节点文献

金属离子对青霉素菌渣厌氧发酵产气模型分析

Different mathematical models in the analysis of the effect of trace elements on the anaerobic digestion of penicillin fermentation residues

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 方楠赵燕肖习彦花刘敬梁文华程辉彩张丽萍

【Author】 FANG Nan;ZHAO Yang-Xiao;XI Yang-Hua;WU Jian;LIANG Wen-Hua;CHENG Hui-Cai;ZHANG Li-Ping;Biology institute,Hebei Academy of Science;

【通讯作者】 程辉彩;

【机构】 河北省科学院生物研究所

【摘要】 为比较响应面法与反向传播神经网络法在厌氧发酵过程中的应用效果,以青霉素菌渣为原料,通过单因素和Box-Behnken法设计试验,在发酵体系中添加不同量的Fe2+、Co2+、Ni2+,以确定其对青霉素菌渣厌氧产气性能的影响.结果表明,Fe2+、Co2+、Ni2+单一最佳添加量为:500mg/L、30mg/L、0.3mg/L,产沼气量较对照分别提高了:102.18%、45.48%、60.12%.其促进作用随添加浓度增大呈现:弱-强-弱趋势.使用响应面法及反向传播神经网络法对金属离子添加量进行建模优化,并使用批式厌氧发酵进行验证.响应面法建模预测Fe2+、Co2+、Ni2+最佳混合添加浓度为:440.94mg/L、16.22mg/L、0.39mg/L,预测累积产沼气量为1314.49mL,R2=0.972,试验与验证相对误差为4.65%;反向传播神经网络法建模Fe2+、Co2+、Ni2+最佳混合添加浓度为495mg/L、21mg/L、0.5mg/L,预测产沼气量为1551.55mL,R2=0.991,试验与验证相对误差为0.47%.反向传播神经网络法建模具有更好的拟合效果且与验证试验误差小,是一种更有效的仿真方法.说明该方法在优化厌氧发酵金属离子添加具有应用潜力,同时也为厌氧发酵条件优化提供新思路.

【Abstract】 In order to testify the element efficiency and compare the application of different models in anaerobic digestion of the penicillin fermentation residues,the optimized test was designed by Box-Behnken,modeled by the response surface methodology(RSM)and the backpropagation neural network(BP-ANN).Meanwhile,the optimal concentration of single and mixture was determined.The optimal concentrations of single trace elements of Fe2+,Co2+ and Ni2+ were 500mg/L,30mg/L and 0.3mg/L,which the accumulated biogas production was increased by 102.18%,45.48% and 60.12% compared with the control.The variation of gas promotion was"weaker-stronger-weaker"with the increasing concentration of different trace elements.The optimal mixture concentration of Fe2+,Co2+ and Ni2+ predicted by the RSM model were 440.94mg/L,16.22mg/L and 0.39mg/L,R2=0.972,which the prediction of accumulated biogas production was 1314.49mL,and the experimental value was 1256mL with the RE(relative error)=4.65%.Meanwhile,the best-adding concentration of Fe2+,Co2+ and Ni2+ predicted by BP-ANN mode was 495mg/L,21mg/and 0.5mg/L,R2=0.991,which prediction of accumulated biogas production was 1551.55mL and the experimental value was 1358mL with the RE=0.47%.Compared with the RSM model,the BP-ANN model had a better fit and robust prediction of adding trace elements in anaerobic digestion of the penicillin fermentation residues,which showed that the BP-ANN model had prudential application in the optimization of trace element in anaerobic digestion,and it also provides a new idea to the optimization of anaerobic digestion.

【基金】 河北省自然基金资助项目(C2020302004);河北省高层次人才资助项目(E2015100006,A201802014);河北省科学院高层次人才培养与资助项目(2018G05);河北省科学院科技计划项目(18301,19301)
  • 【文献出处】 中国环境科学 ,China Environmental Science , 编辑部邮箱 ,2020年07期
  • 【分类号】X787
  • 【被引频次】1
  • 【下载频次】165
节点文献中: 

本文链接的文献网络图示:

本文的引文网络