人工神经网络优化发酵过程中氨气释放条件  

Optimization Condition of Ammonium Release of Aerobic Fermentation by Artificial Neural Network

作　　者：景晓忠[1] 严红[1] 王帆[1] 

机构地区：[1]大连大学环境与化学工程学院,辽宁大连116622

出　　处：《安徽农业科学》2011年第34期21278-21279,共2页Journal of Anhui Agricultural Sciences

摘　　要：在正交试验的基础上,运用人工神经网络对发酵过程中影响氨气释放的条件进行优化。该模型预测的最佳参数条件为:含水率为60%,C/N为37∶1,pH为7,温度为31.3℃,此时总氨气量的预测值为1 149.3 mg/kg。BP神经网络模型的预测值和实测值相差不大,最大相对误差为6.58%,说明该模型具有较高的预测精度。On the basis of orthogonal test,the artificial neural networks were used to predict ammonium release in the aerobic fermentation,The best process parameters obtained by the artificial neural network model were listed as following：moisture 60%,C/N 37∶ 1,pH 7 and temperature 31.3 ℃.Total ammonia amount was 1 149.3 mg/kg.Through the analysis we could see that the forecast value and the actual measured value of BP neural network model had not big difference,the maximum relative error was 6.58%,it indicated that the model had higher forecast accuracy.

关 键 词：人工神经网络 好氧发酵 氨气释放 影响因素 

分 类 号：X171.4[环境科学与工程—环境科学]