An energy-based modeling and prediction approach for surface roughness in turning

Nowadays, the manufacturing industries are continuously challenged by achieving high-quality products and clean production in order to remain competitiveness. Surface roughness is one of the important factors in evaluation of the quality of a machined part. Meanwhile, it is vital to correlate energy usage with operations being performed in the production lines for the green and energy-saving manufacturing. A novel surface roughness prediction model based on the energy consumption is proposed. Specific cutting energy consumption (SCEC) and cutting parameters are employed as inputs to the prediction model. Particle swarm optimization-support vector machine (PSO-SVM) is used to predict the surface roughness value in turning. Furthermore, this method is verified by conducting experiments in turning process and compared with the PSO-relevance vector machine (PSO-RVM). Besides, the prediction performance of the energy based model is compared with the model with different combinations of cutting parameters, energy, and vibration features. The result shows that the proposed model obtains the lowest mean relative error and indicates that the model is effective and straightforward for practical implementation.