34 n MACHINE VISION April 2026 www.drivesncontrols.com Fan-maker uses machine vision to eliminate defects The failure of engine cooling fans in commercial vehicles can result in overheating and catastrophic engine damage. To reduce these risks, a manufacturer of engine components is using an advanced machine vision system to enforce zero-defect quality standards. The manufacturer, a tier-one supplier to many automotive firms, specialises in thermal management systems for heavy-duty applications such as trucks, construction equipment and commercial vehicles. The cooling fan assemblies are critical components in engine thermal regulation, where failure can result in significant downstream costs. Vision On Line, a German specialist in machine vision and robot technologies, was chosen to plan, integrate and commission the installation. It embedded the systems at four balancing stations, creating a synchronised inspection and data feedback loop to the manufacturer's central control infrastructure. “The company wanted to increase the level of automation in its production process and ensure that only flawless fans would be delivered to customers for later installation in engines to avoid consequential costs in the event of a defect,” recalls Vision On Line’s managing director, Andreas Schaarschmidt. The application raised substantial technical challenges: the manufacturer did not have automated fan inspection capabilities, while the size of the fans, with diameters of up to 800mm, required a vision system capable of capturing an unusually wide field-of-view without affecting measurement precision. The inspection process starts with the vision system reading a 2D barcode on the fan’s metallic hub. This code contains modelspecific parameters that the system decodes and transmits to a PC-based supervisory control platform, which responds by retrieving the corresponding tolerance specifications and target values for that fan. The system then performs geometric validation. Primary inspection criteria include verifying the number of blades, and analysing the dimensions of the individual blades. The system detects any material anomalies – such as flash caused by over-moulding and material voids – with a tolerance of ±1mm, ensuring that even minor manufacturing defects are identified before balancing begins. The measurement data is compared in realtime with target parameters. Components that meet all of the specifications proceed to static balancing, before being shipped to OEM customers. Any fans that do not comply are rejected automatically and follow a route that depends on the defect. Marginal deviations are sent for reworking, while out-ofspecification components are scrapped. The main image acquisition device is a SVSVistek Exo183MGE camera with a 20-megapixel sensor that captures full-frame images at 120MB/s using the GigE Vision protocol. A fourchannel strobe controller provides microsecond-level LED synchronisation eliminating the need for external controller hardware, and reducing both complexity and costs, while maintaining the precise illumination coordination needed for high-speed inline inspection. The images are evaluated using MVTec’s Halcon software library. Vision On Line specified a Büchner TLS brightfield array light, for illumination. It has 3,500 blue LEDs across a 600 x 600mm active area, delivering an optical output of 177W. “The required precision could only be achieved by selecting a large brightfield illumination from Büchner and a highresolution 20MP camera from SVS-Vistek, combined with matching optics from OPT,” Schaarschmidt explains. n A company that manufactures fans used to cool heavy-duty engines has implemented a machine vision system to ensure that defective fans do not enter the supply chain, potentially causing engine failures. The project was challenging, not least because the 800mm-diameter fans required an inspection system with an unusually wide field-of-view. The inspection system had to cope with engine-cooling fans up to 800mm in diameter (Image: Vision On Line) The vision system can detect material anomalies, such as over-moulding or material voids, with a tolerance of ±1.0mm (Image: Vision On Line)
RkJQdWJsaXNoZXIy MjQ0NzM=