Visualization Support for Design of Manufacturing Systems and Prototypes – Lessons Learned from Two Case Studies

Summary

This paper presents two case studies in which a framework for classifying the needed Level of Detail, Level of Accuracy and Level of Recognizability for 3D-scanns are used to 1) support installation of a robotic system for cleaning of fish processing lines and 2) support a retrofitting engineering project. Both cases are set in the Norwegian Aquaculture Industry. In Case 1, effort is done to develop a robotic cleaning solution for fish processing plants, due to a need to rationalize and automate the process. The chances of errors in the manual cleaning process is large. 3D-scanning is successfully used to create a solid model of processing equipment which in turn is used to create a cleaning path for the robot. In Case 2, the point cloud from 3D-scanning is used to check a planned layout of a retrofit project against the actual processing plant. Typically, such retrofit projects take more time and costs more money than initially planned because of unforeseen rework is necessary. This often is a result from poor or missing documentation of the existing processing plant. During the project, several errors were discovered in the planned installation due to missing or wrong information about the existing plant. Both cases show that point clouds from 3D-scans greatly enhances communication, can aid in getting rid of design errors in the planning phase and can help shortening installation and commissioning times. 3D-scans are also beneficial when developing robotic simulations in complex environments. The framework helps in classifying the needed amount of work for 3D-scanning projects based on what the needed output is, thus potentially mitigating unnecessary resources being spent on either the scanning itself or post-processing of scan-data.