CEEES - Biological floc characterization via image analysis and machine learning

Wastewater treatment is based on biological flocs that must have good settling properties to work well.  A new treatment approach uses hydrocyclones to retain larger, denser flocs with good settling properties, and reject flocs with poor ones.  Yet hydrocyclone separation of flocs is poorly understood.  We are using flow cytometry, which collects thousands of individual floc images per sample, image analysis, and potentially machine learning to maximize the beneficial effects of hydrocyclones on floc size distributions.  We hypothesize that this can help understand the effect of hydrocyclones on flocs, and also predict floc settling properties, microbial communities, and contaminant degradation potential. There are several possible "spinoff" projects, which could include lab work, computer work, or both.  The research is funded by Veolia, a global water technology corporation.