A comparison of machine learning methods for cutting parameters prediction in high speed turning process

Support vector machines are arguably one of the most successful methods for data classification, but when using them in regression problems, literature suggests that their performance is no longer state-of-the-art. This paper compares performances of three machine learning methods for the prediction of independent output cutting parameters in a high speed turning process. Observed parameters were the surface roughness (Ra), cutting force \((F_{c})\), and tool lifetime (T). For the modelling, support vector regression (SVR), polynomial (quadratic) regression, and artificial neural network (ANN) were used. In this research, polynomial regression has outperformed SVR and ANN in the case of \(F_{c}\) and Ra prediction, while ANN had the best performance in the case of T, but also the worst performance in the case of \(F_{c}\) and Ra. The study has also shown that in SVR, the polynomial kernel has outperformed linear kernel and RBF kernel. In addition, there was no significant difference in performance between SVR and polynomial regression for prediction of all three output machining parameters.     

Friction and wear of sintered metallic brake linings on a C/C-SiC composite brake disc

The friction and wear of sintered metallic brake linings on a C/C-SiC composite brake disc were studied. This paper reports on the friction and wear properties of sintered metallic (MMC) brake linings, which appear to combine well with a C/C-SiC brake disc. The friction characteristics were examined with a dynamometer on two different commercial motorcycle brake systems, differing in terms of the brake caliper and the dimensions of the disc. The influence of the components, such as graphite, and the abrasives in the metallic matrix on the formation of the friction layer was investigated using a scanning electron microscope (SEM) equipped with energy-dispersive X-ray spectroscopy (EDX). The friction layer formed on the pad's sliding surface by oxidation wear, which consisted mostly of iron and copper oxides, was confirmed. The friction properties of the sintered metallic brake pads were determined and related to the composition and structure of the brake lining. This investigation of the friction characteristics of a brake couple comprising (MMC) brake linings and a C/C-SiC composite disc will increase our understanding of this material, which works in a completely different way to classical brakes based on metallic discs.     

Genetic programming approach to determining of metal materials properties

The paper deals with determining metal material properties by the use of genetic programming (GP). As an example, the determination of the flow stress in bulk forming is presented. The flow stress can be calculated on the basis of known forming efficiency. The experimental data obtained during pressure test serve as an environment to which models for forming efficiency have to be adapted during simulated evolution as much as possible. By performing four experiments, several different models for forming efficiency are genetically developed. The models are not a result of the human intelligence but of intelligent evolutionary process. With regard to their precision, the successful models are more or less equivalent; they differ mainly in size, shape, and complexity of solutions. The influence of selection of different initial model components (genes) on the probability of successful solution is studied in detail. In one especially successful run of the GP system the Siebel's expression was genetically developed. In addition, redundancy of the knowledge hidden in the experimental data was detected and eliminated without the influence of human intelligence. Researches showed excellent agreement between the experimental data, existing analytical solutions, and models obtained genetically.     

Velocity anisotropy of an industrial robot

Industrial robots are part of production systems and it is important to place them into the system according to their properties and behaviour. The information, obtained from the technical sheets of robots, about workspace (its dimensions and shape) is insufficient for designing the production system. The information about mobility is missing. To represent the behaviour of the robot in the workspace, velocity anisotropy of the robot is introduced and defined as the length of the shortest velocity ellipsoid axes, which can be constructed for any position of robot in its tool centre point. The position of a tool centre point is equivalent to the point in the workspace. A graphical representation of the 3D workspace with included velocity anisotropy is then performed and an example for a design of a robotised welding production system is given. In this example the benefits of the graphical representation of the workspace with included velocity anisotropy are presented and discussed.     

