毕业论文（设计）紫外连续可见光谱的水质多组分检测和预警.docVIP

摘 要 由于水污染日益加重，水质检测的方法和技术也得到了快速的进步，而新兴的光谱技术由于操作简便，无二次污染，测量精度高等优点加以广泛应用。本文首先阐述了课题的研究背景和意义，其次介绍国内外各种处理连续光谱检测技术的方法，重点归纳总结了多种建模方法，如线性方法中的最小二乘法，岭回归法，偏最小二乘法，非线性方法中的反向传播神经网络法，基于统计学支持向量机法，最小二乘支持向量机法。在此介绍过程中，对各种建模方法进行了一定的论证，并比较相关建模方法的优缺点。 第五章通过实验的方式，用偏最小二乘法检测水中污染物：溶解有机物（TOC）、硝酸盐（NO3）和浊度（NT）的浓度，再利用正交水平实验全面评估各种方法对多种物质混合后的实验效果。最后提出了两种新事物预警方法：主元分析与正交投影预测法及高斯模型的新污染物预警方法来对污水中出现的新事物进行预警。 关键词：光谱分析，水质检测，多种建模方法，未知事物预测 Abstract As the water pollution becomes more and more heavily, the testing methods of water quality and techniques for estimating water parameters have been rapidly progressed. It has many advantages such as easy to operate, no secondary pollution, and high precision measurement ,so the emerging spectrum technology is widely used. This paper first describes the research background and significance, followed by a variety of domestic and foreign processing continuous spectrum detection technology, focusing summarizes a variety of modeling methods. It describes the linear methods such as the least squares, ridge regression and the partial least squares and the nonlinear methods such as the back-propagation neural network method, ,the support vector machine method based on the statistical and the Least Squares Support Vector Machines. In the process of the introduction, the various modeling methods were demonstrated and it compares the advantages and disadvantages of modeling methods. We used the partial least squares method to detects contaminants in water: soluble organics (TOC), nitrate (NO3) and turbidity (NT) in charter 5 and used the concentration levels of recycling orthogonal experiment to evaluate the experimental results of mixed substances in varieties of methods. In the last we proposed two methods of new things warning to predict new things in the water including principal component analysis and orthogonal projection method and a new pollutant warning method based on the Gaussian model. Keywords: spectroscopy, water quality testing, a variety of modeling metho