Hybrid ANFIS-Based Fuzzy-PID Controller Design for Multi-Area Load Frequency Control in the Northwest Ethiopian Power Grid

Abstract The demand for electricity is rising daily, causing a mismatch between generation and demand and leading to frequency deviations from the nominal 50 Hz. Ethiopia's hydropower plants are interconnected to the national power grid and rely on primary load frequency control (LFC) mechanisms; however, a supplementary LFC mechanism is required to maintain system frequency centrally. This paper presents the design of a hybrid ANFIS-based fuzzy-PID (FPID) LFC scheme and compares its performance with those of a fuzzy-PID (FPID) controller, an adaptive neuro-fuzzy inference system (ANFIS)-based LFC controller, and the existing conventional PID controller within a three-area LFC model. The primary aim is to provide a robust and supplementary (centralized) LFC mechanism for the Ethiopian Electric Power grid that can replace the existing primary LFC mechanism and resolve the challenges associated with real-time tuning of PID parameters. The hydro-governor, hydraulic turbine, synchronous generator, tie-line, and load were modeled, followed by the formulation of a three-area LFC system model. The PID, FPID, ANFIS, and hybrid ANFIS-based FPID LFC controllers were designed in MATLAB/Simulink using three different governor models, and their performance metrics were evaluated comparatively. Changes in frequency and tie-line power flow, which together determine the area control error (ACE), serve as feedback signals to the governor for regulating water flow into the turbine. System response was assessed in terms of overshoot, undershoot, steady-state error, and settling time under changes in system frequency and tie-line power flow. The proposed hybrid ANFIS-based FPID controller significantly outperforms the PID, FPID, and ANFIS controllers. In Scenario 1, the frequency deviation settling time was reduced to 12.959 s with an ITAE of 0.00565, compared to 48.8824 s and an ITAE of 0.0838 for the PID controller. Under multi-area disturbances (Scenario 2), the hybrid controller maintained its superiority with the lowest ITAE of 0.02375. Robustness analysis under ± 25% parameter variations further confirmed its stability, with the ITAE remaining within the range of 0.00541 to 0.00580.