pypes-paper

Case studies and software applications for the PyPES journal paper

Optimal Sensor Placement

In this case study, we implemented the algorithm from the paper "Optimal flow sensor placement on wastewater treatment plants" by Villez et al. to demonstrate how PyPES can be used to implement graph-based algorithms in a facility-agnostic fashion. The wastewater model is derived from Benchmark Simulation Model no. 2 (BSM2), which was originally developed by Gernaey et al.

Usage

python optimal_sensor_placement.py [arguments]

Arguments

• --mode or -m: Specifies which Waste Water Treatment Plant (WWTP) to use. It accepts an integer value (1, 2, or 3) and defaults to 1 if not provided.
• --update or -u: A flag that, when set, enables update mode.
• --log_file or -l: Specifies the path to the log file where the results will be saved.
• --original_log_file or -o: Specifies the path to the original log file for comparison purposes in update mode.:

Example

To run the script with the default WWTP (1) and save results to results.txt:

python optimal_sensor_placement.py --log_file results.txt

To run the script with WWTP 1 with update mode enabled and save results to results_update.txt:

python optimal_sensor_placement.py --mode 1 --update --log_file results_update.txt --original_log_file results.txt

Fault Detection

In this case study, we implemented the algorithm from the paper "Advanced monitoring of water systems using in situ measurement stations: data validation and fault detection" by Alferes et al. to demonstrate how PyPES can ease the implementation of complex fault detection algorithms by managing data and metadata. The reverse osmosis treatment train comes from a real world desalination plant.

Usage

python fault_detection.py [arguments]

Arguments

• --tag_file or -t: Specifies the path to the tag file with information about the sensors data.
• --data_folder or -d: Specifies the path to the data folder. Inside this folder, there should be one or more CSV files containing the data to be analyzed.
• --train_test_split or -s: Specifies the train-test split ratio. Default is 0.8.
• --data_plot or -p: Specifies the path to save the data plot. Default is results/fault_detection/data.png.
• --json_path or -j: Specifies the path to the network JSON file. Default is json/desalination.json.
• --T2Q_plot or -t: Specifies the path to save the T2Q plot. Default is results/fault_detection/T2Q_plot.png.
• --PC_plot or -c: Specifies the path to save the PC plot. Default is results/fault_detection/PC_plot.png.

Example

To run the script with the default parameters:

python fault_detection.py

To run the script with a custom tag file and data folder:

python fault_detection.py --tag_file custom_tag.csv --data_folder custom_data

Leakage Detection

In this case study, we implemented the algorithm from the paper "Leakage fault detection in district metered areas of water distribution systems " by Eliades & Polycarpou to demonstrate how a user can automatically convert EPANet files into a PyPES model and then conduct analysis on a large number of nodes and connections.

The distribution system model comes from Dataset of BattLeDIM: Battle of the Leakage Detection and Isolation Methods. The original distribution network, L-TOWN.inp and the converted PyPES model (distribution.json) are included in the json folder under the Creative Commons Attribution 4.0 International license, including the Disclaimer of Warranties and Limitation of Liability. The citation, including full list of authors, can be found in References.

Usage

To convert an EPANet file:

python epyt_utils.py json/L-TOWN.inp json/distribution.json

To run the leakage detection algorithm:

python leakage_detection.py

References

  Alferes, J., Tik, S., Copp, J. and Vanrolleghem, P.A., 2013. Advanced monitoring of water systems using in situ measurement stations: data validation and fault detection. Water Science and Technology, 68(5), pp.1022-1030. https://doi.org/10.2166/wst.2013.302

  Eliades, D.G. and Polycarpou, M.M., 2012. Leakage fault detection in district metered areas of water distribution systems. Journal of Hydroinformatics, 14(4), pp.992-1005. https://doi.org/10.2166/hydro.2012.109

  Gernaey, K.V., Jeppsson, U., Vanrolleghem, P.A., Copp, J.B. (Eds.), 2014. Benchmarking of Control Strategies for Wastewater Treatment Plants. IWA Publishing. https://doi.org/10.2166/9781780401171

  Villez, K., Vanrolleghem, P.A., Corominas, L., 2016. Optimal flow sensor placement on wastewater treatment plants. Water Research, 101, pp.75–83. https://doi.org/10.1016/j.watres.2016.05.068

  Vrachimis, S.G., Eliades, D.G., Taormina, R., Ostfeld, A., Kapelan, Z., Liu, S., Kyriakou, M.S., Pavlou, P., Qiu, M., Polycarpou, M., 2020. Dataset of BattLeDIM: Battle of the Leakage Detection and Isolation Methods. https://doi.org/10.5281/ZENODO.4017658

Acknowledgments

This work is supported by the National Alliance for Water Innovation (NAWI), funded by the U.S. Department of Energy, Energy Efficiency and Renewable Energy Office, Advanced Manufacturing Office under Funding Opportunity Announcement DE-FOA-0001905. The views expressed herein do not necessarily represent the views of the U.S. Department of Energy or the United States Government.

This work is also supported by the Center for Integrated Facility Engineering at Stanford University as a part of CIFE Seed Proposal 2023-02.

We would like to thank the staff of the Charles Meyer Desalination Plant for reverse osmosis data and metadata and our collaborator Akshay Rao for insight into the reverse osmosis system.