AI Program: discover ECable!

How to generate better cable routing plans faster? Discover ECable, one of the twelve projects resulting from Leonard's Artificial Intelligence programme (cohort 2021). Its objective? To help VINCI's teams to become autonomous by implementing AI use cases with a high business impact.

Building a new industrial plant can be exciting, frustrating, overwhelming, but for sure, it is never boring. When it comes to an intensive production site such as steel mills or power stations, several billions of dollars are invested to give birth to a brand-new factory capable of producing a tremendous amount of goods.

Starting from the land acquisition to the production of the first item, many steps must be thoroughly conducted to achieve a successful project that will meet your expectations: obtention of permits & approvals, finding trustworthy partners, scheduling the construction work, purchasing specialized equipment, planning for future expansion, etc.

Nonetheless, there is a particular phase that people are the most excited to get to: design and conception. This phase can take some time: one will work with experts from many areas such as architecture, ventilation, illumination, etc. that will help design a manufacturing plant that meets all of the proper requirements and includes the mandatory features. Still, the results will rarely end up exactly like what one had envisioned…

 

Cable routing as a complex planning challenge

To understand the complexity of each line of work involved in the process, let’s take an example: the wire installation. Imagine one has to build a plan for the installation of 3,000 km of power* and control** cables such that not only must one not interfere with the other installations, but one should also factor in the nature of each cable to prevent electrical interferences, voltage drops, overloaded trays, etc. Now let’s picture that in addition to all of that, when one looks at the costs, dozens of millions of dollars are spent on materials and operations. It then comes as no surprise that so-called cable routing, or the process of optimizing cables’ length and bundling cables together to reduce the number of trays among others, is paramount to both generate significant savings and build a plant that would abide by the necessary regulations it should follow.

Traditionally, the cable routing design is mostly built manually through a trial-and-error methods supported by basic analytical tools. This manual process proves quite painstaking and time-consuming at best, as it locks the qualified workforce of the engineers in repetitive and iterative tasks, when they could spend that precious time finetuning the design or even working on additional projects. At the end of the day, this manual process is quite suboptimal both process-wise and result-wise as the iterative process often makes one ignore some potentially better results than could have been otherwise detected through different methods of tackling the project.

 

How to generate better cable routing plans faster?

On that basis, all experts in the construction field wonder: how could one generate better plans faster? To respond to this strategic concern, the construction sector is paving the way for what has come to be called generative design. An approach that leverages digitalization of processes and AI-methods to give designers and engineers better insight so they can make faster and more optimized design decisions. Evidence of this trend lies in the $240m acquisition of Spacemaker.io, a startup that has developed AI-supported software for urban development, by Autodesk in 2020, or new initiatives by CAD industry leaders like Autodesk’s Dynamo or McNeel’s Generative design modules.

 

AI, data science and graph theory at the rescue

VINCI Energies, always looking for ways to innovate, decided to step into that breach, through a collaboration between two of its business units. With the expertise in electrical and electronic systems of Actemium Large Projects GmbH, the technical capabilities of Axians ICT Austria, and the support of the Leonard AI program, the project eCable was born. A generative design-driven solution that leverages data science and graph theory to automate and optimize cable routing design generation.

To foster the innovation capabilities, the first step was to reorganize the entire data processing pipeline from the initial clients’ data to the final provided cable routing design. Then, through this brand- new structure, a first generative model has been deployed and is able to provide cable routings that respect most of the constraints induced by the building structure. These two steps are setting the ground for an iterative improvement process allowing Actemium to customize their generative-design models taking into account additional constraints. On-going work is conducted to deploy an even better model that would be capable of solving all the constraints, as well as optimizing the cost. In a nutshell, eCable is a step forward in designing greater industrial plants while saving materials and operations.

By Christian Dangl, Axians, and Tobias Kramer, Actemium GmBH

 

*power cable: cable used for transmission of electric power

**control cables: cable used to measure, control and regulate or monitor industrial plants

 

This article is part of a series involving the participants of the AI program by Leonard. The program has been specifically designed to accelerate the adoption of AI technologies within VINCI. It consists in a five-month incubation period where selected VINCI collaborators follow a learning by doing process where they codevelop an AI-based use case under the coaching and mentoring of Leonard’s team and consultants from Eleven Strategy.

> More about Leonard’s AI program

Share this article on