Back to search results
  • Automated Machine Learning results in increased productivity for engineering team
  • AI model calculates accurate lift and brace hardware locations for the erection of enormous tilt-up concrete walls
  • Increases anchor placement calculation speed, allowing engineers  to be reallocated to other complicated challenges
Division(s)
CRH Americas
Product Type(s)
Construction Accessories

Engineers at Leviat, CRH’s construction accessories business, have designed and implemented the use of artificial intelligence (AI) to improve the speed and efficiency of how concrete wall panels are lifted into place.

The initiative involved Leviat engineers using a decade’s worth of panel geometry data to “teach” a machine learning algorithm how to determine the proper locations for lift anchor placement.

Lifting precast concrete wall panels requires acute precision

Site-cast precast concrete wall panels (commonly referred to as tilt-up panels) regularly weighing 50 tons or more - which is the equivalent of a fully-loaded semi-truck – are used extensively for large-scale structures such as logistic warehouses and data centers. Their size and weight provide excellent structural integrity and durability, withstanding environmental stress, heavy loads and long-term wear.

Although tilt-up construction delivers new buildings quickly, the placement of such expansive concrete panels requires extensive analysis and planning.

Typically, engineers calculate the precise locations of lifting anchors and brace connections to ensure these supersized elements remain secured throughout the construction process, protecting the building’s integrity. To expedite the process, Leviat’s North American engineers turned to the fast-growing field of machine learning and AI.

Improving engineering outcomes with efficiency and precision

Though the trial is still under scrutiny, early results  demonstrate the Projects’ AI model accurately predicting the placement of over 80% of the lifting hardware required for nearly 700 test panels in under eight minutes – shaving days, and even weeks off the time it would take a human engineer.

The new model reduces the Projects’ demand on engineers’ time and frees them up to be reallocated to other complex challenges, improving ability to serve customers while reducing overall operating costs.

As the trial continues, engineers are focused on refining the AI model, to achieve an even greater level of accuracy in placement prediction and to scale the model to assist with other, complex lifting designs. This model works hand in hand with engineers to increase their output, although a final quality control is still performed by the engineering team.

AI is poised to become one of the greatest technological revolutions in the construction industry. Leveraging AI to its full potential will enable the industry to complete more work, with greater accuracy and productivity and speed – all while continuing to protect the safety of those who work on the site.