DEVELOPMENT OF AN ANFIS–SQP-BASED ALGORITHM FOR OPTIMAL PARAMETER SELECTION AND CONTROL DECISIONS IN PULSED ELECTRIC TREATMENT OF PLANTS

This study presents an integrated neuro-fuzzy modelling and constrained optimization approach for selecting technological parameters and forming control decisions in pulsed electric treatment of plants. The proposed procedure links an adaptive neuro-fuzzy inference system (ANFIS), used to approximate a nonlinear treatment-response relationship, with sequential quadratic programming (SQP), used to search for an admissible operating point. The source study considers 105 combinations of voltage, capacitor capacitance and treatment duration. A first-order Takagi–Sugeno ANFIS structure with three Gaussian membership functions per input and 27 fuzzy rules is described. The reported model-based optimization result is a voltage of 2528.7 V, capacitance of 517.54 pF and treatment duration of 0.1999 s, with a predicted treatment-response index of 71.3%. The corresponding ideal capacitor energy is approximately 1.655 mJ per pulse. The paper distinguishes calculated predictions from experimentally verified outcomes and identifies the need for independent model validation, a complete equipment-level power model and confirmation of the selected operating point. The contribution is an explicit decision-forming workflow integrating prediction, constrained search and validation rather than an experimentally confirmed optimum.