A Multi Variable Optial Energy Manageent Strategy for Standalone D Microgrids Morthala Praod M.Tech (EPS) Shahjehan ollege of Engineering And Technology, hevella. Abstract: Due to substantial generation and deand fluctuations in standalone green icrogrids, energy anageent strategies are becoing essential for the power sharing and voltage regulation purposes. The classical energy anageent strategies eploy the axiu power point tracking (MPPT) algoriths and rely on batteries in case of possible excess or deficit of energy. In order to realize constant currentconstant voltage (IU) charging regie and increase the life span of batteries, energy anageent strategies require being ore flexible with the power curtailent feature. In this paper, a coordinated and ultivariable energy anageent strategy is proposed that eploys a wind turbine and a photovoltaic array of a standalone D icrogrid as controllable generators by adjusting the pitch angle and the switching duty cycles. EXITING ONEPT: The stability of a dc icrogrid is easured in ters of the stability of its dc bus voltage level which is one of the ain control objectives. The grid voltage source converters (G-VSs) are the priary slack terinals to regulate the voltage level of grid-connected icrogrids. Battery banks, on the other hand, are effective slack terinals for standalone icrogrids their energy absorbing capacities are liited regarding a nuber of operational constraints. Proposed concept: The proposed strategy is developed as an online nonlinear odel predictive control (NMP) algorith. Applying to a saple standalone dc icrogrid, the developed controller realizes the IU regie for charging the battery bank. The variable load deands are also shared accurately between generators in proportion to their ratings. The D bus voltage is regulated within a predefined range, as a design paraeter. INTRODUTION: The icrogrid ay operate as an extension of the ain grid, (grid-connected) or as a standalone grid with no connection to the grid. Standalone dc icrogrids have soe distinct applications autootive or arine industries, rural areas. Since ac systes suffer fro the need of synchronization of several generators, dc icrogrids are ore efficient due to the fact that dc generators and storages do not need ac-dc converters for being connected to dc icrogrids. The few issues regarding voltage regulation, power sharing, and battery anageent, is severe in standalone green icro grids that consist of only interittent solar and wind energy sources, and lead to the necessity of ore controllable strategies. The grid voltage source converters are the priary slack terinals to regulate the voltage level of grid-connected icrogrids. Battery banks, on the other hand, are effective slack terinals for standalone icrogrids. Their energy absorbing capacities are liited regarding a nuber of operational constraints, as explained later in this section. In order to regulate the voltage level of standalone dc icro grids. Present load shedding strategies for the cases in which there is insufficient power generation or energy storage. Present strategies that curtail the renewable power generations of standalone dc icro grids if the battery bank cannot absorb the excess generation. These curtailent strategies restrict the batteries charging Page 100
rate by the axiu absorbing power.standalone dc icrogrids are usually located in sall-scale areas where the power sharing between DGs can be anaged by centralized algoriths which are less affected by two issues: a) Batteries in charging ode are nonlinear loads causing distortions to the grid voltage b) The absolute voltage level of a standalone icrogrid is shifted as the result of the load deand variation. A nuber of phenoena affect the batteries operation during the charging ode [19]: 1) Applying high charging currents, the batteries voltages quickly reach to the gassing threshold. 2) The internal resistor and hence power losses and theral effects increase at high SO levels. 3) Batteries cannot be fully charged with a constant high charging current. Operational constraint, the axiu absorbed power by the batteries in order to protect the fro being overcharged. Therefore batteries act as nonlinear loads during the charging ode. Depending on the proportion of the power generation to the load deand ratio within standalone D icrogrids, three cases are possible. 