[Research on Water Quality Analysis Model with PCA Method and UV Absorption Spectra]

Area of Science:

• Analytical Chemistry
• Environmental Science
• Spectroscopy

Background:

• UV absorption spectroscopy is crucial for real-time water quality monitoring of organic pollutants.
• Complex water matrices introduce uncertainties and offsets in standard measurements.
• Accurate detection of pollutants like Chemical Oxygen Demand (COD) is essential.

Purpose of the Study:

• To develop a robust method for classifying water samples based on UV absorption spectra.
• To accurately quantify organic pollutant concentrations, specifically COD, in water.
• To compare the effectiveness of Principal Component Analysis combined with Partial Least Squares Regression (PCA-PLSR) against direct multi-wavelength absorption models (MWA-PLSR).

Main Methods:

• Water samples were analyzed using UV absorption spectra (200-400 μm).
• Principal Component Analysis (PCA) and Euclidean distance were employed for sample classification.
• PCA-PLSR and MWA-PLSR models were developed and compared for COD concentration measurement.

Main Results:

• PCA effectively classified water samples based on their UV spectra.
• The PCA-PLSR method, utilizing the first and second principal components, achieved the smallest measurement errors (<5%).
• This approach demonstrated superior accuracy in measuring water quality parameters compared to MWA-PLSR.

Conclusions:

• Combining PCA with Euclidean distance enables accurate water sample classification.
• PCA-PLSR offers a highly accurate method for quantifying organic pollutants, such as COD, in water.
• The developed methodology enhances real-time online monitoring of water quality.