Startup supplai has been engaged by Volker Engineering Structures (VES; part of Volker Wessel) to develop a method together with DDC Smart Inspections (Dutch Drone Company) for a smarter inspection at the Haringvliet bridge. Diervoor, DDC carried out a drone inspection and developed an advanced 3D model. Based on this data, supplai applied its real-time video analytics based on artificial intelligence (AI). In this way , a drone can now be used to perform rapid and accurate inspections.
The driving deck of the lifting part of Haringvlietbrug consists of aluminum plates that are attached to the flap of the bridge with spring clamps. Due to the traffic that drives over the bridge, these spring clamps shiver loose. The lifting part and the movement mechanism will be completely replaced, but this will take some time. Until then, the maximum speed on the A29 of the Haringvliet Bridge is only 50 km/h.
In order to determine the safety of the Haringvliet bridge, it is very important to inspect the construction properly and weekly. This time-critical inspection requires high accuracy and must be done safely. At the moment, VES employees have to pass under the bridge in a maintenance cart to inspect each plate. During this inspection it is prohibited to pass under the lifting part.
Edge computing
Supplai is an expert in real-time video analytics through artificial intelligence (AI) on edge computing. Supplai recently received a subsidy for this from the MIT R&D AI fund that was paid out for the first time. This Edge computing is mounted directly on the drone during flight and enables real-time video analytics. Due to the use of drones and real-time video analytics, no people are needed to do manual recordings under the bridge.
Because the photos of each flight can be viewed afterwards, a better data-driven way of inspection is created. The degradation is easier to map and there is always reference work. This makes maintenance inspection safer and more accurate. The use of drones ensures that shipping no longer needs to be blocked. The drone only needs to make flights to fully map the underside of the bridge. Unlike slow artificial intelligence (AI) post-processing techniques of drone data, supplai’s edge-computing real-time video analytics can be performed with the help of supplai. These drones collect a large amount of data. The AI deep learning algorithms ensure that the right data is efficiently extracted from the images immediately during the flight.
The advantage of these new techniques is that the inspection frequency can be increased, and immediate action can be taken if a defect is detected. The data-driven way of inspection reduces the chance of errors. The advantage of these new techniques is that the inspection frequency can be increased, and immediate action can be taken if a defect is detected. The data-driven way of inspection reduces the chance of errors. The advantage of these new techniques is that the inspection frequency can be increased, and immediate action can be taken if a defect is detected. The data-driven way of inspection reduces the chance of errors.
3D Modeling
After the inspection, DDC makes a state of the art 3D model. 3D modeling (3D modeling) is the process of creating a mathematical representation of a three-dimensional object. 3D modeling makes it possible to represent objects such as buildings and infrastructure in 3D. This ensures that an inspector can view an object in detail from different angles.
The result: a highly detailed and extensive artificial intelligence (AI) presorted dataset of detected spring clamps, which was then processed into an advanced 3D model of the Haringvliet bridge.
Fast and data-driven maintenance with less deployment of (scarce) employees and increased safety around the sailing area of the Haringvlietbrug.
