Generation Capacity Design for a Microgrid for Measurable Power Quality Indexes

1 Generation Capacity Design for a Microgrid for Measurable Power Quality Indexes Q. Fu, A. Solanki, L. F. Montoya, Student Members, IEEE, A. Nasiri, Senior Member, IEEE, V. Bhavaraju, Senior Member, IEEE, T. Abdallah, and D. Yu, Senior Member, IEEE Abstract Microgrids are receiving a lot of attention to utilize distributed generations in a sub-system and provide higher efficiency and reliability and support local loads. A high renewable-energy penetrated microgrid is studied in this paper. The distribution system and the loads in the microgrid are represented by a properly scaled 12kV IEEE 34 bus system. The central power is replaced with a 1.875MVA base-load diesel generator and a 250kW solar PV and a 500kW wind generations are added at nodes 848 and 890, respectively. The major loads in this microgrid are then modeled with a load profile. Special variable kw and kvar loads were modeled to include load demand variations. The wind and solar PV plants are modeled with wind and solar power profiles. The microgrid system including the source variations and demand variations is simulated using PSCAD software. The microgrid is monitored at number of buses and the power quality issues are measured and indexes are calculated. This system can be used to determine the capacity requirements for the non-renewable generators to maintain the power quality indexes. Index Terms, Diesel generator, Microgrid, power quality, renewable energy. T I. INTRODUCTION HE global electrical energy demand is growing gradually. Within 20 years, it is expected that the demand will be doubled [1]. Moreover, due to price volatility, limited supply, and environmental concerns of fossil fuels, wind and solar PV power generations are rapidly growing as alternative energy sources in many parts of the world [2]. According to the American Wind Energy Association (AWEA), wind energy is now the largest new source for electricity production. Installed wind energy capacity in the U.S. was at 40,180MW by the end of 2010. PV industry is also experiencing a large growth. Installed grid-tie solar PV capacity reached 2150MW at the end of 2010 [11]. The installed PV capacity in 2010 was eight times of the capacity in 2006. Since deregulation of electrical energy system has been lowering the investment in large This work was supported in part by the U.S. Army Corps of Engineers (ERDC/CERL) under Contract No. W9132T-11-C-0022. Q. Fu, A. Solanki, L. F. Montoya, A. Nasiri, and D. Yu are with University of Wisconsin-Milwaukee, Milwaukee, WI 53201, USA (e-mail: nasiri@uwm.edu). V. Bhavaraju is with Eaton Corporation Innovation Center, Milwaukee, WI 53216, USA (e-mail: VijayBhavaraju@eaton.com). T. Abdallah is with the US Army, Champaign, IL, (e-mail: tarek.abdallah@erdc.usace.army.mil). 978-1-4577-2159-5/12/$31.00 2011 IEEE power plants, the need for new electrical power sources could be very high in the near future. Although renewable energy systems have many benefits, their utilization does not come without cost. The higher penetration of renewable energy systems such as wind and solar PV has introduced many technical and non-technical issues, including power quality, reliability, safety and protection, load management, grid interconnections and controls, new regulations, and grid operation economics [3-5]. Renewable energy systems and other Distributed Generations (DG) can be better utilized in a microgrid concept. Microgrid is a cluster of DGs that are placed in a system to provide power with higher reliability and quality to the local loads [6]. In addition, they have significantly higher energy efficiency compared with utility grid since utility grids suffers from inefficient generation, transmission, and distribution systems. Most microgrids are designed to be connected to the utility grid. In case of grid power outage, they isolate themselves from the grid via a static switch. The local loads, voltage and frequency then need to be managed by local generations. In this paper, the performance of a microgrid in islanded mode is studied and presented. In a microgrid, the