1) Power generation and load deand are balanced; 2) Load deand exceeds power generation causes dc bus voltage to drop in absence of any load shedding; and 3) Power generation is higher than load deand leads batteries to be overcharged and bus voltage to clib. Energy anageent strategy (EMS) is proposed, as its control objectives, three aforeentioned issues corresponding standalone dc icro grids; i.e., dc bus voltage regulation, proportional power sharing, and battery anageent. In contrast to the strategies available in literature in which renewable energy systes (RESs) always operate in their MPPT ode, the proposed ultivariable strategy uses a wind turbine and a PV array as controllable generators and curtails their generations if it is necessary. The proposed EMS is developed as an online novel NMP strategy that continuously solves an optial control proble (OP) and finds the optiu values of the pitch angle and three switching duty cycles. It siultaneously controls four variables of icro grids: 1) Power coefficient of the wind turbine. 2) Angular velocity of the wind generator. 3) Operating voltage of the PV array and 4) harging current of the battery bank. It is shown that, eploying new available nonlinear optiization techniques and tools, the coputational tie to solve the resulting NMP strategy is in perissible range. The proposed strategy ipleents the IU charging regie that helps to increase the batteries life span. Block diagra: Fig. 1.Topology of a sall-scale and standalone dc icrogrid. DESRIPTIONS: The standalone dc icrogrid in above figure is a sallscale icrogrid for reote applications. The wind turbine operates at variable speeds and is connected to the electrical generator directly, that is, the direct-drive coupling. The variable speed operation is ore flexible for the power anageent and MPPT applications. Furtherore, direct-drive coupling is ore efficient Page 101
and reliable and is ore popular for sall-scale wind turbines. In spite of high cost, peranent agnet synchronous generators (PMSGs) are the ost doinant type of direct-drive generators in the arket, chiefly due to higher efficiency. MODELING: Modeling of the Three Syste: A. Wind Branch. B. Battery Branch.. Solar Branch. A. Wind Branch Perforance of the wind turbines is easured as the power coefficient curve with respect to the tip speed ratio and pitch angle. Equation (3) shows the power coefficient curve of three-blade wind turbines: Fig2. Modified version of the syste odel The authors in [20] presented a atheatical odel of standalone green dc icrogrids as hybrid differential algebraic equations (hybrid DAEs). Fig. 2 suarizes a odified version of the proposed odel in [20]. Since this paper focuses on the case in which there is an excess power greater than or equal to the axiu possible absorbing rate of the battery bank, the hybrid nature of the battery bank operation is ignored for the sake of siplicity. The differential and algebraic states, i.e., and, and the anipulated and nonanipulated control variables, naely, and, are detailed later throughout the next sub-sections. In what follows, the following notations are used to odel the standalone dc icrogrid in Fig. 1 as DAEs: Where gaa and bita, respectively, are the tip speed ratio and pitch angle. Rad is the radius of the blades and p,ax is the axiu achievable power coefficient at the optiu tip speed ratio of gaa out. Equation (4) presents the connected PMSG generator: The first two constraints f1 and f2 are due to the fact that in standalone dc icrogrids the su of the generated, stored, and consued powers is always zero: Energy anageent strategies of icrogrids ust estiate the dc bus voltage level deviation fro its set point in about every 5 10 s. It eans that except the angular velocity of the generator (4a) all other fast voltage and current dynaics can be ignored. For energy anageent strategies, the average odel of the buck converter is replaced with the steady-state equations for the continuous conduction ode (M) Page 102
