Michael Rosh
Calibrating a water quality model accurately can be a daunting task, but it doesn’t have to be. We’ve created a way to speed up and automate the process.
Summary
- Hydraulic model calibration is an iterative and time-consuming process, but one that is necessary to ensure that your model is producing reliable outputs that reflect the real world.
- Autodesk’s InfoWater Pro and Info360 Insight have built-in tools to greatly streamline this process using dynamically calibrated dashboards.
- This allows utilities and consultants they work with to focus on the calibration steps, rather than the data collection and processing.
Virtually every water distribution system model requires some level of calibration adjustment so that your model accurately represents the behavior of the real water system you want to portray, with the overall goal of calibration being to adjust the model outputs to better match valid field data.
Importantly, model validation is the comparison of model results with field data from a time period that was not used during model calibration. In cases where the agreement is poor, it is necessary to understand the contributing factors and recalibrate the model – or those portions of the model as necessary – so that it is accurate for both periods. Before using your model, validation should be performed to ensure that the model still accurately reflects the actual system conditions.
This is especially important because hydraulic models typically do not adequately agree with field data when they are first developed or upgraded. In essence, your model must be calibrated by identifying why the agreement isn’t better and making the proper adjustments to achieve agreement. It is also important to recognize that data collection and model calibration can be an iterative process – as the calibration process unfolds, there may be a need to collect additional data.
General guidelines for hydraulic model calibration
Because no two systems are alike, data collection for model calibration is somewhat of a unique exercise. However, there are some general guidelines that can be applied to all systems to assist modelers in collecting useful data that can be used with confidence for hydraulic model calibration.
Some of the basic principles of calibration are to ensure boundary conditions such as tank levels match, an accurate diurnal curve pattern for demands is used and pump and valve status are set correctly. Additionally, pipe roughness may need to be calibrated to ensure pressures and flow rates align to the real world.
An excellent resource on calibration of hydraulic models can be found in AWWA’s M32 manual. (Full disclosure: I have been a contributor to multiple editions of this manual.)
Hydraulic model calibration without resorting to spreadsheets
Data scrubbing and conversions are typical of many model calibration activities. Most modelers use tools like MS Excel to compile data into one location, especially when there are multiple sub-models that comprise the entire network. The data is then run through conversions, so units match and any processing of the data is done. Data processing is done for the user in Info360 Insight using Analytics Tools to set up any calculations, which in turn become a virtual sensor and continuously updates as new data is introduced. This eliminates the need to use spreadsheets for data processing and allows for a more streamlined calibration routine.
SCADA data are typically the best data source to provide insights into observed system behaviors, such as how flow rates, pressures, tanks, pumps and control valves operate on a near continuous basis. Derived data like diurnal patterns are also useful in establishing regular patterns of system behavior. However, SCADA data comes with its own problems: it may have irregular sampling, it may have dirty data, etc. Thus, it is important to have a tool that can account for irregularities in SCADA data to streamline the model calibration process. Info360 Insight provides users with the ability to identify patterns for demands and many additional patterns in the SCADA data. This functionality greatly simplifies the diurnal curve identification process. You simply choose the date range of the calibrated model, and Info360 Insight will determine the demand pattern.
Automation naturally offers significant time savings
Most utilities have an in-house process or work with consultants for model calibration. However, these processes are often cumbersome, require vast amounts of data harvesting and data processing, and typically involve many different data sources coming from different parts of the system. It is typically a very time-consuming process, ranging in time from days to weeks or even months. We believe Info360 Insight’s ability to streamline this process can save users significant amounts of time, perhaps even as much as 40% based on our testing.
When it comes to frequency of hydraulic model calibration, there are different requirements based on the model’s scope and scale of use. Continuous calibration provides a level of calibration that suits operational modeling applications like water quality analysis or energy modeling, whereas periodic calibration may be good enough for applications like master planning or fire flow simulation. Continuous calibration is the process of regularly (daily, weekly, or monthly) comparing a hydraulic model to real world data and adjusting the model accordingly to ensure it accurately reflects observed real world outcomes.
In this process, the model is populated by reading SCADA data at the desired frequency and time frame, updating the network model boundary conditions and operational statuses and generating the corresponding network analysis results. The model results can then be compared to the SCADA data to validate the level of calibration. Multiple iterations may still be needed to establish a calibrated model for the specified timeframe, but part of that estimated 40% time savings comes from the fact that Info360 Insight allows you to easily manage and remove unneeded calculations from your modeling workspace that don’t align with your calibration goals.
A collaborative workflow approach
Using Info360 Insight with InfoWater Pro greatly streamlines this entire process. Info360 Insight – with its connection to live SCADA, and IoT sources – always has the desired data sources ready for comparison and analysis. Info360 Insight also has built-in data scrubbing tools to identify and remove (as per user-configured rules) anomalous or “dirty” system data. InfoWater Pro has direct integration with Info360 Insight, allowing you to easily ingest SCADA data for boundary conditions and diurnal curves as well as to rapidly publish simulation results to calibration dashboards for analysis. Info360 Insight also allows users to monitor SCADA sensor health by analyzing drift in the data and reporting discrepancies for calibration – or if calibration is being performed regularly, planned replacement of the SCADA sensor.
It also helps with collaboration. The web interface of Info360 Insight allows many to users to collaborate in this model performance workflow: modelling teams can provide simulation adjustments and results, operations teams can provide insights into system behaviors and expectations, consultants can easily access all tools and minimize any time lags in accessing and processing SCADA data – all users can have visibility into model performance and collaborate to agree upon acceptance criteria.
This workflow also allows you to very easily compare one or more simulation outputs with others, with real-world SCADA data by creating multiple dynamic dashboards to facilitate a regular calibration cycle. Iteratively adjusting model parameters and re-running will automatically update the calibration dashboards – users don’t need to do any data processing or data manipulation. With old tools, this can often take a week or more for a single model within a system that could be divided into multiple sub models. In the end, this provides a foundation for a digital twin in which the model is routinely calibrated and always ready for real-time analysis.
The end result: accurate data to feed your analytics
Once the model has been calculated, you can take advantage of a multitude of Info360 Insight tools such as Mass Balance. For example, you can import customer billing data as part of Mass Balance, ensuring a high level of accuracy for the given timeframe. (Mass Balance indicating times of Non-Revenue Water (NRW) and changes in NRW over time.) Combining NRW with digital meters such as AMI allows you to create Demand Management Areas (DMAs) so NRW can be brought down to the street level and identified more accurately. Using the Mass Balance results, you can identify leakage using Infrastructure Leakage Index (ILI) tool, which can calculate and monitor daily water loss per zone and help reduce leakage significantly.
Wade deeper into water analytics
- Want to try this yourself? We have a step-by-step guide that shows how to connect your Info360 and InfoWater Pro model and results. Don’t have Info360 Insight? Let us set you up with a custom free trial.
- Analytics are powerful! Read about how Stantec helped Wellington Water save 20% on electricity costs via pump optimisation.
- The ultimate win – a digital twin. Learn how to generate dynamic digital twin hydraulic models with InfoWater Pro and Info360 Insight
This blog post contains excerpts from the Handbook for Water Distribution System Model Calibration, copyright 2023, American Water Works Association.