Adaptive neuro-fuzzy inference system-based grey time-varying sliding mode control for power conditioning applications

This paper develops an adaptive neuro-fuzzy inference system-based grey time-varying sliding mode control (TVSMC) for the application of power conditioning systems. The presented methodology combines the merits of TVSMC, grey prediction (GP), and adaptive neuro-fuzzy inference system (ANFIS). Compared with classic sliding mode control, the TVSMC accelerates reaching phase and guarantees the sliding mode existence starting at arbitrary primary circumstance. But, as a highly nonlinear loading occurs, the TVSMC will undergo chattering and steady-state errors, thus degrading PCS’s performance. The GP is therefore used to attenuate the chattering if the overestimate of system uncertainty bounds exists and to lessen steady-state errors if the underestimate of system uncertainty bounds happens. Also, the GP-compensated TVSMC control gains are optimally tuned by the ANFIS for achieving more precise tracking. Using the proposed methodology, the power conditioning system (PCS) robustness is increased expectably, and low distorted output voltage and fast transient response at PCS output can be achieved even under nonlinear loading. The analysis in theory, design process, simulations, and digital signal processing-based experimental realization for PCS are represented to support the efficacy of the proposed methodology. Because the proposed methodology is easier to implement than prior methodologies and provides high tracking accuracy and low computational complexity, the contents of this paper will be of interest to learners of correlated artificial intelligence applications.