In the future renewable generation sources will have to co-exist with traditional generation sources that can overcome their limitations in providing secure power supply to society, such as gas turbines. In this project we proposed to substitute a simple model of a gas turbine by a detailed model to simulate the turbine´s behavior in a real system. Parameter identification of a simple gas turbine model was performed based on measurements from a detailed model. By this, the project contributes to a library of electric grid component models. The library is meant to allow operators and engineers to predict the system behavior by simulations. This can facilitate to anticipate possible instability in the grid and thereby contribute to the transition to the Smart Grid. Further, another gas turbine model was in the Modelica language was represented using the Unified Modeling Language (UML). This is a standard form of modelling that simplifies exchange of model information between project groups.

At KTH Royal Institute of Technology, a novel type of transverse flux generator aimed for ocean wave energy conversion is being developed. This machine is complex to assemble, and the aim of this project was to develop a tool to aidin mounting neodymium magnets, fixed in rows, next to each otherto create a single magnetic pole. The issue at hand is the orientationof said magnets. They are assembled in such a way as to constantly repel each other, and these induced repelling forces make complete assembly by hand impossible. Simulations were conducted to determine the optimal way to mount the magnets in the desired fashion, and a 3D-model of the tool, with respective drawings was then created.

The increase of wind power in the power gridsystem is becoming more noticeable. This is somewhat due to climate change and the need for more energy. As wind power is considered a green renewable energy resource and its one of the cheapest ways of generating power during operation, it has become more of afavorable energy resource. Although, there are certain benefits of using wind power instead of the alternatives,wind power brings other factors to the power system that must be considered. The power grid system operator strives to always sustain the balance between the load and the supply inthe power grid system. The more wind power that is integrated the more dependent on a resilient powergrid system we become. This is due to the fact that windpower is an unreliable energy resource, as it is always changing, and unlike other methods of generating power it lacks the control mechanisms and cannot be regulated. The more wind power integrated to the power system the more of a challenge sustaining the balance becomes. If large amounts of wind power were integrated to the power grid system, a distinct drop in wind power could correspond to a shortage. Therefore, it is necessary to construct a power system which is reliable and can at any time with stand definite changes in load and supply, if large amount of wind power is to be integrated. We define flexibility as the ability of the power system to always be able to compensate for the changes in load and supply, to sustain balance withinthe power system. In this bachelor thesis, the flexibility of a two-areapower system is investigated by using a mathematic optimization model. The author will also be looking into the impact of carbon policy instruments, such as carbon tax and carbon cap, on the power system together with enhancements to increase the flexibility of the power system.

This project aims to evaluate through simulationsin Simulink whether an average power consuming, 4-person household in Sweden, can be successfully supplied from a microgrid system powered by solar panels. The goal is to achievea 48 V DC system and examine both its voltage stability and long term power balance for stationary and transient conditions. The simulations are based on weather data such as temperature and solar irradiation as well as an estimated consumption pattern fort he household. Loads and sources in the system are identified as well as possible algorithms for communication between nodes. Two scenarios are tested, a stand alone grid with optional energy storage and a system connected to the main grid. The results indicate that a Swedish household can not be supplied from a stand alone microgrid due to low irradiation values, without an unreasonably large energy storage or change in consumption. Although a continuous power supply is missing, voltage stability and power balance is established for this scenario. With connection to the main grid all requirements above are met. The observed limitations of the methodology are analyzed. For future projects a communication algorithm for load variation, synchronized with irradiation inputs, can be implemented in the model. Also, simulations can be done for alternative geographical conditions to yield different results where alternative renewablesources, such as wind and geothermal, can be included.

QC 20101110

Degradation of XLPE insulated power cables by water-trees is a primary cause of failure of these cables. The detection of water-trees and information about the severity of the degradation can be obtained with off-line measurement using dielectric spectroscopy. In many situations only a limited part of the cable may be degraded by the water-trees. In such a situation a method for localization of this water-treed section would be desirable. On-voltage Time Domain Reflectometry (TDR) diagnostics proved to be capable of localizing the water-tree degraded sections of the cable. The possibility of using on-voltage TDR as a diagnostic method opens up as a further step for the development of an on-line TDR method where the diagnostics are performed using pre-mounted sensors on the operating power cable. The benefits with such a method are: ability to perform diagnostics without disconnecting the cable from a power grid; the diagnostics performed during a longer period of time could give an extra information; no need for an external high-voltage supply unit. In this thesis the sensors for the on-line TDR are investigated in terms of sensitivity and bandwidth. High frequency models were built and the simulation results in frequency and time domains were verified by measurements. Results of the on-voltage TDR measurements on the degraded XLPE cables in laboratory as well as on-site are presented. The on-line TDR system and the results of a four-days on-line measurement sequence are presented. Variations due to load cycling of the cable were observed, where an increase in the cable temperature cause an increase of the pulse propagation velocity in the cable. A method has been developed for high frequency characterization of power cables with twisted screen wires, where the measurements are performed using inductive strip sensors. This technique allows the high frequency parameters of the selected section of the cable to be extracted. The high frequency parameters are extracted from frequency domain measurements of S-parameters as well as from TDR measurements. / QC 20100709

