Stephan Rees

New time-discrete modulation scheme for matrix converters

While the known modulation strategies for matrix converters are based on pulsewidth modulation (PWM)-or vector modulation-this paper presents a novel time-discrete modulation method based on real-time prediction calculation to select the switching states. The decision about which switching state is to be set for the following sampling period is made by the use of a predictive quality function. Using this approach, unity displacement factor is seen at the supply side with minimum line current distortion while the load currents follow their reference values with good accuracy. The quality function is derived from a mathematical model of the matrix converter and the controlled system. Measurements taken on a model plant, consisting of a matrix converter and a standard induction machine with a rated output power of 11 kW, show that the matrix converter, equipped with the control method presented here, offers advantages over systems with conventional frequency converters, especially in terms of the input current distortion.

New modulation strategy for a matrix converter with a very small mains filter

While known modulation strategies for matrix converters are based on PWM or vector-modulation, this paper presents a novel time-discrete modulation method by which the switching state of the matrix converter is changed only at equidistant points in time. The decision about which switching state is to be set for the following sampling period, is made by use of a quality function. Using this approach, the switching state is selected in such a way as unity displacement factor is seen at the supply side whilst the load currents follow their reference values with good accuracy. The quality function mentioned above is determined via a mathematical model of the matrix converter and the controlled system. The one switching state that induces the optimum value of the quality function is selected for the next sampling interval. Measurements taken on a model plant, consisting of a matrix converter and a standard induction machine with a rated output power of 11 kW show that the matrix converter, equipped with the control method presented here, offers some advantages over systems with conventional frequency converters.

New cascaded control system for current-source rectifiers

A control method for current-source rectifiers (CSRs), which realizes substantially sinusoidal line currents, unity displacement power factor, and a dc-link current control with excellent dynamic properties is presented. CSRs are usually operated by pulsewidth-modulation (PWM) or space-vector-modulation techniques. However, due to the mains LC filter resonant circuits when using these modulation methods the system stability has to be investigated, resulting in restrictions on the minimum PWM frequency and the minimum size of the LC filter. Furthermore most known dc-link current control loops use dc-link inductors of considerable size. This limits the dynamic performance and, therefore, reduces the attainable efficiency of CSRs. To overcome these problems, a new cascaded dc-link current control system for CSRs is presented. Its inner capacitor voltage controller is based on a time-discrete modulation method, which realizes a fundamentally stable control of the mains LC filter resonant circuits, avoiding the mentioned restrictions. The system controlled by the superimposed dc-link current controller is linearized by a new approach, allowing excellent dynamic performance and, therefore, a comparatively small dc-link inductor to be used. The paper includes guidelines on how to design the mains filter components and the dc-link inductor. The feasibility of the presented cascaded controller is confirmed by measurements taken on a 60-kVA model current-source converter and different loads.

An analytical approach to steady-state current control properties of power converters featuring discrete-time switching

Power converters with semiconductors switching only at predefined, equidistant instants - known as ldquodiscrete-time switchingrdquo, ldquosynchronized switchingrdquo or ldquoclocked commutationrdquo - show considerably different current control behavior compared to pulse-width modulated (PWM) converters with linear current con trollers. This paper presents and explains some of the effects of discrete-time-switching on output current waveforms, especially when using a predictive switching state selection algorithm. Typical characteristics of generated switching state patterns are presented. A single-phase and a three-phase inverter controlled by a predictive discrete-time modulation strategy are considered. Analytical, experimental and simulation results show that average switching frequency and error current RMS are well defined, in spite of the inherent randomness of the control scheme.

A smart synchronous rectifier for 12 V automobile alternators

A new gate control unit for power MOSFETs is presented in this paper, which is especially suited for synchronous rectifiers in 12 V automobile vehicle-nets to replace inefficient silicon diodes. Unlike known control units for such synchronous rectifiers based on comparator circuits or external timing control, the proposed circuitry is inexpensive and insensitive, as it consists of only one operational amplifier and some passive components per MOSFET. Measurements taken on a prototype synchronous rectifier connected to a standard passenger cars alternator prove a major increase of the rectification efficiency as well as a higher output current of the alternator compared to a standard diode rectifier. This synchronous rectifier operates at load currents up to 120 A and up to a maximum alternator speed of 15000 rpm. The entire gate drive circuit is described, providing information about the desired properties of its components. The switching processes of the presented circuit as well as the parameters that determine its operational performance are explained in detail.

Field-oriented control of current-source inverter fed high speed induction machines using steady-state stator voltages

One well suited solution to overcome the problems of voltage-source-fed high speed induction machines such as stress of the stator winding insulation and overheating of the rotor is the use of a current-source inverter, since the machine is then fed by almost sinusoidal currents and voltages. However, such drive systems are only efficient if first of all the DC-link current is adjusted to its lowest possible magnitude and secondly - which is the same with directly-fed machines - the value of the rotor resistance or the magnetizing current respectively is assumed correctly. This paper proposes a new DC-link current set value assignment and a method to improve the models for magnetizing current and field angle by using steady-state stator voltages. Measurements taken on a model plant with an output power of 35 kW and a nominal speed of 30000 min-1 prove the feasibility of the presented methods

Increased efficiency in the calibration process of automotive Li-ion battery systems

This paper discusses the use of model-based calibration for Li-ion batteries. The scope is set on the modeling algorithm, in this case the Gaussian Process Regression (GPR) in combination with an external dynamic structure (NARX), and its capability to describe dynamic battery behavior. A data-driven model is generated based on measurement of a Stuttgart Cycle. Model quality is evaluated on an Artemis Cycle. The achieved results show that the GPR can be used in a future model-based battery calibration process for dynamic applications in order to increase calibration efficiency.

Generation 2 Lithium-Ion battery systems – Technology trends and KPIs

Currently in series produced Lithium-Ion batteries – so called Generation 1 Lithium- Ion batteries – are used in high voltage battery applications in Hybrid, Plug-In-Hybrid and Electric vehicles. These Generation 1 battery systems proved already in series that Lithium-Ion batteries fulfill the high requirements regarding safety and electrical power in automotive applications. The real proof of the requested minimum 10 years life time requirement is not done yet. However, based on results from accelerated aging tests in laboratories one can expect that the Lithium-Ion batteries will fulfill the requested life time requirements of automotive field usage, too. Main targets of the development of the second generation of Lithium-Ion battery systems are decreased specific cost in EUR per kWh as well as increased volumetric and gravimetric energy density. Especially cost down will be the barrier to overcome in order to enable a wide use of electric drives in vehicles. Starting with a short overview of the actual status of Lithium-Ion battery technology this paper gives prognoses how the requirements to and the Key Performance Indicators of Generation 2 batteries will develop. Besides the already established high voltage systems used in Hybrid, Plug-In-Hybrid and Electric vehicles applications also new applications with a voltage level lower than 60 V are described. Such systems will be relevant for Boost-Recuperation-Systems (BRS) or Recuperation-Systems (RS). These systems are in direct competition to optimizing measures of conventional combustion engine powertrains. Hence an increased usage of such systems will only be possible when the demanding cost targets will be met. Focus of the paper will be batteries for the applications in Plug-In-Hybrids (PHEV), Electric vehicles (EV) and Boost-Recuperation-Systems (BRS).