dam_v008

Reservoir version of HydPy-Dam.

dam_v008 is a relatively simple reservoir model, similar to the “TALS” model of LARSIM. It combines the features of dam_v006 (“controlled lake”) and dam_v007 (“retention basin”). Additionally, it allows controlling the stored water volume via defining target values that can vary seasonally.

Like dam_v007, dam_v008 allows for combining controlled, “harmless outflow” (via parameter AllowedRelease) and uncontrolled, “spillway outflow” (via parameter WaterLevel2FloodDischarge), and like dam_v006, it allows to restrict the speed of the water level decrease during periods with little inflow via parameter AllowedWaterLevelDrop (only through reducing the controlled outflow, of course). Before continuing, please first read the documentation on these two application models.

The additional feature of dam_v008 is that it tries to track target volumes that can vary seasonally. Define these target volumes via parameter TargetVolume. The parameters VolumeTolerance, TargetRangeAbsolute, and TargetRangeRelative serve to yield more smooth and realistic reservoir responses for small deviations from the given target values. Setting TargetRangeRelative to 0.2 and both other parameters to zero corresponds to selecting the “TALSPERRE SOLLRANGE” option in LARSIM. Please see the following examples and the documentation on method Calc_ActualRelease_V3 for more information on how to set and combine the individual parameter values for different use-cases.

Integration tests

Note

When new to HydPy, consider reading section How to understand integration tests? first.

We create the same test set as for application models dam_v006 and dam_v007, including an identical inflow series and an identical relationship between stage and volume:

>>> from hydpy import IntegrationTest, Element, pub
>>> pub.timegrids = "01.01.2000", "21.01.2000", "1d"
>>> from hydpy.models.dam_v008 import *
>>> parameterstep("1d")
>>> element = Element("element", inlets="input_", outlets="output")
>>> element.model = model
>>> IntegrationTest.plotting_options.axis1 = fluxes.inflow, fluxes.outflow
>>> IntegrationTest.plotting_options.axis2 = states.watervolume
>>> test = IntegrationTest(element)
>>> test.inits = [(states.watervolume, 0.0)]
>>> test.dateformat = "%d.%m."
>>> watervolume2waterlevel(
...     weights_input=1.0, weights_output=1.0,
...     intercepts_hidden=0.0, intercepts_output=0.0,
...     activation=0)
>>> catchmentarea(86.4)
>>> element.inlets.input_.sequences.sim.series = [
...     0.0, 1.0, 6.0, 12.0, 10.0, 6.0, 3.0, 2.0, 1.0, 0.0,
...     0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

base scenario

First, we again use the quasi-linear relation between discharge and stage used throughout the integration tests of dam_v006 and in the base scenario example of dam_v007:

>>> waterlevel2flooddischarge(ann(
...     weights_input=1.0, weights_output=10.0,
...     intercepts_hidden=0.0, intercepts_output=0.0,
...     activation=0))

Additionally, we set some of the remaining parameter values extremely high or low, to make sure that the reservoir stores all water except the one activating the spillway and thus becoming “flood discharge”:

>>> targetvolume(100.0)
>>> neardischargeminimumthreshold.shape = 1
>>> neardischargeminimumthreshold.values = -100.0
>>> targetrangeabsolute(0.1)
>>> targetrangerelative(0.2)
>>> watervolumeminimumthreshold(0.0)
>>> volumetolerance(0.1)
>>> dischargetolerance(0.1)
>>> allowedrelease(100.0)
>>> allowedwaterleveldrop(100.0)

Due to the same the neural network configuration, the results are identical with the ones of the base scenario example of dam_v006 and the base scenario example of dam_v007:

>>> test("dam_v008_base_scenario")
Click to see the table
Click to see the graph

spillway

When we reuse the more realistic relationship between flood discharge and stage of the spillway example on dam_v007, we again get the same flood discharge time series:

>>> waterlevel2flooddischarge(ann(
...     weights_input=10.0, weights_output=50.0,
...     intercepts_hidden=-20.0, intercepts_output=0.0))
>>> test("dam_v008_spillway")
Click to see the table
Click to see the graph

target volume

During dry periods, application model dam_v007 generally releases all its water until the basin runs dry, as long as parameter AllowedRelease is larger than zero. Application model dam_v008 is more complex, in that it allows to define target storage volumes. dam_v008 tries to control its outflow so that the actual volume approximately equals the (potentially seasonally varying) target volume. However, it cannot release arbitrary low or high amounts of water to fulfil this task due to its priority to release a predefined minimum amount of water (for ecological reasons) and its second priority to not release to much water (for flood protection). In this example, we activate these mechanisms through changing some related parameter values (also see the documentation on method Calc_ActualRelease_V3 for more detailed examples, including the numerous corner cases):

