J. A. Elías-Maxil
Jan Peter van der Hoek
Jan Hofman
Luuk Rietveld
SPN7

Sustainability of the urban water cycle
◦ 80 % of energy input to urban water is heat

Strategies to improve sustainability: Heat
recovery installations
◦ Operates in main sewers

Significant potential for heat recovery in small
sewers


To estimate the potential temperature and
flow data is needed
Flow measurements are some times difficult
to obtain in small sewers
◦
◦
◦
◦
Low flow rates
Intermittent
Difficult access
Costly



Prediction of wastewater flow with little and if
possible no measurements
Possibility to calculate intermittent
wastewater flow
Possibility to use the flow patterns to
calculate wastewater quality (temperature)

Wastewater flow modeling in sewer
(a)
◦ Probability theory to produce expected flow
◦ Intermittent discharges from water consuming
appliances were converted to continuous base flow
◦ The flow rate and arrival time at a certain point of the
sewer was modeled with Saint Venant equations
Flow
Flow
Intermittent inputs
Eq
Eq
Time
(a)
Continuous base flow
Time
Butler, D. and N. J. D. Graham (1995). J. Environ. Eng. 121(2): 161-173.
Background
1.
Methods
Results
Conclusions
Stochastic modeling (Drinking water)
◦ Generation of water pulses
◦ Different for every activity
2.
3.
Adapted to wastewater discharge
Attenuation of intermittent flow
Blokker, E. J. M., et al. (2010). Jour. Water. Res. Plan. and Man. 136(1): 19-26.
Background


Methods
Results
Conclusions
North of Amsterdam
97 household connections
◦ Clustered in 51 connections for
the model

~ 15 days
Geometry

2 Monitoring campaigns

◦ Mean slope < 2%
◦ PVC 250 mm

Flow measurement by pumping time

V [ t off ( n 1 ) , t off ( n ) ] 
C ap  t[ off ( n 1 ) , t on ( n ) ]
[ t off ( n 1)  t off ( n ) ]
Generation of wastewater discharge patterns
0.5
Discharge, l/s

Drinking water at time of consumption
Wastewater after being used
Wastewater at sewer
0.45
0.4
0.35
D1
0.3
I1
0.25
0.2
Dn
0.15
I2
0.1
In
0.05
0
5500
6000
τ1
6500
τ2
7000
Seconds
7500
8000
8500
τn
Background

Methods
Results
Conclusions
Generation of wastewater discharge patterns
Equivalent
Appliance
D, s
I, l/s
t s, s
D s, s
Shower
600
0.123
45
Same as D
Kitchen tap
30
Same as D
Toilet
16|48|15|3 0.083|0.125|0.0
7
83|0.083
45-106
0.042|0.884
180|60
9
Bathroom tap
40 | 15
0.042 | 0.042
0
Same as D
Wash machine
120*
Dish Water
21*
0.167|0.083|0.0
83|0.083
0.19*
3840|1260|114 300*
0|600
1800*
120*
|: Separation of sub-activities or cycles
*: The same parameter was included in the remaining 3 cycles
Generation of wastewater discharge patterns
0.5
Discharge, l/s

Drinking water
water at
at time
time of
of consumption
consumption
Wastewater
being used
Wastewater after
at sewer
Wastewater
sewer
Wastewater at
after
being used
0.45
0.4
0.35
0.3
Ds1
0.25
0.2
Dsn
0.15
0.1
0.05
0
5500
6000
τ1+ts
6500
τ2+ts
7000
Seconds
7500
8000
8000
τn+ts
8500
8500


Mean flow rate / day
Maximum flow rate in time period / day
 O bserved 
&
 M odeled 
 t1

t
 2


 t n 1
 t
 n
Q1 

Q2



Q n 1 
Q n 
Flow patterns
divided
in time
segments
(6s – 1hr)
 O bs  _& _  M od 
 Q1

Q
 x 1


 Q ix
Q2
Qx 

Q2 x



Qn 
Qmean
&
Qmax
Comparison:
RMSE
R2
 O bs  _& _  M od 
 Q m ean _ 1




 Q m ean _ i
Q m ax_ 1 




Q m ax_ i 
Percentiles of
cumulative
results
obtained
Background
Average daily flow, l/s
Methods
Modeled
Observed
0.38
0.36±0.3*
Results
Conclusions
*Expected flow from surveys: 0.4 l/s
Background
Methods
Results
Conclusions
1
0.9
Cumulative probability
0.8
0.7
0.6
0.5
0.4
Observed
Simulated
0.3
0.2
0.1
0
0
1
Qmax(3s), l/s
2
0
1
Qmax(5min), l/s
2
0
1
Qmax(hour), l/s
2
Background
Methods
Results
100
90
80
Percentage
70
Qmax - RSME
60
Qmax - R2
50
Qcumulative - RMSE
40
Qcumulative - R2
30
20
10
6/60 0.5
10
20
30
40
Time scale, min
50
60
Conclusions
Background
Methods
Results
Conclusions

A model that includes

The prediction was stable for time frames from 6
seconds to 1 hour
1. Stochastic simulation of drinking water demand
2. Transformation of pulses to wastewater generation
3. Attenuation of discharge to the sewer
Was found to be adequate to model the wastewater flow
rate of a small sewer
◦ RMSE ~ 20%
◦ R2 > 85%

Future work: Validation of temperature model
Temperature Model
Along the pipe
Along the
water depth
Along the
Distance of
the pipe
Thank you!

Detection of pump intervals
30
2
2

yn
 yn
 y n 1
 0;
 0;
0
O n 
t
t
t

2
2
 yn
 y n 1

y
P um p  O ff 
 0;
 0;
0
t
t
t

O ther



Water level, cm
25
20
15
10
5
0
400
600
800
1000
Seconds
1200
1400
Background

Error analysis of
measurements

◦ Roughness
◦ Pump capacity
Off
•
Effect of time
resolution
Level readings
On
00
05
10
15
Seconds
20
25
Results
Conclusions
Hydraulic model calibration
Sensor 1
Sensor 2
Sensor 3
•
Methods
30
Background
Parameter
Measured
Conf. Int. (ton-toff)
3.2, s
Conf. Int. (toff-toff)
3.1, s
Roughness
15, mm C-W
Pump capacity
8.24 ± 0.47, l/s
Methods
Results
Conclusions