Riku Pollanen

Modeling and Analysis of the Dead-Time Effects in Parallel PWM Two-Level Three-Phase Voltage-Source Inverters

This paper addresses modeling and analysis of the dead-time effects in parallel pulsewidth-modulated (PWM) two-level three-phase inverters. The main objective is to gain understanding of how the effects caused by the necessary blanking time, the finite turn-on and turn-off times of the switching devices, and the forward voltage drops of the switching devices and the antiparallel diodes, i.e., the dead-time effects, influence the circulating current generation between the parallel-connected units. To meet this objective, a circulating current model taking these effects into account is developed for the parallel connection of n units. The model, which is an average model by nature, can be used to study the circulating current behavior with different types of PWM methods and to estimate the resulting circulating current values when there are differences in the dead-time effect parameters. In other words, the model provides an analytical way to consider the significance of these effects. To verify the validity of the developed model, the circuit simulation and experimental results are shown and compared with the analytical results. The illustrations show that the results obtained with the developed model are in good agreement with the circuit simulations and the experiments.

Identification of grid impedance for purposes of voltage feedback active filtering

A voltage feedback active filter is vulnerable to unknown grid impedance. To overcome this problem we propose an identification method, which uses the control system of a frequency selective active filter to measure the grid impedance at selected frequencies. The usefulness of the method is experimentally demonstrated with a 19 kVA active rectifier with a voltage feedback active filtering function. The voltage feedback active filtering is performed in a case in which the active filter control is not stable before the impedance is identified with the method proposed. It is shown that the use of the measured grid impedance in the control system greatly enhances the dynamic stability of the system. Also, the grid impedance measurements are provided in two cases.

Control of zero-sequence current in parallel connected voltage source PWM rectifiers using converter-flux-based control

A new zero vector selection strategy for the converter flux based control (CFC) of voltage source PWM rectifiers is introduced. With this strategy the circulating current in paralleled rectifiers is effectively suppressed and independent operations of the individual converter units are achieved. In addition to normal line filter, no additional passive components such as transformers or current sharing reactors are required to parallel the converters. Simulation and experimental results validate the developed model and the proposed zero vector selection scheme

Analysis of System Architecture of FPGA-based Embedded Controller for Magnetically Suspended Rotor

This paper discusses a design and implementation of an application specific architecture of active magnetic bearings (AMBs) controller in a field programmable gate array (FPGA) circuit. As a first step, a prototyping platform that eases development of control, testing and integration of tested algorithms into the FPGA is built. The FPGA implementation of an AMB current controller of a set of ten H-bridge switching amplifiers is presented. Different architectures of the controller are examined with respect to required performance (response time to measured and control signals, utilization of resources) and design constraints (number of ADC channels, sampling time, applied position control). The implementation of selected architecture is presented with details. The process of designing digital control application in FPGA is revealed. The opportunities and threads of flexible HDL-implementation are discussed.

Physical drawbacks of linear high-speed tooth belt drives

Production rate plays an important role in industrial applications, which means higher demands for accelerations and speeds of the systems. The requirements for accuracy and repeatability are also increasing. A solution for these demands is a high-speed tooth belt linear drive; however, the drawbacks of system of this kind are non-linear friction and flexibility of the belt, which make the precision control of the system difficult. In this paper, the frictions and flexibility of the high-speed tooth belt linear drive are analyzed.

Run-Time Debugging and Monitoring of FPGA Circuits Using Embedded Microprocessor

Field programmable gate arrays (FPGAs) provide a fast and flexible hardware for embedded control systems and signal processing. Despite this, tracing and monitoring of internal signals is awkward. FPGA vendors provide their own tools to solve the debugging problems but they are not sufficient for real time monitoring. Instead, these signal tracing tools are good especially for tracing timing issues. This paper presents a method to monitor the internal signals of FPGA circuits by using an embedded microprocessor. The efficiency of this method is demonstrated with an FPGA-based active magnetic bearing control hardware

Compensating the island network voltage unsymmetricity with DTC-modulation based power conditioning system

An algorithm for compensating island network voltage unsymmetricity with a DTC-modulation based scalar controlled power conditioning system is introduced and experimentally verified. The proposed method consists two separate stages where first the unsymmetricity is detected and then the compensating negative sequence system is calculated. Experimental results show that the proposed method fully compensates the voltage unsymmetricity resulting from unbalanced load. Both the magnitude and the phase angle unsymmetricity are compensated.

Performance, limitations, and control of a high-speed tooth belt drive in a motion control application

Motion control applications often demand fast operation times and high accuracy at the same time. Finding the optimal solution between these demands can be sometimes difficult. A linear tooth belt drive offers inexpensive, fast, and relatively accurate solution for motion control demands. However, if the best possible accuracy and highest speeds must be reached, the control designer should be well aware of the non-linearities and control limitations of the system.

Linearization of Force Characteristics of Active Magnetic Bearings for the FPGA-based LQ-controller

This paper discusses a linear-quadratic (LQ) control and nonlinearities compensation of a rotor suspended by active magnetic bearings (AMBs). The control structure is based on a state-space controller with additional integrative feedback, current estimator and disturbance estimator. An effective compensation of performance variations in dynamic force characteristics of AMBs is performed using two nonlinear principles. Firstly, the usage of the inverse nonlinearity, a well-known method in control engineering is studied. Secondly, the novel model reference based method is introduced and validated. It is shown that the proposed compensation provides certain advantages over conventional linearized model based controllers. The overall control is tested with the flexible rotor model and the nonlinear AMBs. Finally, we present an integration of the proposed compensation into the digital controller. Accurate modeling of the AMB nonlinearities in a field programmable gate array (FPGA) based controller is achieved by using multivariable interpolation and look-up table.

Reluctance network method based dynamic model of radial active magnetic bearings

Active magnetic bearings (AMB) are increasingly used in high-speed electrical machines, compressors, turbomachinery, spindles and energy storage flywheels. In order to investigate the performance of these bearings, and especially to develop a reliable control for them, an accurate and computationally efficient bearing model is required. In the model the magnetic nonlinearities due to saturation of the magnetic materials as well as the impact of the leakage flux paths, cross-coupling effects and the eccentric rotor must be considered. Mainly, the analysis and design of AMB are based on two-dimensional (2-D) finite-element method (FEM). Although this method yields a precise determination of the device performance, it is not attractive for dynamic simulations due to its high computational cost. Alternatively, the reluctance network method (RNM) provides satisfactory accuracy compared to the 2-D FEM but with only a fraction of computations. In this paper, a dynamic model for a radial active magnetic bearing is developed using the RNM.