Transforming wastewater management with advanced data integration and AI-driven automation

The Los Angeles County Sanitation Districts (LACSD) serve a vast population of 5.1 million people across 24 special districts, 78 cities, and unincorporated areas. With an extensive infrastructure that includes 1,400 miles of sewers, 49 pumping plants, 11 wastewater treatment facilities, and 10 solid waste facilities, LACSD faced significant challenges in managing and optimizing its operations. To address these issues, LACSD implemented a strategy to centralize, visualize, and share real-time data, aiming to streamline workflows and provide actionable insights for on-premises data collected from industries and shared with partners.

Overcoming data silos: The roadblocks to proactive operations at LACSD

Before its comprehensive system overhaul, LACSD faced operational and data management challenges that hampered efficiency and proactive decision-making.

Manual processes—such as chemical inventory—were time-consuming, inefficient, and error-prone. Non-operational staff struggled to access real-time operational data, which hindered informed and timely decision-making. The Reuse and Compliance department struggled to track silo tonnage and schedule hauling efficiently, risking non-compliance. Delayed alerts from critical sensors made it difficult to prevent overflows and pollution incidents in time.

In addition, LACSD’s existing infrastructure lacked the flexibility to integrate diverse data sources or scale across facilities. This limited its ability to expand digital capabilities and adopt advanced analytics across operations. LACSD also lacked proactive oversight and suffered from system vulnerabilities because it couldn’t properly monitor the health of its operations data.

The most critical challenge was its inability to predict and address operational issues. This led to reactive troubleshooting and increased downtime. For example, at LACSD’s combined cycle energy plant—which runs on biogas from the plant or natural gas if necessary—certain maintenance issues resulted in both types of gas being burned simultaneously.  Wayne Holder, SCADA Coordinator at LACSD, said: "So, before that, they were kind of blind. We'd find out months later. Now we find out right away."

For daily chemical tank inventories, LACSD utilized various components of AVEVA PI Server’s asset framework (PI AF). The process was fully automated using PI Server’s asset analytics, set to collect total inventory every day at 12:00 AM. Event frames were generated in PI Server, and it sent notifications directly to operators and vendors, which Holder explained: "We automated collection of the data using PI AF templates for different chemical systems, and we took a daily snapshot of the chemical inventories and sent an email out to the users, and eventually we started sending them directly out to the vendors. So, this is all automated. All the ordering is done through PI AF."

This template-based approach was applied across various chemicals with significant results. "We eliminated manual inventory processes, resulting in improved inventory accuracy and vendor interaction,” he said.

To manage sludge storage silo compliance, a similar approach was developed to compare end-of-day tonnage against compliance levels. Notifications showing the severity of the overage and total tonnage were sent to the Reuse and Compliance department and area engineers, which helped them schedule hauling proactively.

At the combined cycle energy plant, the dual gas flow problem was resolved with a PI AF asset analytic that sends a notification when digester gas is being flared while natural gas is consumed for an extended period. With this tool, immediate action can be taken.

To monitor industrial discharge, a notification system alerts inspectors by email and text when multiple sedimentation tank blanket detectors register bad values, enabling proactive investigation of potential pollution. This system has reduced response times and improved wastewater treatment reliability. Similarly, when pumping plant wet well levels stay high, notifications are sent immediately, allowing inspectors to quickly address possible overflows and maintain smooth operations. "And this is kind of a proactive thing that they can do now without having to find out after the fact that somebody has polluted," Holder noted.

The ability to benchmark and snapshot new processes from day one was vital for the Advanced Water Treatment Plant. During commissioning, new PI tags were created for real-time monitoring. Greg Anderson, AVEVA PI System Administrator, LACSD, used AVEVA PI System data to pinpoint issues and assign responsibility. "I could look at the PI data and tell them exactly who did it." He also created comprehensive AVEVA™ PI Vision™ screens for a quick visualization of plant status. LACSD enhanced its AVEVA PI System monitoring with AVEVA's library of templates, and integrated data from Windows performance monitoring data to track the health of AVEVA PI System, including its PI Interfaces, data archives, PI AF servers, and notifications. LACSD can now identify potential problems like “lost events.”

For external data integration, LACSD implemented solutions for both rainfall (Gauge Adjusted Radar Rainfall via REST API) and complex flow meter data. RoviSys played a key role in setting up the interface to pull data from REST API endpoints, creating a custom PowerShell script to query an SQL database made up of thousands of active flow meters. This script automatically created PI tags and PI AF structures based on the database information and then generated JSON payloads processed by PI Connector for Universal File and Stream Loading (UFL). The system, now standardized, could ingest diverse flow meter data from different manufacturers and locations, which allowed for sewer flow analytics, summary data, alarms, and even machine learning applications.

