Bogdan-Adrian Enache

A LiFePO 4 battery discharge simulator for EV applications — Part 1: Determining the optimal circuit based battery model

This paper, the first of a series of papers on developing a LiFePO4 (LFP) battery discharge simulator for EV applications, presents a new strategy for determining the optimal circuit based battery model. The proposed methodology has the fallowing novel features: a new test for establishing the performances of the model and a multi-criteria analysis for determining the optimal number of resistor-capacitor branches. Then the model is validated using data from a LFP battery submitted to discharge in concordance to a driving cycle.

Voltage oriented control of three-phase PWM rectifier using space vector modulation and input output feedback linearization theory

The current energy challenges encourage the researchers to design and optimize the devices allowing the conversion and the control of energy. The power converters consist in adapting the source of energy to the needs for the load. In this paper, we present the study of voltage oriented control with space vector modulation (VOC-SVM) strategies for three-phase PWM rectifiers. This strategy of control founded on the transformation between stationary coordinates α-β and synchronous rotating coordinates d-q, it is based on the input-output feedback linearization controller for which the objective is to control the grid currents d- and q-axis component, while the dc-bus voltage is maintained at the desired level and satisfy the unity power factor (UPF) operation. The strategy VOC-SVM ensures high static performance and fast transient response by using internal current control loop. Moreover, the studies were carried out by using simulation under the environment Matlab/Simulink.

Extending the battery model from a single cell to a battery pack with BMS

This paper presents the means for extending a battery model from a single cell to a battery pack with Battery Management System (BMS. Several modeling strategies are analyzed and a new one based on the cell with the minimum capacity is developed and implemented.

Temperature influence analysis on the hysteresis of a parallel connection of LFP battery cells

This paper analyses the effects of temperature on the hysteresis of a parallel connection of LiFePO4 cells. Within an automotive battery the cells exhibit different temperatures depending on their position. Due to this the cells behave differently on charging and discharging. To quantify the effect of temperature we analyze the Open Circuit Voltage hysteresis for three different temperature differences between two parallel connected cells.

Modeling a PV panel using the manufacturer data and a hybrid adaptive method

This paper presents a new method for establishing the series and parallel resistors for the one-diode model of a PV panel. The proposed method uses only the PV panel data that are provided by the manufacturer to train a feed-forward neural network that computes the values of the two resistors. The developed model is validated by comparing it to a classic one diode model whose resistors are determined through Newton-Raphson method.

Considerations on the command and analysis of the low speed permanent magnet synchronous machines operating

In this paper the authors present an analysis of the command with rectangular (trapezoidal) printed currents of a permanent magnet synchronous motor (PMSM) for driving a bicycle. Constructive details and materials are presented. It also presents the results obtained from numerical simulations of the studied permanent magnet synchronous machine for its operating as a generator. In this operating regime, the magnetic field through machine was determined using 2D finite element method, and were represented the magnetic flux lines in a cross section of the machine and the magnetic induction map. Then the variation curves of the cogging torque of the machine, the phase voltages and the phase currents have been computed. On the studied machine were performed a series of measurements and experimental data acquisition which were then processed in order to validate some characteristics and computing parameters of the permanent magnet synchronous machine operating as a generator.

Modeling aspects of an electric starter system for an internal combustion engine

This paper presents a mathematical model of an electric starter system for an internal combustion engine. For achieving this we started from the main elements that characterize the cranking process: torque, speed, starter motor terminal voltage and the battery terminals voltage depending on the current drawn from it. Based on these values we developed a general model that reproduces the starter motor behavior during cranking.

Determining the Optimal Battery Model for a Specific Application

Determining the optimal battery model for a specific application is a complex process that must take into account several factors: the type of application, the accuracy required, the degree of complexity, etc. This chapter introduces a new method for determining the optimal model and has its starting point in analyzing the discharge profile, and employs a multi-criteria analysis for processing the experimental data. The method presented is validated experimentally for a LiFePO4 battery subjected to discharge after an Urban Dynamometer Driving Schedule (UDDS) cycle.

A LiFePO 4 battery discharge simulator for EV applications — Part 2: Developing the battery simulator

This paper, the second of the series on developing a LiFePO4 (LFP) battery discharge simulator for EV applications, presents an approach for developing the battery simulator. After the optimal battery model was established, in the first paper, the main drawbacks of battery simulator are addressed and a new strategy is proposed. This strategy uses neural networks (NN) for online estimation of the battery parameters during discharge. The proposed simulator is validated by comparisons with a typical simulator and data from a LFP battery submitted to discharge profile in concordance with a driving cycle.

Estimating a battery state of charge using neural networks

This paper presents the means for estimating a battery State of Charge (SoC) using neural networks. Several neural networks such as: radial basis function (RBF), feed forward (FF) and nonlinear autoregressive with exogenous (external) input (NARX) are used for curve fitting and predicting features values. The conclusions are drawn after comparing the values obtained from the models and the data obtained from discharging a LiFePO4 (LFP) battery.