A Day in the Life with Predictive Control

Back to blog home

December 7, 2021 | Javier Cantu

5 minute read

In water treatment, a lot of our daily decisions depend on variables that are constantly changing. What the weather is like, whether a pump is out of service and, perhaps most importantly, the characteristics of the influent waters.

Adverse events usually require us to operate our systems in unusual ways and, naturally, with the risk averse instinct being to over-correct, the cost of operating facilities during these events compounds over time. 30% of utilities' operating budget is pump energy costs. A city the size of Portland, Oregon spends $20M per year on pumping alone. A reduction of just 2% saves $400K per year in energy costs. The costs associated with energy use are independent of the costs of overcompensating in chemical dosing or the effects of potential severe non-compliance penalties when water quality exceeds limits due to unplanned events of consequence.

Predictive control utilizes machine learning to find patterns, forecast scenarios, and evaluate the best control actions to take against cost optimization algorithms, chemical reaction optimization strategies, and so much more.

The following scenarios highlight Emagin's predictive control capabilities:

Scenario Typical Operations Life with predictive control
A professional hockey game is scheduled for tomorrow night, which historically significantly increases the demand for water in its respective pressure zone within the city's network. A major sporting event will occur, all while other system demands continue to affect the city's ability to efficiently meet water and energy demands. The city, being aware of the additional output required of its pumps during peak times, has invested in reliable equipment and control logic to ensure demand is met. The costs to operate during peak demand is considered a standard business expense. Emagin forecasts upcoming demand and evaluates costs and preparation procedures against cost optimization algorithms. Equipment is also monitored to ensure the most efficient equipment is used.

Prior to the event starting, the operations team monitors and reviews Emagin's predictive control recommendations of pumps and enact the control schedules.

With Emagin, evaluation of network adjustments can be made quickly in near real time. Storage is optimized to meet demands and pumps are run in order to most effectively reduce their run times during peak energy tariff periods. Predictive AI recommendations have shown to effectively reduce energy expenses up to 20% in systems with variable tariffs.

Scenario Typical Operations Life with predictive control
A severe storm is approaching a water treatment plant. Upstream water quality has been unusual this week and the effects of the incoming storm are unknown. The city has an effective plan to handle peak and shock loads. In preparation for the storm, operators prepare by opening gates to manage flow and ensuring max inventory for chemicals is available for treatment. Operators have the knowledge that plant headworks usually require additional attention for such events. Acting accordingly, the plant set point limits for chemical dosing stations that are increased and left unattended to account for potential high loading events. Emagin forecasts upcoming shock loads and evaluates treatment operations against predicted water quality parameters. Equipment is also monitored to ensure flow can be managed with online and effective assets.

Prior to the incoming event impacting the plant, the operations team receives an alert of the predicted loading conditions. Best operating practices for the loading conditions are provided in a digital schedule to provide an operator a simple view of the when to operate equipment and optimal dosing points.

With Emagin, a prescriptive approach to handling shock loads while maintaining compliance and operating at the lowest possible cost is used. By simulating chemical reaction curves to most probable water quality scenarios, Emagin can limit excess dosing, optimize storage in a facility, minimize equipment on-off cycles, and manage flow. Predictive AI recommendations have shown to effectively reduce operating costs associated with chemical use, equipment life cycle, and energy expenses across water and wastewater treatment plants.

With Emagin implemented, an operations team can get back what is most valuable: time. Key attention areas of the plant are identified in advance and the guess work of determining optimal set points is taken away. Teams can focus on more critical areas and more effectively identify the source of plant upsets.


To see what a typical day with Emagin looks like at a utility, watch this short, 3-minute video.

For more information on Emagin AI, speak with an expert today!

Find out more

Contact us

Tags: Emagin AI, Artificial Intelligence, Product Knowledge

Javier Cantu

Javier Cantu

Director of AI and Process Optimization

 

Javier is process optimization engineer at Innovyze and is responsible for client service and engineering solutions for industrial applications. He has 10+ years’ experience in the water and wastewater industry designing, managing, and creating optimization strategies and is currently based out of Los Angeles region.

Patrick Bonk

Patrick Bonk

Software Solutions Leader

 

Patrick Bonk is Innovyze’s Asia Pacific team resident Sewer Hydraulics Engineer with the role of Software Solutions Lead. Patrick has a Masters in Engineering Design of Sustainable Infrastructure and joined Innovyze in 2014 with a focus and passion towards the emergence of operational analytics decision support tools and software advancements within the context of Smart Water systems. You can ask me anything at patrick.bonk@innovyze.com