Trevor English
Every built thing in the world has a drainage plan. From housing developments to commercial properties to roads and railways – each of these unique engineering projects requires resilient stormwater infrastructure and sustainable drainage design. For much of human history, engineers and builders took a more rudimentary path towards drainage: dig holes, oversize pipes and canals, and otherwise pour more concrete to get stormwater into sewers – and away from infrastructure as fast as possible.
These design decisions were the most efficient in the past, but how do they hold up to today’s challenges of climate change, population increases, and urban creep? Objectively, not well.
With climate change exacerbating extreme weather events and shifting hydrologic models, our traditional drainage systems have failed us repeatedly. It’s evident that engineers, designers, water modelers and the like need to advance the way that we design drainage systems. Not only to be more efficient and sustainable for the weather events of today, but to better anticipate and manage the extreme deluges of tomorrow. So what tools do we have at our fingertips?
The drainage design tools of tomorrow, today
The Autodesk InfoDrainage team is excited to introduce a new AI-based tool called the Machine Learning Deluge tool. It enables designers to predict flood maps quickly and accurately when applying water on the site surface, all while leveraging new built-in artificial intelligence. InfoDrainage’s machine learning algorithms, which have been trained on more than 10,000 traditional simulations, will identify patterns between the ground model and the resulted flood map and then present the user with the flood map showing hotspots which can be utilized to establish the best locations for storage structures and stormwater controls, like ponds, swales, or rain gardens.
These hotspots can be used to identify locations which infrastructure owners and developers should avoid when deciding where critical buildings like metros, schools, and hospitals should be constructed. It also identifies the water pathways which can inform decisions related to the design of evacuation routes.
This functionality runs without lengthy simulations and provides native intelligence to engineers, enabling their models to be more responsive and efficient to real world conditions.
With this introduced, let’s spend some time understanding how this will impact the designs of today and influence the workflows of tomorrow by leveraging the power of artificial intelligence.
4 ways ML Deluge can help
Right out of the gate, we think this tool will improve the way you work in at least four ways:
- Faster site analysis: InfoDrainage’s ML Deluge tool can analyze and predict channeling and ponding, enabling more informed decision making in the drainage design process.
- Enhanced model stability: Machine learning algorithms are not vulnerable to hydraulic modelling instabilities as no equations are being solved to establish where the water is. ML relies on spotting patterns which is a stable process that will ensure users locate the best possible positioning for Stormwater Controls (SWC).
- Scalability: The tool efficiently handles various project scales, from small residential to large commercial.
- Iterative design: The tool allows for iterative, interactive and real-time modification, encouraging users to make a change, get instant feedback on the flood map, understand the impact, and implement improvements to the design.
Bringing the enhanced AI functionality in InfoDrainage into the real world even further, let’s examine a brief sample use case. Say a drainage designer is at a crossroad trying to choose where to place stormwater controls. To efficiently determine where they need to place them, the designer can load the model into InfoDrainage, alongside the placement of any buildings, and run the ML Deluge tool in seconds, not minutes or hours. What the user is left with is an immediate flood map with hotspots showing them insights on stormwater control placement, assisting and enabling a better design process.
This initial implementation of the machine learning tool marks the start of further integration into our product roadmap of artificial intelligence as a tool for drainage designers inside of InfoDrainage.
You can see a technical preview of this functionality in action in the brief demo video below – this shows the interactivity element and how users can move ponds around to get instant feedback on the flood map as a result of their changes.
Our mission
At Autodesk, specifically within our industry-leading water team, we’re focused on building more resilient and sustainable water systems for all, within the entire scope of water lifecycle management, from cloud to sea. Keeping this mission top of mind inspires our teams to build better tools and build more efficient workflows for the drainage design industry, underscored by these new features within InfoDrainage.
All current subscribers of Autodesk InfoDrainage now have access to the Machine Learning Deluge tool in the latest update. If you’re not a current subscriber, we encourage you to download a 30-day trial of InfoDrainage and try it out.