11/11/2015

1

Joseph Hun-wei Lee

The Hong Kong University of Science and Technology

R&D Forum , Drainage Services Department

November 11, 2015

Field Experiment of Disinfection Dosage Optimization for the Hong Kong Harbour

Area Treatment Scheme

Outline

1. Overview of Harbour Area Treatment Scheme

2. WATERMAN coastal water quality forecast system

- tidal variation of beach water quality

3. Optimal disinfection dosage control strategy – pilot field dosage reduction trial

4. Chlorination of CEPT sewage – field experiment

5. Conclusions

11/11/2015

2

Hong Kong’s beach grading system

香港海灘水質評級系統

Grading

Beach

water quality

泳灘水質

E. coli *

(counts

/100 mL)

大腸桿菌

Minor illnesses

rate **

(cases per 1000

swimmers)

發病率

Water Quality

Objective

Compliance/

Exceedance

1 Good ≤ 24 Undetectable Compliance

2 Fair 25 - 180 ≤ 10

3 Poor 181 - 610 11 - 15 Exceedance

4 Very poor > 610 > 15

*Weekly Beach Grading: Geometric Mean E. coli level of the 5 most recent samplings (ClnEC5) Annual Beach Ranking: Geometric Mean E. coli level of all bathing season (Mar-Oct) samplings ** Skin and Gastrointestinal illnesses (Cheung et al. 1990)

Water Quality Objective: E.coli < 180 counts/100 mL

Harbour Area Treatment Scheme (HATS) Screening

Plants/pumping

stations

Stonecutters Island

Sewage Treatment

Work (SCISTW) Submarine

outfall

23.6 km deep tunnels

(>100m below ground level)

香港島

九龍

Stonecutters

Island STW Chemically Enhanced Primary Treatment (CEPT) since 2001; 23.6 km of deep tunnels Stage 1: Q = 1.4 x 106 m3/d

Stage 2A: Q = 1.8 x 106 m3/d

11/11/2015

3

General Layout of ADF

Chamber 15 Dechlorination

Dosing

Effluent Box Culverts 2 x 2.5 m x 2.5m

Stonecutters Island Sewage Treatment Work

Plan view

Sedimentation Tanks

Main Pumping Station

Flow distribution Chamber

Chlorine Dosing

Influent

Effluent

Chamber 15

Dropshaft

to outfall

Flow Distribution

Chamber

Chamber 9

Dechlorination

chemical storage

HATS

Pumping Station

NaOCl Storage

Chlorine

Contact Channel, 950m ,

2.5 x 2.5m twin box culvert

25m

Chlorine

dosing unitEmergency overflow

3.5m x 3.5m

twin box culvert

25m

Northwest Kowloon

Pumping Station

Rap

id M

ixin

g Tan

ks

Flocc

ulat

ion

Tanks

Flocc

ulat

ion

Tanks

Effluent Channel

Sedimentation

Tanks

Disinfection of CEPT sewage since March 2010 (Removal of 99% E.coli)

11/11/2015

4

Chlorine Dosage Required for Effective Disinfection - Bench scale jar tests

A final TRC of 0.5 mg/L is needed to reduce the effluent E. coli to 200,000 counts/100mL (Huang 2010) In actual operation, the chlorine demand in the CEPT effluent in summer is higher than anticipated. Dosage up to 18 mg/L or more is needed.

0

1

2

3

4

5

6

7

8

9

10

0 1 2 3 4 5 6 7 8 9 10 11 12

Lo

g [

E.

coli

Rm

ain

ing

] (c

fu/1

00

mL

)

TRC Level (mg/L)

Phase I - Simulator (08)

Phase I - Batch (08)

Phase II & III- Simulator ( Feb -- Aug 09)

Phase II & III- Batch (Feb -- Aug 09)

Mean

95% Confidence Boundaries

0.5 1.3 2.0

E. coli Target for HATS -ADF (200,000 cfu/100mL)

E. coli Target for HATS -2A (20,000 cfu/100mL)

95% Confidence Limit - HATS-2A(300,000 cfu/100mL)

