Scholarlink

<p>In certain metallurgical industry, recovery of metal ingot (produced by extraction from a suitable intermediate compound) involves some human involvement after discharging. Following guidelines of Industry 4.0, automation or semi-automation of such processes is recommended to boost output while taking overall safety into consideration. This work presents design and development of a scheme and algorithm for a semi-automated material handling system for detection and retrieval of ingot from the discharge pit. First, Convolution Neural Network (CNN) based program YOLOv3 is customized to detect the ingot in its discharge environment. Using depth sensing cameras, gravity-actuated gripper, and electric overhead traveling (EOT) crane, a computerized remote handling scheme is developed deployment of which can enable human operators to access and retrieve the ingot from its discharge pit from the control room. A graphical user interface (GUI) for the remote handling system is developed which guides the operator in a step-by-step manner from ingot detection to ingot pickup by gripper and dispatch by the EOT crane. A scaled down prototype of the remote handling system was fabricated and tested with a mimic system to demonstrate applicability and reliability of the system. The overall architecture is scalable which ensures that same scheme of sensors, algorithms and scaled up system can be used in industry.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development and Demonstration of Scaled-Prototype for Semi-automated Detection and Retrieval of Ingots from Simulated Post-Discharge Environment

  • 期刊论文

  • 发表日期：

    2026年3月17日

查看全文
  • Amit Kumar Mallick,
  • Sourav Rakshit,
  • G. Saravana Kumar,
  • Swarup Kanti Gupta,
  • Rahul Sakrikar

摘要

In certain metallurgical industry, recovery of metal ingot (produced by extraction from a suitable intermediate compound) involves some human involvement after discharging. Following guidelines of Industry 4.0, automation or semi-automation of such processes is recommended to boost output while taking overall safety into consideration. This work presents design and development of a scheme and algorithm for a semi-automated material handling system for detection and retrieval of ingot from the discharge pit. First, Convolution Neural Network (CNN) based program YOLOv3 is customized to detect the ingot in its discharge environment. Using depth sensing cameras, gravity-actuated gripper, and electric overhead traveling (EOT) crane, a computerized remote handling scheme is developed deployment of which can enable human operators to access and retrieve the ingot from its discharge pit from the control room. A graphical user interface (GUI) for the remote handling system is developed which guides the operator in a step-by-step manner from ingot detection to ingot pickup by gripper and dispatch by the EOT crane. A scaled down prototype of the remote handling system was fabricated and tested with a mimic system to demonstrate applicability and reliability of the system. The overall architecture is scalable which ensures that same scheme of sensors, algorithms and scaled up system can be used in industry.