On-line optical monitoring of activated sludge floc On-line optical monitoring of... · On-line optical…

  • Published on

  • View

  • Download

Embed Size (px)


  • On-line optical monitoring ofactivated sludge floc morphology

    17.9.2015Elisa Koivuranta

    Fibre and particle engineeringUniversity of Oulu


    Professor: Mirja IllikainenFiDiPro-professor: Kristiina OksmanStaff: 30

  • Background

    Activated sludge process is the most commonmethod in municipal and industrial wastewatertreatment

    The efficiency of the ASP is highly dependent onthe solid-liquid separation which is affected by thestructure and the strength of the flocs

    Evaluation and modeling of the wastewatertreatment process by image analysis methods willbring new horizons in the monitoring and controlapproaches

    Background: Activated sludge process


    Municipal and industrialwastewater contain usually highamounts of COD, nitrogen andphosphorous, which can beremoved by biologicalprocesses

    Biological flocs are formed inaeration tanks and removed bysettling in settling tank

  • Background: Floc formation

    The floc formation is complex phenomenon whichmechanisms are still poorly understood

    Flocs are made up of a variety of microorganismsas well as organic and inorganic particles

    Different disturbances in the process has effectson the floc structure

    Unsatisfactory floc formation is a common problemin activated sludge plant worldwide which haveboth environmental and economic consequences

    Background: Floc morphology indifferent situations

    Lots of filaments Small flocs Big flocs, enoughfilaments

  • Background

    Floc formation and settling properties of activated sludge are criticalfor the purification process

    There is still lack of a proper on-line measurement of flocculation

    Optical monitoring of flocs wouldenable better control of the wholeactivated sludge process


    Floc imaging device - MOFI

    From laboratory scale toin situ scale

    Basic principle is thesame in both:Only needed pre-treatment is dilution

    Image size is 5.01mm x3.74mm and imaged pixelsize is 3.6m

  • Image analysis

    (1) the original image; (2) background removal; (3) adjustment of image intensityvalues to increase contrast; (4) creating negative image; (5) median filtering; (6)thresholds to create binary images; (7) filling holes; (8) dilation; (9) filamentimage; (10) floc image; (11) processed floc image; and (12) processed filamentimage.

    Image analysis

    Blue: Accpeted flocs Red: Small particleGreen: Too light Purple: Air bubble

  • Image analysis

    Blue: Flocs Red: Filaments

    Image analysis

    Analyzing:- Form factor- Roundness- Equivalent diameter- Aspect ratio- Total filament

    length/total floc area- Mean area of

    particles- Amount of small

    particles (

  • Preliminary tests

    Two different situation were compared Bigger flocs, less filaments better settling properties

    Preliminary tests

    Evaluation of the flocmorphology in the industrialwastewater treatment plant

  • On-line optical monitoringAt Viikinmki

    Sludge is taken from theprocess line into deairingcontainer

    From the container thesample is diluted andpumped through cuvetteof imaging unit

    The sample dilution,pumping, image grabbingand cuvette flushing iscontrolled by PC softwareand control electronics

    On-line optical monitoring

    - 1000 images/film (4 films/day)- 100 individual flocs/image

    400 000 individual flocs are analyzed daily

  • On-line optical monitoring

    18On-line optical monitoring

    Imaging unit consists of flow throughcuvette, standard machine visioncamera and LED lighting

    The imaging unit captures flocimages 30 frames/second

    Image size is 5.01mm x 3.74mmand imaged pixel size is 3.6m

    Floc images are analyzed andseveral morphological featurescalculated in floc analysis

  • Results

    Possibilities and challenges

    Would enable better control of theactivated sludge process

    Early warning system

    Challenges: Outdoor conditions Representative sampling

  • Thank you!

    Elisa Koivurantaelisa.koivuranta@oulu.fi


View more >