wheredw is the switching duty cycle of the converter and all reaining paraeters are as depicted in Fig. 1. The average dc output voltage of the rectifier, Vwt, in presence of the non-instantaneous current coutation is calculated as follows.. Solar Branch: The equivalent electrical circuit of the PV odule [27], [28] is used to atheatically odel the solar branch, consisting of a PV array and a boost converter [29]. Eq. (9) shows the characteristic equations of a PV array, consisting Npvp*Npvs of PV odules: B. Battery Branch: harging operation of a lead acid battery bank, consisting of (Nbatp*Nbats) batteries, is odeled as: Where Iph denotes the photocurrent and I0 is the diode reverse saturation current. Rs and Rsh, respectively, are the series and parallel equivalent resistors of each PV odule and all other paraeters are as follows: the voltage, current, and state of charge of the battery bank. If is the filtered value of the battery current with the tie constant of Ts and Qact is the actual battery capacity. The experiental paraeter P1 requires being identified for each type of battery while the axiu aount of the battery capacity, ax, internal resistor of battery, Rbat, and the battery constant voltage, V0, are given by anufacturers. By ignoring the discharging ode of the battery bank operation, the bi-directional converter acts as a boosttype converter [(8d) (8e)]. Siilar tothe wind branch,the average odel ofthe boost converter is replaced with the steady-state equations for M Page 103
ONTROLLER DESIGN: Optial ontrol Probles (OPs) OPs, as (11), ake explicit use of the syste odel, given by (11b), in order to find an optial control law u*(.), which eets nuber of equality and inequality constraints. The ter optial here is defined with respect to a certain criterion that iplies the control objectives. This criterion is specified with a cost functional J, consisting of the Lagrangian ter labda and the terinal cost ter M. While the Lagrangian ter indicates the cost function during the period of tie, the terinal cost penalizes final values. Equations (11d) and (11e), respectively, forulate the final and initial constraints which ust be aintained by the optial solution. Moreover, (11g) represents boxing constraints on the states and control variables: 1) Dynaic prograing ethod based on the Bellan's optiality principle. 2) Indirect ethod based on the Pontryagin iniu principle. 3) Direct ethods that convert OPs into nonlinear optiization probles (NLPs) which are then solved by NLP solvers. ontrol Syste: Since it focuses on the charging ode of the battery operation, The proposed EMS successively gets the estiated syste states,, as inputs and calculates the optial solution,, as outputs. The external state estiator and the predictor of the non-anipulated variables are out of the scope of this paper. step ahead predictions of the solar irradiance, wind speeds, and load deands are extracted either fro a eteorological center or an external predictor using autoregressive-oving-average (ARMA) technique [37]. The bus voltage level of the icrogrid,, is set externally and hence the developed controller can act as the secondary and priary levels of the hierarchical architecture [13]. The developed NMP controller consists of three entities: 1) The dynaic optiizer that successively solves OP at each sapling tie h, Nonlinear Model Predictive ontrol (NMP): OPs are open-loop strategies and are wrapped by a feedback loop to construct NMP strategies [30]. NMP strategies, which are also called as the receding horizon control, continuously solve an OP over a finite-horizon T using the easureents obtained at t as the initial values. Then the first optial value is applied as the next control signal. oparing with the conventional ethods, NMPs are inherently nonlinear and ultivariable strategies that handle constraints and delays. There are three different techniques to discretize and solve OPs: 2) The atheatical odel of the syste to predict its behavior 3) The cost function and constraints of the relevant OP. The optial pitch angle is applied as a set point to an inner closed-loop controller. Moreover, the optial values of the switching duty cycles are applied to the pulse width odulators (PWMs) of the dc-dc converters. 1) ontrol Objectives: Three aforeentioned control objectives, i.e., dc bus voltage regulation, proportional power sharing, and ipleenting the IU regie to charge batteries, are Page 104
