Detection of machine failure conditions on industrial labeling machinery using laser displacement sensors
Document typeMaster thesis
Rights accessRestricted access - author's decision
All rights reserved. This work is protected by the corresponding intellectual and industrial property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public communication or transformation of this work are prohibited without permission of the copyright holder
This project focuses on the use of laser displacement sensor to have a better control of the labeling process of plastic bottles of dish wash soap. As the rotatory labelers on the manufacturing lines have multiple variables to adjust and present frequent operator interactions, this complexity of the machinery is a main area of losses as a wrong adjustment of the machine will most certainly produce badly labeled bottles. A laser displacement sensor will be used to measure the movement of the wipe over time. The wipe is the part of the machinery that pushes the label against the plastic bottle, the wipe is placed just after where the label is dispensed and has not yet contacted the bottle. As the wipe is flexible to a certain extent, its movement over time can be measured with the sensor. Firstly, some tests at a lab scale will be executed, where the laser displacement sensor will be integrated in an isolated rotatory labeler, where different induced failure tests will be performed to see how the movement of the wipe changes when a parameter of the machine is not properly adjusted. That will align expectations on how a failure condition can be detected using the signal from the sensor and how consistent the movement of the wipe is when labeling consecutive bottles. Secondly, after the tests executed at the lab, the next step is to temporarily integrate the sensor at the real manufacturing line to collect data from the wipe during normal production as well as on induced failure tests. The manufacturing data from the sensor will show the amount of noise that the manufacturing environment presents, as the machinery at manufacturing is bigger and connected to other machinery that the lab does not have. The tests from manufacturing will uncover some problems with the machine that were not known, such as patters found in the sensor’s data that the machinery is causing. Concepts from process improvement as the DMAIC approach will be used to analyze the problem, find its root cause, and define actions to solve or mitigate the problem. Finally, once having proven the potential of the sensor at a manufacturing scale, the recommended actions and future work will be defined, considering the permanent integration of the sensor at the line and defining how the data should be processed in order to identify issues of the machinery and which plots of the data from the sensor can be useful to follow and control the situation on the line.