2 log E. coli reduction can be achieved with 12 mg/L chlorine dosage with TRC less than 1 mg/L (DSD 2002)

Total Residual Chlorine

Optimal Chlorine Dosage Control

1. Excess Total Residual Chlorine (TRC) is toxic to marine organisms

2. Chlorinated organic compounds are harmful to marine environment

3. Complex interaction between chlorine dosage and CEPT sewage; chlorine demand and disinfection efficiency at high chlorine concentration

4. Enhance sustainability: energy and operation costs

5. Environmental protection should be set at an adequate but not unnecessarily severe level

11/11/2015

5

Bacterial decay modeling

Bacterial Loading Current, Turbulent Mixing

E. coli concentration

Far-field Hydrodynamic

Model

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

20.0 25.0 30.0

Salinity (ppt)

Dep

th (

m)

Neap

Spring

Near-field model JETLAG

Submarine outfalls

Dynamic coupling (DESA)

Project WATERMAN 3D Hydrodynamic Model of Hong Kong waters (Water Research, Chan et al. 2013)

Application of WATERMAN system in coastal water quality management

Daily beach water quality forecast

pollution incident response

Disinfection dosage optimization

Environmental impact assessment/setting of effluent discharge standard

Emergency sewage bypass

11/11/2015

6

Validation - Diurnal Beach E.coli Variation

18-Aug-2012, sunny day, semi-diurnal 27-Aug-2012, sunny day, diurnal

Lido beach

LIDO

180

24

610

1

10

100

1000

10000

0:00 6:00 12:00 18:00 0:00

E.coli

(#/100mL)

0.0

0.5

1.0

1.5

2.0

2.5

3.0Tide (m)

Model

Field data

Tide

LIDO

180

24

610

1

10

100

1000

10000

0:00 6:00 12:00 18:00 0:00

E.coli

(#/100mL)

0.0

0.5

1.0

1.5

2.0

2.5

3.0Tide (m)

Model

Field data

Tide

18-Aug-2012, sunny day, semi-diurnal 27-Aug-2012, sunny day, diurnal

GEM

180

24

610

1

10

100

1000

10000

0:00 6:00 12:00 18:00 0:00

E.coli

(#/100mL)

0.0

0.5

1.0

1.5

2.0

2.5

3.0Tide (m)

Model

Field data

Tide

GEM

180

24

610

1

10

100

1000

10000

0:00 6:00 12:00 18:00 0:00

E.coli

(#/100mL)

0.0

0.5

1.0

1.5

2.0

2.5

3.0Tide (m)

Hindcast

Field data

Tide

Gemini beach

Diurnal Beach E.coli Variation

•Beach WQ depends on tidal and solar radiation

•During diurnal tides beach WQ is relatively better than during semi-diurnal tides

• Longer travel time (from HATS to TW beaches) during diurnal tide, allows for more bacterial decay by solar radiation (Chan et al, 2013).

11/11/2015

7

Field Experiment of Disinfection Dosage Optimization for HATS during non-bathing season December 2014 – February 2015

Optimal real-time controlled disinfection dosage

Optimize chlorine dosage by maximizing

self-purification capability of receiving

water while still protecting Tsuen Wan beaches

11/11/2015

8

Coastal Tide Prediction

• The tide level variation can be represented by the harmonic sum of tidal constituents

• Tides in Hong Kong can be characterized by four major tidal constituents (M2 & S2 – semi-diurnal, K1 & O1 - diurnal). The daily tidal form can be predicted using M2, S2, O1, K1.

