A COD measurement method with turbidity compensation based on a variable radial basis function neural network.

Area of Science:

• Environmental Science
• Analytical Chemistry
• Spectroscopy

Background:

• Ultraviolet-visible (UV-vis) spectrometry is commonly used for measuring chemical oxygen demand (COD) in sewage.
• Existing UV-vis methods struggle to eliminate turbidity's influence, affecting measurement accuracy.

Purpose of the Study:

• To develop a new, accurate, and efficient method for measuring sewage COD using UV-vis spectroscopy.
• To address the challenge of turbidity interference in spectroscopic COD analysis.

Main Methods:

• A novel turbidity compensation algorithm utilizing principal component analysis (PCA) to identify characteristic wavelengths.
• Development of a turbidity compensation model to remove spectral interference from turbidity.
• Application of a variable radial basis function (VRBF) neural network for COD concentration measurement.

Main Results:

• The proposed method significantly improved the coefficient of determination (R²) from 0.27 to 0.88 compared to traditional partial least squares regression.
• Root-mean-square deviation decreased from 5.56 to 1.69, indicating enhanced accuracy.
• The VRBF neural network approach demonstrated faster, more direct, and accurate COD measurements than improved bagging and MLP algorithms.

Conclusions:

• The new UV-vis spectroscopy method with turbidity compensation and VRBF neural network provides a superior approach for sewage COD measurement.
• This method overcomes the limitations of existing techniques by effectively handling turbidity.
• The findings suggest a promising advancement for real-time and reliable water quality monitoring.