Materials Today: Proceedings
Genetic programming approach to predict the performance characteristics of WEDM taper cutting process
Introduction
Many complicated manufacturing processes exist in today’s industrial environment where the cost and quality are the main driving forces. Among them, taper cutting utilizing wire electrical discharge machining process is a sophisticated engineering process. In this process the wire is deformed at the time of angular cutting, resulting variations in the actual angle of machined items [1]. As a result, the accuracy of the machined component is compromised. Hence, the realm of taper cutting operation in WEDM, effective process parameter selection is critical challenge. However experimental process optimization of any machining process is expensive, time consuming and demands high degree of technical expertise owing to the complex, nonlinear nature of input–output variables of the process. In the other way we may acquire output values for preset input parameter values either empirically, mathematically, or numerically. However, employing mathematical and numerical models to map the behavior and interrelation of process factors is also difficult. Because this procedure is time intensive and demands a high degree of technical competence, solving such numerical and mathematical models is not always a straightforward task. Hence, statistical models are unable to forecast performance measurements due to the multiple factors involved and the non-linearity of the connection between the parameters. Only when the number of variables is minimal and their effects can be characterized by simple connections will statistical approaches be appropriate [2]. To overcome this constraint, multiple academics have suggested several artificial intelligence strategies for correlating input parameters with WEDM performance measures. [3], [4], [5]. Saha et al. proposed a multi variable regression model and a feed- forward neural network model for WEDM process to predict the performance measures such as cutting speed and surface roughness considering pulse on time, pulse off time, peak current and capacitance as input variables [6]. Aich and Banarjee used support vector machine regression to create a process model for the EDM process and used particle swarm optimization to find the optimal parametric combination [7]. Conde et al. proposed a way to predict the accuracy of components produced by WEMD by using an Elman-based Layer Recurrent Neural Network (LRNN). They have also suggested an algorithm for designing wire paths of variable radius to correct the deviations in the machined part through software [8]. Majumdar and Maity [9] predict and optimize the WEDM performance measures such as surface roughness and micro hardness using general regression neural network (GRNN) and MOORA fuzzy approach considering nitinol as workpiece material. Caydas et al. developed an adaptive neuro-fuzzy inference system (ANFIS) model for the white layer thickness (WLT) and the average surface roughness of WEDM process considering pulse duration, open circuit voltage, dielectric flushing pressure and wire feed rate as model’s input feature [10]. Nayak and Mahapatra have developed a support vector regression (SVR) model to predict the angular error in taper cutting using wire electrical discharge machining process [11]. However, the use of artificial intelligent techniques in the field of taper cutting employing the WEDM is quite limited. As a result, geometric programming (GP) model is proposed in this paper to predict angular error and surface roughness during taper cutting in WEDM process. The root meat square error (RMSE), mean absolute percentage error (MAPE) and co-efficient of determination (R2) are used to asses the model’s performance.
Section snippets
Proposed methodology
In the present work, an intelligent approach is presented for prediction of angular error and surface roughness using genetic programming approach (GP). The difference between GP and traditional techniques is that it makes no assumptions regarding the formulation to be made. The model that is developed also aids in the interpretation of the parameters that affect the process. The flow chart of the proposed methodology is shown in Fig. 1.
Experimental details
The taper cutting was done in this study using an AC Progress V2 high precision CNC WEDM. For the experiment coated Broncocut-W (by Bedra) wire of diameter 0.2 mm has been used due to its low yield strength and high elongation property. The di-electric medium was deionized water. Inconel 718 was used as workpiece material with 25 mm diameter and thickness of 20 mm, 30 mm and 40 mm respectively. Table 1, lists the input and fixed parameters utilized in the current investigation. These were
Results and discussions
In the present study, the experimental data sets as shown in Table 3 were group into two sets: training data set and testing data set. Twenty (about 75 percent) of the experimental data is utilized for training, while the remaining seven (around 25 percent) is used for testing or validation. For model development, GPTIPS, an open-source MATLAB tool box, is employed. Parameter settings are crucial for a successful implementation of the GP model. The parameter settings are chosen using the
Conclusions
The purpose of this study is to determine the motivation for applying a genetic programmed strategy to forecast angular error and surface roughness during taper cutting in the WEDM process prior to machining. It is a biologically inspired artificial intelligence technique that uses evolutionary algorithms to solve challenging problems. Root mean square error (RMSE), mean absolute percentage error (MAPE), and coefficient of determination are used to evaluate the suggested model's performance (R2
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (12)
- et al.
Optimization of WEDM process parameters using deep cryo-treated Inconel 718 as work material
Eng. Sci. Technol. Int. J.
(2016) - et al.
WEDM process variables investigation for HSLA by response surface methodology and genetic algorithm
Eng. Sci. Technol. Int. J.
(2015) - et al.
Modeling of EDM responses by support vector machine regression with parameters selected by particle swarm optimization
Appl. Math. Model.
(2014) - et al.
High-accuracy wire electrical discharge machining using artificial neural networks and optimization techniques
Rob. Comput. Integr. Manuf.
(2018) - et al.
Prediction and optimization of surface roughness and micro-hardness using grnn and MOORA-fuzzy-a MCDM approach for nitinol in WEDM
Measurement
(2018) - et al.
An adaptive neuro-fuzzy inference system (ANFIS) model for wire-EDM
Expert Syst. Appl.
(2009)