Manufacturing processes like face milling that involve controlled removal of material are energy intensive [1]. Here, the digital twin of a precision machining process is developed that can be used to minimize the cost and power consumption of machining operations and for model-based maintenance. For the latter, a methodology for simultaneous selection of sensors and test settings is developed for the digital twin to determine faults in the machining operations. Robust maintenance algorithms and standard work target energy consumption reduction through early detection of faults in the system that prevent losses from downtime. The method is applied on a simple case study of face milling of metals. Face milling is a precision machining process that involves material removal, scrap generation from metal to metal interaction and cutting fluid utilization for lubrication and heat removal [2]. The digital twin of face milling was formulated in this work as a physics-based model based on [3] to compute power consumption and scrap generation. The model of a Mazak Variaxis 630-5X II T Mill/Turn Machine 40 hp 5-axis CNC machine was validated against experimental data, shown in Fig.1, of the power consumption for one pass dry face milling. The model was formulated as a system of algebraic equations with inputs, outputs, parameters and faults as detailed in [5]. The experimental facility is instrumented to acquire process data for machine power, spindle power, spindle vibration, and to record audio/video and is MTConnect capable. Moreover, the machine can be controlled to perform milling at various speeds, widths of cut, depths of cut and feed rates, which provides a wide space for maintenance testing. In this work, we explore the issue of maintenance testing as a non-linear program that selects maintenance test conditions and sensors on the basis of information criteria. Common uncertainty in the milling process include the material hardness of the workpiece, friction factor, depth of cut and the width of cut. The most common fault (target of maintenance testing) is the tool wear [4]. Optimal selection of test settings and sensors for the purpose of tool wear prediction is performed by applying the model-based method proposed in [5] for fault detection and isolation (FDI). The method formulates the selection of test settings and sensors as discrete and continuous variables, respectively; and the selection is done based on the contribution of each to the information gain with respect to the estimation of tool wear. The problem is formulated as a constrained mixed integer non-linear problem and solved for all possible combinations of available sensors. The objective of the optimization problem is to maximize a measure of the normalized (for the number of sensors used) Fisher Information Matrix (FIM), which is calculated from the sensitivities of the outputs with respect to system faults and uncertainties. The isolation capacity of the test settings is determined from the Kullback-Leibler divergence of healthy and faulty system measurements and the confusion matrix of tool wear classification. Classification of the tool wear is performed using the k-nearest neighbor algorithm for the optimal maintenance test with the most informative set of sensors. It is shown that classification accuracy is variable within the operating space of the machine, and that noisy sensors can further deteriorate tool wear condition assessment. Therefore, a maintenance test for the digital twin of the precision machining equipment that optimizes test conditions and sensor network is imperative.
Acknowledgment: This material is based upon work supported by the U.S. Department of Energyâs Office of Energy Efficiency and Renewable Energy (EERE) under the Advanced Manufacturing Office Award Number DE-EE0007613.
Disclaimer: This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.â
References
- EIA, 2011. Annual Energy Review. http://www.eia.gov/totalenergy/data/annual/index.cfm.
- https://www.sandvik.coromant.com/en-us/knowledge/milling/pages/face-milling.aspx
- Shao, H., Wang, H. L., & Zhao, X. M. (2004) A cutting power model for tool wear monitoring in milling. International Journal of Machine Tools and Manufacture, 44(14), 1503â1509. https://doi.org/10.1016/j.ijmachtools.2004.05.003.
- Altintas, Y., Yellowley, I., & Tlusty, J. (1988) The detection of tool breakage in milling operations. Journal of Manufacturing Science and Engineering, Transactions of the ASME, 110(3), 271â277. https://doi.org/10.1115/1.3187881
- U. Awasthi, K.A. Palmer, G.M. Bollas (2020) Optimal test and Sensor Selection for Active fault diagnosis using Integer programming, Journal of Process Control, (under review).
- U. Awasthi and G.M. Bollas (2020) Sensor network design for smart manufacturing { Application on precision machining, 21st IFAC World Congress, Berlin, Germany.
![](https://www.aiche.org/sites/default/files/aiche-proceedings/1783961/papers/610913/Paper_610913_abstract_170296_0.png)