Calibration of hydrodynamic model-driven sewer maintenance

Figures & data

Figure 1. Layout and monitoring network ‘Tuindorp’ catchment.

Table 1. Main characteristics ‘Tuindorp’ catchment.

Characteristics	Tuindorp catchment
Area use	Residential
System type	Combined
System structure	Looped
Ground level/surface level (m above average sea level)	0.75–2.25
Minimum and maximum surface slope (m/m)	0–0.045
Minimum and maximum pipe slope (m/m)	0–0.075
Contributing area (ha)	56.9
Number of CSO structures (–)	5
Storage volume (m^3)	4669 (=8.2 mm)
Volume storage settling tank (m^3)	822 (=1.4 mm)
Number of pumping stations (–)	1
Pumping capacity (m^3/h)	800^*
Number of inhabitants (–)	10,656

^* Based on flow measurements. According to the municipal administration the pumping capacity is 540 m³/h.

Figure 2. Storm events and works ‘Tuindorp’ sewer system for 1 October 2008–6 October 2015.

Figure 3. Calibration process (after Hemker, 1996).

Figure 4. Rainfall runoff model (Stichting RIONED, 1999).

Figure 5. Outgoing discharge ‘Tuindorp’ pumping station based on flow and water level measurements at the pumping station for June 2014–May 2015.

Figure 6. Observed storm events for model calibration (11/07/2010, 26/08/2010, 05/05/2015 and 04/09/2015) and obtaining the reduced parameter set (24/09/2014).

Table 2. Parameter set for calibration of the hydrodynamic model.

Full parameter set		Reduced parameter set
Parameter	Description	Parameter	Description
B2 (mm)	Initial losses on flat, impervious areas	B2 (mm)	Initial losses on flat, impervious areas and inclining roofs
B5 (mm)	Initial losses on flat, semi-pervious areas	B5 (mm)	Initial losses on flat, semi-pervious areas and flat roofs
B6 (mm)	Initial losses on large flat, semi-pervious areas	B6 (mm)	Initial losses on large flat, semi-pervious areas and large flat roofs
B7 (mm)	Initial losses on inclining roofs	F2 (s)	Routing coefficient for flat, impervious areas and inclining roofs
B8 (mm)	Initial losses on flat roofs	F5 (s)	Routing coefficient for flat, semi-pervious areas and flat roofs
F2 (s)	Routing coefficient for flat, impervious areas	F6 (s)	Routing coefficient for large flat, semi-pervious areas and large flat roofs
F5 (s)	Routing coefficient for flat, semi-pervious areas	I5 (mm/h)	Maximum infiltration capacity flat, semi-pervious areas and flat roofs
F6 (s)	Routing coefficient for large flat, semi-pervious areas	I6 (mm/h)	Maximum infiltration capacity large flat, semi-pervious areas and large flat roofs
F7 (s)	Routing coefficient for inclining roofs	K (mm)	Hydraulic roughness for all conduits
F8 (s)	Routing coefficient for flat roofs	CC (m^0.5/s)	Overflow coefficient for all CSOs
I5 (mm/h)	Maximum infiltration capacity semi-pervious areas
I6 (mm/h)	Maximum infiltration capacity large, semi-pervious areas
KZ_1 (mm)	Hydraulic top roughness for all conduits
K2_1 (mm)	Hydraulic bottom roughness for all conduits
CC (m^0.5/s)	Overflow coefficient for all CSOs

Figure 7. Local hydrological circumstances on roofs affecting total amount of runoff water (vegetation and storage).

Figure 8. Average daily dry weather flow on the ‘Tuindorp’ catchment.

Notes: The dry weather flow is determined from measurement data during periods in which no precipitation occurred and increased groundwater level (after a storm event) no longer affects the dry weather flow because of infiltration. Furthermore, days with (partly) missing data and data affected by maintenance or construction activities are also ignored in the analysis.

Figure 9. Daily pattern dry weather flow on the ‘Tuindorp’ catchment for 2010 and 2012.

Figure 10. Correlation matrix full parameter set storm event 24/09/2014.

Figure 11. Correlation matrix reduced parameter set storm event 24/09/2014.

Figure 12. Measured and calibrated water levels Lev7 storm event 04/09/2015.

Notes: Lev7 is situated downstream in the ‘Tuindorp’ sewer system, nearby the pumping station and a CSO (see also Figure 1 for the location). NAP is reference level. Negative measured and calibrated values represent water levels below this reference.

Figure 13. Measured and calibrated water levels Lev10 storm event 04/09/2015.

Notes: This monitoring location is located more upstream in the ‘Tuindorp’ sewer system (see also Figure 1 for the monitoring location). NAP is reference level. Negative measured and calibrated values represent water levels below this reference.

Figure 14. Measured and calibrated water levels Lev19 storm event 04/09/2015.

Notes: This monitoring location is situated in the centre of the ‘Tuindorp’ sewer system (see also Figure 1 for the monitoring location). NAP is reference level. Negative measured and calibrated values represent water levels below this reference.

