Enrique Kremers

A Complex Systems Modelling Approach for Decentralised Simulation of Electrical Microgrids

The structure and behaviour of Electrical Grids share many of the properties of Complex Computer Systems, with microgrids and other decentralised electrical systems attached to them, so they can be interpreted as Systems of Systems. Furthermore, the evolution of future electrical systems will bring a higher degree of decentralisation, especially concerning production and control. To deal with this paradigm change, new models and tools are necessary. In this paper a model of an electrical microgrid is presented. The approach used in the development of the model is agent-based in combination with system dynamics modelling. By mixing these approaches the different entities of the electrical system (production, demand, storage, etc.) have been represented. Through the individual behaviour of the agents it is possible to reproduce the complex behaviour of the system as a whole. This can produce expected and unexpected emergent effects on the interconnected system that are analysed through the simulation. A case study is presented to analyse the capabilities of such models. The example shows the simulation of an integrated microgrid system, where different components such as renewable energy sources and storage have been implemented. The simulation results of this case study are discussed.

Towards a 3D Spatial Urban Energy Modelling Approach

Todays needs to reduce the environmental impact of energy use impose dramatic changes for energy infrastructure and existing demand patterns e.g. buildings corresponding to their specific context. In addition, future energy systems are expected to integrate a considerable share of fluctuating power sources and equally a high share of distributed generation of electricity. Energy system models capable of describing such future systems and allowing the simulation of the impact of these developments thus require a spatial representation in order to reflect the local context and the boundary conditions. This paper describes two recent research approaches developed at EIFER in the fields of a geo-localised simulation of heat energy demand in cities based on 3D morphological data and b spatially explicit Agent-Based Models ABM for the simulation of smart grids. 3D city models were used to assess solar potential and heat energy demand of residential buildings which enable cities to target the building refurbishment potentials. Distributed energy systems require innovative modelling techniques where individual components are represented and can interact. With this approach, several smart grid demonstrators were simulated, where heterogeneous models are spatially represented. Coupling 3D geodata with energy system ABMs holds different advantages for both approaches. On one hand, energy system models can be enhanced with high resolution data from 3D city models and their semantic relations. Furthermore, they allow for spatial analysis and visualisation of the results, with emphasis on spatially and structurally correlations among the different layers e.g. infrastructure, buildings, administrative zones to provide an integrated approach. On the other hand, 3D models can benefit from more detailed system description of energy infrastructure, representing dynamic phenomena and high resolution models for energy use at component level. The proposed modelling strategies conceptually and practically integrate urban spatial and energy planning approaches. The combined modelling approach that will be developed based on the described sectorial models holds the potential to represent hybrid energy systems coupling distributed generation of electricity with thermal conversion systems.

A neural network based wind speed estimator for a wind turbine control

Variable speed wind generation systems are more attractive than fixed-speed systems because of the more efficient energy production improved power quality, and improved dynamic performance during grid disturbances. In this sense, to implement maximum wind power extraction, most controller designs of the variable-speed wind turbine generators employ anemometers to measure wind speed in order to derive the desired optimal shaft speed for adjusting the generator speed. In this paper it is proposed a new Neural Network Based Wind Speed Estimator for a wind turbine control. The design uses an feedforward Artificial Neural Network (ANN) to implement a rotor speed estimator, and simulated results show that the proposed observer provides high-performance dynamic characteristics.

A robust position control for induction motors using a load torque observer

The design of a robust position control scheme for an induction motor drive using the field oriented control theory is proposed. The proposed sliding-mode control law incorporates an adaptive sliding gain in order to adjust the sliding gain to the system uncertainties. Moreover, the sliding gain adaptation avoids having to calculate the upper limit for the system uncertainties. The design also incorporates a load torque observer in order to obtain the load torque applied to the induction motor without the use of the load torque sensor. The proposed observer is based on the system dynamical equation and uses the rotor speed and the stator current in order to obtain the load torque. The stability analysis of the proposed controller under parameter uncertainties and load torque variations is provided using the Lyapunov stability theory. Finally experimental results show that the proposed controller with the proposed observer provides high-performance dynamic characteristics and that this scheme is robust with respect to plant parameter uncertainties and load torque variations.

Adaptive robust control to maximizing the power generation of a variable speed wind turbine

The actual wind turbines are provided with adjustable speed generators, like the double feed induction generator, that are capable to work in variable speed operations. One of the main advantage of adjustable speed generators is that they improve the system efficiency compared to fixed speed generators because turbine speed is adjusted as a function of wind speed to maximize output power. However this systems requires a suitable speed controller in order to track the optimal wind turbine reference speed. In this work, an adaptive robust control for variable speed wind power generator is described. The proposed robust control law is based on a sliding mode control theory, that presents a good performance under system uncertainties. The proposed sliding-mode control law incorporates an adaptive switching gain, which avoids having to calculate an upper limit of the system uncertainties that is necessary in the traditional sliding-mode control laws. The stability analysis of the proposed controller under disturbances and parameter uncertainties is provided using the Lyapunov stability theory. Finally simulated results show, on the one hand that the proposed controller provides high-performance dynamic characteristics, and on the other hand that this scheme is robust with respect to plant parameter variations and external disturbances.

