johnkeating1997 / pyswmm5

A Python package to manipulate EPA SWMM5 like a data scientist.

PySWMM5

This package aims to provide an easy and efficient way to interact with Storm Water Management Model (SWMM), developed by US EPA and widely used in urban stormwater management practices.

Specifically, you can do the following things with the help of this program:

• Modify the data in the input file (*.inp)
• Run simulation
• Save simulation results in SQLite database.
• Manipulate input and output data through DataFrame.
• Visualize urban drainage systems and design the color, size, and style of any objects.
• Export input file to GIS format (.shp) and create input file from GIS format.

It should be noted that storing results in a database is not only for ease of operation but also for future consideration. With the rapid growth of IoT and smart city, the future simulation will involve millions of pipes/nodes, which could result in huge file sizes (in order of gigabytes) and slow I/O operation if we still use text files to store data. To this end, a database like SQLite can perfectly handle these issues. For now, the database generated by pyswmm5 can save around 50% storage space.

In addition, the database is a good friend of Data Science, which means pyswmm5 is closely related to Python Data Science packages, especially for packages based on Pandas, because most data are stored as the Pandas DataFrame type after being loaded into memory.

To take full advantage of this program, I strongly recommend you use this program in Jupyter Notebook, which allows you to run your programs in an interactive way.

How do we use PySWMM5?

All you need is to move the pyswmm5folder into your working directory. Then, you can import it as follow:

import pyswmm5 as ps

To read a .inp file:

swmmInp = SwmmInp('swmm.inp')

To run simulation:

databaseName = swmmInp.run_swmm('output/', filename='test') # save result in a database

To retrieve the simulation results:

swmmRpt = SwmmRpt(databaseName) # all results are stored in swmmRpt

More details can be found in this tutorial.

How does PySWMM5 work?

The PySWMM5 is implemented mainly in two steps. In the first step, the source code of SWMM5 was modified to save simulation results in a database (.db) rather than a text file (.rpt). The modified code was built into a dynamic-link library (DLL) that can be used by many programming languages, such as C++, Python, MATLAB, etc. The next step is to utilize Python and useful packages (Pandas, NetworkX, etc.) to wrap this DLL so that we can achieve mentioned functionalities.