Compensation of friction and force ripples in the estimation of cutting forces by neural networks

Heydarzadeh, Mohammad S. and Rezaei, Seyed Mehdi and Mardi, Noor Azizi and Kamali E., Ali (2018) Compensation of friction and force ripples in the estimation of cutting forces by neural networks. Measurement, 114. pp. 354-364. ISSN 0263-2241, DOI https://doi.org/10.1016/j.measurement.2017.09.032.

Full text not available from this repository.
Official URL: https://doi.org/10.1016/j.measurement.2017.09.032

Abstract

Estimated cutting forces are usually mixed up with disturbing forces such as friction and need to be compensated. In common compensation methods, such forces are firstly recorded along machining contours under air-cutting conditions. Then, recorded disturbing forces are recalled for the compensation during the main machining process. This method doubles the process time and needs a precise synchronization. This problem is addressed in this paper. A novel method based on neural networks is introduced to compensate of friction and force ripples during cutting force estimations when signals of permanent magnet linear motors (PMLMs) are used. To this end, a Kalman filter observer was designed and experimentally verified for measuring of friction and force ripples. It was then used to provide target series required for training a neural network. Time series of the translator position along some sinusoidal trajectories were selected as training inputs. Taguchi experimental design method was used to determine the structure of the network (number of layers, nodes, and delays). It can be seen that increasing the complexity of the network does not necessarily lead to a more precise network, and a neural network with a hidden layer,16 nodes in the hidden layer and two time-delays can well model such forces. Experiments showed that the results of both methods are very similar and therefore, the proposed method can be used as well as the recording method. Finally, the designed method was successfully applied to the precise estimation of micro milling forces in order to estimate tool deflections.

Item Type: Article
Funders: Research project SF025-2013 and UM.C/625/1/HIR/MOE/ENG/33 from “ University of Malaya ”
Uncontrolled Keywords: Linear motors; Friction and force ripples; Neural networks; Kalman filter; Cutting force estimation
Subjects: T Technology > TJ Mechanical engineering and machinery
Divisions: Faculty of Engineering
Depositing User: Ms. Juhaida Abd Rahim
Date Deposited: 08 Aug 2019 03:58
Last Modified: 08 Aug 2019 03:58
URI: http://eprints.um.edu.my/id/eprint/21899

Actions (login required)

View Item View Item