Process Dynamics Engine
Project description
PDEngine - Process Dynamics Engine
Process Dynamics Engine (PDEngine) is an online, iterative simulator for process control models described by transfer functions or state space representations. PDE is implemented in Python and uses the Python Control Systems Library. Users can interact with the simulation by using an API to send control actions (e.g. increasing a reflux rate or shutting off a valve) and query process variables like temperature and pressure. PDE is intended to be an educational tool in chemical engineering process control courses.
Usage
Models are defined by inheriting from the Model
parent class. An example using the Wood-Berry distillation model (3 inputs - 2 MVs, 1 FF & 2 outputs) is provided in /src/WoodBerryModel.py
.
A PDEngine Model must include:
- Name (e.g. Wood-Berry Distillation)
- Input names (e.g. $R$ - Reflux flow rate, $S$ - Steam flow rate, $F$ - Feed flow rate)
- Ouput names (e.g. $X_D$ - top composition, $X_B$ - bottom composition)
- Steady-state values for inputs and outputs
- Validity limits for inputs and outputs
- Typical moves for inputs
A driver program is then created to set up the simulation and initialize the model. Users must define:
- The simulation step size, $dt$
- Simulation speed, in terms of a time delay, $T_D$ between steps
wm = WoodBerryModel()
dt=0.02 # in minutes, for the Wood-Berry model
woodberry = wm.create_system(dt=dt) # make the actual generator with step size = 0.02
At each iteration of the simulation, users can send values to the simulation by using send()
with a dictionary with the Input names as keys:
# the main engine loop
for i in range(N):
y = woodberry.send({'R': 0.01, 'S': 0.01, 'F': 0.05}) # send this value to the model
time.sleep(0.001) # this is the time delay T_D for the simulation speed
An example driver program is provided in the Example.ipynb
notebook that uses the Wood-Berry distillation example.
Development
Recommended environment setup:
conda create -n pdengine python=3.10
conda activate pdengine
conda install -c conda-forge ipykernel slycot
python -m pip install control jupyter
python -m ipykernel install --user --name=pdengine
For an editable development environment, navigate to the pd-engine
folder, then run:
python -m pip install -e .
If everything is working correctly, you should be able to run the example notebooks.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.