A map of the San Francisco Public Utilities Commission's assets
By Max Chung, Intelligence System Engineer, San Francisco Public Utilities Commission, with insights from Gary Wong, Principal, Global Water Industry, OSIsoft
Tasked with managing a century-old, mixed-use system of reservoirs, tunnels and pipelines to deliver water and sewer services to over 2.6 million residential, commercial and industrial customers, the San Francisco Public Utilities Commission (SFPUC) is no stranger to the ever-mounting pressure to do more with less.
Because of this fact, a few years ago SFPUC began looking at ways to decrease costs while still meeting high standards in and management. The utility recognized that with advances in sensor-tracking technology, there was a big opportunity to create a more data-driven management system. In the age of the Internet of Things, with sensors being more diverse, smaller, cheaper and easier to deploy, many systems that were once simple hunks of metal now report rich data back to operators thousands of times per minute.
With this, the SFPUC realized that they already possessed big data waiting to be extracted and interpreted. To begin to leverage this real-time data, they turned to the OSIsoft PI System, which, in conjunction with IBM’s Maximo, enabled them to ingest operational data tagged to each asset and develop actionable insights to address core aspects of the enterprise.
Greater Operational Visibility and Data De-siloing
While water utilities and large water consumers have employed SCADA systems and automation for years to control operations, this shift to digital analytics is marking a fundamental turning point in the industry.
The system has allowed SFPUC to integrate data from many sources normally considered separate and siloed in order to derive better predictive insights. For example, they can now pull weather services’ data such as area rainfall directly into the system, geotag it and derive insights into what kinds of storm water volumes are being faced. Whereas before that data was entered manually and could only be approximated, the utilty now knows exactly what is happening which makes planning and reacting to events far easier.
The system is also allowing SFPUC to hone in on process standardization in its treatment plants by giving them a real-time picture into what is taking place. Doing this gives them an edge in regulatory compliance, ensures they aren’t under- or over-treating water and enables them to better track their energy usage, among other advantages.
However, the greatest benefit of developing a real-time, data-driven enterprise is from the efficiency gains and cost avoidances in maintenance.
From Scheduled to Condition-Based Maintenance
One of the biggest benefits of de-siloing data and increasing visibility across the network of operational technology is enabling smarter maintenance. The reality is valves jam, pumps need oiling and motors break.
Ongoing maintenance is easily one of the most significant challenges SFPUC faces, especially given their limited staff. Plus, many of the more than 325,000 assets are at or nearing the end of their operational lifespan, making a solution to the maintenance challenge truly mission critical.
Before transitioning to a data-driven, condition-based maintenance model, upkeep on the utility’s assets was a significant headache. On a calendar-scheduled basis, SFPUC managers would dig into the database, look at what assets were up for routine maintenance, create work orders and send crews out to complete them. This method was highly inefficient, as not all assets require maintenance at the same rate.
By pulling operational data from individual assets into the system, SFPUC was able to see exactly which assets were being used enough to warrant maintenance and which were not. Analytics allows the commission to perform live analyses of the complex process and automatically create daily reports. Whereas previously the utility was driving blind, they now had an odometer to see how far the “car” had traveled and when it is due for an oil change.
The results have enabled SFPUC to create a condition-based maintenance program where only the assets that are in need of maintenance receive work orders. This program is expected to save more than $400,000 and about 9,500 hours of unneeded labor on the system’s 100 pumps, and ensure even better service to customers.
Alexandre Miot, one of the lead engineers at the Oceanside Plant, recently said, “Ultimately, the improvement we make to process monitoring and control will standardize our operation. With that, we hope to reduce energy and chemical costs as well as labor time through proactively identifying otherwise hard-to-spot issues.”
In many ways, SFPUC has only just begun to scratch the surface of the data opportunity. In the future, they aim to integrate system data with Esri’s ArcGIS platform to enable visualization of assets and real-time data mapping. This will allow the utility to identify where issues are arising and dispatch work crews with even greater speed. Right now, SFPUC is only using the system in its , but there will also be opportunity to integrate power data in the future.
For more information on the commission's use of big data, click here.