农药残留微流控光度检测恒程误差补偿方法研究

发布时间：2018-06-21 09:00

本文选题：农药残留 + 恒光程　；参考：《江苏大学》2017年硕士论文

【摘要】：目前主流的农药残留检测手段所涉及的检测设备普遍存在着体积庞大、检测精度低、稳定性差的缺陷。为此,课题基于微流控芯片搭建了便携式农药残留检测装置。然而所搭建的便携化农药残留检测装置存在下述两个问题:(1)存在着恒程误差的干扰;(2)缺少相应的硬件或软件误差补偿模型。针对上述问题,课题从以下几个方面展开工作:(1)基于朗伯比尔定律建立了恒程误差σ与光程长l之间的关系模型,并从影响其关系模型的噪声角度对所建立的关系模型进行了修正;(2)利用CorelDraw设计了旋转蝶式微流控芯片模型并借助COMSOL对其进行了仿真,确定了芯片的材料和制作方法,搭建了农药残留光度检测装置并优化了实验温度、光源光强、光源波长三个参数,利用农药残留光度检测装置验证了均匀指数仿真模型并建立恒程误差硬件补偿模型;(3)利用最小二乘法、BP神经网络以及支持向量机三种算法分别建立了相应的恒程误差软件补偿模型,选取最优模型作为最终恒程误差软件补偿模型;(4)搭建了基于微流控芯片便携式农药残留检测装置,介绍了便携式农药残留检测装置硬件电路设计,优化了底物的浓度值、酶抑制时间以及溶液的pH值三个参数。借助便携式农药残留检测装置并应用所建立的硬件和软件补偿模型对草莓和西红柿中敌百虫、呋喃丹、久效磷3种常见的有机磷或氨基甲酸酯类农药残留进行检测以验证所建立补偿模型的正确性。实验结果表明:经硬件和软件补偿模型补偿后的吸光度均方根误差RMSE分别为0.3901、0.0883,基本实现了农药残留检测恒程误差的补偿;与传统微流控农残光度检测方法相比,课题所述方法使草莓和西红柿中的敌百虫等3种常见有机磷或氨基甲酸酯类农药农残检测误差降低24.3%以上,基本实现较为精确的农药残留检测;在应用所建立的补偿模型后,课题所述的农药残留方法对草莓和西红柿中农药残留检测值与农药加标浓度值的相对误差分别为3.99%、4.58%、4.03%,基本验证了所建立补偿模型的正确性。
[Abstract]:At present, the detection equipment involved in the mainstream pesticide residue detection methods generally has the defects of large volume, low detection precision and poor stability. Therefore, a portable pesticide residue detection device based on microfluidic chip is built. However, the portable pesticide residue detection device has the following two problems: 1) there is constant range error interference / 2) there is a lack of hardware or software error compensation model. In order to solve the above problems, the following work is done: 1) based on Lambert's law, a model of the relationship between constant path error 蟽 and optical path length l is established. At the same time, the relationship model is modified from the angle of noise which affects its relation model. The model of rotating butterfly microfluidic chip is designed by CorelDraw and simulated by COMSOL. The material and fabrication method of the chip are determined. The experimental temperature, light intensity and wavelength of the light source were optimized. The uniform exponent simulation model and the hardware compensation model of constant range error are established by using the pesticide residue photometric detection device.) the corresponding constant range is established by using the least square BP neural network and the support vector machine respectively. Error software compensation model, Selecting the optimal model as the final constant range error software compensation model, the portable pesticide residue detection device based on microfluidic chip is built. The hardware circuit design of the portable pesticide residue detection device is introduced, and the concentration value of the substrate is optimized. The enzyme inhibition time and pH value of the solution were three parameters. With the help of portable pesticide residue detection devices and using the established hardware and software compensation models, furan, trichlorfon and furan in strawberries and tomatoes, Three common organophosphorus or carbamate pesticide residues were detected to verify the correctness of the compensation model. The experimental results show that the root mean square error (RMSE) of absorbance after compensating by hardware and software compensation model is 0.3901 / 0.0883respectively, which basically realizes the compensation of the constant path error of pesticide residue detection, and compared with the traditional microfluidic method, the RMSE of RMSE is 0.3901 / 0.0883.Compared with the traditional microfluidic method, The method can reduce the detection error of pesticide residues of three common organophosphorus or carbamate pesticides, such as trichlorfon in strawberry and tomato, by more than 24.3%. The relative error between the detection value of pesticide residue in strawberry and tomato and the concentration of pesticide in strawberry and tomato is 3.99 and 4.58 and 4.03, respectively, which basically verifies the correctness of the compensation model.
【学位授予单位】：江苏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S481.8;TP18

【参考文献】