372
G.A. Schultz
sible for all time intervals. The quadratic objective function guarantees that during
all time increments this flow exceeding AQ is kept as low as possible. Figure 16.9
shows the result of the optimization of the joint operation of the two parallel reservoirs given in Fig. 16.8a for the flood given in Fig. 16.8b. As can be seen by
comparing the two diagrams above each other on the right hand side of Fig. 16.9
the floods in the Danube river are reduced from about 130 m 3 /s to about 100 m 3 /s
by the reservoir operation. This considerable reduction could be achieved with the
two rather small reservoirs only by the fact that this flood was controlled individually by the DP algorithm. During a flood event in real-time a new flood forecast
and a new optimization computation has to be carried out for each new time increment due to the fact that such forecasts are never perfect. More information on
this technique is given in the literature (Schultz, 1994).
16.4 Flood Forecasting and Control in an Urban Environment
The examples of flood forecasting and flood control given in this chapter have
dealt with meso-scale rural catchments. In recent times flood forecasting - particularly with the aid of weather radar - is becoming rapidly more important also
in the field of urban drainage systems. Since radar is capable to detect the high
variability of precipitation intensity in space and time it is very suitable for computing flood forecasts in the various canals of an urban drainage system. On the
basis of such forecasts it is possible to operate underground reservoirs in the canal
system as well as releases into rivers in an optimum way such that the sewage
treatment plant will not be overloaded during floods.
a (rn%) BREG RIVER
80
60'
40
20
INFLOW
o
t
3.2.704.2.7052.706.2.70
a (rn 3 {s)
80
BREG RESERVOIR
60
40
20
o ,
x
OUTFLOW
3.2.70 4.2.705 .2.7062.70
a (rn%)
a (rn 3 {s)
BRIGACH RIVER
80
60
INFLOW
3.2.704.2.70' 52.7() 6.2.70
a (rn 3 /s)
80- BRIGACH RESERVOIR
60
OUTFLOW
80
60
40
40
20
20
DANUBE RIVER
O"""~L+'...L..I.+LL..I..1i-J-L..J..... t
O ·L.fo-J...L..I.~I..L.f...L..I..l.f-L.LJ- t
3.2.704.2.705.2.706.2.70
3.2.7042.705.2.706.2.70
- - joint reservoir operahon
----- independent reservoir
operation
Fig. 16.9. Optimal release policy for two parallel reservoirs. Flood of Feb. 1970:
- joint reservoir operation ----- independent reservoir operation
G.A. Schultz
sible for all time intervals. The quadratic objective function guarantees that during
all time increments this flow exceeding AQ is kept as low as possible. Figure 16.9
shows the result of the optimization of the joint operation of the two parallel reservoirs given in Fig. 16.8a for the flood given in Fig. 16.8b. As can be seen by
comparing the two diagrams above each other on the right hand side of Fig. 16.9
the floods in the Danube river are reduced from about 130 m 3 /s to about 100 m 3 /s
by the reservoir operation. This considerable reduction could be achieved with the
two rather small reservoirs only by the fact that this flood was controlled individually by the DP algorithm. During a flood event in real-time a new flood forecast
and a new optimization computation has to be carried out for each new time increment due to the fact that such forecasts are never perfect. More information on
this technique is given in the literature (Schultz, 1994).
16.4 Flood Forecasting and Control in an Urban Environment
The examples of flood forecasting and flood control given in this chapter have
dealt with meso-scale rural catchments. In recent times flood forecasting - particularly with the aid of weather radar - is becoming rapidly more important also
in the field of urban drainage systems. Since radar is capable to detect the high
variability of precipitation intensity in space and time it is very suitable for computing flood forecasts in the various canals of an urban drainage system. On the
basis of such forecasts it is possible to operate underground reservoirs in the canal
system as well as releases into rivers in an optimum way such that the sewage
treatment plant will not be overloaded during floods.
a (rn%) BREG RIVER
80
60'
40
20
INFLOW
o
t
3.2.704.2.7052.706.2.70
a (rn 3 {s)
80
BREG RESERVOIR
60
40
20
o ,
x
OUTFLOW
3.2.70 4.2.705 .2.7062.70
a (rn%)
a (rn 3 {s)
BRIGACH RIVER
80
60
INFLOW
3.2.704.2.70' 52.7() 6.2.70
a (rn 3 /s)
80- BRIGACH RESERVOIR
60
OUTFLOW
80
60
40
40
20
20
DANUBE RIVER
O"""~L+'...L..I.+LL..I..1i-J-L..J..... t
O ·L.fo-J...L..I.~I..L.f...L..I..l.f-L.LJ- t
3.2.704.2.705.2.706.2.70
3.2.7042.705.2.706.2.70
- - joint reservoir operahon
----- independent reservoir
operation
Fig. 16.9. Optimal release policy for two parallel reservoirs. Flood of Feb. 1970:
- joint reservoir operation ----- independent reservoir operation