The Use of Neural Networks and Genetic Algorithms to Control Low Rigidity Shafts Machining

Abstract

The article presents an original machine-learning-based automated approach for controlling the process of machining of low-rigidity shafts using artificial intelligence methods. Three models of hybrid controllers based on different types of neural networks and genetic algorithms were developed. In this study, an objective function optimized by a genetic algorithm was replaced with a neural network trained on real-life data. The task of the genetic algorithm is to select the optimal values of the input parameters of a neural network to ensure minimum deviation. Both input vector values and the neural network’s output values are real numbers, which means the problem under consideration is regressive. The performance of three types of neural networks was analyzed: a classic multilayer perceptron network, a nonlinear autoregressive network with exogenous input (NARX) prediction network, and a deep recurrent long short-term memory (LSTM) network. Algorithmic machine learning methods were used to achieve a high level of automation of the control process. By training the network on data from real measurements, we were able to control the reliability of the turning process, taking into account many factors that are usually overlooked during mathematical modelling. Positive results of the experiments confirm the effectiveness of the proposed method for controlling low-rigidity shaft turning.

Recent Advances on Machine Learning Applications in Machining Processes

“…In the same way, a Convolutional Neural Network structure was implemented with ResNet configuration, considering the responsive fixtures and process data as input, in order to allow the Bidirectional LSTM model to predict the part deformation with an error equal to 10.61% [29]. LSTM was also applied to model the dependency between the deviation, tensile force and eccentricity of low-rigidity shaft machining with a MSE of 1.5456 × 10 −5 [90].…”

Application of an Artificial Neural Network in the Modelling of Heat Curing Effects on the Strength of Adhesive Joints at Elevated Temperature with Imprecise Adhesive Mix Ratios

“…However, a large number of experimental tests is often infeasible, for example, because of the constraints of time and costs. One of the solutions can be the application of computer methods, such as the finite element method (FEM) [25][26][27][28][29][30], the boundary element method (BEM) [31][32][33][34], predictive modelling [35][36][37][38][39], and data analytics [40][41][42][43][44][45]. Mathematical modelling with a modest dataset acquired may help to determine the relationships between the individual parameters and mechanical properties, to identify the most promising direction of research, to reduce the number of physical tests, and to markedly reduce the time and costs of research.…”

Wear Resistance Improvement of Cemented Tungsten Carbide Deep-Hole Drills after Ion Implantation

“…Coordination of the guide pads, cutting edges and machining parameters can produce exclusive surface features [ 30 ], but there is need to monitor fracture-related damage of both the workpiece and machine tool, and to perform recondition of the tool suitability for further use [ 31 ]. A cutting tool wear may be of a mechanical, adhesive, chemical and thermal nature [ 32 , 33 , 34 , 35 , 36 , 37 ]. Geometric changes are caused by friction related to the loss of the edge material and the changing local properties by plastic deformations, high temperatures and the chemical effect of the interacting medium [ 38 ].…”