Monitoring Wastewater Treatment Systems
Abstract: This work considers various techniques to extract information from the vast amount of data collected at a modern wastewater treatment plant. If the information extracted is to be considered reliable is highly dependent on the data screening. Data screening includes validation and quality improvement of data. Adequate methods for validation, noise reduction and other forms of quality improvements of wastewater treatment data are discussed. In order to detect deviations and disturbances, the measurement variables can be investigated individually or many variables simultaneously. Single variable detection involves investigation of the basic signal characteristics such as amplitude, mean and spread. Usable methods are discussed and examples are given. In order to detect synergetic effects, techniques capable of investigating several variables simultaneously, are needed. Multivariate statistics based methods, such as principal component analysis (PCA), principal component regression (PCR) and projection to latent structures (PLS), are considered and their applicability discussed. Some possibilities to adapt the methods to the dynamic situation in a wastewater treatment plant are also outlined.
CLICK HERE TO DOWNLOAD THE WHOLE DISSERTATION. (in PDF format)