Urbio’s AI software accelerates the heat transition in urban areas

This article, authored by Sébastien Cajot, was originally published in the Euroheat & Power Magazine in July 2024.

While tools like ChatGPT have changed the game for tasks like copywriting and image generation, similar breakthroughs in AI are redefining how the wave of low-carbon energy systems in cities throughout Europe are planned. In 2024, Urbio released a patent-pending solution to enable utilities and consultants to keep up with the unprecedented rate of change.

Accelerated heat planning

Planning the energy systems of tomorrow – heat pumps, district heat networks (DHN), solar panels, and more – is a highly complex task, consuming large amounts of time and resources. What was once a relatively straightforward framework of predictable energy flows from central providers to consumers is seeing roles and boundaries blurred with the emergence of distributed “prosumers” and the reliance on intermittent resources (Figure 1). In addition, fragmented and sparse data sources cast uncertainty and cause financial risks. With trillions of euros pouring into the energy transition and a desperately scarce workforce to take on the workload ahead, effective planning is vital to ensure local stakeholders get their priorities straight, and so every euro invested maximizes both its financial returns and its CO2 reduction to comply with increasing regulations.

Figure 1: Evolution of urban energy systems towards increasingly complex and interconnected systems. 

The ball is already in motion, and pressure from regulators is in effect. In Germany for example, 11,000 municipalities are now obliged to implement heating plans within the next 4 years. To realize these plans, expected costs are upward of €500m annually.

Strong tailwinds are shaping the transition in Europe, and experts predict up to €70 billion in annual investments in district heating networks by 2030. Given its high energy efficiency and reliability, thermal networks are becoming a cornerstone of net-zero buildings, as a superior alternative to heat pumps and solar thermal in highly dense urban environments, or in historic neighborhoods where refurbishment options are limited.

The heating network market is therefore growing fast, as governments set firm deadlines for reaching their targets. France, for example, aims to triple the number of households connected to heating networks by 2035, and utilities and municipalities are rushing to roll out new networks and expansions strategically.

Currently, we lack the human resources to achieve these plans at the required pace and scale. Traditional methods are highly labor-intensive, yet skilled labor is precisely what is missing. Today, actors in the utility space are already noticing first-hand the limits of conventional manual processes – and the potential benefits of digitized workflows.

“Feasibility studies are vital to our work, but also time-consuming given the larger amount of data and scenarios to consider.” - Rafael Mesey, Head of Department New Energy, CKW

Furthermore, data used to inform investment decisions must be structured and readily available, as producing and maintaining dynamic business plans is so time-consuming that engineers often only have time and budget to work on one scenario, even though market conditions constantly evolve throughout a project’s lifetime.

So how can energy providers und cities, ensure they reach their net-zero targets while remaining profitable? Generative design, reliable data, and intuitive visualization of maps are the key ingredients Urbio deploys to speed up the decarbonization of Europe’s building stock.

Generative design for heat planning

Powered by artificial intelligence, generative design automates the creation of georeferenced and quantified scenarios based on the user’s objectives and restrictions. By incorporating large datasets in a structured manner, and by prioritizing the user’s most trusted data sources, the model delivers more comprehensive and precise results than could be achieved before, thus de-risking the investment process. Under the hood, algorithms and data are the cornerstones of Urbio’s generative design solution (Figure 2).

Figure 2: Urbio’s data solution ensures the most accurate data sources are leveraged to inform and de-risk decisions. 

On the surface, Urbio also innovates on how users interact with the underlying algorithms. While similar in name, Urbio’s generative design functionality differs from “traditional” large language models. Tools like ChatGPT excel at verbal interactions with users and at creating textual outputs, but outcomes remain largely unexplainable and limited when handling numbers and spatial information. Instead, Urbio’s patent-pending generative design approach produces quantified, georeferenced, and explainable net-zero scenarios, while enabling smooth user interaction with the scenarios, directly on a map. The interactions result in a synergistic human-computer collaboration: rather than relying on algorithms to produce a final “solution” – which rarely lives up to the complexity of real-world contexts – users iteratively create their plan, incorporating their know-how and context, while algorithms supercharge workflows by crunching numbers on the fly and optimizing CO2 emissions, investment costs, or other goals. As a result, the user remains in the driver’s seat, and their AI “companion” supports their actions in the background (Figure 3).

Figure 3: Three district heating scenarios created with generative design based on user inputs.

By leveraging generative design, decision-makers in the energy space can screen larger territories, much faster than was previously possible. Outcomes are improved as they can compare options at little extra costs, thus increasing ROI. Consequently, engineers and commercial teams spend more time adding value to projects and clients rather than preparing data and scenarios, leveraging their industry expertise, and selecting what options are best suited for the specific project.

Planning district heating networks with Urbio

Swiss utility company CKW wanted to know which areas to prioritize next for new district heating networks. Facing the need for quick action in a highly competitive environment, they were committed to making their planning processes more efficient.

However, their assessments implied significant manual effort: collecting building attributes, applying spreadsheet models for estimating demands, and mapping data to footprint for visualization and prospection purposes in desktop GIS tools. As a result, processes suffered from non-exhaustive data and several manual repetitions, costing time and resources for prospectors.

With Urbio, CKW automated the entire process, saving valuable work hours and letting the team focus on exploring scenarios with the support of maps and key metrics.

Figure 4: CKW leveraged the full suite of Urbio’s 3-in-1 platform to streamline and improve feasibility studies of district heating networks

Using Urbio’s 3-in-1 solution, Rafael Mesey’s team at CKW completed the feasibility studies 5x faster, leveraging 3x more scenarios at higher data accuracy than previously achieved. These factors accumulated in more rigorous and reliable assessments, improving the speed and accuracy of the resulting investment decisions.

“I use Urbio to accelerate the development of our renewable energy activities.” - Rafael Mesey, Head of Department New Energy, CKW

Outlook

Urbio is not only used by CKW, but also by European companies such as the German energy supplier NEW, the Belgian heating network specialist Resolia and and industry-leading energy consulting firms like Energielenker and WSP. Urbio, which was founded in January 2020 as a spin-off of the climate-tech university EPFL, is constantly expanding its product boundaries on multiple fronts with continuous refinements to UX allowing even more user control, enrichment of data sources and processes, and improvements to the platform’s scale capabilities. As such, Urbio remains committed to digitizing and accelerating heat-planning processes across Europe to support the achievement of decarbonization targets.