In English

Comparison of Simulation Programs for Supercapacitor Modelling

Patrik Johansson ; Björn Andersson
Göteborg : Chalmers tekniska högskola, 2008.
[Examensarbete på avancerad nivå]

To meet increasing demands for fuel efficiency, alternative drivetrains are being developed. Electric hybrid vehicles are one such alternative and an important part of them is their energy storage. To provide this storage the supercapacitor is an interesting component because of its high power density compared to batteries. The increasing use of this component provides the base for this project. This report consists mainly of three parts. The first part contains information about seven different simulation programs, which has been gathered through brief evaluations of the respective programs. In this part general information about the programs and their capabilities is presented. The analysed programs are Simulink, SimPowerSystems, OrCAD Capture, PSCAD, Saber, PLECS and Dymola. Advantages and disadvantages are summarised for the program alternatives. The second part of the report presents the supercapacitor modelling. During the initial tests of all simulation programs a basic model is developed. After initial program tests using the basic model, the search for a more advanced model is started. A part of this effort is to conduct several measurements on various supercapacitors in varying temperatures. The measurement results are used to develop an advanced model which includes temperature dependency and is able to accurately model the behaviour of different supercapacitors. After the advanced model development is finished, it is implemented in the five programs which seemed most interesting during the initial round of modelling. The advanced model is given the name DTDSM which stands for Dynamic Temperature Dependent Supercapacitor Model. These programs are Simulink, SimPowerSystems, Saber, PLECS and Dymola. The third part of the report covers the simulation results. It is found that it is possible to do multi-domain modelling in the programs Saber and Dymola using predefined components. Furthermore it is observed that the electrical modelling available in all program alternatives except for Simulink, can improve the model overview and reduce problems with causality. Regarding simulation performance in the program alternatives it is found that for simulations resulting in comparable accuracy, Saber performs fastest with a time of 5 seconds and Dymola has the longest time with 41 seconds. Moreover the model error of the DTDSM Simulink model seems to be relatively stable during simulation and the magnitude of the error towards the measurement is also low, about 2.1 % of the rated voltage. For the basic model the same error is at about 5 %.

Publikationen registrerades 2008-12-24. Den ändrades senast 2013-04-04

CPL ID: 82856

Detta är en tjänst från Chalmers bibliotek