Power Control in Isolated Microgrids with Renewable Distributed Energy Sources and Battey Banks

2nd International Conference on Renewable Energy Research and Applications Madrid, Spain, 20-23 October 2013 Power Control in Isolated Microgrids with Renewable Distributed Energy Sources and Battey Banks José Gomes de Matos, Luiz Antonio de Souza Ribeiro Evandro de Carvalho Gomes Institute of Electrical Energy Federal University of Maranhão São Luís, Brazil gomes@dee.ufma.br, luiz_ribeiro@dee.ufma.br Federal Institute of Education, Science, and Technology of Maranhão São Luís, Brazil evandrogomes@ifma.edu.br Abstract This paper presents a new strategy to control the power generation from existing energy sources in autonomous and isolated Microgrids. In this study, the Microgrid is composed of a power electronic converter, supplied by a battery bank, which is used to form the AC network (grid-forming power converter), an energy source based on a wind turbine and its respective power electronic converter (grid supplier converter), and the loads. The primary subject of the proposed control strategy is to keep the energy balance into the Microgrid, in order to control the battery bank state of charge even when more power can be generated than loads can consume. This goal is achieved controlling the generated power into the Microgrid, without using dump load or any physical communication with the power electronic converters or individual energy source controls. The electrical frequency of the Microgrid is used for dictating the amount of power the energy sources need to generate in order to maintain the battery-bank state of charge below its maximum permissible value. A modified droop control to implement this work is proposed. Keywords isolated microgrids; power control; state of charge; battery banks; parallel inverters; renewable energy sources; I. INTRODUCTION Worldwide, a significant number of villages still have no access to electricity due to its remoteness. Fortunately, in many villages and other places, such as in oceanic islands, there are renewable energy sources, especially solar radiation and wind. These energy resources can be used to form isolated Microgrids to meet local energy needs, as reported in [1]-[3]. The intermittent nature of renewable energy sources (RES), the required operation availability, and the standalone operation of these systems need some kind of energy storage system (SES). Although there are several different SES, in practical applications battery banks have been used mainly due to economical reasons. A diesel generator set (DGS) is generally necessary as a backup source to supply power when renewable energy resources are limited and the battery bank state of charge (SOC) is in a critical situation. There are several types of topologies which have been used to assemble these isolated Microgrids. The connection of all energy sources with a concentrated DC bus, including the battery banks, is one possibility. The resultant DC voltage needs to be converted to AC voltage to form the local power distributed network. In this configuration, an AC/DC converter (controlled rectifier) is necessary as an interface between the DGS and the DC bus if the DGS is required to supply the loads and also charge the battery bank. This is a disadvantage of this topology, since this AC/DC converter increases the system cost and makes its operation and maintenance more complicated. Another possibility of isolated Microgrid topology is connecting all the existing power sources to a concentrated AC bus, instead of a common DC bus. In that case, at least one bidirectional power electronic converter, supplied by a battery bank, is necessary to form the AC bus and the local power distribution AC network. The DGS, when existing, can be synchronized with the AC bus using a conventional power unit control, not requiring an AC/DC converter. The topologies with concentrated or centralized generation, previously described, are not the most appropriate configurations for communities with consumers or energy sources spatially distributed. For these situations, an improvement can be obtained if the AC bus is distributed rather than concentrated, as illustrated in Fig. 1. The GFC is a grid-forming converter, GSC is a grid supplier converter, that is, a converter connected to a primary energy source, WT is a wind turbine, PV is a photovoltaic panel, BB is a battery bank, and. is the line impedance between the and the buses. This topology allows connecting the power sources to the grid anywhere next to the generation area, reducing the costs with cabling to connect the generators with a distant and centralized AC bus. Figure 1. Microgrid with the power sources distributed along an AC bus. 978-1-4799-1464-7/13/$31.00 2013 IEEE

A crucial issue related to these autonomous and distributed Microgrids is to control the power internally generated to maintain the system energy balance in order to keep the terminal voltage of the battery bank limited to a safe value and consequently its state of charge (SOC) under control. This is not a simple task if the system has distributed energy sources located far from the battery bank. One possible solution to this drawback consists of using centralized or distributed resistive dump loads that burn the eventually excess power generated, as presented in [4]-[5]. This solution has the disadvantage of using additional components, generally with the same rated power of the renewable energy sources, which increases the system costs, reducing its useful lifetime, making its operation and maintenance more complicated. Another possibility is to use physical communication (wiring) among the converter control systems in order to control the energy balance into the Microgrid, and the amount of power that each of them must generate to keep this energy balance under control. Despite its simplicity, this solution has the disadvantage of reducing the system reliability since it is dependent on the operation of a physical system of communication. This paper proposes an alternative strategy to control the power generated from an isolated Microgrid with distributed RES. The proposal is controlling the terminal voltage of the existing battery banks below or equal to its maximum permissible value. This is done limiting the amount of power that each energy source can generate at each instant, using a modified droop control strategy. The battery banks terminal voltage control implies indirectly in their SOC control. This solution does not use neither dump loads to dissipate the excess power generated, nor wired communication with the energy sources and the energy storage systems. II. SYSTEM DESCRIPTION Fig. 2 illustrates the simplified diagram of a standalone Microgrid used to explain the control strategy proposed in this paper. It is composed of a grid-forming converter (GFC), a grid supplier converter (GSC), a battery bank, and the loads. The renewable energy source, in this particular study, is a variable speed wind turbine coupled to a permanent magnet synchronous generator (PMSG) [6]. The simplicity of this system is useful to show the feasibility of the proposed control strategy without loss of generality. The grid-forming converter is a bidirectional converter composed of a PWM inverter, and a DC-DC converter that works in a buck or boost mode. The grid supplier converter is used to control the power that is generated from the renewable energy source. In this particular example, the grid supplier converter is composed of a grid side PWM inverter (GSI) and a wind turbine side PWM inverter (TSI). A. Grid Voltage and Frequency Controls In the grid-forming converter, the PWM inverter controls voltage and frequency of the Microgrid. The grid voltage controller uses a traditional configuration implemented on a synchronous reference frame, synchronized with the grid voltage vector. The model of the LC filter in delta side of the transformer (see Fig. 2), implemented in the dq synchronous reference frame, is used to design the control loops of the grid-forming converter. The block diagram of this model is presented in Fig. 3, where is equivalent series resistance of the inductor. Based on this model, an inner current loop and an outer voltage loop were designed, as depicted in Fig. 4. In this figure is the frequency of the local grid, the superscript denotes synchronous reference frame, ^ denotes estimated quantity, is used to decouple the disturbance caused by the load currents and, and the cross coupling disturbance of and, which is introduced by the model, is the per phase equivalent filter capacitance which value is 3, and means Zero Figure 2. The simplified diagram of studied Microgrid.

Order Holder. The current on the inductance is controlled to regulate the voltage on the capacitance [7]. The current and voltage controllers were tuned for 1 khz and 200 Hz bandwidths, respectively. converter can be equivalent to connect a controlled voltage source, with average value, between the terminals of the converter circuit, as shown in the Fig. 5(a) and Fig. 5(b). If the losses in the converter are neglected, the voltage on will only depends on the difference between the power on the battery bank terminals ( ) and the power on the PWM inverter terminals ( ), as shown in Fig. 5(c). So, the dynamic equation of can be written as (1), where is a state variable defined as. 1 2 1 2 (1) From (1) and Fig. 5(c), the DC Bus voltage controller can be designed using an inner loop to control the battery bank current, and an outer loop to control the voltage level on the capacitor, as depicted in Fig. 6. The bandwidths for the inner and outer loop are 500 Hz and 50 Hz, respectively. Figure 3. Block diagram of the LC filter in the synchronous reference frame. Figure. 5. The DC-DC converter: a) original circuit, b) equivalent average circuit model, and c) DC bus average model. Figure 4. Block diagram of the microgrid voltage controller (GFC voltage controller). The grid frequency reference value is calculated using the power control strategy into the Microgrid and it will be explained later. The voltage reference value can be equal to the rated Microgrid voltage or it can be calculated as a function of the reactive power at the grid-forming converter terminals, as reported in [7], [8]. The reference frame is aligned with grid terminal voltage vector, thus the value of is imposed to be zero, while is equal to the grid magnitude voltage value previously explained. B. Bidiretional DC-DC Converter Control In the grid-forming converter, the DC-DC buck or boost converter is used to control the voltage in capacitor. The control action of the voltage controller in the DC-DC Figure 6. Block diagram of the bus voltage controller in the DC-DC converter. The controller output voltage ( ) is the reference value for the PWM block which is used to generate the control signals to and switches, as shown in Fig.7 [4]. If is positive, the battery bank will supply the load and the DC- DC converter operates on boost mode using switch and diode. On the other hand, the DC-DC converter operates on buck mode using switch and diode. C. Grid Supplier Converter Control In the grid supplier converter, the GSI is used to control the grid current and the DC link voltage (see Fig. 2). The controller is implemented on a synchronous reference frame, with an inner current loop to control the grid current, and an outer voltage loop to control the voltage on the capacitance C [7]. This controller is designed and implemented in a similar way as the controller depicted in Fig. 3. The inner and outer loop bandwidths were set to 500 Hz and 10 Hz, respectively.

The TSI controls the power generated, using an algorithm to extract the maximum power from the wind turbine (MPPT) [9]. The TSI controller is implemented on synchronous reference frame aligned with the rotor of the turbine generator. Based on the references adopted in this work, the d-axis current of the generator ( is controlled to be zero. Thus, in steady state, the generator torque is proportional to q-axis current ( ) component, as shown in (2) [10], where is given by (3), is the number of the generator poles, and is the generator permanent magnet flux. Figure 7. Block diagram of the voltage controller of DC bus in the grid former converter. The proposed control strategy uses the grid frequency value to characterize whether the battery bank SOC has the tendency for its maximum permissible value or not. Based on that information, the amount of power that can be generated from the primary energy sources connected with the system should be restricted. The grid frequency control is implemented in the grid-forming converter. When the battery bank SOC is not in its maximum limit, the grid frequency is determined according to the conventional droop control principle, described by line in Fig. 8 [7], [8]. In that case, there are no restrictions on the amount of power that can be generated, and the existing renewable energy sources operate on their maximum power point. Obviously, this is true only if the battery bank has been designed with sufficient capacity to absorb all power that the renewable sources can produce in a particular instant. If the battery bank SOC is in its maximum limit, the grid frequency is imposed to be always greater than the value, and the droop control will track the line shown in Fig. 10. In that case, restriction on the amount of power that can be generated from the renewable sources is necessary, otherwise the integrity of the battery bank is at risk. The amount of power that must be reduced from the maximum power that can be generated is proportional to the difference between and (see Fig. 8). (2) 3 2 2 (3) When the wind turbine works on its maximum power point, the mechanical torque ( ) is proportional to the square of the rotational speed ( ), given by (4), where is a constant that depends on the physical and operational characteristics of the wind turbine, and the air density. The value of may be obtained experimentally or by computational simulations based on mathematical models [11], [12]. From (3), (4), and, the reference value for -axis current controller of the generator ( ) is given by (5). The reference value for -axis current component ( ) is imposed to zero. The current controller bandwidths for axes are equal to 10 Hz. (4) (5) III. STRATEGY TO CONTROL THE GENERATED POWER AND BATTERY STATE OF CHARGE In standalone or isolated Microgrids there is no available commercial grid. Therefore, the power generated from distributed energy sources needs to be controlled when the load demand is less than the amount of power that could be generated from the energy sources. This control is necessary to keep the energy balance in the Microgrid and to keep the battery bank voltage less than or equal to its maximum permissible value. Figure 8. Frequency versus power in the grid-forming converter according to proposed power control strategy. The battery bank SOC is detected using battery voltage ( ). A PI controller is used to regulate the battery terminal voltage less than or equal to its maximum value V. The output of this controller is the incremental frequency ( ) that must be added to in order to get the new reference value of the grid frequency. The value is proportional to the amount of power that must be decreased from the power generated in order to control the battery bank terminal voltage. This is illustrated in Fig. 9, where is the slope of or lines in Fig. 8. The grid frequency is measured in the controller of grid supplier converter. If its value is great than, the battery bank SOC is close to its upper limit, the GSC controller decreases the current reference that has been calculated on the maximum power point tracker algorithm, reducing the power that could be generated from wind turbine, and controlling the

SOC and the battery bank terminal voltage. This is illustrated in Fig. 10, where is the amount of power that must be reduced, is the generator rotational speed, is the generator torque variation, and is the generated power. Figure 12. Experimental wave forms of voltage and current during the tests with a 30 Ah 12 V lead-acid battery: a) Currente and b) Voltage. Figure. 9. Block diagram of the frequency control in the grid former-forming power converter The variation in the grid frequency, shown as the output of the SOC controller in Fig. 9, means a decrease equal to in the overall power generated in the microgrid, as depicted in Fig. 10. Assuming that the load power does not change and the variation in the system losses can be neglected during the control action, can represent the amount of power that will not be delivered to the battery bank in order to keep its terminal voltage ( ) less than or equal to its maximum value ( ). A variation equal to in the battery input power means a variation equal to in battery current, which is approximately given by (6). (6) Figure. 10 Block diagram of the power control in the grid supplier converter The tuning of the SOC controller in Fig. 9 needs to take into account the dynamic of the battery bank. One possible model for lead-acid battery is shown in Fig. 11, where is the battery open circuit voltage, is the equivalent series internal resistance, and model the overvoltage or undervoltage arising from the battery charging or discharging, is the resistance due to natural losses, and models the battery capacity to storage energy [13]. Normally, the natural losses has a very high time constant, such that its effect can be neglected for the purpose of this work. The experimental voltage and current waveforms of a 30 Ah/12 V lead-acid battery during the tests to estimate the parameters of the battery equivalent circuit are presented in Fig.12. The estimated values of the battery parameters at the end of the charge cycle are 8.7 mω, 430 mω, 60 F, and 44.000 F. From these data it can be seen that the lowest time constant related to the battery dynamic behavior is approximately 25 seconds ( ). Figure 11. Battery equivalent circuit. Thus, the block diagram for analyses and tuning of the SOC controller can be depicted in Fig. 13, where is the transfer function /, and is the transfer function that relates the current and voltage at the battery terminals. Normally, is much faster than, so it can be equivalent to 1. The relationship between and is given by (7). (7) Figure 13. Block diagram for analyses and design of the controller of power genareted and battery bank state of chage.. IV. SIMULATION RESULTS Simulation results of the Microgrid shown in Fig. 1 are presented in Fig.14. The system rated power is 15 kw. The battery bank is compound of 60 lead-acid batteries of 12 V - 220 Ah, connected in 3 parallel lines, each line with 20 batteries in series. The maximum permissible terminal voltage of the battery bank is 280 V. The wind speed is defined equal to 11 m/s. The PI controller in Fig. 13 was tuned with a bandwidth of 0.036 Hz. The grid frequency range is 59.4 Hz up to 61.2 Hz and the SOC controller is activated when the frequency is greater than 60.6 Hz. The simulation was implemented using MATLAB/SIMULINK software. The converters are modeled as ideal controlled voltage sources. In

the beginning of the simulation, the battery bank voltage is equal to 275 V (see Fig. 14a). and the system is supplying a load of 10.9 kw (see Fig. 14e and 14f). At the instant 100 s, the wind turbine is turned on and produces 10.1 kw (see Fig. 14g). Thus, the battery is still supplying the difference between the power consumed by the load and the power produced from the wind turbine. At the instant 210 s, the load is turned off and all power produced from the wind turbine is used to charge the battery bank (see Fig. 14a). The battery terminal voltage increases quickly and gets 280 V at 310 s. At this time, the power controller starts to operate and the battery bank terminal voltage stops growing. The Fig. 14b presents a continuous voltage at the bidirectional converter terminals showing that it was perfectly controlled by the voltage controller depicted in Fig. 6. The Fig. 14c presents the frequency variation. From 0 s to 300 s, the frequency changes according to the conventional droop control law, and after 300 s, it changes according to the proposed control strategy. Figure 14. Simulation results: (a) the battery bank voltage; (b) the DC bus voltage of the GFC; (c) the grid frequency; (d) the generated power by the turbine; (e) the output power of the GFC; and (f) the power at the battery bank terminals. sources. Simulation results have demonstrated the effectiveness of the strategy. This strategy need neither wired communication with the distributed renewable sources nor surplus power dissipation into dump loads. These technical advantages make the proposed strategy a promising tool to increase the viability and reliability of renewable power generation system in isolated and remote communities. Although a wind turbine has been used to demonstrate the validity of the proposed strategy, it is also valid regardless the type and number of power sources existing in the isolated Microgrid. The proposed strategy calculates the amount of power that must be generated from all sources, at each time, to keep the balance of energy in the grid. REFERENCES [1] R. E. Foster, R. C. Orozco, and A. R. P, Rubio, Lessons llearned from the Xcalak Hybrid System: a Seven Year Retrospective, in Proc. of 1999 Solar World Conference, International Solar Energy Society, vol. I, pp. 319-328, Jerusalem, Jul. 1999. [2] L. A. S. Ribeiro, O. R. Saavedra, J. G. Matos, S. L. Lima and, G. Bonan, and A. S. Martins, Isolated Micro-Grid With Renewable Hybrid Generation: The case of Lençóis Island, IEEE Transactions on Sustainable Energy, vol. 2, no. 1, pp. 1-11, Jan. 2010. [3] L. A. S. Ribeiro, O. R. Saavedra, J. G. Matos, S. L. Lima and, and G. Bonan, Making Isolated Renewable Energy System More Reliable, Renewable Energy, vol. 45, pp. 212-231, 2012. [4] N. Mendis, K. M. Muttaqi, and S. Pereira, A Novel Control Strategy for Stand-alone Operation of a Wind Dominated RAPS System, in IEEE 2011 Industry Applications Annual Meeting, pp. 1-8, 2011. [5] M. Singh and A. Chandra, Control of PMSG Based Speed Wind- Battery Hybrid System in an Isolated Network, in IEEE 2009 Power & Energy Society General Meeting, pp. 1-6, 2009. [6] H. Li and Z. Chen, Overview of different wind generator systems and their comparison, IET Renewable Power Generation, vol. 2, no. 2, pp. 123-138, Jun. 2008. [7] J. Rocabert, J. A. Luna, F. Blaabjerg, and P. Rodrígues, Control of Power Converters in AC Microgrids, IEEE Transactions on Power Electronics, vol. 27, no. 11, pp. 4734-4749, Nov. 2012. [8] K. de Brabandere, B. Bolsens, J. Van Keybus, A. Woyte, J. Driesen and, R. Belmans, A Voltage and Frequency Droop Control Method for Parallel Inverters, IEEE Trans. on Power Electronics, vol. 22, no. 4, pp. 1107 1115, July 2007. [9] Z. Chen, J. M. Guerrero, and F. Blaabjerg, A Review of the State of Art of Power Electronics for Wind Turbines, IEEE Transactions on Power Electronics, vol. 24, no. 8, pp. 1859-1875, Aug. 2009. [10] D. W. Novotny and T. A. Lipo, Vector Control and Dynamics of AC Drives, Clarendon Press, Oxford, 1996. [11] S. Morimoto, H. Nakayama, M. Sanada, and Y. Takeda, Sensorless Output Maximization Control for Variable-Speed Wind Generation System Using IPMSG, IEEE Transactions on Industry Applications, vol. 41, no. 1, pp. 60-67, Jan./Feb. 2005. [12] H. Siegfried Heier, Grid Integration of Wind Energy Conversion Systems, John Wiley & Sons, Chichester, England, 2009. [13] Z. M. Salameh, M. A. Casacca, and W. A. Lynch, A Mathematical Model for Lead-Acid Batteries Parallel, IEEE Transaction on Energy Conversion, vol. 7, no. 1, pp. 93-98, March 1992 V. CONCLUSIONS This paper presented a strategy to control the power generated in order to regulate the battery bank state of charge in standalone Microgrids with distributed renewable energy