Innovation: Artificial Intelligence for quality control of concrete components

Fractures, cracks and spallings within concrete components affect the security of construction projects and must be minimized. Therefore, manufacturers are controlling the components’ quality at different times throughout the production process to avoid damages. However, supervision by the human eye is potentially prone to error due to possible eye fatigue caused by monotonous tasks. Furthermore, there are major differences between the reviewing process performed by different employees. In some cases, a complete manual control cannot be provided because of short cycle times. To simplify this process for the manufacturing of tunnel-segments and to exclude possible sources of error, Preml invented a new system using their own AI.

Additionally to the AI, the system consists of different cameras aligned with the components to be tested. The cameras offer high-resolution images, which can be analyzed by computer, within cycle time. An algorithm detects possible damages and decides whether the components can be processed or need to be sorted out for repair or discard. The analyzation takes approximately 15 seconds and includes the storage of image- and metadata for a complete quality documentation. By means of a dashboard, the decisions made by the AI can be visualized, to guide the employee through the repair- or sorting process. Due to the AI, the system is mostly protected against tough conditions on site, such as dusty environments. Preml also promises an easy adaption to changes. The system, which was invented for the production process of tunnel segments, can also be used for small concrete components and other precast elements.