Hyundai Construction Equipment (HCE) has partnered with AWS to unveil an AI-enabled solution to diagnose failure in construction equipment. The solution is expected to drastically reduce the downtime of construction equipment (suspended operations time), HCE said.
The AI-enabled construction equipment failure diagnostics technology developed by HCE in collaboration with AWS plays a pivotal role in the smart construction roadmap, built around AI, which facilitates maintenance service with remote failure diagnostics to significantly enhance the durability, lifecycle and safety of the equipment.
Combining AIoT with offline services
HCE Managing Director Pan-young Kim said: “We expect to combine AIoT services with separate offline services to improve customer’s equipment uptime after the free product warranty period. In addition, we will shift to a business model that offers not only physical equipment but also operational uptime time services.”
The two companies have adopted AIoT modules to harness remote data. Using AIoT technology that combines AI and IoT, HCE collects operating status data from numerous pieces of equipment running on IoT platforms and applies the unsupervised learning technique (a machine learning technique that finds patterns and features of specific datasets) to massive data collected in such a way. As a result, HCE secures data on normal status.
Theis normal status data are fed to AI model learning operations to develop a technique that can detect anomaly out of the normal status range.
The AI-enabled anomaly detection solution jointly developed by HCE and AWS helps effectively responding to potential failure issues in early stage which were difficult to detect with legacy solutions. It prevents greater accidents or operational downtime.
To achieve that objective, cloud technology is used in accumulating failure data. Cloud is essential in gathering massive data and upgrading AI performance, thereby an AI solution that can distinguish the types of equipment failures, identify the root causes of such failures, and predict future failures, HCE is operating machine learning training by using a variety of failure diagnostics data on cloud platform, can be developed.
HCE also showcased an AI model that detects failures in hydraulic components. The AI model was trained based on the data of normal hydraulic operations. HCE trained the AI model using idling test data in various modes for technology demonstration, and mounted the AI model on applicable equipment as an AIoT module.
During the technology demonstration proceeded in collaboration with Yongin Technology Innovation Center, the AIoT module detected irregular pump regulator pressure patterns in equipment and transferred the data to a monitoring system over an LTE network.
Subsequently, HCE will further upgrade the AIoT cloud service to provide optimum power mode, protection of workers and pedestrians, voice recognition and remote control based smart interface, autonomous work, and predictive maintenance features.