Analys av spänningsstabilitetsberäkningar i det svenska elsystemet

Sammanfattning: Numerical simulations play a vital role in both static and dynamic stability assessment of electric power systems and the foreseeing of voltage collapse. Simulations performed by Svenska Kraftnät (SvK) are mainly done in the specialized voltage collapse predicting software tool, SPICA and the more all‐purpose software tool, ARISTO. The grid operators at SvK use SPICA to predict the maximum transfer limit at specified sections (in this report the section in concerns is the southern part of Sweden i.e. interface1 4). SPICA is mainly used as an online application by using actual states from the grid, these states update every quarter of an hour. This thesis is a continuation of the master thesis SPICA – ARISTO JÄMFÖRELSE by Jakob Katzman and Johan Fält. Royal Institute of Technology, Stockholm, May 2004. The task in this thesis is to seek alternative applied methods that can enhance the existing methodologies that are used by SPICA and ARISTO, to make the power system more efficient and reliable. The alternative methods involve alternative load modeling, voltage control capability of the sub‐transmission network and its impact on voltage stability and changing of SPICA parameters. Investigation was done in SPICA and the results were later validated in ARISTO. As it will show in this thesis the different load characteristics did not have an impact in SPICA due to SPICA’s Load Tap Changer model which keeps the voltage at the sub‐transmission level constant. On the other hand due to different load models and ARISTO’s time delay dynamic model ability, a lot of dynamic phenomena were presented in the ARSTIO simulations, phenomena like rotor angle instability. It is also shown that a voltage insensitive load model for both the reactive and active load power is not a recommended load alteration in ARISTO. An analysis is made to enhance SPICA calculations by tuning the programs parameters and later validate the results in ARISTO. The results show that by changing some parameters, divergence in SPICA calculations, which is a huge problem for grid operators, can be avoided. Also the load model has a role in SPICA’s ability not to diverge from a solution. Other parameters that had an impact on the SPICA calculations were where the geographical positioning of the continually increasing loads. Export to other countries is a common operation. This thesis has studied transmission capacity when only the export is increased. The results differ between SPICA and ARISTO. In SPICA the transmission ability to other systems remains basically the same as for load increase in company 4 for intact grid albeit when the grid is weakened2 there is quite a difference between the two increasing methods. Note that the results are independent of the load model. Where as in ARISTO if only the export is 1 The word ”interface” is used as the English translation of ”snitt” which is a term used to describe a set of transmission lines (or transformers) that carry the main power transfer between two areas, thus, snitt 4 is translated to interface 4. Also, the area between two interfaces is called “company” which is the English translation of “snittområde”. 2 A grid fault that has the most impact on the transmission capacity is applied to the system. increased then the system has an increased transmission capacity compared to if the loads are onlyincreased in company 4 independent of the condition of the grid, the transmission capacity differswith the load model. Note that neither of the tools have a complete DC‐link model or transmissionsystem model for the neighboring countries. During the simulations, weak system buses could be identified at the south‐west part of Sweden, this comes as no surprise to the people at SvK. Other conclusions that are presented are the actual operator process when using a SPICA calculated limit. It is important that the operator is aware of the reactive state of the system, that is, how much reactive power reserve in the form of not activated reactors, condensers, generators and other voltage regulator devices there are in the system. If there are a lot of reserves left then one has to be very careful of the transmission limit predictions made by SPICA calculations.