Working parameter optimization of strengthen waterjet grinding with the orthogonal-experiment-design-based ANFIS
Zhongwei Liang,Shaopeng Liao,Yiheng Wen,Xiaochu Liu
DOI: https://doi.org/10.1007/s10845-016-1285-z
IF: 8.3
2016-12-02
Journal of Intelligent Manufacturing
Abstract:Abstract In this paper, the working parameter optimization of strengthen waterjet grinding by employing the orthogonal-experiment-design-based ANFIS (Adaptive Neural Fuzzy Inference System), was conducted to obtain an optimal result of bearing ring machining. An improved ANFIS system based upon orthogonal experiment design, was proposed to optimize the working parameters in grinding practices, which increases the surface hardness of ring surface from 49.0 to 72.0 HRC, topography elasticity variance from 330.0 to 670.0, texture energy from 24.5 to 88.0, decreases the surface roughness from 0.65 to 0.25 (upmu ) m, and loading deviation from 1860.5 to 1320.0, thereafter an optimal grinding quality can be obtained. The optimization approach proposed involve the following steps: Preparation of experimental environment; Measure index determination for ring surface; Orthogonal experiment design for making fuzzy logic rules; Establishment of ANFIS system; Working parameter optimization for waterjet grinding; and Performance verification for actual grinding. Objective of this research is determining the optimal working parameters with fewer experimental iterations compared to other alternative approaches, such as Genetic parameter optimization, SA–GA parametric prediction, Taguchi parameter estimation, ANN–SA parametric selection, and GONNs parameter selection method. Statistical analysis and result comparisons support its efficiency and reliability in machining practices, a stable and reliable grinding process can be achieved for typical conditions by using waterjet pressure at around 310ṀPa, flow rates of water mass at about 5.8 kg/min, attack angle by 60–75 ({^{circ }}) , mass rate of abrasive grit by about 0.4 kg/min, and traverse speed by 60 mm/min. It was concluded that this proposed ANFIS system can be used as a suitable and effective tool, to investigate the complicated influential correlation between waterjet working parameters and grinding effectiveness in bearing manufacturing, and to give a better machining performance compared to other experimental practices.
engineering, manufacturing,computer science, artificial intelligence