Design and Optimization of Control System for More Electric Aircraft Power Systems Using Adaptive Tabu Search Algorithm Based on State-Variables-Averaging Model

Abstract

An important reflection of the overall efficiency, reliability, and passenger safety of a more electric aircraft (MEA) is the output performance of its electrical power system (EPS) controller. This output performance encompasses the rise time, setting time, and percent undershoot of the voltage across the capacitor bank. Therefore, this article presents an optimal controller design using an artificial intelligence method called the adaptive tabu search (ATS) algorithm. The state-variables-averaging model is applied with the ATS algorithm to reduce computational time. Moreover, stability analysis based on the eigenvalue theorem is used as the penalty condition during the searching process to avoid unstable operation. The output performance of the proposed controller design is superior to that of the conventional controller design. All design results are verified by good agreement with MATLAB and hardware-in-theloop (HIL) simulations.INDEX TERMS More electric aircraft, vector control, state-variables-averaging model, adaptive tabu search algorithm. I. INTRODUCTIONAccording to published concepts for more electric aircraft [6]. aircraft performance optimization, flight reliability improvement, and passenger safety improvement are essential tasks. Altering the MEA controller design is an approach with strong potential to accomplish these tasks. Artificial intelligence (AI) techniques are required to achieve the optimal output performance of the MEA's electrical power system (EPS) controller in terms of the rise time, setting time, and percent undershoot of the voltage across the capacitor bank. AI can be used for controller design and many other functions. For example, AI has been applied to the design of active power filters for the adaptive tabu search (ATS) algorithm [7] and genetic algorithm [8], power flow optimization based on particle swarm optimization [9], ant colony search algorithm [10], antenna array design using artificial bee colony algorithm [11], and the application of the ATS algorithm to instability mitigation [12]. However, one of the crucial problems that arises when applying AI techniques to the EPS of MEA is the simulation time. This is because the simulation of EPS using software packages such as PSIM and MATLAB causes huge computational time due to switching behavior. Thus, a time-invariant model is necessary and sufficient for the controller design. There are several reasonable methods for deriving the timeinvariant model of EPS. The generalized state-space averaging modeling technique [13], [14] has been universally used for single-phase rectifiers of AC distribution systems and the power converters of DC distribution systems. The direct quadrature (DQ) approach [1], [2], [6], [12], [15], [16] is suitable for power converters in three-phase AC distribution systems. The nonlinear average-value method modeling technique [17] has been applied to analyze 6-and 12-pluse diode rectifiers.Based on a literature review, the AI technique called the ATS algorithm is selected for use in the present stud...