News

Horse Powertrain has formally integrated the Siemens Inspekto AI visual inspection platform at its Skövde engine manufacturing facility, marking a decisive shift from manual quality verification toward automated defect classification. Operating under the Aurobay Technologies division, the system interfaces directly with the plant’s existing material handling infrastructure, where engines arrive via automated guided vehicles before undergoing rapid surface and structural analysis. A collaborative robot-mounted vision array executes multi-angle imaging and real-time machine learning inference, identifying dimensional deviations, surface irregularities, and assembly misalignments within seconds. Each captured image and classification decision is archived within the plant’s manufacturing execution system (MES), establishing a fully traceable, searchable quality ledger that supports root-cause analysis and continuous process improvement. Unlike traditional fixed-vision systems that rely on rigid rule-based thresholds, the AI engine adapts dynamically to variations in lighting, material reflectivity, and component geometry across plastic, metallic, and composite powertrain assemblies.

The architectural simplicity of the deployment aligns with modern Industry 4.0 integration standards, eliminating the need for dedicated machine vision engineers or prolonged commissioning cycles. Production technicians can deploy, calibrate, and scale inspection checkpoints directly from the shop floor, adjusting detection parameters without modifying underlying robot kinematics or PLC ladder logic. Within two hours of physical installation, the system undergoes live product evaluation, allowing operators to validate go/no-go thresholds and integrate seamlessly with the Simatic S7 control network and HMI dashboards. This rapid time-to-value directly addresses a persistent challenge in discrete manufacturing: the gap between prototype validation and full-scale production line deployment. By decoupling vision configuration from complex software stacks, the platform empowers existing engineering teams to iterate inspection protocols as product designs evolve or new defect categories emerge, maintaining alignment with stringent automotive quality management standards such as IATF 16949.

Operational leadership at the Skövde facility emphasizes that the technology functions as a knowledge multiplier rather than a replacement for human expertise. Decades of accumulated process intuition from veteran line engineers are codified into the AI’s training datasets, ensuring that nuanced defect recognition scales consistently across shifts, operator rotations, and seasonal demand fluctuations. The collaborative nature of the deployment maintains human oversight while removing the ergonomic strain associated with repetitive manual inspections. By offloading high-frequency visual scanning to the AI system, skilled personnel can redirect their attention toward predictive maintenance scheduling, process optimization, and cross-training initiatives that elevate overall equipment effectiveness (OEE). The flexibility to position inspection nodes at intake, mid-line assembly verification points, or final quality gates allows the plant to implement a tiered quality control architecture, catching non-conformances earlier in the production flow and minimizing scrap generation.

The integration of no-code AI vision into established automation environments signals a broader industry transition toward democratized quality assurance. Traditional machine vision deployments often require extensive optical calibration, specialized lighting rigs, and proprietary programming languages that create vendor lock-in and prolong ROI timelines. The Inspekto framework bypasses these constraints through standardized Ethernet/IP communication protocols, native TIA Portal connectivity, and modular mounting architectures that adapt to constrained footprint requirements. As powertrain manufacturers face increasing pressure to accelerate production cycles while maintaining zero-defect tolerances for low-emission engine platforms, scalable inspection solutions become critical infrastructure rather than auxiliary tools. The Skövde deployment establishes a replicable blueprint for brownfield modernization, where legacy control systems remain intact while AI-driven analytics operate as an intelligent overlay, delivering immediate throughput gains and reducing reliance on subject-to-variation manual verification processes. Future iterations will likely incorporate edge computing enhancements and federated learning capabilities, enabling multiple facilities to share defect recognition models while preserving localized data sovereignty.

Written by: David Edwards | With over fourteen years of experience in industrial automation and machine vision architecture, David specializes in deploying AI-driven quality control systems, integrating collaborative robotics into legacy PLC environments, and optimizing production traceability for discrete manufacturing operations.