Using an Adaptive Network-based Fuzzy Inference System to Estimate the Vertical Force in Single Point Incremental Forming

Abstract

Manufacturing processes are usually complex ones, involving a significant number of parameters. Unconventional manufacturing processes, such as incremental forming is even more complex, and the establishment of some analytical relationships between parameters is difficult, largely due to the nonlinearities in the process. To overcome this drawback, artificial intelligence techniques were used to build empirical models from experimental data sets acquired from the manufacturing processes. The approach proposed in this work used an adaptive network-based fuzzy inference system to extract the value of technological force on Z-axis, which appears during incremental forming, considering a set of technological parameters (diameter of the tool, feed and incremental step) as inputs. Sets of experimental data were generated and processed by means of the proposed system, to make use of the learning ability of it to extract the empirical values of the technological force from rough data.

Emerging Trends in Single Point Incremental Sheet Forming of Lightweight Metals

“…In order to evaluate the forming forces applied by the forming punch in incremental forming, different optimisation concepts are used, which are dominated by prediction using FEM analyses [98,101,102] as well as the use of different Taguchi analyses [103,104], response surface methods (RSM) [105,106], Pareto optimisation [107], analysis of variance (ANOVA) [108,109], various types of artificial neural networks (ANN) [110][111][112], and genetic algorithms (GA). The statistical methods mentioned are used with results obtained by the finite element method (FEM) or by experiment.…”

“…The incremental forming process is influenced by several process parameters leading to the dynamic and fast-changing forming load being difficult to predict and control. To overcome this problem, Racz et al [112] have used an adaptive network-based fuzzy inference system to estimate the vertical forming force in advance. In the fuzzy inference system developed, several technological influential parameters served as the inputs, including the diameter of the tool, feed rate, and incremental step.…”

Optimization study of single point incremental forming process of low carbon steel/CP-Titanium bimetallic sheets using grey relational analysis and response surface methodology

“…Response Surface Methodology (RSM) [24][25][26], Artificial Neural Networks (ANN) [27,28], Genetic Algorithms (GA) [29,30,32], Fuzzy Logic (FL) [33,34] and Grey Relational Analysis (GRA) [35][36][37]. In fact, sometimes it is needed to produce components with specific characteristics, for example: minimum surface roughness and maximum wall thickness.…”