Diagnosis Prediction of Construction Vehicles and Model Explainability Industry 4.0 Implementation
Main Article Content
Abstract
With the developing technology, the production and use of computerized machines and vehicles have gathered momentum and consequently the amount of information obtained has increased gradually. The conversion of continuously increasing and streaming data stacks into logical and useful information becomes more and more important. Various methods and algorithms have been developed for this purpose. This group of methods and algorithms, which are called data mining in the most general terms, have been combined with statistics and turned into methods of more comprehensible and logical solutions.
Today, many devices produced are equipped with electronic circuit elements and software. Thus, the operation and control of these devices becomes convenient and it is possible to detect any failure that may occur in the devices in such a way that it does not stop the flow of the process. This has resulted in significant returns in terms of cost and time.
BorusanCAT, one of the world's leading companies in the production and maintenance of construction vehicles, aims to use electronic circuits in its equipment in the most effective way. Construction vehicles of Caterpillar Inc. sold by BorusanCAT are equipped with sensors used to detect major faults that require replacement. With the help of these sensors, the data related to the operation of the construction vehicles are provided instantly via satellite over GPS and GPRS.
In this study, an anomaly detection model has been developed to provide early fault detection and vehicle maintenance needs by using instant data obtained from Caterpillar Inc. construction machinery (vehicles). With the Early Warning System (EWS), primarily, the selected sensor data coming from the satellite related to the vehicles is used to predict the failure possibility of the vehicles in a certain time ahead remotely by using the methods of machine learning methods and using the internet of things and cloud technology. After all, prediction data is integrated into business process decision-making with the addition of model explainability.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
References
Internet: What is Industry 4.0?, (2018), from www.armolis.com/akademi/endüstri-40
Internet: Recommendations for implementing the strategic initiative INDUSTRIE 4.0, (2013), from http://alvarestech.com/temp/tcn/CyberPhysicalSystems-Industrial4-0.pdf
Achim Bertsche, & Thorsten Engelhardt, (2001). Monitoring of The State of a Motor Vehicle Using Machine Learning, US.
Claude-Nicolas Fiechter, & Mehmet H. Göker, (2001). Method and System For Condition Monitoring of Vehicles, US.
Shi Jinyuan, Yang Yu, Deng Zhicheng, & Wang Xingping, (2007). Service Life Predicting Method and System for Machine and Vulnerable Component, CN.
Richards, W. G., Martin, K. D., Lieurance, W. L., & Kosler, M. M. (2013). U.S. Patent No. 8,494,826. Washington, DC: U.S. Patent and Trademark Office.
Yasukawa, K., Adachi, K., Uwatoko, K., Yamada, N., Nakagawa, E., & Satonaga, T. (2012). U.S. Patent No. 8,132,049. Washington, DC: U.S. Patent and Trademark Office.
Discenzo, F. M., Chung, D., Kendig, M. W., & Loparo, K. A. (2009). U.S. Patent No. 7,581,434. Washington, DC: U.S. Patent and Trademark Office.
Rumi, E. M., Zepf, P. J., Baird, J., Norris-Lue, C., Rintjema, T., & Wu, Y. (2007). U.S. Patent No. 7,218,974. Washington, DC: U.S. Patent and Trademark Office.
Quinlan, J. R. (1996). Improved use of continuous attributes in C4. 5. Journal of artificial intelligence research, 4, 77-90.
Chen, T. (2014). Introduction to boosted trees. University of Washington Computer Science, 22(115), 14-40.
Friedman, J. H. (2001). Greedy function approximation: a gradient boosting machine. Annals of statistics, 1189-1232.
Internet: Gradient Boosting, (2018), from https://devhunteryz.wordpress.com/2018/07/11/gradyan-arttirmagradient-boosting/
Mertler, C. A., & Vannatta, R. A. (2005). Advanced and multivariate statistical methods: Practical application and interpretation (third edition). United States: Pyrczak Publishing.
Emre YAKUT, & Eray GEMİCİ, (05, 2017). Predicting Stock Return Classification through LR, C5.0, CART and SVM methods, and Comparing the Methods Used: An Application at BIST in Turkey
Internet: Classification: ROC Curve and AUC, (2019), from https://developers.google.com/machine-learning/crash-course/classification/roc-and-auc
Lundberg, Scott M., Gabriel G. Erion, and Su-In Lee. “Consistent individualized feature attribution for tree ensembles.” arXiv preprint arXiv:1802.03888, 2018.