Decision tree for identification and prediction of filamentous bulking at full-scale activated sludge wastewater treatment plant

Nashia Deepnarain, Mahmoud Nasr, Sheena Kumari, Thor A. Stenström, Poovendhree Reddy, Kriveshin Pillay, Faizal Bux

Research output: Contribution to journalArticle

Abstract

This study attempted to model sludge bulking in a full-scale wastewater treatment plant operated as a 3-stage Phoredox process. Principal component analysis and a regression tree model were employed to describe the correlations between influent wastewater characteristics and operational conditions (as inputs) and sludge volume index (SVI) as an output. A classification tree model was used to determine the environmental factors that affected the proliferation of filamentous microorganisms. Fluorescent in situ hybridisation analysis identified filamentous species of Microthrix parvicella, Thiothrix I & II, and Eikelboom Types 0041, 0092, and 021 N. It was found that SVI increased with an increment in sludge retention time, but it negatively correlated with soluble chemical oxygen demand (sCOD) and ammonium-nitrogen. The dominance of Microthrix parvicella was observed with a decline in temperature below 15.5 °C, causing an increase in SVI during the winter and spring seasons. The overgrowth of Thiothrix could be linked to the unbalanced ratio between readily biodegradable COD and nutrient species. The filament Type 0092 contributed to high SVI, and it prevailed with a decrease in food-to-microorganisms ratio below 0.08 1/d. Based on the satisfactory training, validation, and generalization procedures, the proposed models could be applied for the prediction of sludge bulking episodes.

Original languageEnglish
Pages (from-to)25-34
Number of pages10
JournalProcess Safety and Environmental Protection
Volume126
DOIs
Publication statusPublished - 1 Jun 2019

Fingerprint

Decision trees
Wastewater treatment
activated sludge
sludge
prediction
Microorganisms
Chemical oxygen demand
Ammonium Compounds
Principal component analysis
Nutrients
microorganism
Wastewater
Nitrogen
spring (season)
decision
wastewater treatment plant
chemical oxygen demand
principal component analysis
environmental factor
ammonium

Keywords

  • Classification and regression trees
  • Filamentous detection
  • Sludge bulking
  • Sludge volume index
  • Wastewater characteristics

ASJC Scopus subject areas

  • Environmental Engineering
  • Environmental Chemistry
  • Chemical Engineering(all)
  • Safety, Risk, Reliability and Quality

Cite this

Decision tree for identification and prediction of filamentous bulking at full-scale activated sludge wastewater treatment plant. / Deepnarain, Nashia; Nasr, Mahmoud; Kumari, Sheena; Stenström, Thor A.; Reddy, Poovendhree; Pillay, Kriveshin; Bux, Faizal.

In: Process Safety and Environmental Protection, Vol. 126, 01.06.2019, p. 25-34.

Research output: Contribution to journalArticle

Deepnarain, Nashia ; Nasr, Mahmoud ; Kumari, Sheena ; Stenström, Thor A. ; Reddy, Poovendhree ; Pillay, Kriveshin ; Bux, Faizal. / Decision tree for identification and prediction of filamentous bulking at full-scale activated sludge wastewater treatment plant. In: Process Safety and Environmental Protection. 2019 ; Vol. 126. pp. 25-34.
@article{4975cb7779384f5d8a8d323dda4c6236,
title = "Decision tree for identification and prediction of filamentous bulking at full-scale activated sludge wastewater treatment plant",
abstract = "This study attempted to model sludge bulking in a full-scale wastewater treatment plant operated as a 3-stage Phoredox process. Principal component analysis and a regression tree model were employed to describe the correlations between influent wastewater characteristics and operational conditions (as inputs) and sludge volume index (SVI) as an output. A classification tree model was used to determine the environmental factors that affected the proliferation of filamentous microorganisms. Fluorescent in situ hybridisation analysis identified filamentous species of Microthrix parvicella, Thiothrix I & II, and Eikelboom Types 0041, 0092, and 021 N. It was found that SVI increased with an increment in sludge retention time, but it negatively correlated with soluble chemical oxygen demand (sCOD) and ammonium-nitrogen. The dominance of Microthrix parvicella was observed with a decline in temperature below 15.5 °C, causing an increase in SVI during the winter and spring seasons. The overgrowth of Thiothrix could be linked to the unbalanced ratio between readily biodegradable COD and nutrient species. The filament Type 0092 contributed to high SVI, and it prevailed with a decrease in food-to-microorganisms ratio below 0.08 1/d. Based on the satisfactory training, validation, and generalization procedures, the proposed models could be applied for the prediction of sludge bulking episodes.",
keywords = "Classification and regression trees, Filamentous detection, Sludge bulking, Sludge volume index, Wastewater characteristics",
author = "Nashia Deepnarain and Mahmoud Nasr and Sheena Kumari and Stenstr{\"o}m, {Thor A.} and Poovendhree Reddy and Kriveshin Pillay and Faizal Bux",
year = "2019",
month = "6",
day = "1",
doi = "10.1016/j.psep.2019.02.023",
language = "English",
volume = "126",
pages = "25--34",
journal = "Process Safety and Environmental Protection",
issn = "0957-5820",
publisher = "Institution of Chemical Engineers",