sources and customers are within the microgrid and some of the sources could be renewable sources such as wind and solar. In the presence of renewable sources, there is an opportunity to reduce the fossil fuel consumption within a microgrid. However, renewable power sources are intermittent and can cause power reliability or quality concerns [7]. Providing the right size of diesel or natural gas generation along with renewables is addressed in this paper. In addition, the power reliability and voltage quality of the system is analyzed. Utilities have been using the Power reliability indexes namely System Average Interruption Duration Index (SAIDI), System Average Interruption Frequency Index (SAIFI), and Customer Average Interruption Duration Index (CAIDI) to evaluate the reliability of power provided to their customers. An analysis of SAIDI and SAIFI is provided in this paper for the microgrid system by monitoring the voltage at key load locations. II. MICROGRID CONFIGURATION In order to accurately study the behavior of the renewable energy systems and diesel generator and their effects on the voltage and frequency in a microgrid, a standard IEEE 34 bus system is adopted in this paper [8-9]. Figure 1 shows the configuration of the microgrid. The original system is a 60Hz,

2 250kW Solar PV 1.875MVA Diesel Gen 400kW 24.9kV, 12MVA with different types of loads connected to a utility main at bus 800. The load types include constant active/reactive power loads and constant impedance loads, in forms of three-phase, single-phase, and two-phase. In order to match the properties of the system with a microgrid under construction, the nominal voltage of the system is changed to 12kV and other components of the system including loads and line impedances have been scaled accordingly. The base parameters of the system are changed to 12kV, 6MVA. The transformer on bus 832 is scaled down to 12kV/4.16kV and the two voltage regulators at bus 832 and 814 are also scaled to 6.9282 kv, phase voltage. The power ratings of the fixed PQ loads are reduced to half of their original values. The same also applies to the single-phase PQ loads. To scale the constant impedance loads, their impedances are reduced to half. Since the voltage is also half of the original value, their power rating is reduced to half. There are two types of the transmission lines in this system namely, lumped line impedance and distributed line impedance. Below is the description on scaling these lines. a. Lumped Line Impedance To keep the same voltage drop, the line impedances should be halved since the line carries the same current as described by (1) and (2). P / 2 I = V / 2 Vdrop / 2 V % = V / 2 Figure 1. The configuration of the microgrid studied in this paper in islanded mode. I( Rline / 2) drop = V / 2 (2) When the voltage level is scaled down from 24.9kV to 12kV, the distance between phase to phase and phase to earth is reduced, which means that the equivalent capacitance increases. In this case, we have reduced resistances and inductances to half and doubled the capacitance for each lumped line impedance. Figure 2 shows an example for the scaling of line 302 connected to bus 822. (1) (a) (b) Figure 2. Scaling lumped line impedance for line 302; a) with the original impedance, and (b) with the revised impedance. b. Distributed Line Impedance The case for distributed line impedance is different. We have considered three ways to change the line impedances, when scaling from 24.9kV to 12kV system: (i) halving the R/L matrix, (ii) halving the length of lines, and (iii) halving the length of line and quadrupling the capacitance matrix. Methods (i) and (ii) yield similar results but the voltage drop is larger than the original case. Method (iii) cuts the line power flow in half and at the same time keeps the nodal voltages in per unit the same. Therefore, we have used method (iii) to scale the distributed line impedances. Figure 3 shows the original case and all the three methods for the line. The number 10 and 20 in the figure shows the length of the line. Case (d) provide similar per unit voltage as case (a) while halving active and reactive power.