)cos(10 j

N

j jj φtωaHH

Diurnal tide

Quarry Bay

Use of Daily Tidal Form (Fd) Factor

Fd = (Tidal Range of K1+O1) (Tidal Range of M2+S2) = z1/ z2

Fd represents the relative importance of the diurnal and semi-diurnal components on any particular day

11/11/2015

9

Daily Tidal Form (Fd) Factor correlates with Observed Beach Water Quality

Beach

grading

Gemini (GEM)

Fd 0.5

Semi-

diurnal

0.5 < Fd ≤ 1.0

Mixed

Fd > 1.0

Diurnal

Good 3 (15.0%) 2 (15.4%) 11 (34.4%)

Fair 8 (40.0%) 5 (38.4%) 13 (40.6%)

Poor 1 (15.0%) 4 (30.8%) 6 (18.7%)

Very poor 8 (40.0%) 2 (15.4%) 2 (6.3%)

Total N 20 13 32

EPD water quality data (2010 – 11) - significantly better water quality is observed for Fd > 1

Month Range of applied dosage concentration (mg/L)

Dec 2014 6 – 10

Jan 2015 6 – 9

Feb 2015 6 – 8

Field Experiment of Disinfection Dosage Optimization

Pilot dosage reduction experiment based on the daily tidal form factor - December 2014 to February 2015.

Chlorine (Cl) dosage reduction criteria:

• If Fd > 1 Cl = 6 mg/L is applied (all day)

• If Fd < 0.5 Cl = 8-10 mg/L is applied (all day)

• If 0.5 < Fd < 1.0 a 6-hour period reduction starting 6.5 hours before the Lower Low Water (LLW) is applied.

11/11/2015

10

Pilot Dosage Reduction Trial (Dec 2014 – Feb 2015)

Compared with the dry season dosage level of 12 mg/L, the dosage reduction plan can decrease the monthly use of disinfection chemical by 27-39 %.

The beach water quality with the dosage reduction plan is similar to that observed prior to reduction (2010 – 2013)

Dosage reduction

11/11/2015

11

Summary

• The beach water quality in Hong Kong harbour is highly dependent on tidal conditions. Based on the WATERMAN system and the use of a daily tidal form factor, a chlorine disinfection dosage control strategy has been developed.

• Continuous field experiments for the non-bathing season confirm that with carefully planned dosage reduction, the water quality at the Tsuen Wan bathing beaches would still meet the Water Quality Objectives.

• The optimization of chlorine dosage to 6-10 mg/L for the Harbour Area Treatment Scheme results in savings of 30-40% of chemicals while still protecting public health.

Part 2: Chlorination of CEPT

sewage – Field experiment

11/11/2015

12

Emergency bypass chamber

Flow distribution chamber12m (L) x 7.5m (W)

50m twin box culvert

3.5x3.5m twin box culvert

Emergency bypass chamber

Flow distribution chamber12m (L) x 7.5m (W)

50m twin box culvert

3.5x3.5m twin box culvert

Inclined Weir

To 1km box

culvert and

Chamber 15

The problem of high chlorine demand at SCISTW Previous study show significant

chlorine demand is exerted in the FDC

chamber.

Reasons of rapid chlorine demand in

FDC are unknown.

Detailed on site measurement is

constrained by safety concern and

operation constraints

Vertical section view

1.7 m opening

Submerged flow

Free surface flow

Dense Jets

Dosing

Unit

2.65mPD

0.85mPD

Flow Distribution Chamber

(Chlorine dosing unit)

• An inclined weir of 1.8m height in the middle of FDC

• 10% chlorine solution is injected to the sewage flow through an array of dense jets in two layers

Chlorine

dosing

unit

Weir

2 3.5m 3.5m

inlet culvert from

sedimentation

tanks

Plan view

12.5m

11/11/2015

13

Chlorine dosing

unit (2 rows of

jets)

Chlorine dosing in

flow distribution

chamber

CEPT

Effluent

Dosing of 10%

sodium hypochlorite

Dosing of sodium

bisulphite

Dechlorinated

Effluent

Existing 1km box

culvert as the chlorine

contact system

Discharge

to Outfall

Chamber 15 Flow

Distribution

Chamber

SCISTW Advance Disinfection Facility (ADF)

- Original design concept

Cl2 conc. (ppm)

x (m) 30m 1km

105 Initial

mixing

2 order E.coli kill

in contact time of

~10 min

TRC > 1

FDC

Chamber

9

Average chlorine Concentration = 12 mg/L (dry) = 18 mg/L (wet)

after disinfection:

107 → 105 count/100mL

Required Dilution ~ 10000!