Figure 15. Normal probability plots of residuals after calibration of the clean system (storm events 05/05/2015 and 04/09/2015) and the system including defects (storm events 11/07/2010 and 26/08/2010).

Note: Residuals of all measuring locations.

Figure 16. Bias and MSE for location Lev19 during storm event 04/09/2015.

Note: In this case, the system is considered to be clean.

Table 3. Correlation between measured water levels up and downstream the observed root intrusion at Professor Jordanlaan and Jan van Galenstraat for storm event 11/07/2010.

	Professor Jordanlaan
	Lev6	Lev12	Lev13	Lev14	Lev22
Lev6	–	0.85	0.78	0.81	0.80
Lev12	0.85	–	0.94	0.96	1.00
Lev13	0.78	0.94	–	0.99	0.93
Lev14	0.81	0.96	0.99	–	0.96
Lev22	0.80	1.00	0.93	0.96	–
	Jan van Galenstraat
	Lev15	Lev16	Lev17	Lev25	Lev26	Lev27
Lev15	–	0.992	0.940	0.992	0.941	0.865
Lev16	0.992	–	0.957	0.999	0.971	0.902
Lev17	0.940	0.957	–	0.966	0.989	0.982
Lev25	0.992	0.999	0.966	–	0.974	0.913
Lev26	0.941	0.971	0.989	0.974	–	0.977
Lev27	0.865	0.902	0.982	0.913	0.977	–

Notes: Monitoring locations located in the upstream part are in bold. See Figure 1 for the monitoring locations.

Table 4. Correlation between residuals after calibration up and downstream observed root intrusion at Jan van Galenstraat for storm event 11/07/2010.

	Jan van Galenstraat, downstream of root intrusion
	Lev15	Lev17	Lev27
Lev15	–	0.98	0.98
Lev17	0.98	–	0.99
Lev27	0.98	0.99	–
	Jan van Galenstraat, upstream of root intrusion
	Lev16	Lev25	Lev26
Lev16	–	1.00	0.93
Lev25	1.00	–	0.94
Lev26	0.93	0.94	–

Note: See Figure 1 for the monitoring locations.

Table 5. Correlation between residuals after calibration up and downstream observed root intrusion at Jan van Galenstraat for storm event 26/08/2010.

	Jan van Galenstraat, downstream of root intrusion
	Lev15	Lev17	Lev27
Lev15	–	0.87	0.81
Lev17	0.87	–	0.99
Lev27	0.81	0.99	–
	Jan van Galenstraat, upstream of root intrusion
	Lev16	Lev25	Lev26
Lev16	–	0.99	0.92
Lev25	0.99	–	0.88
Lev26	0.92	0.88	–

Note: See Figure 1 for the monitoring locations.

Table 6. Correlation between residuals after calibration up and downstream observed root intrusion at Professor Jordanlaan for storm event 11/07/2010.

	Professor Jordanlaan, downstream of root intrusion
	Lev6	Lev12	Lev22
Lev6	–	0.81	0.83
Lev12	0.81	–	0.97
Lev22	0.83	0.97	–
	Professor Jordanlaan, upstream of root intrusion
	Lev13	Lev14
Lev13	–	0.98
Lev14	0.98	–

Note: See Figure 1 for the monitoring locations.

Table 7. Correlation between residuals after calibration up and downstream observed root intrusion at Professor Jordanlaan for storm event 26/08/2010.

	Professor Jordanlaan, downstream of root intrusion
	Lev6	Lev12	Lev22
Lev6	–	0.83	0.74
Lev12	0.83	–	0.97
Lev22	0.74	0.97	–
	Professor Jordanlaan, upstream of root intrusion
	Lev13	Lev14
Lev13	–	0.87
Lev14	0.87	–

Note: See Figure 1 for the monitoring locations.

Figure 17. Mean value of residuals per event after calibration of the clean system (events 05/05/2015 and 04/09/2015) and the system including in-sewer defects (events 11/07/2010 and 26/08/2010).

Figure 18. Mean value of residuals per sensor after calibration of the clean system (events 05/05/2015 and 04/09/2015) and the system including in-sewer defects (events 11/07/2010 and 26/08/2010).

Figure 19. Measured and modelled (calibrated) water levels upstream (Lev13) root intrusion at Professor Jordanlaan for event 26/08/2010. NAP is reference level.

Note: Negative measured and calibrated values represent water levels below this reference.

Hemker, C. J. (1996). The reliability of parameters in automatic calibration (in Dutch). In Modelkalibratie: Het automatisch ijken van grondwatermodellen [Model calibration: The automatic calibration of groundwater models], The Dutch Hydrological Association, NHV, Special Issue 2, 39–51.

 Google Scholar

Stichting RIONED. (1999, July). Leidraad Riolering. Module C2100. Rioleringsberekeningen, hydraulisch functioneren (in Dutch). Kluwer, Alphen aan de Rijn.

 Google Scholar