Semi-physical neural network model for fretting wear estimation

Abstract

In this paper it is shown that we can estimate the fretting wear evolution via an artificial neural network (ANN) model without making use of the back-propagation learning algorithm and without using any regularization method. This can be done by integrating in the ANN model all the available knowledge about the wear mechanism. This kind of model is referred to as a semi-physical neural model. One of the main advantages in building a semi-physical neural model is that its number of parameters is reduced compared with a standard ANN model. This is a very favourable property against the over-fitting inconvenience. In addition, via appropriate nonlinear transformations, the semi-physical neural model can be rendered linear with respect to the parameters that are to be determined. Consequently, a simple least square approximation can be used to determine the unknown parameters.

Artificial Neural Networks (ANNs) as a Novel Modeling Technique in Tribology

“…Recently, Li et al (2017) have shown that a hybrid neural network (HNN) model, which combines two types of ANNs, possesses remarkable capability for modeling tribological properties (see Figure 4). Specifically, along with a backpropagation (BP) network (this is another name for a MLP) it is FIGURE 5 | Schematic of the ANN-based semi-physical fretting wear model (Haviez et al, 2015).…”

“…Recently, a semi-physical NN model was developed by Haviez et al (2015) for fretting wear (see Figure 5). Their model introduces a two-level function approximation for the wear volume and utilizes the dissipated energy approach for getting insight into the mechanisms of fretting damage.…”

Artificial neural network modeling of sliding wear

“…15,16 To address this so-called ''black box'' issue of ANN modeling approach, a number of problem-specific ANN-based models of hybrid type have been developed. [17][18][19] In the present paper, we consider the main problem associated with the application of multilayer perceptron (MLP) for analyzing experimental data obtained from pin-on-disk sliding wear tests. Notwithstanding that MLPs, which is a standard neural network technique, the novelty of our approach, in particular, lies in the way how the coefficient of wear is determined.…”

Modeling and prediction of wear rate of aluminum alloy (Al 7075) using power law and ANN

“…Several modeling techniques have been used for modeling and optimization of wear, which include regression, ANN. The ANN application in the field of tribology has attracted the increasing interest of the researchers [8,9]. The main reason is that analytical and computational modeling is difficult to use for tribological phenomena, which are highly nonlinear in nature [10].…”