10-20 ppm

11/11/2015

14

What happens actually … • Most of the chlorine consumption has already taken place before Chamber 9; very few TRC remains at Chambers 9 and 15. • Significant E.coli reduction may have achieved at Chamber 9; insignificant E.coli reduction between Chamber 9 and 15 • The 1km culvert does not act as a chlorine contact chamber as expected.

Cl2 conc. (ppm)

x (m)

10m 1km

105

10-20 ppm

Few E.coli kill

TRC << 1

FDC

1km culvert

Chamber 15 Chamber 9

80% Chlorine

consumed

Expected

Required Dilution ~ 10000!

Layout of 1:2 FDC model

Sewage

inflow

Head

tank

Dosing

unit

10% NaOCl solution

Test flume

Outflow

tank

Weir

11/11/2015

15

Scale Value

Length

Velocity

Time

Sewage flow rate

1:2 Froude scale model of a “1/16 slice”

of Flow Distribution Chamber (FDC)

414.1r

rr

U

LT

2rL

414.1 rr LU

91165.2

rr LQ

Sewage flow rate for FDC model =100 L/s, corresponding to

low sewage flow rate of 9 m3/s

The similar time scale between model and prototype (1:1.4) allows the

realistic simulation of chemical process happening in the prototype FDC

On the roof top between sedimentation tanks no. 32 & 34

Site of Physical model in SCISCW

Before model installation

With the model installed

Head tank

Test flume

Outlet tank

Chemical storage tank

11/11/2015

16

Chlorine dosing system

Diaphragm pump

Air chamber

Storage tank

Safety valve

To dosing unit

Dosing unit

0.0

25

0.24 m

0.2

5 m

¦¤h

=

Low port ( 2.5

mm diameter)

Upper port ( 5

mm diameter)

Upper port

Low port

11/11/2015

17

Dosing sugar solution ( = 1.168 g/mL) into air

Q = 25 mL/s Q = 35 mL/s

Dosing sugar solution ( = 1.168 g/mL) into cross flow

11/11/2015

18

Mixing of chlorine dosage and tap water crossflow

• Dosage chlorine flow rate= 20 mL/s

• Tap water crossflow rate=100 L/s

• Chlorine dosage=20 mg/L

• Average TRC concentration above the weir plate=18 mg/L

d

NaOCl solutionq

y

x

(0,0)

x=2.5 m

9 mg/L44 mg/L

0 mg/L1

.26

5

Q0=100 L/s

=20 mL/s

Reasonable mass balance is

observed;

Incomplete mixing feature

near the weir plate.

Operation of test flume in CEPT sewage

Head tank FDC Weir

End Weir

11/11/2015

19

Sampling of sewage and measurement of TRC

• 9 sections along the flume centerline are sampled with a vertical interval of 0.025m

• Sewage is extracted from the flume using a sampler with plastic tubing.

• TRC is measured in situ using colorimeter.

Sampling tube (Inside flume)

Vertical traverse

Portable TRC colorimeter

meter

Reagent

powder

Vials

Measured flume centerline TRC concentration field (Sewage flow rate=100 L/s, dosage flow rate=20 mL/s, dosage=21 mg/L)

Sampling sections

Only 20-30% of sewage is in contact with chlorine before the weir

Ua = 0.35m/s

11/11/2015

20

Measured typical TRC concentration distribution at the centerline

section of the FDC model

x=0.5 m x=1.0m x=1.5 m x=2.0 m

y=1.265 m

=20 mL/sd

NaOCl solutionq

y

x(0,0)

x=2.5 m

Qs= 100 L/s

Above weir: average

TRC= ~ 4 mg/L

About 80% chlorine

dosage is exerted.

