兰州大学机构库 >数学与统计学院
基于混合数学模型的巢湖水质预测与评价
Alternative TitlePrediction and Evaluation of Chaohu Water Quality Based on Hybrid Mathematical Model
王静
Thesis Advisor李维德
2018-03-30
Degree Grantor兰州大学
Place of Conferral兰州
Degree Name硕士
Keyword水质预测 水质评价 灰狼优化算法 互补型集成经验模态分解 支持向量机
Abstract

近几年来,随着我国经济发生跨越式发展,水资源污染情况表现严重,管理和保护好各个地段的水质有着重要的意义。对水质进行正确评估是水资源保护的重要手段。 本文以长江流域安徽巢湖裕溪口断面和淮河流域合肥湖滨断面2008年至2017年上半年前26周的主要污染性水质参数溶解氧(DO)、氨氮(CODMn)、高锰酸盐(NH3-N)数据为研究对象,对水质进行预测和综合评价研究。考虑到水环境系统是复杂的非线性系统,研究过程中预测与评价指标会包含大量的非线性相关特征,因此提出了基于互补型集成经验模态分解和灰狼优化算法优化支持向量机的组合水质预测模型与T-S模糊神经网络的水质评价方法进行研究。数据实验证实了本文提出的算法在适用性及误差控制等多种性能上具有优越性。 在提出水质预测模型之前,文中介绍了互补型集成经验模态分解方法(CEEMD)、灰狼优化算法(GWO)以及支持向量机(SVM)。针对SVM建模时对预测精度影响的因素除原始序列自身噪声的影响外另一个关键因素是SVM中参数选择的问题,因此在建模过程中首先采用CEEMD方法解决由原始信号产生的白噪声对结果造成的影响,其次,采用灰狼优化算法对SVM进行参数优化。实证研究结果表明,所构建的算法较传统算法收敛速度更快,且GWO算法在SVM参数寻优中较其它算法效率更高。 在提出水质评价模型之前,文章介绍了人工神经网络(ANN)与模糊数学,并将二者有效结合起来,构建了T-S模糊神经网络水质评价模型。该评价模型将DO, CODMn, NH3-N作为模糊系统输入,得出实际观测值属于各类水质的隶属度,最终获得测试值的水质类别。由于模糊神经网络的每个节点都有实际含义,该模型很大程度克服了ANN理论的短板。将所构建模型运用到水质评价实证研究中具有较满意的解释性及可靠性。

Other Abstract

In recent years, with the rapid development on economy of China, there is a serious problem on water quality, which need to be monitored and protected. The correct assessment of water quality is important for the maintenance of water resources. In this paper, the Yuxikou section of Chaohu in the Yangtze river basin and the Binhu section of Heifei in the Huai river basin were used as the study areas to explore the issue of water quality. The main pollutants are dissolved oxygen (DO), ammonia nitrogen (CODMn) and permanganate (CODMn). The used data are weekly collected from the beginning of 2008 to the half of 2017. Considering that the water environment system is a complex nonlinear system, the prediction and the evaluation indicators in the research process will contain a large number of relevant nonlinear features. Therefore, a combined water quality model based on complementary integrated empirical mode decomposition and a support vector machines optimized by gray wolf optimization algorithm is proposed for prediction, and a T-S fuzzy neural network method is proposed for water quality assessment. The experimental results show that the algorithms presented in this paper are superior in applicability and error control. Before the water quality prediction model was put forward, the complementary integrated empirical mode decomposition (CEEMD) method, the gray wolf (GWO) optimization algorithm and the support vector machine (SVM) were introduced. Aimed at modeling using SVM, we must not only consider the noise influence of original sequence, but also consider the parameter selection. In the modeling process, CEEMD method is first used to solve the effect of the white noise generated by the original signal on the result. Then, using the gray wolf optimization algorithm to optimize the parameters of SVM, this algorithm converges faster than the traditional algorithms. Finally, we concluded that the GWO algorithm is more efficient than other algorithms in the SVM parameter optimization. Before presenting the water quality assessment model, the article introduced the artificial neural network (ANN) and fuzzy set theory, and effectively combined the two methods to build a T-S fuzzy neural network model for water quality assessment. The evaluation model uses DO, CODMn and NH3-N as a fuzzy system input to obtain the degree of membership of the actual observed values belonging to various types of water quality, and finally the water quality categories of test data are obtained. Because each node of the fuzzy neural network has practical meanings, it largely overcomes the shortcomings of ANN theory. Applying the proposed method to the assessment of water quality, it has satisfactory interpretability and reliability.

URL查看原文
Language中文
Document Type学位论文
Identifierhttps://ir.lzu.edu.cn/handle/262010/225162
Collection数学与统计学院
Recommended Citation
GB/T 7714
王静. 基于混合数学模型的巢湖水质预测与评价[D]. 兰州. 兰州大学,2018.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Altmetrics Score
Google Scholar
Similar articles in Google Scholar
[王静]'s Articles
Baidu academic
Similar articles in Baidu academic
[王静]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[王静]'s Articles
Terms of Use
No data!
Social Bookmark/Share
No comment.
Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.