Mladen Knezic

Power loss model for efficiency improvement of boost converter

The analytical expressions for boost converter loss are presented in this paper. All significant sources of conduction and dynamic losses within the converter elements, including losses in choke magnetic material and switching losses which occur due to diode recovery time, are discussed. Sources of losses in continuous and discontinuous current mode of converter are examined. The equivalent input resistance of the generator is included in converter loss model. In the paper, the case when single-phase rectifier is input circuit of the converter is presented. Loss model of converter and its efficiency was compared with experimental measurements on a prototype of the boost converter.

Increasing EtherCAT performance using frame size optimization algorithm

EtherCAT protocol is a popular Real-Time Ethernet solution that offers the highest communication efficiency in a number of operating conditions. However, for huge networks with several hundreds of devices distributed in the field, EtherCAT performance can become a critical factor. In this paper, we propose a solution for increasing EtherCAT performance using a frame size optimization algorithm. The optimization algorithm is described and analyzed for a simple scenario present in typical automation systems.

Topology aspects in EtherCAT networks

Fieldbus networks based on Ethernet become an inevitable part of factory automation systems. The number of installed nodes has increased over the last few years. Therefore, efficient network planning sets new challenges for todays engineers. This paper shows major aspects of topology planning in industrial networks based on EtherCAT protocol.

Boost converter efficiency optimization in wind turbine

A method for efficiency optimization of boost converter applied in wind turbine with permanent magnet generator (PMG) is proposed in this paper. This method gives the best results when output power of generator is less than rated value. Its validation is verified through simulation and experimental results.

Towards extending the OMNeT++ INET framework for simulating fault injection in ethernet-based Flexible Time-Triggered systems

Traditional distributed embedded systems are configured using static environment information and thus do not support dynamic behavior of the system. The necessary flexibility in the system may be provided by the Flexible Time-Triggered (FTT) communication paradigm. If, in addition, it is required that the system operates continuously, the suitable fault tolerance mechanisms that provide high reliability have to be developed and deployed in the system. To be able to successfully assess those mechanisms, it is reasonable to develop simulation models that support injection of various types of faults. This paper describes an OMNeT++ simulation model for distributed systems that are based on the Hard Real-Time Ethernet Switching (HaRTES) implementation of the FTT paradigm. The contribution of the paper is twofold. First, we provide a library of components that are required for modeling FTT networks with arbitrary number of FTT slaves connected to a HaRTES switch, and second, we used the developed components to build an FTT system that is suitable for assessing some of recently proposed mechanisms for tolerating certain transient faults in the communication channel.

Algorithm for efficiency optimization of the boost converter in wind turbine

Implementation of a method for efficiency optimization of boost converter in wind turbine with permanent magnet generator is described in this paper. This method is based on variable switching frequency technique suitable in conditions when input power of the converter is less than rated. The proposed concept is verified through computer simulation of the boost converter and experimental measurements on the prototype.

Performance analysis of the Ethernet Powerlink PollResponse Chaining mechanism

Ethernet Powerlink (EPL) is an industrial Ethernet networking solution that enables hard real-time communication in distributed control and automation systems. The PollResponse Chaining (PRC) mechanism is an extension of the EPL protocol aimed at increasing the network performance when nodes exchange small amount of data, especially if they are connected in line topology. In this paper, we conduct performance assessment of the PRC mechanism for two basic network configurations (line and star topology). In addition, we provide some experimental results obtained by measuring relevant timing parameters of the real EPL network prototype operating in the PRC mode.

Theoretical and Experimental Evaluation of Ethernet Powerlink PollResponse Chaining Mechanism

Ethernet Powerlink (EPL) is an industrial Ethernet networking solution commonly used as a communication network in distributed control and automation systems ranging from simple I/O to highly complex motion control applications. The PollResponse Chaining (PRC) mechanism is a new EPL standard feature aimed at increasing the network performance when nodes exchange small amount of data, especially if they are connected in line topology. In this paper, we present a theoretical model suitable for assessing the real-time performance of the EPL protocol operating in both the standard and PRC mode for two basic topologies commonly found in real installations, line, and star. Moreover, we carried out a series of experiments on prototype networks in order to acquire the relevant timing parameters of the EPL network components required for the development of an OMNeT++ simulation model, which can be further exploited to evaluate the EPL protocol in the case of more complex scenarios. Finally, we propose a modification of the original PRC solution to improve its flexibility while allowing the same (or even higher) performance level. The feasibility of the proposed approach was demonstrated on a real prototype, whereas a certain performance gain over the original PRC mechanism was proved through simulations conducted on a more complex network structure.

A first performance analysis of the Admission Control in the HaRTES Ethernet switch

There is a growing interest in developing embedded systems capable of being deployed in dynamic environments that may change in unpredictable manners. When such systems are Distributed Embedded Systems (DESs) they must exhibit flexibility at all levels of their architecture, including the network. On the other hand, there is a clear trend in industry towards using Ethernet-based protocols at the network level of DESs. Nevertheless, Ethernet lacks appropriate support for real-time (RT) communications, mixing different RT traffic and on-line management of the Quality of Service (QoS). Several implementations of the Flexible Time-Tiggered (FTT) protocol over Ethernet were proposed to cope with these drawbacks. FTT is a master/multi-slave protocol that is able to simultaneously convey real and non-real-time traffic and provides mechanisms for dynamically changing the QoS of the network, including Admission Control (AC). The AC is a fundamental component for on-line network management, since it guarantees that each participant gets the required QoS. This paper presents the implementation in OMNeT++ of a simulation model of the AC in the FTT HaRTES switch as well as a preliminary performance study using that model.

An algorithm for boost converter efficiency optimization

In this paper, an algorithm based on the technique of variable switching frequency is applied, so that working point of boost converter is at the boundary between continuous and discontinuous working mode aiming at achieving maximum efficiency of the converter. Controller is based on variable switching frequency and measuring the voltage on the main converter switch. The proposed algorithm is verified by the simulations and experimental measurements on a converter prototype.