Chlorine dosage=21 mg/L

56 mg/L 34 mg/L 23 mg/L

20 mg/L

13 mg/L

Ua = 0.35m/s

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

0 0.5 1 1.5 2 2.5x (m)

TRC mass

flux (g/s)

Total Residual Chlorine mass flux estimation

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

-100 100 300 500

x = 0.2m

mg/L

Chlorine is reduced by 80% during the initial contact with sewage!

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

-20 0 20 40 60 80

x = 0.4m

mg/L

z1

z2

h = 1.26m

Measured TRC fitted with a double peak Gaussian profile

Mass flux is determined by integrating the concentration against flume area

TRC profiles

11/11/2015

21

CFD modeling of the 1:2 scale flume

• 900,000 grid cells

• Free surface determined by

volume of fluid (VOF)

method

• Time step = 0.01s

Dosing jet flow = 20mL/s

C0 = 100,000 mg/L

Δρ/ρ = 0.2

Ua = 0.35m/s

X

Z

0 1 2 3

0

0.5

1

1.5

Streamlines of flow

Weir (0.9m high)

Dosing jets

Y

Z

0

0

0.5

1

1.5

x=0.5m

Y

0

x=1.0m

Y

0

x=1.5m

Y

0

x=2.0m

Y

0

UDS0

2000

1000

500

200

100

50

10

x=2.5m(weir)

• Only 10-20% of the sewage are in contact with chlorine

• Significant concentration variation vertically and laterally above the weir

• Strong mixing after the weir

X

Z

0 1 2 3

0

0.5

1

1.5

UDS0

10000

5000

1000

500

100

50

10

JETLAG

Concentration of conservative tracer - Channel centerline

x-sections

11/11/2015

22

Simultaneous sampling of TRC and

E.coli above the weir (6 Nov) • Nine samples were taken above the weir and measured for

TRC in situ. E.coli was measured in lab.

• Only 18% of the total dosed chlorine remains. E.coli kill efficiency (~ 2 order) is high.

Y

Z

0

0

0.5

1

1.5

x=0.5m

Y

0

x=1.0m

Y

0

x=1.5m

Y

0

x=2.0m

Y

0

UDS0

2000

1000

500

200

100

50

10

x=2.5m(weir)

+

+ +

+ +

+ + + +

TRC mg/L (E.coli cfu/100mL)

y = -0.073m (L)

y = 0m (M)

y = 0.073m (R)

z = 1.15m 0.25

(6.9E+06)

3.52

(3.8E+05)

2.86

(5.1E+05)

z = 1.05m 11.5

(< 100)

6.85

(< 100)

7.12

(< 100)

z = 0.95m 0.31

(7.7E+06)

1.62

(3.9E+06)

0.23

(1.2E+06)

+ sample points

TRC Mass flux: Dosing unit = 2g/s

Weir = 0.369 g/s (18% remaining)

E.coli conc.: Upstream = 5.3e+6 cfu/100mL (log averaging) Weir = 8.61e+4 cfu/100mL (disinfection efficiency = 98%)

L M R

1

1.5

0.5

Conclusions

• Chlorine dosage at SCISTW can be optimized by making use of a daily tidal form factor.

• The dense chlorine jets cannot achieve full mixing of chlorine dosage with the CEPT sewage in the flow distribution chamber. Only 20-30% of sewage is in contact with chlorine at the weir.

• Due to the fast exertion of chlorine demand which includes ammonia oxidation, 80% of chlorine is reduced in the FDC.

• At locations where sewage is in contact with chlorine, E.coli kill is rapid within the FDC. The original concept of using the 1km twin culvert as a contact chamber may not be valid.

• Further experimental study is required to confirm the chlorine demand and disinfection efficiency, and to test improvement measures.

11/11/2015

23

Research Team Members - Prof J H W Lee - Prof Howard Huang

- Prof Stanley Lau

- Dr David K W Choi

- Dr Tree S N Chan

- Dr Q S Qiao

- Dr A W Thoe

- Ms Mary Anne Borigas

- Mr M F Poon