A B g E g g E E - forulated by two slack variables in (12) and (13) and the cost function in (14) TABLE II:WIND TURBINE, PMSG, BATTERY STAK, AND PV PARAMETERS IN THIS STUDY EXPERIMENTAL RESULTS: Siulation circuit design in Matlab: Discrete, Ts = 1e-06 s. powergui 1.2 Generator speed (pu) 3 Generator speed (pu) Pitch angle (deg) T (pu) 1 z Pulse Generator NOT Logical Operator [S1] Goto [S2] Goto1 - v V1 44.65 Vdc 1 z Pulse Generator1 NOT [S3] Goto2 [S4] Product Vpv Ppv Pitch angle Wind speed ( /s) Wind Turbine [S1] Fro g [S4] Fro3 g Logical Operator1 Goto3 PVA 12 E E Vpv PMSG S1 L1 S4 L2 Ipv T A B a A b B c BWF A B - DBR [S2] Fro1 S2 1 B R Rload [S3] S3 Fro2 2) Box onstraints: Equation (15) adds the pitch angle control feature to the developed EMS in order to liit the produced aerodynaic power by the wind turbine: Three-Phase Parallel RL Load Vabc_WF Vabc Mag_V_I Scope4 Fro6 Iabc_WF Iabc P_Q Fro7 Discrete 3-phase Positive-Sequence Active & Reactive Power Scope1 L3 [S6] Fro5 S6 Pulse Generator2 [S5] Fro4 1 z NOT Logical Operator2 g E S5 2 Product1 [S6] Goto4 [S5] Goto5 Battery _ <SO (%)> <urrent (A)> <Voltage (V)> Pbat Bat The other box constraints on the anipulated variables and the syste states are forulated as follows: TABLE I: DESIGN PARAMETERS AND THE OMPUTATIONAL TIME OF THE DEVELOPED NMP ONTROLLER Page 105
In order to address these objectives, the developed EMS siultaneously controls the pitch angle of the wind turbine and the switching duty cycles of three dcdc converters. It has been shown that the developed controller tracks the MPPs of the wind and solar branches within the noral conditions and curtails their generations during the underload conditions. REFERENES: [1] J. M. Guerrero, M. handorkar, T. Lee, and P.. Loh, Advanced ontrol Architectures for Intelligent Microgrids-Part I: Decentralized and Hierarchical ontrol, IEEE Trans. Ind. Electron., vol. 60, no. 4, pp. 1254 1262, 2013. [2] R. S. Balog, W. W. Weaver, and P. T. Krein, The load as an energy asset in a distributed D sartgrid architecture, IEEE Trans. Sart Grid, vol. 3, no. 1, pp. 253 260, 2012. [3] J. M. Guerrero, P.. Loh, T. L. Lee, and M. handorkar, Advanced ontrol Architectures for Intelligent Microgrids-Part II: Power quality, energy storage, and A/D icrogrids, IEEE Trans. Ind. Electron., vol. 60, no. 4, pp. 1263 1270, 2013. [4] N. Eghtedarpour and E. Farjah, ontrol strategy for distributed integration of photovoltaic and energy storage systes in D icro-grids, Renew. Energy, vol. 45, no. 0, pp. 96 110, 2012. Fig shown: Output wave fors of the siulation ONLUSION AND FUTURE WORKS: In this paper, we developed a novel optial EMS that anages the energy flows across a standalone green dc icrogrid, consisting of the wind, solar, and battery branches. A coordinated and ultivariable online NMP strategy has been developed to address, as the optial EMS, three ain control objectives of standalone dc icrogrids. These objectives are the voltage level regulation, proportional power sharing, and battery anageent. [5] D. hen and L. Xu, Autonoous D voltage control of a D icrogrid with ultiple slack terinals, IEEE Trans. Power Syst., vol. 27, no. 4, pp. 1897 1905, Nov. 2012. [6] L. Xu and D. hen, ontrol and operation of a D icrogrid with variable generation and energy storage, IEEE Trans. Power Del., vol. 26, no. 4, pp. 2513 2522, Oct. 2011. [7] S. Anand, B. G. Fernandes, and M. Guerrero, Distributed control to ensure proportional load sharing and iprove voltage regulation in low-voltage Page 106
D icrogrids, IEEE Trans. Power Electro., vol. 28, no. 4, pp. 1900 1913, 2013. [8] B. Zhao, X. Zhang, J. hen,. Wang, and L. Guo, Operation optiization of standalone icrogrids considering lifetie characteristics of battery energy storage syste, IEEE Trans. Sustain.Energy, to be published. [9] T. Zhou and B. Francois, Energy anageent and power control of a hybrid active wind generator for distributed power generation and grid integration, IEEE Trans. Ind. Electron., vol. 58, no. 1, pp. 95 104, 2011. [10] X. Liu, P. Wang, and P.. Loh, A hybrid A/D icrogrid and its coordination control, IEEE Trans. Sart Grid, vol. 2, no. 2, pp. 278 286, 2011. [11] H. Kanchev, D. Lu, F. olas, V. Lazarov, and B. Francois, Energy anageent and operational planning of a icrogrid with a PV-based active gen. for sart grid applications, IEEE Trans. Ind. Electron., vol. 58, no. 10, pp. 4583 4592, 2011. [12] H. Ghoddai, M. B. Delghavi, and A. Yazdani, An integrated windphotovoltaic-battery syste with reduced power-electronic interface and fast control for grid-tied and off-grid applications, Renew. Energy, vol. 45, no. 0, pp. 128 137, 2012. [13] J. M. Guerrero, J.. Vasquez, J. Matas, L. G. de Vicua, and M. astilla, Hierarchical control of droopcontrolled A and D icrogrids-a general approach toward standardization, IEEE Trans. Ind. Electron., vol. 58, no. 1, pp. 158 172, 2011. [14] P. H. Divshali, A. Aliardani, S. H. Hosseinian, and M. Abedi, Decentralized cooperative control strategy of icrosources for stabilizing autonoous VS-Based icrogrids, IEEE Trans. Power Syst., vol. 27, no. 4, pp. 1949 1959, Nov. 2012. Page 107