טכניון מכון טכנולוגי לישראל
הטכניון מכון טכנולוגי לישראל - בית הספר ללימודי מוסמכים  
M.Sc Thesis
M.Sc StudentFrolich Levi David
SubjectSpectral Methods for Analysis of Bacteria Counts in Water
Wells
DepartmentDepartment of Civil and Environmental Engineering
Supervisor Professor Barak Fishbain
Full Thesis textFull thesis text - English Version


Abstract

Maintaining water quality is critical for any water supply company. One of the critical issues in water quality assurance is bacteria growth in water wells. To date, bacteria growth models cannot predict with sufficient accuracy when a bacteria outburst will occur. Often, time series of bacteria count are sparse, which means that they contain many zero elements. This is a result of both the natural growth of bacteria and the limited capabilities to frequently sample the water. The sparsity of the time series hinders the observation of deviations from normal behavior. This essentially precludes the application of many statistical quality control methods for the detection of high bacteria counts before contamination occurs. This results in a subjective process where the appointed engineer must make an educated guess as to how the well behaves. This research developed a new cost-effective method that controls water quality in wells through objectively determining a water well’s normative state, under certain conditions. Looking at a well’s bacteria time series as a signal allows for the use of signal processing methods to analyze the time series. The hypothesis of this research is that a small number of frequency coefficients will express changes in counts and deviations from the norm in a vivid way. By performing a dimension reduction and utilizing spectral methods, noise can be removed and computation time can be improved. This research developed signal processing (i.e., spectral analysis) and machine learning (i.e., Support Vector Machine (SVM) and Minimum Volume Ellipsoid (MVE)) methods for predicting bacteria outbursts from the bacterial counts history of a well. The results show that these tools can be implemented by the water quality engineering community to create more robust quality control techniques to ensure safer water supply.