Ivo Pazdera

This paper is part of an effort to bridge the gap between energetic and equivalent-circuit motor models used in system-level electric-powertrain simulations, balancing model fidelity with computational efficiency. The goal is to investigate the use of lookup tables as tools to abstract low-level motor control into a numerical optimization problem requiring few input data and minimal specialized knowledge, thereby making high-fidelity equivalent-circuit models more suitable for early-stage system-level powertrain studies. The approach was experimentally validated on a 600-W induction motor. Results demonstrate that the control based on lookup tables achieves torque, flux, stator current, and efficiency levels comparable to a traditional field-oriented control scheme. While the two methods use different current trajectories, the overall performance metrics remain similar, confirming the effectiveness of lookup tables as a simulation tool for electric-powertrain development, control strategy evaluation, and early-stage electric vehicle system studies.

This paper presents a comparison of commonly used induction machine control methods for loss minimization (maximum torque per ampere, maximum torque per ampere with iron loss consideration and search algorithm) with the presented loss minimization algorithm. The efficiency is compared by simulation in Matlab/Simulink environment on an induction machine model with real parameters for different speed and torque. The results show that the proposed loss minimization algorithm achieves very good efficiency over a wide speed and torque range compared to the other methods.

In this paper, a comparison of two loss minimization algorithms for induction motor drives derived for the Gamma network and the inverse Gamma network is performed. The basic principle of the algorithm is to find the optimal flux at which the motor has the lowest power losses at the desired torque and speed. The optimal flux is then given as the desired flux to the flux controller in the d-axis. To validate the algorithms, motor models in a stationary frame were first built with non-constant magnetizing inductance and estimation of iron losses and mechanical losses. The proposed algorithms were verified by simulation in Matlab/Simulink environment with real motor parameters. The simulation results confirmed that both algorithms achieve the same efficiency. In terms of derivation complexity, the inverse Gamma network-derived model is simpler.

This paper deals with the design and operational evaluation of a timber tractor-trailer unit with a hybrid trailer drive. The source of electrical energy for the two induction motors driving the front wheels of the tandem trailer axle is a battery, which is recharged by an induction machine operating as a generator during periods of a lower demand for power from the tractor diesel engine. An electric drive was designed for the defined working cycle of the tractor-trailer unit, and its loading characteristics were tested in the laboratory. The parameters measured on the field tests during timber forwarding were battery voltage and power, and the energy balance. Three adjustment levels of the potentiometer controlling the trailer hybrid drive (50, 75 and 100%) were tested at three different forwarding distances of 100, 500 and 1000 m. Additionally, any slippage of the prime mover wheels and trailer was measured. The maximum peak power taken from the battery was ca. 33 kW during the field tests, whilst the drive was able to deliver a peak output of up to 72 kW for 10 s and permanently up to ca. 50 kW. Even in harsh terrain conditions, the electric drive assisted the combustion engine only when the loaded tractor-trailer unit was travelling uphill. The hybrid drive operation was sustainable for the whole working shift, without the need for recharging when the potentiometer was set to 50%. This appropriate setting of the potentiometer controlling the trailer’s hybrid drive reduced the slipping of the driven wheels of the tractor-trailer unit whilst travelling uphill.

This paper deals with the estimation of depth of discharge for Li-ion batteries. Estimation is based on the knowledge of discharging curves measured for discrete values of loading currents. The estimator of the depth of discharge is a form of feedforward neural network which is trained with the measured data of discharge curves. Accuracy of estimation of the depth of discharge is shown for arbitrary generated and measured loading characteristics, where the depth of discharge is estimated by the designed neural network and measured by using the Coulomb counting method.

This paper discusses the simulation of the cascaded H-bridge (CHB) multilevel converter for power coolant pump with the demand of high reliability and robustness. Multi-level converters are designed to control medium-voltage motors, energy conversion, active power filtering, and many other applications. First, an overview of the CHB converter, the method of control, and operation under fault conditions is described. The next part deals with the requirements of the application of the converter and the parameters of the simulation system. Then the 13-level multilevel converter is simulated during operation under normal conditions and during operation under abnormal conditions. If the cell is damage, it is bypassed by an auxiliary switch. The simulated voltages and torques were compared with each other. As a result, the converter is able to operate even with three faulty cells without interrupting the operation of the cooling pump.

This paper deals with a mathematical description of an alkaline battery impedance dependence on frequency. This mathematical description is done in two different ways. In the first case, a general fractional transfer function is used and in the second case an artificial neural network is used. Both approaches are discussed and compared with real measurement.

This paper deals with problems with the estimation of the state of charge of Li-ion batteries. There are shown and discussed two basic methods for estimation of the state of charge. There are discussed essential problems with an estimation based on the so-called Coulomb counting method and based on an equivalent circuit of Li-ion battery.

The paper describe design of battery powered drive unit intended for tree harvester application. Battery powered hybrid drive unit in working machine is currently quite common (1). The battery capacitance is not sufficient to realize harvester as an only battery powered machine. Nevertheless, the hybrid concept of electric drive is useful to cover short power peaks demands during the harvester standard operation. In harvester application there is typical high ratio between short power peak and average demanded power. The development in batteries increases of significance and possibility of electric drive utilization in three harvesters.

This paper deals with a measurement of alkaline and Li-ion battery impedances in wide-range frequency from about one Hz to a few hundred kHz. Measured battery admittances are used for the construction of battery transfer functions, i.e., the dependences of admittance on frequency. Comparison of transfer functions for different battery types is shown. It is shown that transfer functions for both types have the same form and this transfer function is only the third-order.