Enligt riksdagsbeslut skall fler företag än SJ få möjlighet att bedriva järnvägstrafik. Hur skall då Banverkets energikostnader för kontaktledningsnätet fördelas mellan de olika trafikerande före­tagen? Individuella mätningar på järnvägsfordonen kan vara ett sätt att skapa ett underlag för hur energikostnadema skall fördelas. Syftet med detta examensarbete var ursprungligen att utarbeta ett pr.ktis:t systemförslag till ett energimatningssystem för mätning av elenergiförbrukningen i järnvägsfordon. Det visade sig dock inte vara så lätt att fastställa en lämplig mätstorhet. Därfor kom arbetet med detta. att ta mycket längre tid an beräknat. Dessutom skulle först en diskussion föras huruvida det ar moti­verat att över huvud taget utföra mätningar. När en mätmetod väl hade bestamts, blev det nödvändigt att utföra åtskilliga matematiska beräkningar för att motivera den valda mätmetoden. I korthet kan sägas att metoden innebar följande: Den aktivt forbrukade lokenergin mäts i en energimätare. Dessutom mäts strömavtagarens effektivvärdesström i kvadrat också i en energimätare. På så sätt är det inte tvunget att använda dyrbara datorutrustningar för att registrera den momentana effektförbrukningen i loken. Med hjälp av det första värdet kan operatören debiteras för den aktiva energi som förbrukats. Den aktiva energin kommer att utgöra största delen av operatörens energikostnad. Den andra storheten, kvadratvärdet av strömmen, används som ett mått på hur fordonet har framförts och därmed vilka förluster det har orsakat på kontaktledningen och i omformarstationer. Eftersom det är fråga om det sanna effektivvärdet kommer hänsyn att tas till både reaktivt effektuttag samt till övertoner i uttagen ström. Körsättets inverkan på effektuttaget kommer där­for att avspeglas i denna storhet. Med k:unskap om värdet av detta förlustmått för de respektive fordonen, kan en uppdelning av de gemensamma nätforlusterna göras. Med verskap om vad som totalt matats in till kontaktledningsnätet, och vad som har förbrukats som aktiv energi, kan således de totala förlustema i nätet beräknas. Dessa ar nämligen omöjliga att mäta direkt. Detta sätt att angripa problemet med hur förlusterna skall delas upp, benämns i rapporten totalförlustmodellen. For att kvadraten av strömmens effektivvärde skall vara ett bra förlustmått, krävs det att kontaktledningsimpedansens variationer i genomsnitt inte har för stor inverkan på strömmens vär­de. Det är denna problemställning som den utvecklade beräkningsmodellen behandlar. Slutsatsen som kan dras från beräkningarna är, att om effektuttagen är stora kommer ovannämnda förlustmått att ge en for liten andel av förlusterna för det aktuella loket, dock understi­gande 10 procents fel även vid höga effektuttag. Man kan korrigera för detta fel genom att an­vända sig av de två mätta storheterna. Genom att sätta förlustmåttet i relation till hur mycket aktiv energi som har förbrukats, fås en kvot som på ett utmärkt sätt avspeglar hur loket har framförts. Med hjälp av denna kvot fås en korrektionsfaktor, med vilken en korrigering av andelen kan ske. / The Swedish Parliament has come to a decision that other companies than SJ, the Swedish State railways, will be given the opportunity to carry on railway traffic. The question then occurring is how to distribute the energy costs of Banverket, the Swedish National Rail Administration, among the companies operating on the railway network. Individual measurements may be a way of founding a basis for the energy cost distribution. The purpose of this work was originally to draw up a practical suggestion how to measure the electricitiy consumption of railway vehicles. The first step was to decide what to measure. Sol­ving this task took far more time than had been calculated for initially. To begin with, the que­stion whether measurements should be performed at all, had to be discussed. When the method of measuring had been chosen, several mathematical calculations had to be carried out, in order to justify the chosen method. The method can be described as follows: The active power consumption of the engine is mea­sured by an electrical energy meter. Furthermore, the square of the effective current in the cur­rent collector is measured by a similar energy meter. Using this method, there is no need for expensive computer equipment to registrate the instantaneous power consumption of the vehic­le. The first of these two measured values is used to charge the traffic operator for the active ener­gy being consumed. This will be the main part of the total energy cost. The second value , the square of the effective current, will represent in what way the vehicle has been driven, and thereby the losses being caused in the contact lines and in the frequency con­verters. Since we are dealing with the true rms current, both the reactive influence of the engine, and the current harmonics will be regarded. Firstly, knowing what has been fed into the power supply network, and secondly, what has been consumed as active power in the vehicles, it is possible to calculate the total losses in the net­work. These losses are in fact impossible to measure directly. This way of treating the problem how t() distribute the losses, is called totalförlustmodellen (model of total losses) in this work. To be able to use the square of the effective current as a measure of losses, it is necessary that the current value in average is not greatly affected by variations of the contact line impedance. The developped calculation model deals with this matter. The conclusion that can be drawn is that great power consumption will lead to a share too small for the vehicle, however giving us an error less than ten percent. It is possible to correct this deviation by forming a correction quotient from the two measured values. This quotient results in a correction factor that is used to correct the share. / <p>QC 20180306</p>