As part of AVEVA’s Lighthouse Program, LACSD also embarked on a transformative initiative to modernize operations at the Valencia Water Reclamation Plant. The project focused on two high-impact use cases: Flow Equalization and UV Disinfection Optimization. These were historically managed through manual adjustments and static dosing, but with the deployment of CONNECT and the Industry Accelerators for Water, LACSD transitioned to a predictive, data-driven model. The accelerators—designed as pre-configured service layers within CONNECT—enabled rapid deployment of advanced analytics and visualization tools, cutting implementation time by 75% and delivering immediate operational value.

To optimize flow equalization at the Valencia Water Reclamation Plant, LACSD implemented AI-driven automation through CONNECT. Historically, operators manually adjusted basin setpoints based on experience and daily conditions. With CONNECT’s calculation services and integration of weather forecast data from OpenWeatherMap, LACSD deployed machine learning models that predict influent flow and recommend optimal setpoints for the secondary treatment process. These models account for diurnal and seasonal fluctuations, storm events, and historical flow patterns. The result was a significant reduction in peak flow magnitudes, improved biological efficiency in aeration tanks, and decreased reliance on chemical additives like coagulants and disinfectants.

For UV disinfection, LACSD faced challenges with overdosing and inconsistent water quality outcomes. Using AVEVA Advanced Analytics on CONNECT and the Industry Accelerator for Water, they transitioned from static dosing to dynamic UV dose control. AI/ML models were trained to correlate water quality parameters—such as turbidity, total coliform counts, and influent characteristics—with effective UV intensity levels for disinfection. These models, fed by real-time data from LIMS and the AVEVA PI Server, enabled operators to adjust dosing precisely, ensuring regulatory compliance while minimizing energy consumption. Acceptance testing previously required 80 mJ/cm² doses to meet standards; with AI-guided control, LACSD is now approaching the design target of 50 mJ/cm², demonstrating improved disinfection performance and operational efficiency.

The Lighthouse implementation also showcased the scalability of CONNECT’s architecture. LACSD’s on-premises Lab Information Management System (LIMS) and PI Server were seamlessly integrated into CONNECT data services, allowing for centralized data access and visualization. With CONNECT visualization already enabled, LACSD operators could interact with intuitive dashboards tailored for the water industry, built on best-of-breed UI/UX designs. As the project progresses, LACSD plans to expand into additional use cases such as leak detection, predictive maintenance, and energy optimization—further validating the power of AI/ML and the accelerators in driving digital transformation across municipal water operations.

Reduced downtime and improved efficiency

The implementation of these solutions significantly enhanced LACSD’s operations. One of the most notable benefits for LACSD was the improvement of its operational efficiency. Automating chemical inventory processes improved inventory accuracy and streamlined vendor interactions.

By identifying and addressing issues before they became critical, proactive system health monitoring was achieved, which substantially reduced unplanned downtime and critical incident response time. Holder stated, "We are able to catch potential operational issues ahead of time and deal with them in real time instead of after the fact."

This proactive capability extended to detecting unauthorized discharges at pumping plants and industrial waste issues from sedimentation tanks immediately, which allowed for prompt investigation and pollution prevention. Compliance for sludge hauling was improved through automated notifications, which helped the Reuse and Compliance department maintain optimal levels. LACSD also avoided disputes because it received immediate insights into performance and accountability when it benchmarked new facilities from their commissioning stage.

The system centralized and standardized the visualization of complex flow monitoring data from the sewer system, which was previously a major pain point. The use of CONNECT also enabled efficient and secure data sharing with third-party analytics platforms and industries, which fostered collaborative problem-solving and optimization that went beyond internal operations.

Non-operational staff were also empowered with easy access to relevant operational data on PI Vision dashboards, which facilitated better decision-making across the organization. The system now includes over 30,000 PI tags, 810 PI Vision displays, and nearly 1,400 asset analytics. This provided comprehensive insights and enabled better-informed decisions across departments. For example, the PI Vision screens provided a clear and comprehensive view of the processes so that staff could better monitor equipment performance and identify potential issues. The ability to send notifications via email and text further ensured that staff were promptly informed about issues and could take immediate action.

Looking ahead, LACSD is focused on continued innovation and aims to further enhance its unified, real-time data system. To continue to reduce energy usage and maintain water quality, future initiatives include improved GIS integration, treatment plant optimization through data-driven modeling to identify sensor drift, and machine learning models to assist plant operators in predictive maintenance and process optimization. LACSD also plans to detect and respond to blockages and identify areas for inflow and infiltration sources in its sewer system.
 

The implementation of the PI System and CONNECT has revolutionized our operations. By automating processes and providing real-time data, we've significantly improved efficiency, reduced downtime, and empowered our staff to make informed decisions. This technological advancement has been a game-changer for LACSD, enabling us to achieve our goals and maintain compliance with ease.

– Wayne Holder, SCADA Coordinator, Warren Wastewater Resource Facility
 

All of the AVEVA solutions used by LACSD are available through the AVEVA PI Data Infrastructure offering.