Q-Learning-based fuzzy energy management for fuel cell/supercapacitor HEV

A novel fuzzy energy management strategy (EMS) based on improved Q-learning controller and genetic algorithm (GA) is proposed for the real-time power split between fuel cell and supercapacitor of hybrid electric vehicle (HEV). Different from driving pattern recognition–based method, Q-Learning controller takes actions by observing the driving states and compensates to fuzzy controller, that is, no need to know the driving pattern in advance. Aimed to prolong the fuel cell lifetime and decrease its energy consumption, the initial values of Q-table are optimized by GA. Moreover, to enhance the environment adaptation capability, the learning strategy of Q-learning controller is improved. Two adaptive energy management strategies have been compared, and simulation results show that current fluctuation can be reduced by 6.9% and 41.5%, and H2 consumption can be saved by 0.35% and 6.08%, respectively. Meanwhile, state of charge (SOC) of supercapacitor is sustained within the desired safe range.

Important powertrain dynamics for developing models for control of connected and automated electrified vehicles

Connected and Automated Vehicles (CAV) technology presents significant opportunities for energy saving in the transportation sector. CAV technology forecasts vehicle and powertrain power needs under various terrain, ambient, and traffic conditions. Integration of the CAV technology in Hybrid Electric Vehicles (HEVs) provides the opportunity for optimal vehicle operation. Indeed, Hybrid Electric Vehicle powertrains present high degrees of flexibility and possibility for choosing optimum powertrain modes based on the predicted traction power needs. In modeling complex CAV powertrain dynamics, the modeler needs to consider short-time scale powertrain dynamics, such as engine transients, and hysteresis of mode-switching for a multi-mode HEV. Therefore, the powertrain dynamics essential for developing powertrain controllers for a class of connected HEVs is presented. To this end, control-oriented powertrain dynamic models for a test vehicle consisting of full electric, hybrid, and conventional engine operating modes are developed. The resulting powertrain model can forecast vehicle traction torque and energy consumption for the specified prediction horizon of the test vehicle. The model considers different operating modes and associated energy penalty terms for mode switching. Thus, the vehicle controller can determine the optimum powertrain mode, torque, and speed for forecasted vehicle operation via utilizing connectivity data. The powertrain model is validated against the experimental data and shows prediction error of less than 5% for predicting vehicle energy consumption. The model is used to create energy penalty maps that can be used for CAV control, for example fuel penalty map for engine torque changes (10–40 Nm) at each engine speed. The results of model-based optimization show optimum switching delays ranging from 0.4 to 1.4 s to avoid hysteresis in mode switching.

Combined Vector and Direct Controls Based on Five-Level Inverter for High Performance of IM Drive

The goal of this study was to figure out how to regulate an induction motor in a hybrid electric vehicle. Conventional combined vector and direct control induction motors take advantage of the advantages of vector control and direct torque control. It is also a method that avoids some of the difficulties in implementing both of the two control methods. However, for this method of control, the statoric current has a great wealth of harmonic components which, unfortunately, results in a strong undulation of the torque regardless of the region speed. To solve this problem, a five-level neutral point clamped inverter was used. Through multilevel inverter operation, the voltage is closer to the sine wave. The speed and torque are then successfully controlled with a lower level of ripple in the torque response which improves system performance. The analysis of this study was verified with simulation in the MATLAB/Simulink interface. The simulation results demonstrate the high performance of this control strategy.

Design and lumped parameter magnetic network model of hybrid excited consequent pole flux switching machine

Purpose The purpose of this paper is to analyze electromagnetic performance and develop an analytical approach to find the suitable coil combination and no-load flux linkage of the proposed hybrid excited consequent pole flux switching machine (HECPFSM) while minimizing the drive storage and computational time which is the main problem in finite element analysis (FEA) tools. Design/methodology/approach First, a new HECPFSM based on conventional consequent pole flux switching permanent machine (FSPM) is proposed, and lumped parameter magnetic network model (LPMNM) is developed for the initial analysis like coil combination and no-load flux linkage. In LPMNM, all the parts of one-third machine are modeled which helps in reduction of drive storage, computational complexity and computational time without affecting the accuracy. Second, self and mutual inductance are calculated in the stator, and dq-axis inductance is calculated using park transformation in the rotor of the proposed machine. Furthermore, on-load performance analysis, like average torque, torque density and efficiency, is done by FEA. Findings The developed LPMNM is validated by FEA via JMAG v. 19.1. The results obtained show good agreement with an accuracy of 96.89%. Practical implications The proposed HECPFSM is developed for high-speed brushless AC applications like electric vehicle (EV)/hybrid electric vehicle (HEV). Originality/value The proposed HECPFSM offers better flux regulation capability with enhanced electromagnetic performance as compared to conventional consequent pole FSPM. Moreover, the developed LPMNM reduces drive storage and computational time by modeling one-third of the machine.