摘要
为预测船舶压载水微生物数量,以压载水的温度、导电度、浊度、p H值、溶氧量、含盐量、总固体溶解量等7个常规数据为输入信号,以大肠杆菌量和肠球菌量为预测参数构建BP神经网络预测模型,并通过遗传算法对BP神经网络初始权值和阈值实施优化,加快模型的收敛速度,减少迭代次数,增强全局搜索能力.预测结果与实际样本的比较结果表明:基于遗传算法改进的BP神经网络有效提高了模型的收敛速度和预测精度,为船舶压载水处理系统的开发和维管提供了技术支撑.
In order to predict the quantity of the microbial indicators in marine ballast water, a BP neural network prediction model is constructed in this study. In the model, seven routine data, which are the temperature, conductivity, turbidity, pH, dissolved oxygen, salinity, total dissolved solids, are taken as the input signals, and the amount of Escherichia coli and Enterococcus as the predicting parameters. By means of the genetic algorithm, the initial weights and threshold of BP neural network are optimized. The convergence rate of the model is accelerated, the number of iteration steps is reduced, and the global search ability is enhanced. The actual test results show that the method has high accuracy and precision, and provide technical support for the development and maintenance of the ship ballast water treatment system.
作者
王琪
乔红宇
WANG Qi QIAO Hongyu(Department of Marine Engineering, Nantong Vocational ＆ Technical Shipping College, Nantong 226010, China)
出处
《江苏科技大学学报（自然科学版）》
CAS
北大核心
2016年第6期529-537,共9页
Journal of Jiangsu University of Science and Technology:Natural Science Edition
基金
江苏省交通运输科技项目(2014C03)
江苏省航海学会科研项目(2015B02)
关键词
船舶压载水
遗传算法
BP神经网络
微生物
数量预测
ballast water, genetic algorithm, BP neural network, microbial indicators, quantity prediction
