Development of a Tool to Determine the Correlations between Transmission Losses and Influencing Factors

Investigations on which factors that affect the power transmission losses have been performed several times before. However, increasing the size of the studied system rapidly increases the complexity and the problem becomes computationally heavy to solve. In this thesis a statistical approach has been used in order to be able to investigate what factors that affect the power transmission losses in a large and complex network. The area investigated is Regionnät Syd, a medium voltage network with over 200 substations, owned by E.ON Elnät Sverige AB. Special focus in the thesis is on how losses are affected by the increasing share of distributed generation, such as wind power production, in the system. The statistical calculations were performed in Lavastorm, a data analysis software which can support statistical models written in the programming language R. The result gave the correlations between the power losses and the parameters included in the study and was presented as a marginal effect analysis. The main finding is that distributed wind power production affects the power losses significantly. For the majority of the data an increasing share of wind power decreases the power losses, but after a critical value the losses start increasing. To use a statistical approach proved to be a feasible method to study a power network. However, it is important to keep in mind that a correlation between a factor and a variable are not necessarily the same thing as causation. The final model presented in the results has some defects that are accounted for in the discussion. Possible improvements, in order to get reliable results, are presented as recommendations for future work