>>> targetvolume(0.5)
>>> neardischargeminimumthreshold(0.1)
>>> allowedrelease(4.0)
>>> allowedwaterleveldrop(1.0)

Compared with the allowed release results of dam_v007, dam_v008 dampens the given flood event less efficiently. dam_v007 releases all initial inflow, while dam_v008 stores most of it until it reaches the target volume of 0.5 million m³. After peak flow, dam_v008 first releases its water as fast as allowed, but then again tries to meet the target volume. The slow negative trend away from the target value at the end of the simulation period results from the lack of inflow while there is still the necessity to release at least 0.1 m³/s:

>>> test("dam_v008_target_volume")
Click to see the table
Click to see the graph

sharp transitions

Due to smoothing, the above results deviate from the ones one would expect from LARSIM simulations to some degree. However, if we set both “target range” parameters to zero (like one does not set the LARSIM option “TALSPERRE SOLLRANGE”) and both “tolerance” parameters to zero (to disable any smoothing), we should get more similar results:

>>> targetrangeabsolute(0.0)
>>> targetrangerelative(0.0)
>>> volumetolerance(0.0)
>>> dischargetolerance(0.0)
>>> test("dam_v008_sharp_transitions")
Click to see the table
Click to see the graph

higher accuracy

The first water volume calculated in the sharp transitions example is negative, which is the result of the limited numerical accuracy of the underlying integration algorithm. We can decrease such errors through defining smaller error tolerance values, but at the risk of relevant increases in computation times (especially in case one does apply zero smoothing values):

>>> solver.abserrormax(1e-6)
>>> test("dam_v008_higher_accuracy")
Click to see the table
Click to see the graph

target range

In the last example, the behaviour of the reservoir always changes abruptly when the actual volume transcends the target volume. According to its documentation, LARSIM then predicts unrealistic jumps in discharge. To solve this issue, LARSIM offers the “TALSPERRE SOLLRANGE” option, which ensures smoother transitions between 80 % and 120 % of the target volume, accomplished by linear interpolation. dam_v008 should never output similar jumps as it controls the correctness of its results. As a drawback, correcting these jumps (which still occur “unseeable” and possibly multiple times within each affected simulation time step) costs computation time. Hence, at least when using small smoothing parameter values, dam_v008 can also benefit from this approach. You can define the range of interpolation freely via parameters TargetRangeAbsolute and TargetRangeRelative depending on your specific needs. Setting the latter to 0.2 corresponds to the original “TALSPERRE SOLLRANGE”-configuration:

>>> targetrangerelative(0.2)
>>> test("dam_v008_target_range")
Click to see the table
Click to see the graph

minimum volume

In all examples above, the dam would run dry entirely after a certain amount of time to fulfil the downstream demand defined by parameter NearDischargeMinimumThreshold. Usually, this is neither desired nor technically possible. The following example shows that the parameter WaterVolumeMinimumThreshold allows setting a minimum amount of water below which no release occurs:

>>> watervolumeminimumthreshold(0.45)
>>> test("dam_v008_minimum_volume")
Click to see the table
Click to see the graph
class hydpy.models.dam_v008.Model[source]

Bases: hydpy.core.modeltools.ELSModel

Version 8 of HydPy-Dam.

The following “inlet update methods” are called in the given sequence at the beginning of each simulation step:
The following methods define the relevant components of a system of ODE equations (e.g. direct runoff):
  • Pic_Inflow_V1 Update the inlet link sequence.

  • Calc_WaterLevel_V1 Determine the water level based on an artificial neural network describing the relationship between water level and water stage.

  • Calc_SurfaceArea_V1 Determine the surface area based on an artificial neural network describing the relationship between water level and water stage.

  • Calc_AllowedDischarge_V2 Calculate the maximum discharge not leading to exceedance of the allowed water level drop.

  • Calc_ActualRelease_V3 Calculate an actual water release that tries to change the water storage into the direction of the actual target volume without violating the required minimum and the allowed maximum flow.