}

TY - JOUR

T1 - Decision tree for identification and prediction of filamentous bulking at full-scale activated sludge wastewater treatment plant

AU - Deepnarain, Nashia

AU - Nasr, Mahmoud

AU - Kumari, Sheena

AU - Stenström, Thor A.

AU - Reddy, Poovendhree

AU - Pillay, Kriveshin

AU - Bux, Faizal

PY - 2019/6/1

Y1 - 2019/6/1

N2 - This study attempted to model sludge bulking in a full-scale wastewater treatment plant operated as a 3-stage Phoredox process. Principal component analysis and a regression tree model were employed to describe the correlations between influent wastewater characteristics and operational conditions (as inputs) and sludge volume index (SVI) as an output. A classification tree model was used to determine the environmental factors that affected the proliferation of filamentous microorganisms. Fluorescent in situ hybridisation analysis identified filamentous species of Microthrix parvicella, Thiothrix I & II, and Eikelboom Types 0041, 0092, and 021 N. It was found that SVI increased with an increment in sludge retention time, but it negatively correlated with soluble chemical oxygen demand (sCOD) and ammonium-nitrogen. The dominance of Microthrix parvicella was observed with a decline in temperature below 15.5 °C, causing an increase in SVI during the winter and spring seasons. The overgrowth of Thiothrix could be linked to the unbalanced ratio between readily biodegradable COD and nutrient species. The filament Type 0092 contributed to high SVI, and it prevailed with a decrease in food-to-microorganisms ratio below 0.08 1/d. Based on the satisfactory training, validation, and generalization procedures, the proposed models could be applied for the prediction of sludge bulking episodes.

AB - This study attempted to model sludge bulking in a full-scale wastewater treatment plant operated as a 3-stage Phoredox process. Principal component analysis and a regression tree model were employed to describe the correlations between influent wastewater characteristics and operational conditions (as inputs) and sludge volume index (SVI) as an output. A classification tree model was used to determine the environmental factors that affected the proliferation of filamentous microorganisms. Fluorescent in situ hybridisation analysis identified filamentous species of Microthrix parvicella, Thiothrix I & II, and Eikelboom Types 0041, 0092, and 021 N. It was found that SVI increased with an increment in sludge retention time, but it negatively correlated with soluble chemical oxygen demand (sCOD) and ammonium-nitrogen. The dominance of Microthrix parvicella was observed with a decline in temperature below 15.5 °C, causing an increase in SVI during the winter and spring seasons. The overgrowth of Thiothrix could be linked to the unbalanced ratio between readily biodegradable COD and nutrient species. The filament Type 0092 contributed to high SVI, and it prevailed with a decrease in food-to-microorganisms ratio below 0.08 1/d. Based on the satisfactory training, validation, and generalization procedures, the proposed models could be applied for the prediction of sludge bulking episodes.

KW - Classification and regression trees

KW - Filamentous detection

KW - Sludge bulking

KW - Sludge volume index

KW - Wastewater characteristics

UR - http://www.scopus.com/inward/record.url?scp=85064001618&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064001618&partnerID=8YFLogxK

U2 - 10.1016/j.psep.2019.02.023

DO - 10.1016/j.psep.2019.02.023

M3 - Article

VL - 126

SP - 25

EP - 34

JO - Process Safety and Environmental Protection

JF - Process Safety and Environmental Protection

SN - 0957-5820

ER -