Agent-Based Simulation of Wind Farm Generation at Multiple Time Scales

Since the past decade, energy systems are undergoing a deep paradigm shift, caused by the liberalisation of energy markets, the introduction of renewable energies, and the emergence of new, distributed producers that feed into the grid at almost every level of the system. The general trend towards the introduction of renewable energy sources in the industrialised countries implies one of the greatest changes in the structure of energy systems. These systems are moving away from a centralised and hierarchical energy system, where the production follows a top-down principle under the strict control of the electricity supply companies towards a new system where diverse actors influence the energy supply. The production is no longer limited to large energy providers, as small decentralised producers now exist and inject energy at much lower voltage levels than before. These energy systems are suffering the consequences of such a paradigm change. This change basically consists in new regulations and the introduction of new energy production technologies that transform traditional centralised systems into decentralised ones. This whole process is part of the framework of the fight against the causes of climate change, which is mostly due to CO2 emissions. This paradigm change encompasses new tools and methods that can deal with decentralised decision-making, planning and self-organisation. The large amount of new actors and technologies in the energy production chain requires a shift from a top-down to a more bottom-up approach. Multi-scale simulation systems offer several advantages over classical models. The ability to run simulations on different time scales using the same model is an important issue for the upcoming modelling of energy systems. The main advantages are that there are fewer models and no need to port data between platforms. This leads to a more efficient simulation run and decision-making support. The challenges of these kind of simulations are that a multi-scale model for the moment will not be as accurate as a purpose made model. So, the modelling method, the parameters, etc. included must be carefully chosen to ensure both flexibility and accuracy. The work presented in this chapter concerns the wind generation module of an agent-based model for integral energy systems (developed at the European Institute for Energy Research 14

Agent-based modelling of electric vehicle driving and charging behavior

Electromobility lies on the crossroad between mobility and energy systems. The individual heterogeneous behaviours, and especially the spatial distribution and dynamism of the system make it a complex one. In this work, it is proposed an agent-based model to reflect this complexity and create a bottomup model which addresses specifically driving and charging behaviours of the individual agents. The model was implemented in a simple network which included the commonly used facilities in a city. This allows the computation of the generated load curve in a geographical context for any network. Different technical parameters were varied, as well as the driving and charging behaviours. The load curve as an aggregated result showed emergent patterns such as non-trivial effects when increasing the charging power. The model provides qualitative results from an exploratory point of view, which help to better understand electromobility systems by relating its causes and effects.

Optimization applied with agent based modelling in the context of urban energy planning

The inherent complexity of urban energy systems, and related decision making on system configuration and system operation strategies requires appropriate energy modelling, simulation and optimization means and tools. Due to its character, optimization applied with agent based modeling can be used to tackle problems whose nature is distributed and complex. In this work we present the insights gained through the optimization model building process in the area of urban energy planning, which deals with multi-scale, domain transversal and largely heterogeneous systems. Optimization is applied to urban energy infrastructure planning and energy system operation planning.

Modelling Lifestyle Aspects Influencing The Residential Load-Curve.

Using the results of a representative survey for the city of Stuttgart, household load-curves are simulated through an agent-based modelling approach. Aggregated household load-curves for different lifestyle groups are presented and the effect of their corresponding behavioral profiles on the residential load-curve is evaluated.

Integrated Urban-Energy Planning for the Redevelopment of the Berlin-Tegel Airport

In order to achieve their sustainable targets, cities are today looking for better solutions for integrating infrastructure systems into their urban planning. A large variety of tools exists for decision support both in energy planning and in city planning, but few of them combine detailed multi-energy modelling and a user-centered collaborative development process in the early phases of an urban project. With the opening of the Berlin Brandenburg Airport, the Berlin-Tegel Airport (Berlin TXL) will be redeveloped as an innovative hub for cutting-edge research and industry under the umbrella of Berlin TXL – The Urban Tech Republic (UTR). The European Institute for Energy Research (EIFER), the energy provider Electricite de France (EDF), the Department of Urban and Regional Planning (ISR) of the University of Technology Berlin (TU Berlin) and the Drees and Sommer Advanced Building Technologies company for energy design started in 2014 a collaboration with Tegel Projekt GmbH, the agency in charge of the development of the site, in order to unify urban and energy planning for the redevelopment of Berlin TXL. Based on an innovative modelling approach coupling both spatial and multi-energy systems, they developed a simulation prototype illustrating the interrelation between different technologies, land uses, and planning decisions. Several collaborative workshops were conducted as TU Urban_Labs led by the TU Berlin in order to integrate relevant actors into the planning process. This paper describes the integrative and collaborative approach developed by the participants to answer the needs and questions of Tegel Projekt GmbH regarding the energy planning of the future redevelopment of Berlin TXL according to the spatial setting.