摘要
通过对养殖池中影响溶氧变化因子的分析,选择溶氧、水深(cm)…水温度等9项因子作为输入参数,建立了溶氧预测模型。在该模型中,采用LMBP算法对BP神经网络进行优化,解决了BP神经网络的训练存在陷入局部最小点或训练速度慢等问题,提高了网络训练速度、保证了预测精度,具有较好的实用价值并可应用于其它水质因子的预测。
The prediction model of the dissolved oxygen is erected after analyzing the influencing factor of the dissolved oxygen in the culture pond,the value of dissolved oxygen,the depth & the temperature of the water etc.be chosen as the input parameter.Using the LMBP algorithm,the BP neural network is optimized,and the puzzles which the training of neural network in slow convergence and local minima is solved.The model is in application in other influencing factor of water quality with its fast convergence and high prediction accuracy.
出处
《自动化技术与应用》
2013年第10期1-3,41,共4页
Techniques of Automation and Applications
基金
辽宁省教育厅科研项目资助(编号L2010073)
辽宁省海洋与渔业厅科研项目资助(编号201006)
辽宁省海洋与渔业厅科研项目"先进控制技术在集约化水产养殖中的应用研究"资助(编号2011003)
关键词
溶解氧
神经网络
预测模型
优化算法
dissolved oxygen
neural network
prediction model
optimize algorithm