The emergence of smart and integrated industrial ecosystems is replacing traditional manufacturing, where operations are more digitalized and automated. For communication and cooperation, these operations require seamless integration of the devices to ensure interoperability in a secure, reliable, and adaptable environment. Designing a solution with these capabilities raises concerns about automation and engineering optimization. More recently, the rapid progression toward Industry 5.0 (I5.0) is further reshaping the landscape of smart industry ecosystems, necessitating innovative engineering and management solutions based on its core values of resilience, sustainability, and human-centricity. This thesis investigates these challenges and requirements by emphasizing adaptable, secure, reliable, composable, and scalable communication in complex industrial ecosystems. Central to this work is a local cloud-based collaboration approach to the design and development of composable ecosystems using microservices, which facilitate the integration of various information technology and operational technology (IT/OT) systems prevalent in smart industries. These encompass industrial and smart home Internet of Things (IoT)-based smart industry ecosystems and smart energy systems. Using the microservice-oriented Eclipse Arrowhead framework, this research provides scalable and adaptable solutions that adhere to the core values of I5.0. This research also bridges the integration gap between smart manufacturing ecosystems and smart home IoT technologies, laying the foundation for interconnected smart factories and improved energy management systems. Collaboration between IT/OT components and stakeholders in smart industry and smart energy ecosystems improves competitiveness, productivity, and informed decision making, thereby filling a critical research gap. The thesis presents a cloud-based collaborative learning (CCL) approach for automated condition monitoring in smart industry ecosystems. The thesis exemplifies the use cases of wind farms and smart manufacturing ecosystems that use CCL to address the issues of dynamic learning and real-time data sharing between various IoT-based IT/OT systems. Unlike traditional smart manufacturing models that focus primarily on automation and cost efficiency, CCL-based and I5.0 core value-driven ecosystems support human-centricity, sustainability, and resilience. Lastly, the thesis investigates the optimization of demand response based on collaboration among stakeholders in smart energy systems using edge-based automation clouds. The proposed approach promotes resilient and sustainable smart city demand response strategies by ensuring human comfort, security, data privacy, and all stakeholder integration in smart energy systems.

In response to the rapid growth of electric vehicles (EVs), our research focuses on the critical need for efficient management of charging infrastructure to facilitate the widespread adoption of EVs. Thisresearch leverages historical charging data as a foundation for predicting charging station availability. The focus of our approach lies in the utilization of autoregressive models and Long Short-Term Memory (LSTM) algorithms, which play a crucial role in enhancing the accuracy of our predictions. We combine the results to build the ensemble model, improving prediction accuracy. Additionally, we enhance our model by incorporating transfer learning (TL) techniques to adapt it to new stations. The resulting predictive methodology empowers users with accurate insights into charging station availability one hour in advance, serving as a valuable tool for effectively planning EV charging activities.Our research goes beyond mere forecasting; it aims to contribute significantly to optimizing EV usage. By supporting informed decision making and encouraging efficient charging practices, we strive to pave the way for the seamless integration of EVs into modern transportation systems. Through these efforts, our research seeks to advance sustainable and eco-friendly mobility solutions in the realm of electric transportation.

The advancement of Artificial Intelligence technologies has paved the way for the increasing use of autonomous systems in various fields, including air, land, and sea. Maritime is an important domain for applying autonomous systems, as collisions at sea can result in significant losses. How- ever, the development of autonomous systems in the maritime domain involves the collection of vast amounts of data from sensors on vehicles, which poses significant challenges in understanding the types of data collected from vehicle sensors and the corresponding data system requirements due to the large amount of data collected. To address this issue, this thesis proposes the use of Oper- ational Design Domain (ODD), inspired by the autonomous car domain, to structure and manage the data. An ODD taxonomy was designed with the help of BSI PAS 1883 and ASAM OpenODD concepts. This thesis presents an initial draft of the ODD taxonomy for the maritime domain, followed by the implementation of ODD samples on actual data. The work involved converting unstructured data into ODD-friendly data. The results of this thesis include an ODD taxonomy, a framework, and a use case demonstrating the application of these concepts to actual data. These results provide an example of how to manage large amounts of data in the autonomous maritime environment and facilitate further testing of autonomous systems in this domain.