Multiple characterisation modelling of friction stir welding using a genetic multi-objective data-driven fuzzy modelling approach

Abstract

Friction Stir Welding (FSW) is a relatively new solidstate joining technique, which is versatile, environment friendly, and energy and time efficient. For a comprehensive understanding of the effects of process conditions, such as tool rotation speed and traverse speed, on characterisations of welded materials, it is essential to establish prediction models for different aspects of the materials' behaviours. Because of the high complexity of the FSW process, it is often difficult to derive accurate and yet transparent enough mathematical models. In such a situation, a systematic data-driven fuzzy modelling approach is developed and implemented in this paper to model FSW behaviour relating to AA5083 aluminium alloy, consisting of microstructural features, mechanical properties, as well as overall weld quality. This methodology allows constructing transparent fuzzy models considering both accuracy and interpretability attributes of fuzzy systems. The elicited models proved to be accurate, interpretable and robust and can be further applied to facilitate the optimal design of process parameters, with the aim of finding the optimal combinations of process parameters to achieve desired welding properties.

A real-time quality monitoring framework for steel friction stir welding using computational intelligence

“…In general, the models with more linguistic rules (hence more parameters to be optimised) often lead to better accuracy due their ability to capture more information (with the risk of over-fitting). However, this leads to lack of interpretabiltiy and simplicity, while models with fewer linguistic rules and parameters result in simpler models and easier to interpret, albeit with lower performance in accuracy [46]. One of the objectives of this research work was to keep the overall system structure as simple as possible in order to create a computational framework that can be used in real-time, inline to the process with modest computational demands.…”

Optimization of Process Parameters to Maximize Ultimate Tensile Strength and Hardness of Underwater Friction Stir Welded Aluminium Alloys using Fuzzy Logic

“…Dewan MW, et al [27] found that the number of the membership functions (MF) and their locations on the universe of discourse influenced the fuzzy algorithm, compared to the shape variations of MFs. Zhang Q, et al [28] developed a systematic datadriven fuzzy modelling approach to AA5083 Aluminum alloyrelated FSW behaviour with microstructural features, mechanical properties, and overall weld quality. The extracted models have proven to be accurate, interpretable, and resilient, and can be applied to facilitate the optimal design of process parameters to achieve the desired welding properties.…”

Multiobjective optimal design of friction stir welding considering quality and cost issues

“…The modelling method was designed based on fuzzy rule based systems, 22, 23 which are very practical to be applied into the nonlinear, data driven leaning context. An improved version of the data driven fuzzy modelling approach, with a representative data selection method, was further implemented to develop dynamic models for predicting internal process attributes, 24 as demonstrated in Fig. 3.…”