Lukas Exel

Component-Oriented Modeling of Thermoelectric Devices for Energy System Design

Thermoelectric (TE) devices are used in the form of Peltier coolers and as TE generators, with the latter producing electrical energy from waste heat, based on the Seebeck effect. In both cases, modeling of the TE device is a prerequisite for the design and control verification of the resulting overall energy system. To this end, the model has to be integrated seamlessly in an overall system model containing other electrical, thermodynamic, or even mechanical components. Following this premise, this paper presents a component-based model for TE devices described in the Modelica language. The model incorporates the temperature dependences of decisive material properties (Seebeck coefficient, thermal conductivity, and electrical resistivity) in 1-D spatial resolution. With the help of few additional geometrical parameters, e.g., the thickness of TE legs, the model is capable of describing the dynamic behavior of the TE device in accordance with the experimental results.

Re-use of existing simulation models for DCS engineering via the Functional Mock-up Interface

The Functional Mock-up Interface (FMI) is the result of a research program initiated by Daimler AG, further industrial companies, and several institutes, with the aim of enabling the exchange of dynamic simulation models between different simulation tools as well as their co-simulation. The upcoming release of the second version of this standard in 2014 is the motivation for the current investigation into possible uses in the process industry. The standard is therefore presented and discussed in general, but with focus on the needs of the process industry. This article also illustrates how a Functional Mock-up Unit (FMU) can be coupled to widely-used simulation tools and presents a concept for its coupling and implementation. The concept uses the open and well-defined Shared-Memory-Gateway provided by the SIMIT Simulation Framework.

Component-oriented ORC plant modeling for efficient system design and profitability prediction

This contribution presents the modeling of compact Organic Rankine Cycle (ORC) power plants intended to generate electrical energy from waste heat. The component-oriented models are illustrated as a valuable means in system design and efficiency/profitability prediction of a potential product in consideration of varying boundary conditions. First, we demonstrate the modeling of two crucial components, one in a bottom-up manner and another one in a top-down manner, as well as the effective construction of a closed-loop system model. Secondly, all physical component models are systematically extended by their respective economic quantities (costs and profit-relevant energy production and consumption). On the system level, these quantities are generically merged with all legal subsidies and compensation that are applicable to the considered plant (with the German renewable energy source act serving as an example). Finally, using the integrative physico-economic models in dynamic system simulations, we show efficiency and profitability predictions of an exemplary ORC plant in two scenarios.

Modelica® Library for Dynamic Simulation of Thermoelectric Generators

The contribution presents a new modeling library for the dynamic simulation of thermoelectric generators (TEG) in 1D spatial resolution. The core of the library is a model of the thermoelectric legs (TEL), which has already been published by the authors. In the submitted work, this model is expanded to an overall Modelica® library for complete TEGs. The library is open source and can be extended. It is also usable by end users without deeper knowledge through a graphical user interface (GUI). The use of the library is illustrated by the example of an electronic thermostat valve powered by a TEG.

Multi-domain modeling of distributed energy systems — The MOCES approach

We present a holistic approach of modeling complex energy systems (MOCES). It follows the need for multi-domain models and appropriate tools for the design and evaluation of new concepts and ideas related to the upcoming Smart Grid. As an adequate approach, MOCES is capable of modeling physical systems as well as the mainly event-driven behavior of interacting entities and feedback loops across the domains. In addition, it covers timescales from seconds to days. Contrary to other approaches following an agent or co-simulation concept, MOCES is monolithic, based on the modeling language Modelica. To meet the requirements imposed on a simulation tool, we use the possibility to enhance Modelica within its language specification. The basic idea of MOCES is to extend the physical model of each entity by a connected representative in a parallel simulation. To show the feasibility of the concept, a simulation example investigates the benefits and drawbacks of local energy markets.

Design of distributed energy systems: Role and requirements of modeling and simulation

The interconnection of distributed generation units has a significant impact on the performance of an energy system. In this paper, challenges in designing grid-connected and islanded distributed energy systems (DES) are described and a concept of simulation models in the component-oriented modeling language Modelica® is presented, wherein simulation models are divided into physically detailed and physically abstracted models.

Modeling and simulation of local flexibilities and their effect to the entire power system

The transformation of the energy system requires new methods for its modeling and simulation, as established methods are facing problems in representing the increasing complexity and flexibility. We present the modeling and simulation framework MOCES that is based on the modeling language Modelica. The novelty of MOCES is the explicit modeling of the processes that ensure the balancing of production and consumption and which link the physical system with the energy markets. In this contribution, we show how MOCES can be used to model and simulate the German power system with detailed spatial representation by adapting and extending the ELMOD-DE model. The adapted model is modified to investigate the effect of local storage systems that aim at minimizing the exchange of electricity with the upstream grid. We provide findings on their effect to the grid usage, the grid stability, and the economic performance of the storage systems. In addition, we examine the influence of an increasing number of local storage systems on their economic performance.

Efficient Models of Partially Shaded PV Modules for Energy System Design

To simulate the power output of a photovoltaic module (PM) under partial shading conditions each cell has to be simulated individually. The following contribution will present a method how the simulation effort can be reduced with little or even no loss in simulation accuracy under typical shading conditions. This reduction is required if the PM is part of a heterogeneous distributed energy system which shall be efficiently analyzed on system level. The core idea is a grouping model approach which maps each strand of a PM on only three, two or one grouping cell(s) whose respective parameters are scaled by a factor. These parameters, as well as the irradiation values given to the individual grouping cells(s), are determined dynamically by a pure algebraic clustering algorithm during simulation runtime. The proposed method is implemented and demonstrated in Modelica/Dymola.

Toward a decentralized forecast system for distributed power generation

A general forecasting concept for power systems is developed based on a detailed requirement analysis concerning a forecasting system for supply-dependent feed-in units within the present and future electrical grid. The fundamental idea is an agent-based system in which each single supply entity determines and communicates its own feed-in forecast. The contribution describes the proposed architecture and defines a corresponding data structure which includes information about the uncertainty of the individual forecast. To demonstrate the feasibility, the design of the forecast agent at entity level is presented in more detail and discussed in relation to a solar energy system.