3 Figure 3. Scaling transmission lines with distributed line impedance, (a) original case, (b) Impedance reduced in half, (c) line length is reduced to half, and (d) line length is reduced to half and capacitance is quadrupled. c. Power Sources After scaling the microgrid, three types of power sources are added including, 250kW solar PV, 400kW wind turbine, and 1.875MVA diesel generator. The study in this paper is focused on the islanded mode of operation. Therefore, the grid connection at bus 800 is replaced with a 1.875MVA diesel generator. Solar PV and wind turbines are modeled in PSCAD in current mode. They need a reference voltage from the diesel generator to provide power. III. SYSTEM POWER PROFILE The power data for load, wind and solar PV are actual measured data of existing systems, which are scaled for the microgrid in this paper. The system includes a total of 52 loads. They include fixed and variable PQ loads and fixed impedance loads. The load data for a single load on bus 848 and total microgrid load are presented in Figure 4. The peak load occurs around 7PM and it is 1721kW. The minimum load occurs at 3AM and it is 1438kW. The solar PV system is modeled using solar irradiation data from Solar Advisor Module (SAM) for the city of Milwaukee, WI. The inverter is modeled as a current source connected to the micro grid or grid. The PV power model contains a 24 hours insulation profile for the summer of 2002. Maximum Power Point Tracking (MPPT) for the panels was developed and simulated using PSCAD software. Figure 5 shows the output power profile of the 1MW system. A control method is developed to curtail the PV power when the total renewable generation is more than the total load demand. In this case, the diesel power is at minimum and it only establishes the voltage reference for the microgrid. The wind turbine power profile is also modeled using measured wind speed data near city of Milwaukee WI. The turbine is modeled using PSCAD software considering the turbine efficiency factor (C P ) and the mechanical and electrical efficiencies. The model is current source similar to the solar PV model. Figure 6 shows the power profile for a 400kW turbine in a 24 hour period. Power (kw) 250 200 150 100 50 Load on Bus 848 (kw) 0 0 4 8 12 16 20 24 Time (Hr) (a) (b) Figure 4. Load power profile for a single day, (a) single load on bus 848 and (b) total load for the microgrid.

There is one diesel generator in our microgrid system. The properties of this 1.875MVA machine are provided in Table 1. Table 1. Parameters of the 1.875MVA diesel generator. 4 Figure 5. The power profile for a 1MW solar PV plant. Figure 6. The power profile for a 400kW wind turbine. IV. DIESEL GENERATOR MODELING Diesel generator plays a very important role in Microgrid. It controls the voltage and frequency in the microgrid in islanded mode. Other sources use it as a reference for frequency. Whenever a load is applied to or removed from the microgrid, the voltage and frequency experience a transient before settling at the steady state values. The magnitude and duration of this transient depends on the generator exciter and engine governor controls. During sudden changes in the load, the diesel generator must be able to maintain the voltage and frequency within the limits. The same is also true when there is sudden change in the renewable energy generations. Figure 7 shows the basic block diagram of a diesel generator connected to a grid or microgrid. The exciter is in Figure 7. Basic block diagram of a diesel generator connected to a gird/microgrid. charge of regulating output voltage and the governor adjusts engine speed, which is translated to output frequency. Rated RMS Line to Neutral Voltage 6.9282 [KV] Rated RMS Line Current 0.09 [KA] Base Angular Frequency 376.991118 [rad/sec] Armature Time Constant [Ta] 0.332 [pu] Poitier Reactance [Xp] 0.011 [pu] D: Unsaturated Reactance[Xd] 0.13 [pu] D: Unsaturated Transient Reactance[Xd ] 0.03 [pu] D: Unsaturated Transient Reactance 5.2 [s] Time(open)[Td0 ] D: Unsaturated Sub Transient Reactance[Xd ] 0.022 [pu] D: Unsaturated Sub Transient Reactance 0.029 [s] Time(open)[Td0 ] Q: Unsaturated Reactance[Xq] 0.510 [pu] Q: Unsaturated Transient Reactance[Xq ] 0.228 [pu] In order to accurately study the behavior of the synchronous machine for the power system stability studies, it is essential that the excitation system of the machine is modeled with sufficient details. The desired model must be suitable for representing the actual excitation equipment performance for large, severe disturbances as well as for small perturbations. IEEE Standard 421.5 recommends three distinctive types of excitation systems including DC type excitation systems, AC type excitation systems, and static type excitation systems. Due to fairly small size of the machines in this paper, we have chosen AC8B: Alternator Supplied Rectifier Exciter with Digital Control #2 exciter type. Figure 8 shows the block diagram of this exciter. Figure 8. Block diagram of the AC8B AC Exciter. Where, V S is the terminal voltage transducer and load compensation elements [pu], V REF is the voltage regulator reference (determined to satisfy initial conditions) [pu], V C is the combined power system stabilizer and possibly discontinuous control output after any limits or switching, as summed with terminal voltage and reference signals [pu], and E FD is the Exciter output voltage [pu]. Figure 9 shows the simulations results for the diesel generator when load steps occur at 10 and 20 seconds. The machine starts at no load. At 10s, a 100% step load with a power factor of 0.8 is applied. The speed and terminal voltage of the machine drop as indicated. At 20s, the load is removed and the voltage and frequency spike. It should be noted that these results are only for the diesel generator and a load. Results for the IEEE 34 bus configuration is discussed in the next sections.

