Type-2 Takagi-Sugeno-Kang fuzzy logic system and uncertainty in machining

The main objective of this thesis is to apply the type-2 TSK identification algorithm based on subtractive clustering method to real problem in uncertainty estimation in machining, and develop a high order type-2 fuzzy system to solve the problem of dimensionality. This thesis includes the theoretical studies on type-2 TSK FLS and experimental studies on application of type-2 FLS on uncertainty estimation of dynamics in high precision machining process.;This thesis is composed of three papers published in international journals. They cover the following topics: type-1 TSL FLS, first order interval type-2 FLS and high order Interval FLS modelling based on one experimental data set. The data set is a time series AE signal voltage for a high precision turning process.;For the theoretical contributions, this thesis introduces the development of fuzzy logic sets and systems, especially, TSK FLS—from type-1 TSK FLS to type-2 TSK FLS including subtractive clustering based type-2 identification algorithm. To solve the curse of dimensionality in type-2 interval systems, the generalized type-2 TSK FLS and its architecture, inference engine and design approach are proposed.;For the experimental studies, the same data sets recorded from the physical machining system is used to identify the type-1, first-order type-2 and high order type-2 fuzzy systems. The fuzzy models of the experimental results are compared to each other in order to show the differences between type-2 TSK system and its type-1 counterpart and demonstrate that type-2 modeling performs better and that the high order interval type-2 TSK FLS has the capability to overcome the problem of dimensionality AE. The results obtained in this thesis show that the type-2 fuzzy estimation not only provides a simpler way to arrive at a definite conclusion without understanding the exact physics of the machining process, but also assesses the uncertainties associated with the prediction caused by the manufacturing errors and signal processing. It is possible to establish a reliable type-2 fuzzy tool condition estimation method based on information of uncertainty in AE signal because AE uncertainty scheme corresponds to the complex tool wear state development. A second order IT2 TSK FLS has less rules than that of first order FLS, it can modeling the AE signal with similar RMSE. That prove that high order interval type-2 TSK FLS has the capability to overcome the problem of dimensionality.;Type-2 fuzzy logic is a new research direction and there exists no applications to machining. The main contributions of this thesis are applying the type-2 TSK identification algorithm based on subtractive clustering method on uncertainty estimation in dynamic in machining processes and proposing the generalized interval type-2 FLS which establishes the theoretical basics for high order type-2 FLS.