Application of a neural network integrated with the internet of things sensing technology for 3D printer fault diagnosis
- 2 Downloads
As a rapidly developing and maturing technology, 3D printing has substantially simplified the manufacturing of complex components. To improve the safety and practicability of machinery, fault diagnostic systems have become essential. This study employed a neural network algorithm to implement fault diagnosis. This study proposed a 3D printer fault diagnostic system that incorporated a neural network with a graphic interface. The MPU6050 accelerometer was combined with an Arduino microcontroller to monitor the status of a 3D printer. The neural network simulated the relationships of faults and training, and adopted different transfer functions to compare their training and convergence performance. The data were transferred to Laboratory Virtual Instrument Engineering Workbench after training and were displayed on a human machine interface, enabling users to explicitly identify the potential location of faults in a machine. This system ameliorates the work efficiency of managers and achieves real-time fault diagnosis. Moreover, the proposed 3D printer fault diagnostic system identified neural network parameters (e.g., the number of neurons, the learning rate, and the number of training) suitable for this system through various static and dynamic simulation experiments. Various shapes were adopted for testing the detection rate of the 3D printer fault diagnostic system. Finally, the real-time dynamic printing status data were recorded for approximately 10 min consecutively, yielding 6000 pieces of data, and that the transfer functions in the hidden and output layers were adopted for static simulation. The results showed that the overall fault detection rate of the system was as high as 83.5%.
This study was supported in part by the Ministry of Science and Technology MOST 106-2622-E-150-018-CC3 and National Formosa University 107AF06.
- Chang HC, Lin SC, Kuo CC, Hsieh CF, Lin CY (2016) Fuzzy theory-based hybrid decision-making system for rotating electrical machinery fault diagnosis. In: 2016 international conference on fuzzy theory and its applications (iFuzzy), pp 1–6. https://doi.org/10.1109/ifuzzy.2016.8004928
- Huang HP, Hsiao ZY (2002) Development of remote control system of a semiconductor cluster tool. In: Proceedings of the IEEE international conference on systems, man and cybernetics, pp 1–6. https://doi.org/10.1109/icsmc.2002.1173222
- Jaai R, Challa V, Pecht M (2010) Prognostics and health management based qualification. In: 4th Asia–Pacific international symposium on advanced reliability and maintenance modeling, Victoria University of Wellington, pp 265–272Google Scholar
- Montes JO (2016) Impacts of 3D printing on the development of new business models. In: 2016 IEEE European technology and engineering management summit (E-TEMS), pp 1–9. https://doi.org/10.1109/e-tems.2016.7912605
- Tao S, Zhang T, Yang J, Wang X, Lu W (2015) Bearing fault diagnosis method based on stacked autoencoder and softmax regression. In: 2015 34th Chinese control conference (CCC). https://doi.org/10.1109/chicc.2015.7260634
- Tsai YT, Chen C, Liao Z (2012) A study of performance degradation predictions and fault diagnoses for mechanical systems. J Technol 27:121–129Google Scholar
- Weng PH, Yang JP, Liu FT, Huang HC, Yu KW, Hwang RC (2008) An intelligent fault diagnostics for turbine generator by modified neural model. In: 2008 3rd international conference on innovative computing information and control, Dalian, Liaoning, pp 296–296. https://doi.org/10.1109/icicic.2008.151
- Zhao W, Wang Z, Lu C, Ma J, Li L (2016) Fault diagnosis for centrifugal pumps using deep learning and softmax regression. In: 2016 12th world congress on intelligent control and automation (WCICA). https://doi.org/10.1109/wcica.2016.7578673