18 th Power Systems Computation Conference Wroclaw, Poland – August 18-22, 2014 An Improved Modeling for Microturbines and Fuel Cells to the Energy Management Problem of Microgrids Daniel Tenfen Prof. Erlon C. Finardi Victor S. S. Fernandez Thomas Ober Universidade Federal de Santa Catarina Florianópolis, Brazil Grenoble INP - Ense3 Grenoble, France tenfendaniel@gmail.com erlon.finardi@ufsc.br sfernandezvictor@gmail.com thomas.ober@etu.grenoble-inp.fr Abstract— This paper presents a mathematical model that presents improvements for the gas microturbines (MT) and fuel cells (FC) in the energy management problem of microgrids. In this problem, the objective is to determine a generation policy that minimizes, over a planning horizon, an objective function subject to economical and technical constraints. We propose a detail modeling for MTs and FCs, where the constraints associated with the ramps, minimum up and downtime, generation limits, etc., take into account some peculiarities that have not been adequately considered in literature. As a highlight, we include in MT modeling the effect of ambient temperature in the operating cost and maximum output generation. To analyze the influence of the proposed modeling, it is used a microgrid with a MT, a FC cell, a battery, wind and photovoltaic generators connected to the grid. The results indicate that the proposed modeling issues have significant influence over the generation policy, especially in relation to the optimal operating cost and the amount of reserve necessary to deal with the uncertainties associated with intermittent generation and load. Keywords—Microturbines; fuel cells; energy management; microgrids I. INTRODUCTION The modern electrical energy industry is dealing with more affordable electronic technologies. The integration of small Distributed Energy Resources (DERs), such as Microturbines (MTs), Fuel Cells (FCs), batteries, wind and photovoltaic generators is a trend that is currently in progress. The presence of DERs can reduce fossil fuel consumption and technical losses, as well as postpone investments in new transmission and distribution lines [1]. In this new environment, it is important to highlight the Microgrids (MGs), which are emerging as an additional element to maintain the growth and sustainability of the modern electric energy industry. Roughly speaking, a MG consists of a group of DERs and loads that operates synchronized with the main grid, although it can be disconnected and operates autonomously. Despite several advantages of MGs, it is also important to notice that new challenges are inherent, which are associated with the existence of DERs with different operating characteristics in relation to large hydro and thermal generation resources [2]-[3]. In this context, a methodological challenge with fundamental importance to the economic and technical support of a MG is the Energy Management (EM) problem. In general, this problem aims to determine a generation policy that minimizes, over a planning horizon, an objective function subject to economical and technical constraints. The generation policy is given by the on/off status, and the respective output active power, of each DER. This policy is used as a reference for the voltage and frequency control in MG real-time operation. Therefore, since it is necessary to minimize an objective function subject to constraints, the EM is usually performed based on the solution of an optimization problem. Consequently, it is essential a precise DERs modeling to achieve a high-quality generation policy. In general, MTs and FCs can be modeled similarly to the large thermal units, i.e., it is necessary to represent maximum rates of power variation between two consecutive stages, minimum time that a unit must be on or off, output generation limits, etc.[4]. However, since they have different technologies of large thermal units, FCs and MTs possess peculiarities in their modeling that has not been adequately considered in the literature. For instance, a typical MT (a few dozens of kW), may vary the all power range in few dozens of seconds. Consequently, it is not necessary to represent operating ramp rate (also known as ramp-down and ramp-up rate) in the EM problem. However, the start-up ramp rate, which depends on auxiliary power source until the MT achieve the nominal temperature, must be accurately represented because this process may takes few minutes 1 . On the other hand, the shutdown ramp rate also has a peculiar feature, because initially the rotation of a MT is decreased to a specified value to cool down the unit, and during this time, the power decreases linearly before finishing shutdown cycle. Another particularity of MT modeling, which is not presented in literature, concerns with the variation of the maximum power and the electrical efficiency as a function of the ambient temperature. Regarding FC, this kind of DER has the same modeling requirements aforementioned, exception of the ambient temperature influence and the shutdown ramp rate. In this work, we are interested in Solid Oxide FC (SOFC), which works at high temperature and requires constant power consumption during a few hours until reaching the temperature 1 Due to intermittent generation, it is crucial to discretize the planning horizon in steps of one or two minutes. Thus, any operational characteristic of DERs that takes more time than this must be precisely modeled in the EM problem. This project had the financial support from CPFL Paulista and RGE R&D Program, regulated by ANEEL, PD-0063-0026/2011 Project - PB0026 Estudo dos Impactos da Inserção de Microrredes e Microgeração em Sistemas de Distribuição - and Conselho Nacional de Desenvolvimento Cientıfico e Tecnológico (CNPq).