| 赵润,牟美睿,王鹏,孙迪,刘海学,张克强,杨仁杰.基于近红外漫反射光谱的规模化奶牛场粪水氮磷定量分析及模型构建[J].农业环境科学学报,2019,38(8):1768-1776. |
| 基于近红外漫反射光谱的规模化奶牛场粪水氮磷定量分析及模型构建 |
| Quantified analysis and model structure of nitrogen and phosphorus in the slurry of large-scale dairy farms based on near-infrared transmission-diffuse reflectance spectroscopy |
| 投稿时间：2019-04-16 |
| DOI：10.11654/jaes.2019-0415 |
| 中文关键词: 规模化奶牛场 粪水 近红外漫反射光谱 主成分分析 偏最小二乘 总氮 总磷 定量分析模型 |
| 英文关键词: large-scale dairy farm slurry near-infrared diffuse reflection spectrum(NIDRS) principal component analysis(PCA) partial least squares (PLS) total nitrogen (TN) total phosphorus (TP) quantified analysis model |
| 基金项目:国家重点研发计划项目（2018YFD0800100）；天津市现代奶牛产业技术体系创新团队建设专项（ITTCRS2017006）；中央级公益性科研院所基本科研业务费专项项目（Y2019GH14）；国家自然科学基金项目（41771357，21607114，81471698）；天津市自然科学基金项目（18JCYBJC96400，16JCQNJC08200） |
|
| 摘要点击次数: 1916 |
| 全文下载次数: 1780 |
| 中文摘要: |
| 为建立规模化奶牛场粪水中氮磷含量现场快速检测方法，以实现准确预测的同时替代常规监测程序，选取23家天津市典型种养结合模式的规模化奶牛场，围绕粪水处理全过程环节依次开展样品采集、实验室常规化学检测、近红外漫反射光谱采集，并进行主成分分析和偏最小二乘分析，建立多种动态复合影响因素条件下的全局、全程快速检测定量分析模型。结果表明：主成分分析不仅反映出同一奶牛场粪水有机组分随处理环节的变化，而且也反映出不同奶牛场粪水样品的差异性，以及在粪水处理过程中各因素对后续模型的影响程度。建立的全过程环节定量分析模型对总氮含量预测结果与实际含量的线性拟合相关系数R为0.96，预测均方根误差RMSEP为187.80；对总磷含量预测结果与实际含量的线性拟合相关系数R为0.91，预测均方根误差RMSEP为3.59。建立的全局定量分析模型对总氮含量预测结果与实际含量的线性拟合相关系数R为0.96，预测均方根误差RMSEP为238.59；总磷含量预测结果与实际含量的线性拟合相关系数R为0.91，预测均方根误差RMSEP为6.56。研究表明，基于近红外漫反射光谱和偏最小二乘法对规模化奶牛场粪水处理全过程环节粪水样品中氮、磷含量进行定量分析是可行的；纵向模型比横向模型能提供更好的预测结果；近红外漫反射光谱技术可实时、快速、高效地对规模化奶牛场粪水处理全过程总氮和总磷进行跟踪和监控。 |
| 英文摘要: |
| In order to establish an on-spot rapid testing method for determining the nitrogen and phosphorus contents in the slurry produced from large-scale dairy farms, for accurate prediction meanwhile replacing the regular monitoring procedure, we selected 23 typical large-scale dairy farms in the combination of growing and breeding for experimental analysis. Samples were collected throughout the entire course of slurry treatment, at both the vertical and horizontal levels, and then subjected to chemical and statistical analyses, i.e. near-infrared diffuse reflection spectrum (NIDRS), principal component analysis (PCA), and partial least squares (PLS) analysis. Under the conditions of multiple dynamic composite impacting factors, we established a rapid testing qualified analysis model for the global and entire process of slurry treatment. The results indicated that the organic components of slurry not only varied among the different stages of treatment on individual dairy farms, but also differed among the different farms, which were given by the PCA. Furthermore, we found that factors associated with the different stages of the slurry treatment procedure, which influenced the models, were reflected by the PCA. The linear fitting correlation coefficient (R) for the relationship between predicted and actual content of total nitrogen (TN) and total phosphorus (TP), during the establishment of the quantified analysis model and through the entire course and links, were 0.96 and 0.91, respectively. The root mean square errors of the predictions were 187.80 and 3.59, respectively. The R values for the relationship between the predicted and actual contents of TN and TP, during the establishment of the quantified analysis model and through the entire course and links, were 0.96 and 0.91, respectively, whereas the root mean square errors of the prediction were 238.59 and 6.56, respectively. In this study, we therefore demonstrated the feasibility of performing a quantitative analysis of the TN and TP content in slurry samples collected at the vertical and horizontal levels of the entire slurry treatment process based on NIDRS and PLS. Better prediction result was provided by the entire process model compared to the global model. Accordingly, TN and TP concentration over the entire course of the slurry treatment process on largescale dairy farms can be rapidly and efficiently traced and monitored in real time via the NIDRS technology. |
| HTML
查看全文
查看/发表评论 下载PDF阅读器 |
|
|
|
