APP下载

基于改进正余弦优化神经网络的水体余氯预测

2021-04-27安小宇鲁奎豪崔光照赵复兴

人民黄河 2021年4期

安小宇 鲁奎豪 崔光照 赵复兴

摘 要:为保证居民用水安全,搭建更加准确、稳定的水质余氯预测模型,提出了一种改进正余弦算法优化神经网络的水体余氯预测模型。在正余弦算法(SCA)位置变化中加入非线性权重,平衡全局勘探和局部开发能力;同时在粒子更新后融入Levy飞行,丰富种群多样性,强化局部搜索能力;将改进正余弦算法用于BP神经网络参数优化,搭建ISCA-BP水体余氯预测模型。利用在线水质监测设备的监测数据进行余氯含量预测,不同预测模型结果对比表明,ISCA-BP模型预测结果相对误差的平均值为4.04%,均方根误差为0.011 3 mg/L,与BP、RBF神经网络模型相比,误差最小,模型预测结果与实际值最接近,泛化能力更强,预测结果精确度高,稳定性好。

关键词:余氯;ISCA-BP神经网络;正余弦算法;Levy飞行;非线性权重

中图分类号:TP183;X824 文献标志码:A

doi:10.3969/j.issn.1000-1379.2021.04.020

引用格式:安小宇,鲁奎豪,崔光照,等.基于改进正余弦优化神经网络的水体余氯预测[J].人民黄河,2021,43(4):111-116.

Abstract: In order to ensure the safety of residents water use, a more accurate and stable prediction model of residual chlorine in water quality was established. In this paper, an improved sine cosine algorithm was presented to optimize the prediction model of neural network. Add nonlinear weight to the position change of Sine Cosine Algorithm (SCA) to balance the global exploration and local development capability; at the same time, Levy flight was integrated after particle updating to enrich the diversity of the population and enhancing local search capabilities; then the improved sine and cosine algorithm was used for BP network parameter optimization to build the ISCA-BP prediction model. Finally, using the online water quality testing equipment for prediction of residual chlorine content and comparing different results of forecasting models, it found that the results of relative error of ISCA-BP model predicted is 4.04%, the average root mean square error is 0.011 3 mg/L, the error is lower than BP and RBF model. The prediction results of the model are the closest to the true value, the generalization ability of the ISCA-BP model is stronger, the accuracy of the prediction results is high and the stability is good.

Key words: residual chlorine; ISCA-BP neural network; sine and cosine algorithm; Levy flight; nonlinear weight

人口數量逐年递增、用水量日益增加使得用水量增加,与此同时人类生产和生活中污水排放导致水体污染、水质恶化,水资源更加紧张。供水公司通常利用二氧化氯杀死水体中的有害微生物达到消毒目的,在消毒的同时二氧化氯与水管壁上附着的生物膜发生反应,导致管网中因余氯含量降低而不能达到消毒目的。但余氯含量过高时,会对人体带来一定损害[1]。因此,通过预测余氯变化情况,精确控制加氯量,确保日常饮用水安全是当前研究的热点。……

登录APP查看全文