Surface roughness prediction by extreme learning machine constructed with abrasive water jet
Introduction
Need to achieve higher quality of products along with demand for effective management of materials and information in a production system, as well as strong competition on the market, propel development and improvement of production technology. Abrasive water jet machining (AWJ) belongs to the group of non-conventional treatments, and is widely used for cutting various materials such as metal, fibreglass, wood, all types of plastics, rubber, stone and others. Due to the absence of heat and side effects, application of the AWJ method is especially suitable for cutting glass and other brittle crystalline materials.
The quality and efficiency of AWJ cutting process depend on several machining parameters, such as hydraulic parameters, abrasive material, workpiece material and cutting regime parameters [1]. Roughness of surface is one of the main attributes of quality of product derived from AWJ processing, and is directly depending on the process parameters such as thickness of the workpiece, abrasive flow rate, cutting speed and others [2]. Therefore, any modelling strategy that establishes a link between surface roughness of the product and machining parameters is of significant practical importance.
The roughness of the surface machined with abrasive water jet analysis needs accurate on-line identification. In this article, we motivate and introduce the estimation model of roughness of the surface machined with abrasive water jet using the soft computing approach, namely extreme learning machine (ELM). Some researchers applied soft computing methods for surface roughness prediction in many different fields. In article [3] estimation of surface roughness from texture features of the surface image using an adaptive neuro-fuzzy inference system was investigated. Surface roughness prediction using artificial neural network with minimal cutting fluid application was investigated in [4]. Another predictive model of surface roughness in hard turning using regression and neural networks was established in [5]. One common property of the investigated models for prediction of machining surface roughness is that fluid applications are rare. Therefore, in this paper the main aim was to model water jet application for surface machining [6].
Nowadays, application of modern computational approach in solving the real problems and determining the optimal values and functions are receiving enormous attention by researchers in different scientific disciplines. Neural network (NN), as a major computational approach, has been recently introduced and applied in different engineering fields. This method is capable of solving complex nonlinear problems which are difficult to solve by classic parametric methods. There are many algorithms for training neural network such as back propagation, support vector machine (SVM), hidden Markov model (HMM). The shortcoming of NN is its learning time requirement. Huang et al. [7] introduced an algorithm for single layer feed forward NN which is known as extreme learning machine (ELM). This algorithm is capable to solve problems caused by gradient descent based algorithms like back propagation which applies in ANNs. ELM is able to decrease the required time for training a neural network. In fact, it has been proved that by utilizing the ELM, the learning process becomes very fast and it generates robust performance [8]. Accordingly, a number of investigations have been carried out related to application of ELM algorithm successfully for solving the problems in various scientific fields [9], [10], [11], [12], [13], [14].
In general, ELM is a powerful algorithm with faster learning speed comparing with traditional algorithms like back-propagation (BP). It also has a better performance too. ELM tries to get the smallest training error and norm of weights.
In this study, a predictive model of roughness of the surface machined with abrasive water jet was developed using ELM method. ELM results were also compared with genetic programing (GP) and artificial neural networks (ANNs) results. The given inputs are: cutting speed (V [mm/min]), material thickness (S [mm]), abrasive flow (Q [g/min]) and measurements position and the single output is: predicted surface roughness.
Section snippets
Experimental setup
Experimental investigation was carried out on a machine for abrasive water jet cutting, type ByJet 4022, a product of Bystronic AG from Switzerland. As workpiece material, aluminium alloy AA-ASTM 6060 (EN: AW-6060; ISO: Al MgSi0.5) was used. A typical application of this alloy includes profiling and parts for the automotive industry, and it is one of the most popular of the 6XXX series alloys. Such material is chosen because it is very attractive, possess resistance to corrosion and can provide
Evaluating accuracy of proposed models
Predictive performances of proposed models were presented as root means square error (RMSE) and coefficient of determination (R2). These statistics are defined as follows:
- (1)
root-mean-square error (RMSE)
- (2)
coefficient of determination (R2)where Pi and Oi are known as the experimental and forecast values of surface roughness, respectively, and n is the total number of test data.
Performance evaluation of proposed ELM model
In this section, performance results of ELM roughness of the
Performance comparison of ELM, ANN and GP
To demonstrate the advantages of the proposed ELM approach, ELM model prediction accuracy was compared with prediction accuracy of GP and ANN methods, which were used as a benchmark. Conventional error statistical indicators, RMSE, r and R2, were used for comparison. Table 4 summarizes the prediction accuracy results for test data sets since training error is not credible indicator for prediction potential of particular model.
ELM model outperforms GP and ANN models according to the results in
Conclusion
The study carried out a systematic approach to create the ELM roughness of the surface machined with abrasive water jet predictive model. Surface roughness is the most important variable to evaluate influence of machining parameters on the surface quality and thus many authors claim that the understanding of surface topography can provide key data about the mechanism of abrasive water jet machining. In conducted experiments, machined material was aluminium alloy EN AW 6060, while tracking of
Acknowledgment
The authors would like to thank the University of Malaya for the research grants allocated (UMRG-RP015C-13AET and High Impact Research Grant, HIR-D000015-16001). Special appreciation is also credited to the Malaysian Ministry of Education, MOE for the Fundamental Research Grant Scheme (FP053-2013B).
References (28)
- et al.
Effect of feed rate on surface roughness in abrasive water jet cutting applications
J Mater Process Technol
(2004) - et al.
Accurate estimation of surface roughness from texture features of the surface image using an adaptive neuro-fuzzy inference system
Precis Eng
(2005) - et al.
Prediction of machined surface evolution in the abrasive jet micro-machining of metals
Wear
(2012) - et al.
Bankruptcy prediction using extreme learning machine and financial expertise
Neurocomputing
(2014) - et al.
Online sequential extreme learning machine with kernels for nonstationary time series prediction
Neurocomputing
(2014) - et al.
Mobility prediction in mobile ad hoc networks using extreme learning machines
Procedia Comp Sci
(2013) - et al.
Fast prediction of protein–protein interaction sites based on extreme learning machines
Neurocomputing
(2014) - et al.
Extreme learning machine towards dynamic model hypothesis in fish ethology research
Neurocomputing
(2014) - et al.
Modeling and optimization of biodiesel engine performance using kernel-based extreme learning machine and cuckoo search
Renew Energy
(2015) - et al.
Extreme learning machine: theory and applications
Neurocomputing
(2006)
Principles of abrasive water jet machining
Surface roughness prediction using artificial neural network in hard turning of AISI H13 steel with minimal cutting fluid application
Procedia Eng
Predictive modeling of surface roughness and tool wear in hard turning using regression and neural networks
Procedia Eng
Extreme learning machine: a new learning scheme of feedforward neural networks
Int Jt Conf Neural Netw
Cited by (73)
Optimization of Abrasive Water Jet (AWJ) cutting process accuracy
2023, Procedia Computer ScienceWASPAS Optimization in Advanced Manufacturing
2022, Procedia Computer ScienceEnhancement of cutting quality of abrasive waterjet by using multipass cutting strategy
2020, Journal of Manufacturing ProcessesCitation Excerpt :The results show that best surface roughness was 4.32 μm. Ćojbašić et al. [9] used Extreme-Learning-Machine (ELM) method to evaluate the surface roughness resulted by AWJ. They controlled the cutting parameters to improve the cutting quality.
Smart manufacturing through machine learning: A review, perspective and future directions to machining industry
2023, AIP Conference Proceedings