Significant investments are spent on building and calibrating hydraulic models for planning. Unfortunately, these models often sit isolated from real-time sensor data and provide no value to operations and maintenance divisions. However, we have a solution.
Real-time modeling systems can bridge this gap by leveraging sensor and model investments as a real-time decision support system for operations teams. However, they remain relatively rare throughout North America. While software capabilities have been available for decades, the lack of experience and evidence of benefits leads to reluctance in live modeling adoption.
Innovyze has had the privilege of working with pioneering utilities worldwide that have elected to deploy live modeling solutions. In this blog, we’re sharing some of the lessons learned and the key benefits gained in using live models to help ensure successful real-time modeling programs.
Let’s first begin by understanding what live modeling is and why it is so valuable.
Live modeling integrates hydraulic models, SCADA historians, and weather forecast databases. It provides a near-real-time view of network performance and alerts your team to adverse events. It reduces the time needed to run a model from days to minutes by automating the data ingestion, simulation, and analysis process.
Live modeling automatically records performance data and checks key performance metrics, allowing data and insight-driven improvements to be evaluated and tested daily. Providing benefits for water utility operators, analysts, and managers such as:
- Real-time conditions and forecasts help make better decisions
- Testing and applying insights from the operators using real-time modeling
- Recording and evaluating the system operation performance using an automated centralized IT system
- Allows you to respond with proactive measures rather than reactive measures
Live modeling can also be used as a digital twin for monitoring diagnostics and prognostics to optimize asset performance and utilization. In this field, sensor data can be combined with historical data, human expertise, and simulation learning to improve the outcome of prognostics. Therefore, complex prognostics and intelligent maintenance systems platforms can use digital twins in finding the root cause of issues and improve productivity.
Here are some lessons learned from handling sewer overflows, road closure, sensor issues, and pipe blockage and how live modeling can provide additional improvements.
No trust, no action.
Take the time to earn the trust of operators. Your first project should focus on planting the seeds for growing confidence that doesn’t require operators to change the way they work or manage in a way that requires them to involve other departments or teams.
Design the system with simplicity.
It takes a team effort to keep live modeling up and running correctly. Some systems span multiple servers pulling data from various sources, and every process in the system must do its job reliably to produce accurate forecasts and alerts. Simplicity means keeping the number of mission-critical resources and teams to a minimum. It can be tempting when modeling to try and employ highly detailed numerical models with pre-processing or post-processing, but the best live modeling systems are as simple as possible while achieving the business objectives. In some cases, existing models should be intentionally simplified, if the additional detail will not be used.
Make frequent model updates and calibrations.
Real-time decision support systems rely on accurate predictions. Fortunately, real-time models are often equipped with sensor integration at comparison points within the model, so every automated simulation provides real-time feedback on the status of calibration. In addition, real-time sensor data can be incorporated as boundary conditions to help drive a closer behavior between the model and the real world. Together, this enables modeling teams to go from rare large-effort calibration projects to continuous ongoing feedback and minor corrections over time. While updating real-time models, it’s essential to remember that a lower calibration tolerance might be allowed compared with traditional calibration requirements of detailed planning models; it all depends on the business objectives of the system.
Be prepared for emergencies.
Identify blockages in the system faster than ever before. Live modeling helps with both the detection of problems in the system and simulation-based planning around how and when to respond to issues in your system. Proactive planning and scheduling of maintenance work can drastically improve the water system’s efficiency.
Live modeling, and digital twins can help optimize asset performance and utilization by making daily improvements in identifying deficiencies in the water system and executing solutions, allowing you to combat typical failures.
Reliability goes a long way.
Live modeling can support real-time intelligence to inform operations and management with simulation-based awareness. This type of output may require nearly 24/7 uptime and reliability. One approach that can help in achieving this reliability is to conduct system failure testing. This is achieved by testing the system by temporarily turning off its connection to various components like SCADA data or forecast rainfall. During these tests, failover strategies can be employed where necessary. Commercial software like ICMLive comes with features to support data outage alerts and backup data feeds. For instance, if a critical rain gauge stops reporting, an average of the surrounding reporting rain gauges can be used as a backup. Achieving these improvements requires building the IT systems, engineering processes, and operation procedures around a team dedicated to continuously improving the system performance, one alert at a time.
Thames Water Case Study
In a recent case study using ICM Live, Thames Water identified a sewer collapse and needed to be prepared for any further damage while fixing the pipe. The incident involved a brick-lined pipe underneath a railroad that started to collapse and cave in, causing a blockage and only allowing for partial flow. When the incident was reported, it was initially estimated that it would take nine weeks before the relining of the pipe would be completed, which would have been a long time to go with the only partial flow in the pipe. For the pipe to be influx for such a long period in those conditions, it could have created even more damage if the flows had gotten too high and could have created overflows or backup flows.
The modeling team was asked about rerouting options and expected flows. For instance, considering what level thresholds would be safe to get by with because they might need to deploy teams for bypassing pumping and responding to overflows at a certain point. The modeling team looked at this case and already had a real-time model connected to live data that enabled them to deploy a scenario and see what would happen if the pipe was blocked partially for long periods. ICMLive could forecast potential incidents using four different level thresholds, with each alert type assigned to specific action plans and dispatched to crews. Over a period of nine weeks, ICMLive issued 14 different alerts, triggering two instances of sending out crews to respond to temporary over-pumping of flows. Moreover, due to the intervention, no flooding or pipe damage occurred.
Live modeling provides a real-time view of your water systems performance, enabling you to alert your team faster than ever before in adverse events, such as pipe failures, blockages, breakages, or maintenance repairs. Working in real-time allows you to plan for significant storm or flooding events, helping to protect lives, reduce damage, control spills, and overall manage your water systems more effectively.
See how live water modeling can help you optimize the full potential of your water system today!
To learn more, visit: InfoWorks ICM.