diesel cannot meet its demand, load shedding must be applied to balance the system. 5 Figure 9. Simulation results for the diesel generator when load steps are applied at 10s and 20s; from top, electrical frequency (Hz), terminal voltage (pu), and rotor speed (pu). V. GENERATION CAPACITY REQUIREMENT A first order approach is used in this paper to size the generation capacities. The minimum power demand of the system, as shown in Figure 4, is 1438kW. The diesel generator rating is selected at this minimum load for 1.875MVA considering 0.8 power factor. The rest of the power demand should be met by the renewable energy sources. Wind speed and solar irradiation pattern near city of Milwaukee has been used to calculate the capacity factors for both wind and solar PV systems. The capacity factors for location and device specific data have been calculated at 0.29 for solar PV and 0.34 for wind turbines. Considering an average energy cost of $210/MWh for solar PV and $90/MWh for wind energy [12], required capacities for PV and wind energy are determined to meet the peak load demand. Capacities of 250kW for solar PV and 400kW for wind turbine are calculated. Figure 10 shows the power profile of the wind turbine and solar PV system for the proposed system. The wind power varies considerably during a 24-hour period. The generations of wind and solar PV is deducted from the total demand to calculate the demand for the diesel generator. Figure 11 shows the total demand and rating for the diesel generator. When the Figure 11. The total demand and rating of the diesel generator considering the renewable generation. VI. POWER QUALITY ASSESSMENT The IEEE 34 bus system shown in Figure 1 is simulated in PSCAD with 52 loads and three generations. Average models for wind and solar PV, and detailed diesel generator model without droop control are implemented as discussed in previous sections to reduce the computation time. The voltage at load terminal and system frequency drops when the generation cannot meet the load demand. Programable breakers are placed at each load to perform the load shedding when the voltage drops below 0.92 per unit. The breaker is reclosed when the voltage climbes to 0.98 per unit. One second delay has been considered before opening or closing the breakers to filter transients. Wind and solar PV are considered as energy sources. That means that the system absorbs all the energy produced by these sources as indicated in Figure 10. The remainder of the demand is supplied by the diesel generator as shown in Figure 12. During early evening, the load demand is at maximum while the solar PV generation is down to zero and the average wind energy production is reduced. The diesel generator runs at maximum power but still cannot meet the load demand. Although the exciter regulates the voltage at the diesel generator terminal, as shown in Figure 12, but the voltage at other buses experience voltage dips. Figure 14 shows the voltage at bus 822. This is the worst case voltage since this bus is located at the end of line. When the load voltage drops under 0.92, the load breaker opens and removes that load. This happens frequently before 8AM and Figure 10. The 24-hour power profile for 400kW wind turbine and 250kW solar PV. Figure 12. The active and rective power delivered by the diesel generator.

6 Figure 13. Voltage waveform at diesel generator terminal. VII. CONCLUSIONS Generation capacity sizing and power quality evaluations for a microgrid in islanded mode have been presented in this paper. Standard IEEE 34 bus system is used as a microgrid and diesel, solar PV and wind generations have been added to the system. Average models for the sources have been implemented to run the whole system using PSCAD software for 24 hours. The voltages at different nodes have been monitored to perform load shedding when the load voltage falls under 0.92 per unit. Quality indexes for the system have been calculated and presented. The models provided in this paper can be used to properly size the renewable generations to reach at certain power quality indexes. VIII. ACKNOWLEDGMENT This material is based upon work supported by the U.S. Army Corps of Engineers (ERDC/CERL) under Contract No. W9132T-11-C-0022. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the U.S. Army Corps of Engineers. Figure 14. The voltage at bus 822 of the IEEE 34 bus system. around 8PM. SAIDI and SAIFI are the parameters used by the utility companies to evaluate the power quality and reliability. SAIDI is the average outage duration for each customer served. SAIDI is measured in units of time, often minutes. It is usually measured over the course of a year, and the median value for North American utilities is approximately 1.50 hours. It is described as follows. Sum of all customer interruption SAIDI = (3) Number of customers served SAIFI is the average number of interruptions that a customer would experience. SAIFI is measured in units of interruptions per customer. It is usually measured over the course of a year, and the median value for North American utilities is approximately 1.10 interruptions per customer. Total number of customer interruptions SAIFI = (4) Total number of customers served In our system, by assuming this 24 hours load and power profile as the average daily data in a year, the total duration for all interruption is 2022 hours per year. SAIDI is calculated as 2022/1100 =1.8383 hours. There are totally 56940 interruptions for 1100 loads during the year. SAIFI is calculated as 56940/1100 =51.76. The values calculated for SAIDI and SAIFI are larger than average utility grid numbers since (i) the diesel is only rated for the minimum load value and there is high penetration of intermittent renewable energy, (ii) reactive power management is not performed to regulate the bus voltages, and (ii) no energy storage device is added to support the renewable energy systems. REFERENCES [1] F. Blaabjerg, F. Iov, R. Teodorescu, Z.Chen, Power Electronics in Renewable Energy Systems Power Electronics and Motion Control Conference, pp. 1 17, 2006. [2] Behera, R.K. Parida, S.K. Wenzhong Gao, Design of a Conditioner for smoothing wind turbine output power Electrical Machines and Systems, pp. 2390-2395, 2008. [3] T. Degner, J. Schmid, and P. Strauss, Distributed Generation with High Penetration of Renewable Energy Sources, Final Public Report of Dispower Project, ISBN 3-00-016584-3, March 2006. [4] Renewable Energy 2000: Issues and Trends, Energy Information Administration Office of Coal, Nuclear, Electric and Alternate Fuels, February 2001. [5] O. Anaya-Lara, Tutorial: Transmission Integration of Wind Power Systems: Issues and Solutions Modeling and Control of Wind Generation Systems, 2nd International Conference on Integration of Renewable and Distributed Energy Resources, Napa, CA. [6] J. Huang, C. Jiang, and R. Xu, "A review on distributed energy resources and microgrid". Renewable and Sustainable Energy Reviews, vol. 12, no. 9, pp. 2472-2483, 2008. [7] N. Saito, T. Niimura, K. Koyanagi, R. Yokoyama, "Trade-off analysis of autonomous microgrid sizing with PV, diesel, and battery storage," IEEE Power & Energy Society General Meeting, 2009, Tokyo, Japan. [8] R.C. Dugan, W. H. Kersting, "Induction machine test case for the 34-bus test feeder description," IEEE Power Engineering Society General Meeting, 2006, Montreal, Quebec. [9] Samaan, N.; McDermott, T.; Zavadil, B.; Li, J.; "Induction machine test case for the 34-bus test feeder - steady state and dynamic solutions" IEEE Power Engineering Society General Meeting, 2006, Montreal, Quebec. [10] M. Saejia and I. Ngamrooa, "Stabilization of microgrid with intermittent renewable energy sources by SMES with optimal coil size," in press, Elsevier publishing. [11] Larry Sherwood, U.S. Solar Market Trends, 2010, Integrated Renewable Energy Council, June 2011. [12] Energy Information Administration, Annual Energy Outlook 2011, http://www.eia.doe.gov/oiaf/aeo/index.html.