Integrated Genetic Programming and Genetic Algorithm Approach to Predict Surface Roughness

In this article we propose a new integrated genetic programming and genetic algorithm approach to predict surface roughness in end-milling. Four independent variables, spindle speed, feed rate, depth of cut, and vibrations, were measured. Those variables influence the dependent variable (i.e., surface roughness). On the basis of training data set, different models for surface roughness were developed by genetic programming. The floating-point constants of the best model were additionally optimized by a genetic algorithm. Accuracy of the model was proved on the testing data set. By using the proposed approach, more accurate prediction of surface roughness was reached than if only modeling by genetic programming had been carried out. It was also established that the surface roughness is most influenced by the feed rate, whereas the vibrations increase the prediction accuracy.     

Laser deposition of Metco 15E, Colmony 88 and VIM CRU 20 powders on cast iron and low carbon steel

In the tooling industry, there is often the need to repair damaged or worn out components, or to apply coatings on functional surfaces of tools. When depositing a material which properties are very different from the properties of the substrate material, difficulties, such as cracks, can occur. In this paper, we investigate the deposition of three different powders (Metco 15 E, Colmony 88 and VIM CRU 20) on cast irons and low carbon steel with the laser engineered net shaping (LENS?) technology. Coatings with a maximum of four layers were deposited with different process parameters. Although most bonds itself were of good quality, some coatings had numerous cracks. Preheating of the cast iron samples with the laser beam was used to reduce the number of cracks. In the preheating process, the surface of samples was partially melted in order to dissolve graphite nodules which were often the starting point of cracks.     

Modeling of impact toughness of cold formed material by genetic programming

In the paper, an approach completely different from the conventional methods for determination of accurate models for the change of properties of cold formed material, is presented. This approach is genetic programming (GP) method which is based on imitation of natural evolution of living organisms. The main characteristic of GP is its non-deterministic way of computing. No assumptions about the form and size of expressions were made in advance, but they were left to the self organization and intelligence of evolutionary process. First, copper alloy rods were cold drawn under different conditions and then impact toughness of cold drawn specimens was determined by Charpy tests. The values of independent variables (effective strain, coefficient of friction) influence the value of the dependent variable, impact toughness. On the basis of training data, different prediction models for impact toughness were developed by GP. Only the best models, gained by genetic programming were presented in the paper. Accuracy of the best models was proved with the testing data set. The comparison between deviation of genetic model results and regression model results concerning the experimental results has showed that genetic models are more precise and more varied then regression models.     

Emergence of intelligence in next-generation manufacturing systems

In the paper we propose a fundamental shift from the present manufacturing concepts and problem solving approaches towards new manufacturing paradigms involving phenomena such as emergence, intelligence, non-determinism, complexity, self-organization, bottom-up organization, and coexistence with the ecosystem. In the first part of the paper we study the characteristics of the past and the present manufacturing concepts and the problems they caused. According to the analogy with the terms in cognitive psychology four types of problems occurring in complex manufacturing systems are identified. Then, appropriateness of various intelligent systems for solving of these four types of problems is analyzed. In the second part of the paper, we study two completely different problems. These two problems are (1) identification of system in metal forming industry and (2) autonomous robot system in manufacturing environment. A genetic-based approach that imitates integration of living cells into tissues, organs, and organisms is used. The paper clearly shows how the state of the stable global order (i.e., the intelligence) of the overall system gradually emerges as a result of low-level interactions between entities of which the system consists and the environment.     

Comparison Between Genetic Algorithm and Genetic Programming Approach for Modeling the Stress Distribution

This article compares genetic algorithm (GA) and genetic programming (GP) for system modeling in metal forming. As an example, the radial stress distribution in a cold-formed specimen (steel X6Cr13) was predicted by GA and GP. First, cylindrical workpieces were forward extruded and analyzed by the visioplasticity method. After each extrusion, the values of independent variables (radial position of measured stress node, axial position of measured stress node, and coefficient of friction) were collected. These variables influence the value of the dependent variable, radial stress. On the basis of training data, different prediction models for radial stress distribution were developed independently by GA and GP. The obtained models were tested with the testing data. The research has shown that both approaches are suitable for system modeling. However, if the relations between input and output variables are complex, the models developed by the GP approach are much more accurate.