Scholarlink

<p>This study aims to measure the concentrations of chemical oxygen demand (COD), nitrate, ammonia nitrogen, total nitrogen (TN), and total phosphorus (TP) in the main treatment stages of an urban wastewater treatment plant (WWTP) using a fused spectral technique integrating three-dimensional fluorescence spectroscopy and UV-Vis absorption spectroscopy, combined with a single-parameter feature selection method. Three-dimensional fluorescence spectra (Excitation–Emission–Matrix Spectra, EEMs) and UV-Vis absorption spectra of water samples were collected at the inlet, anaerobic tank, anoxic tank, aerobic tank, biochemical tank, secondary sedimentation tank, high-efficiency sedimentation tank, deep-bed filter, and effluent of the WWTP. Then, pretreatment methods, including standard normal variate (SNV), total normalization (Total), maximum normalization (Max), and minmax normalization (Min-Max) were applied to the original spectra, including raw three-dimensional fluorescence spectra and raw UV-Vis absorption spectra, to facilitate spectral data fusion. Given the different responses of various water quality parameters to fluorescence and absorption spectra, multiple feature selection methods — including competitive adaptive reweighted sampling (CARS), random frog (RF), uninformative variable elimination (UVE), and successive projections algorithm (SPA) — were employed to extract the characteristic fluorescence and absorption wavelengths for each water quality parameter. The results show that spectral preprocessing based on SNV and a characteristic wavelength extraction method based on CARS had the best prediction results for multi-parameter water quality in the key processes of the WWTP. Among them, the coefficients of determination (R2) between the predicted values and the actual values of COD, nitrate, ammonia nitrogen, TN, and TP all exceed 0.95. This method can provide theoretical guidance for the rapid diagnosis of operational stability in WWTPs.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-Parameter Water-Quality Monitoring in Wastewater Treatment Processes Based on Multi-Source Spectroscopy and Single-Parameter Analysis

  • 期刊论文

  • 发表日期：

    2026年7月22日

查看全文
  • Xiaowei Chen,
  • Siru Li,
  • Fei Wang,
  • Zhao Cheng,
  • Nanjing Zhao,
  • Gaofang Yin

摘要

This study aims to measure the concentrations of chemical oxygen demand (COD), nitrate, ammonia nitrogen, total nitrogen (TN), and total phosphorus (TP) in the main treatment stages of an urban wastewater treatment plant (WWTP) using a fused spectral technique integrating three-dimensional fluorescence spectroscopy and UV-Vis absorption spectroscopy, combined with a single-parameter feature selection method. Three-dimensional fluorescence spectra (Excitation–Emission–Matrix Spectra, EEMs) and UV-Vis absorption spectra of water samples were collected at the inlet, anaerobic tank, anoxic tank, aerobic tank, biochemical tank, secondary sedimentation tank, high-efficiency sedimentation tank, deep-bed filter, and effluent of the WWTP. Then, pretreatment methods, including standard normal variate (SNV), total normalization (Total), maximum normalization (Max), and minmax normalization (Min-Max) were applied to the original spectra, including raw three-dimensional fluorescence spectra and raw UV-Vis absorption spectra, to facilitate spectral data fusion. Given the different responses of various water quality parameters to fluorescence and absorption spectra, multiple feature selection methods — including competitive adaptive reweighted sampling (CARS), random frog (RF), uninformative variable elimination (UVE), and successive projections algorithm (SPA) — were employed to extract the characteristic fluorescence and absorption wavelengths for each water quality parameter. The results show that spectral preprocessing based on SNV and a characteristic wavelength extraction method based on CARS had the best prediction results for multi-parameter water quality in the key processes of the WWTP. Among them, the coefficients of determination (R2) between the predicted values and the actual values of COD, nitrate, ammonia nitrogen, TN, and TP all exceed 0.95. This method can provide theoretical guidance for the rapid diagnosis of operational stability in WWTPs.