Skip to content

Latest commit

 

History

History
174 lines (148 loc) · 6.05 KB

Settings.md

File metadata and controls

174 lines (148 loc) · 6.05 KB
Raw

pyDMPC settings

All required inputs are inserted into the Init file. The required selection is described in detail in the following. In order to just run the demo without any changes, you can skip this part.

Global paths

There are 3 global library paths, namely

  • pyDMPC Modelica Models
  • Modelica Buildings
  • AixLib.

Moreover, specify the result path, where the working directories will be created automatically.

Finally, there is the Dymola path, where the .egg file is stored.

glob_lib_paths = [r'C:\Git\pyDMPC\pyDMPC\ModelicaModels\ModelicaModels',
             r'C:\Git\modelica-buildings\Buildings',
             r'C:\Git\AixLib\AixLib']
glob_res_path = r'C:\TEMP\Dymola'
glob_dym_path = r'C:\Program Files\Dymola 2018 FD01\Modelica\Library\python_interface\dymola.egg'

Working directory

The working directory is automatically created out of the current time stamp.

import time
timestr = time.strftime("%Y%m%d_%H%M%S")
name_wkdir = r'pyDMPC_' + 'wkdir' + timestr

Identifiers

The unique identifier is an integer. Additionally, a subsystem name can be given as a string.

sys_id.append(0)
name.append("Field")

Controlled system

The controlled system type can either be "Modelica" or "PLC". There is only one controlled system for each main system. In case of the PLC controlled system, an ADS address and a port have to be specified. In case of a Modelica model, which is provided as an FMU model, the name of the FMU file has to be provided. The file should be stored in the global result path. The time increment is the minimal time step that is used for communication with the FMU.

contr_sys_typ = "Modelica"
ads_id = '5.59.199.202.1.1'
ads_port = 851
name_fmu = 'pyDMPCFMU_Geo.fmu'
orig_fmu_path = glob_res_path + '\\' + name_fmu
dest_fmu_path = glob_res_path + '\\' + name_wkdir + '\\' + name_fmu
time_incr = 120

Neighbors

Each subsystem can have upstream and downstram neighbors. To assign a subsystem to its neighbors, provide the IDs of the neighboring subsystems or None.

ups_neigh.append(None)
downs_neigh.append(0)

Inputs

The input names are given as a list of strings. These are the names of the variables in the FMU model or the data point names on the PLC. By contrast, the inputs_variables are the names in the subsystem model. Be sure to use the same order in both lists so that the mapping is correct. Finally, a list of inputs can be provided using the range function. The algorithms will use these values to vary the inputs to the models and create lookup tables. Alternatively, use a list with only the entry "external" to use an external input. In that case, the algorithms do not vary the inputs but, instead, inputs should be provided in the model. This can be useful, if the model has a much longer prediction horizon than the other models and relies on some kind of long-term prediction, e.g. weather or load prediciton.

input_names.append(["supplyTemperature.T"])
input_variables.append([r"variation.table[1,2]"])
inputs.append(range(280,310,10))

Commands

The commands are speicified just like the inputs, with the names referring to the controlled system and teh variables to the subsystem models. Again, the values to be used can be given using the range function. All these commands are evaluated in a brute-force manner. In future versions, there will be more sophisticated optimization algorithms.

command_names.append(["heatShare"])
command_variables.append(["decisionVariables.table[1,2]"])
commands.append(range(0,105,5))

Outputs

The names of the outputs are a list of strings and represent the variable names as used in the subsystem models.

output_names.append(["returnTemperature.T"])

Set points

The set points are given as a list of floats. The list should have the same order as the outputs list as the algorithms compare the set points to the corresponding output.

set_points.append([287])

State variables

The state variables are used to initialized the subsystem models. There are currently no state estimators. Therefore, use the state_var_names to provide the names of the variables in the controlled system and model_state_var_names as their counterpart in the subsystem model.

state_var_names.append(["supplyTemperature.T"])
model_state_var_names.append(["vol.T_start"])

Prediction times

Specify when a prediction should start (usually a relative time 0) and when it should stop. Also indicate the increment, which will be the sampling time of the result time series. The opt_time is the interval (referring to the global system time), in which the subsystem is optimized. The samp_time is the interval, in which a subsystem communicates with the controlled system.

start.append(0.)
stop.append(3600.0*24*365.25*3)
incr.append(3600.)
opt_time.append(600)
samp_time.append(10)

Subsystem paths

Most of the paths can remain unchanged as they are formed based on the global paths. The path that should be specified, yet only in case of a Modelica subsystem model, is the path to the model in the Modelica-specific package structure.

mod_path.append(r'ModelicaModels.SubsystemModels.DetailedModels.Geo.Field')

Trajectories

Some subsystems require their neigbors to follow certain trajectories. To get an upstream subsystem to provide the trajectory, the subsystem assigns costs to deviations from the trajectory that is calculated in the model. The name of the variable is provided in the traj_var. The trajectory points are used to approximate the costs that are cause if the upstream system does not follow the trajectory.

traj_points.append(range(278,310,1))
traj_var.append(["supplyTemperature.T"])

Cost factors

The cost factors are essential to select which types of costs should be considered and how they should be weighted. The first element in the list is the cost related to the control effort. The second is the cost that is received from the dowmstream neighbor and the third is the cost due to deviations form the set point.

cost_fac.append([0.0, 0.0, 1.0])