010.Yang L.Research on N-nitrosamines formation during disinfection and removal methods of their precursors[D].Harbin:Harbin Institute of Technology,2010.(in Chinese)
[22]郭宁,张建,孔强,等.温度对亚硝化及氧化亚氮释放的影响[J].环境工程学报,2013,7(4):1308-1312.Guo N,Zhang J,Kong Q,et al.Effect of temperature on ammoxidation and nitrous-oxide emission in partial nitrification[J].Chinese Journal of Environmental Engineering,2013,7(4):1308-1312(in Chinese).
[2]黄翔鹄,李长玲,郑莲,等.亚硝酸盐氮对凡纳滨对虾毒性和抗病相关因子影响[J].水生生物学报,2006,30(4):466-471.Huang X H,Li C L,Zheng L,et al.The toxicity of NO2-N on Litopenaeus vannamei and effects of NO2-N on factors relating to the anti-disease ability[J].Acta Hydrobiologica Sinica,2006,30(4):466-471(in Chinese).
[3]王鸿泰,胡德高.池塘中亚硝酸盐对草鱼种的毒害及防治[J].水产学报,1989,13(3):207-214.Wang H T,Hu D G.Toxicity of nitrite to grass carp(Ctenopharyngodon idellus)in ponds and its way of prevention[J].Journal of Fisheries of China,1989,13(3):207-214(in Chinese).
[4]张武肖,孙盛明,戈贤平,等.急性氨氮胁迫及毒后恢复对团头鲂幼鱼鳃、肝和肾组织结构的影响[J].水产学报,2015,39(2):233-244.Zhang W X,Sun S M,Ge X P,et al.Acute effects of ammonia exposure on histopathology of gill,liver and kidney in juvenile Megalobrama amblycephala and the post-exposure recovery[J].Journal of Fisheries of China,2015,39(2):233-244(in Chinese).
[5]陈家长,臧学磊,胡庚东,等.氨氮胁迫下罗非鱼(GIFTOreochromis niloticus)机体免疫力的变化及其对海豚链球菌易感性的影响[J].生态环境学报,2011,20(4):629-634.Chen J Z,Zang X L,Hu G D,et al.The immune response of GIFT Oreochromis niloticus and its susceptibility to Streptococcus iniae under stress in different ammonia[J].Ecology and Environmental Sciences,2011,20(4):629-634(in Chinese).
[6]韩英,张辉,王琨.亚硝态氮对鲤鱼种血液SOD及GSH-Px的影响[J].淡水渔业,2007,37(1):66-68.Han Y,Zhang H,Wang K.Activity of SOD and GSH-Px in the blood of fingerling Cyprinus carpio under different nitrite concentration[J].Freshwater Fisheries,2007,37(1):66-68(in Chinese).
[7]Dunier M,Siwicki A K,Dema?l A.Effects of organophosphorus insecticides:effects of trichlorfon and dichlorvos on the immune response of carp(Cyprinus carpio):Ⅲ.In vitro effects on lymphocyte proliferation and phagocytosis and in vivo effects on humoral response[J].Ecotoxicology and Environmental Safety,1991,22(1):78-87.
[8]洪美玲,陈立侨,孙新谨,等.亚硝酸盐急性胁迫对中华绒螯蟹幼体相关免疫指标和应激蛋白(HSP70)表达的影响[J].应用与环境生物学报,2011,17(5):688-693.Hong M L,Chen L Q,Sun X J,et al.Effects of acute nitrite exposure on immunity indicators and HSP70expression in Chinese mitten-hand crab(Eriocheir sinen- sis)[J].Chinese Journal of Applied&Environmental Biology,2011,17(5):688-693(in Chinese).
[9]周鲜娇,邱德全.亚硝酸氮和副溶血弧菌对凡纳滨对虾部分免疫指标的影响[J].水生态学杂志,2009,2(1):49-53.Zhou X J,Qiu D Q.Effects of nitrite nitrogen and Vibrio parahaemolyticu on some immunity indicators of Litopenaeus vannamei[J].Journal of Hydroecology,2009,2(1):49-53(in Chinese).
[10]邓嫔,刘威,李小明,等.pH、ORP监控在亚硝酸型生物脱氮过程中的应用[J].环境科学与技术,2007,30(3):97-99.Deng P,Liu W,Li X M,et al.Application of pH,ORPon-line monitor in shortcut biological nitrogen removal process[J].Environmental Science&Technology,2007,30(3):97-99(in Chinese).
[11]Pagacova P,Blstakova A,Drtil M,et al.Continually measured orp and Ph signal for control of nitrogen removal[M]//Hlavinek P,Bonacci Q,Marsalek J,et al.Dangerous Pollutants(Xenobiotics)in Urban Water Cycle.Dordrecht:Springer,2008:287-296.
[12]Bastian T,Brondum J.Do traditional measures of water quality in swimming pools and spas correspond with beneficial oxidation reduction potential?[J].Public Health Reports,2009,124(2):255-261.
[13]Hanaki K,Wantawin C,Ohgaki S.Nitrification at low levels of dissolved oxygen with and without organic loading in a suspended-growth reactor[J].Water Research,1990,24(3):297-302.
[14]Hellinga C,Schellen A A J C,Mulder J W,et al.The SHARON process:an innovative method for nitrogen removal from ammonium-rich waste water[J].Water Science and Technology,1998,37(9):135-142.
[15]杨琴,谢淑云.BP神经网络在洞庭湖氨氮浓度预测中的应用[J].水资源与水工程学报,2006,17(1):65-70.Yang Q,Xie S Y.Application of BP neural network into predicting NH3-N concentration of Dongting Lake[J].Journal of Water Resources&Water Engineering,2006,17(1):65-70(in Chinese).
[16]乔俊飞,马士杰,许进超.基于递归RBF神经网络的出水氨氮预测研究[J].计算机与应用化学,2017,34(2):145-151.Qiao J F,Ma S J,Xu J C.The water ammonia nitrogen prediction research based on a recursive RBF neural network[J].Computers and Applied Chemistry,2017,34 (2):145-151(in Chinese).
[17]冯红利,刘秀红,杨庆,等.低溶解氧下氨氧化过程神经网络预测控制模型[J].中国环境科学,2017,37(1):139-145.Feng H L,Liu X H,Yang Q,et al.Neural network prediction and control model for ammonia oxidizing process under low DO concentration[J].China Environmental Science,2017,37(1):139-145(in Chinese).
[18]管秋,王万良,徐新黎,等.基于神经网络的污水处理指标软测量研究[J].环境污染与防治,2006,28(2):156-158.Guan Q,Wang W L,Xu X L,et al.Soft-sensing method development based on neural network for wastewater treatment[J].Environmental Pollution and Control,2006,28(2):156-158(in Chinese).
[19]汪翔,何吉祥,佘磊,等.基于NAR神经网络对养殖水体亚硝酸盐预测模型的研究[J].渔业现代化,2015,42(4):30-34.Wang X,He J X,She L,et al.Establishment of aquaculture water nitrite prediction model based on NARneural network[J].Fishery Modernization,2015,42(4):30-34(in Chinese).
[20]林志贵,姚芳琴,冯林强,等.结合自适应遗传算法与弹性BP神经网络的亚硝酸盐预测模型[J].天津工业大学学报,2015,34(3):67-72.Lin Z G,Yao F Q,Feng L Q,et al.Nitrite prediction model based on adaptive genetic algorithm and elastic BP neural network[J].Journal of Tianjin Polytechnic University,2015,34(3):67-72(in Chinese).
[21]杨磊.水中亚硝胺类消毒副产物生成规律及其前质去除方法研究[D].哈尔滨:哈尔滨工业大学,2010.Yang L.Research on N-nitrosamines formation during disinfection and removal methods of their precursors[D].Harbin:Harbin Institute of Technology,2010.(in Chinese)
[22]郭宁,张建,孔强,等.温度对亚硝化及氧化亚氮释放的影响[J].环境工程学报,2013,7(4):1308-1312.Guo N,Zhang J,Kong Q,et al.Effect of temperature on ammoxidation and nitrous-oxide emission in partial nitrification[J].Chinese Journal of Environmental Engineering,2013,7(4):1308-1312(in Chinese).