Sany welders train robots for Ningxiang workshop

Seven intelligent welding robots are now working in Sany’s Ningxiang crane-boom workshop, according to Xinhua’s November 14 factory visit. One worker can oversee two machines instead of watching one, chiefly handling loading and unloading. Experienced welders helped train the system for roughly eighteen months, supplying demonstrations that the team evaluated. Their craft has become part of the machines’ training material.

The site has a longer automation history. In March 2025, Sany described a separate hybrid laser-and-arc process intended to reduce deformation in crane-boom panels. Its August manufacturing meeting included an AI-welding demonstration in Ningxiang. These are different layers of factory improvement: changing how a joint is made and changing how a machine responds while making it. A factory can improve one layer while continuing to refine the other.

Supplier IROOTECH’s earlier automotive case offers another comparison. In 2023, the company described collecting electrical resistance, current and heat data to flag abnormal welds and connect records to individual vehicles. That application concerned quality monitoring. The Ningxiang report concerns adjustment during welding. Together, the examples illustrate why “AI in a factory” can mean several distinct jobs rather than one interchangeable capability.

TWI, the British welding research organization, explains that robotic welding sensors can monitor either process conditions or joint geometry. Current and wire-feed measurements serve a different purpose from information about a seam’s position or gap. Think of tracing lettering on a sheet of paper: a memorized movement works only while the sheet stays where expected. If it shifts, the pen must follow. Sensors and feedback control provide that correction; this engineering foundation predates the current enthusiasm for AI.

The useful question is how work changes around the machine. A skilled operator can contribute examples and judgment rather than continuously steering every movement. Material handling remains a human task at the reported workshop. For manufacturers, the next test is consistent operation across a wider range of components. For customers buying the finished crane, an impressive robot demonstration is less meaningful than a dependable load-bearing structure. Greater automation changes the production process, while the obligation to manage quality remains. Patterns learned from experienced people still have to earn their place on a working production line.

Original source published: Xinhua, November 14, 2026, 16:32:33 Beijing time / 3:32:33 AM EST. Sany background: March 13 and August 30, 2025. IROOTECH case: November 28, 2023.