  • Calc_FloodDischarge_V1 Calculate the discharge during and after a flood event based on an SeasonalANN describing the relationship(s) between discharge and water stage.

  • Calc_Outflow_V1 Calculate the total outflow of the dam.

The following methods define the complete equations of an ODE system (e.g. change in storage of fast water due to effective precipitation and direct runoff):
The following “outlet update methods” are called in the given sequence at the end of each simulation step:
numconsts: hydpy.core.modeltools.NumConstsELS
numvars: hydpy.core.modeltools.NumVarsELS
class hydpy.models.dam_v008.AideSequences(master: hydpy.core.sequencetools.Sequences, cls_fastaccess: Optional[Type[FastAccessType]] = None, cymodel: Optional[hydpy.core.typingtools.CyModelProtocol] = None)

Bases: hydpy.core.sequencetools.ModelSequences[AideSequence, hydpy.core.variabletools.FastAccess]

Aide sequences of model dam_v008.

The following classes are selected:
class hydpy.models.dam_v008.ControlParameters(master: hydpy.core.parametertools.Parameters, cls_fastaccess: Optional[Type[hydpy.core.parametertools.FastAccessParameter]] = None, cymodel: Optional[hydpy.core.typingtools.CyModelProtocol] = None)

Bases: hydpy.core.variabletools.SubVariables[hydpy.core.parametertools.Parameters, Parameter, hydpy.core.parametertools.FastAccessParameter]

Control parameters of model dam_v008.

The following classes are selected:
class hydpy.models.dam_v008.DerivedParameters(master: hydpy.core.parametertools.Parameters, cls_fastaccess: Optional[Type[hydpy.core.parametertools.FastAccessParameter]] = None, cymodel: Optional[hydpy.core.typingtools.CyModelProtocol] = None)

Bases: hydpy.core.variabletools.SubVariables[hydpy.core.parametertools.Parameters, Parameter, hydpy.core.parametertools.FastAccessParameter]

Derived parameters of model dam_v008.

The following classes are selected:
class hydpy.models.dam_v008.FluxSequences(master: hydpy.core.sequencetools.Sequences, cls_fastaccess: Optional[Type[FastAccessType]] = None, cymodel: Optional[hydpy.core.typingtools.CyModelProtocol] = None)

Bases: hydpy.core.sequencetools.OutputSequences[FluxSequence]

Flux sequences of model dam_v008.

The following classes are selected:
class hydpy.models.dam_v008.InletSequences(master: hydpy.core.sequencetools.Sequences, cls_fastaccess: Optional[Type[FastAccessType]] = None, cymodel: Optional[hydpy.core.typingtools.CyModelProtocol] = None)

Bases: hydpy.core.sequencetools.LinkSequences[InletSequence]

Inlet sequences of model dam_v008.

The following classes are selected:
  • Q() Discharge [m³/s].

class hydpy.models.dam_v008.OutletSequences(master: hydpy.core.sequencetools.Sequences, cls_fastaccess: Optional[Type[FastAccessType]] = None, cymodel: Optional[hydpy.core.typingtools.CyModelProtocol] = None)

Bases: hydpy.core.sequencetools.LinkSequences[OutletSequence]

Outlet sequences of model dam_v008.

The following classes are selected:
  • Q() Discharge [m³/s].

class hydpy.models.dam_v008.SolverParameters(master: hydpy.core.parametertools.Parameters, cls_fastaccess: Optional[Type[hydpy.core.parametertools.FastAccessParameter]] = None, cymodel: Optional[hydpy.core.typingtools.CyModelProtocol] = None)

Bases: hydpy.core.variabletools.SubVariables[hydpy.core.parametertools.Parameters, Parameter, hydpy.core.parametertools.FastAccessParameter]

Solver parameters of model dam_v008.

The following classes are selected:
  • AbsErrorMax() Absolute numerical error tolerance [m3/s].

  • RelErrorMax() Relative numerical error tolerance [1/T].

  • RelDTMin() Smallest relative integration time step size allowed [-].

  • RelDTMax() Largest relative integration time step size allowed [-].

class hydpy.models.dam_v008.StateSequences(master: hydpy.core.sequencetools.Sequences, cls_fastaccess: Optional[Type[FastAccessType]] = None, cymodel: Optional[hydpy.core.typingtools.CyModelProtocol] = None)

Bases: hydpy.core.sequencetools.OutputSequences[StateSequence]

State sequences of model dam_v008.

The following classes are selected: