Monitoring cone crusher wear parts : visual tools for decision making
Talvitie, Topi (2023)
2023
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2023112832287
https://urn.fi/URN:NBN:fi:amk-2023112832287
Tiivistelmä
The purpose of this theses was to create ways to indicate the predicted performance and lifetime of a given crusher wear part set while it is still in operation. Goal being to enable customer’s process operators and maintenance management to plan their respective actions proactively according to reliable information.
The data used in this thesis was collected from 2017 to 2021 by customer’s process management. The theoretical part of the work aimed to understand the crushing phenomenon and the influence of different crusher parameters to the performance of the crushing. The data was then enriched and made uniform for the analysis part of the work. Correlations and groupings between datasets were analyzed to create indicators that could be used to predict the performance.
No groups differentiating the wear part sets could be found in the analyses. Data sets were evenly distributed along the variance, with poor and good performance datasets populating same areas. These results suggest that either the labeling to poor and good is unreliable, or the perceived performance is not related to the crushers performance.
The findings indicate that the quality of this kind of data collection is not good enough to generate reliable results. Higher resolution of data points is needed to allow either finding differentiating short term events, or to conclude that such are not present. A more holistic approach encompassing the whole process at crushing plant level, would also be beneficial. Such approach allows for more data driven evaluation of the performance. Currently the presence of operator’s subjective opinion can be influenced by factors or events unrelated to crusher performance.
The data used in this thesis was collected from 2017 to 2021 by customer’s process management. The theoretical part of the work aimed to understand the crushing phenomenon and the influence of different crusher parameters to the performance of the crushing. The data was then enriched and made uniform for the analysis part of the work. Correlations and groupings between datasets were analyzed to create indicators that could be used to predict the performance.
No groups differentiating the wear part sets could be found in the analyses. Data sets were evenly distributed along the variance, with poor and good performance datasets populating same areas. These results suggest that either the labeling to poor and good is unreliable, or the perceived performance is not related to the crushers performance.
The findings indicate that the quality of this kind of data collection is not good enough to generate reliable results. Higher resolution of data points is needed to allow either finding differentiating short term events, or to conclude that such are not present. A more holistic approach encompassing the whole process at crushing plant level, would also be beneficial. Such approach allows for more data driven evaluation of the performance. Currently the presence of operator’s subjective opinion can be influenced by factors or events unrelated to crusher performance.