The accuracy, ease of use, and execution time of the finite element method based software Maxwell coupled to the system simulation software Simplorer was evaluated for simulation of hydropower generator - power grid interaction. A generator test rig were modelled in Maxwell and coupled to Simplorer with a strong circuit coupling as a single machine infinite bus system. The accuracy of the model was measured by comparing the simulated output power oscillation frequency and damping characteristics to the measured ones after a torque step. The result shows that the difference in output power oscillation frequency between the model and the generator test rig was small, and that the difference in damping characteristics was significant. The usability of the software package was found to be fair, as were the execution times.

Photovoltaic solar power is an increasing source of energy and part of the renewable energy generation which is needed in the near future to achieve the set climate goals. When planning new photovoltaic installations, parameters which affect the design are both local conditions (e.g. weather) and system parameters such as tilt and azimuth angles. Commercial areas often have high loads during the day when solar power is available and are therefore interesting for photovoltaic installations. In order make a quick estimation of photovoltaic power potential in an area, a simulation tool which handles load profiles from multiple buildings would be desirable. The aim of this thesis project is therefore to create a tool which can simulate multiple photovoltaic systems and for each of them estimate system sizes, grid interactions, and area requirements. The simulation tool is based on Python programming with the aid of System Advisor Model, a simulation software for photovoltaic and other renewable energy tech-nologies. Optimization of orientation angles was made for clear sky with the goal of high load-generation match. Different system sizes were estimated and simulated based on different degrees of self-sufficiency, net-zero consumption, and the existing transfer capacity of the building in question. When the simulation result was compared to a detailed photovoltaic design project, some agreements between the results were found, as well as further development needs such as refining area estimation. To further develop the usability of the tool, a more user-friendly interface is needed. Other improvements could be to enable simulations of multiple direction systems and integration of the local grid structure and limitations.

Technological improvements along with falling prices on photovoltaic (PV) panels and electric vehicles (EVs) suggest that they might become more common in the future. The introduction of distributed PV power production and EV charging has a considerable impact on the power system, in particular at the end-user in the electricity grid. In this PhD thesis PV power production, household electricity use and EV charging are investigated on different system levels. The methodologies used in this thesis are interdisciplinary but the main contributions are mathematical modeling, simulations and data analysis of these three components and their interactions. Models for estimating PV power production, household electricity use, EV charging and their combination are developed using data and stochastic modeling with Markov chains and probability distributions. Additionally, data on PV power production and EV charging from eight solar charging stations is analyzed. Results show that the clear-sky index for PV power production applications can be modeled via a bimodal Normal probability distribution, that household electricity use can be modeled via either Weibull or Log-normal probability distributions and that EV charging can be modeled by Bernoulli probability distributions. Complete models of PV power production, household electricity use and EV home-charging are developed with both Markov chain and probability distribution modeling. It is also shown that EV home-charging can be modeled as an extension to the Widén Markov chain model for generating synthetic household electricity use patterns. Analysis of measurements from solar charging stations show a wide variety of EV charging patterns. Additionally an alternative approach to modeling the clear-sky index is introduced and shown to give a generalized Ångström equation relating solar irradiation to the duration of bright sunshine. Analysis of the total power consumption/production patterns of PV power production, household electricity use and EV home-charging at the end-user in the grid highlights the dependency between the components, which quantifies the mismatch issue of distributed intermittent power production and consumption. At an aggregate level of households the level of mismatch is shown to be lower.