VOLTAGE CONTROL IN DISTRIBUTION SYSTEMS AS A LIMITATION OF THE HOSTING CAPACITY FOR DISTRIBUTED ENERGY RESOURCES

VOLTAGE CONTROL IN DISTRIBUTION SYSTEMS AS A LIMITATION OF THE HOSTING CAPACITY FOR DISTRIBUTED ENERGY RESOURCES C. Schwaegerl*, M.H.J. Bollen, K. Karoui #, A. Yagmur + *Siemens AG, # Tractebel STRI AB + Tubitak PTD SE NC Engineering Germany Sweden Belgium Turkey Christine.Schwaegerl@siemens.com ABSTRACT The hosting capacity for distributed energy resources (DER) identifies the acceptable degree of DER penetration under given circumstances. It depends on various parameters such as the characteristics of the generation units, the configuration and operation of the network, the requirements of the loads as well as national and regional requirements. To determine the hosting capacity and to find ways to expand it by overcoming existing barriers is one subject of the ongoing European research project EU- DEEP. This paper describes the impacts of non-controlled DER units on the resulting voltage profiles as these impacts showed to be a main limitation for the hosting capacity for DER in existing distribution networks. The methodology for the analysis is introduced and basic results of the case studies performed are demonstrated. Generally the hosting capacity for DER can be extended applying different technical measures all leading to additional investments into the network. In future active distributions networks will provide a cost-effective solution to increase DER penetration. IMPACT OF DER ON VOLTAGE PROFILE Voltage control in existing distribution systems is based on radial power flows from the substation to the loads and on the fact that real power magnitude decreases with increasing distance from the substation. But, the introduction of DER in distribution systems impacts the methods of voltage control. DER may change the direction of power flows. This may lead to a reduced performance of the voltage control. It remains required that the system operates with adequate security and within statutory limits,. despite the fact that the Distribution System Operator (DSO) does not exert control over the generation.. Depending on the spatial relationship of loads and generation units, on the type of DER, its active and reactive power output, feeder parameters and loading power flow and therefore voltage can increase or decrease along a feeder. At periods of high generation output and low load, the upper statutory voltage level may be violated, while during periods of low generation and high load problems with the lower statutory limit may be reached. The variations in power output of some types of distributed generation will further lead to an increase in voltage variations. HOSTING CAPACITY FOR DER The hosting capacity for DER identifies the acceptable degree of DER penetration under given circumstances. It is possible to integrate DER units into existing networks as long as its performance is acceptable. The performance depends on different parameters such as the characteristics of the generation units, the operation of the network and requirements of the loads. As soon as the performance drops below the acceptable limit the hosting capacity is exceeded (Figure 1). Performance indices may vary between different countries or even between different DSOs depending on network configuration and operation as well as on the requirements of the loads connected. performance index hosting capacity Figure 1. Definition of the hosting capacity limit penetration level (% DER) If the performance index is not impacted by increasing DER penetration the hosting capacity of the network is unlimited. There are various issues that define the hosting capacity for DER like reliability, power quality, short circuit contribution and fault levels, or transient stability aspects, partly with

strong dependencies between different impacts. Actually figure 1 should be a multi-dimensional graph considering all dependencies. But, constraints for the maximum hosting capacity are set as soon as one aspect of the performance index exceeds acceptable limits. Keeping the voltage within acceptable or statutory bandwidths is a major limitation for the introduction of higher amounts of DER into existing networks. The concept of hosting capacity has been developed for cases in which DER integration potentially leads to a reduction of the system performance, mainly to identify barriers for DER integration. There are however situations in which DER may actually bring advantages to system performance, i.e. alleviate heavy loading in distribution systems or improve system reliability. To determine the hosting capacity for all considerable dependencies for both steady state and dynamic behaviour of DER and to find ways to expand it overcoming existing barriers is one subject of the ongoing European research project EU-DEEP (The birth of a EUropean Distributed EnErgy Partnership that will help the large-scale implementation of distributed energy resources in Europe [1]) that is funded by the European Commission under the 6 th framework programme for research and development. Within this project a partnership of 39 utilities, manufacturers, research organisations, professionals, national agencies and a bank, implements with a 29 million budget a demand-pull rather than technology-push approach for massive deployment of DER in Europe. Final results are scheduled for end 2008. PERFORMANCE INDICES To quantify the impact of DER units on voltage suitable performance indices and limits have to be defined. It is possible to use overall performance indices as well as indices dedicated to specific customers or to a specific phenomenon that is affected by DER. Further performance indices can be of general nature or express small deviations from an ideal or nominal value thus leading to values starting from 0 % or % increasing or decreasing with increasing DER penetration. In any case hosting capacity is reached if a given limit is exceeded. The main source for indices concerning steady state operation is the European voltage characteristics standard EN 50160. A more general approach is to study voltage variations in three different time scales: The percentile of the 10-minute rms voltage averages during one week as in EN50160; Voltage flicker as in IEC 60-4-15; Voltage variations in the timescale between 3 seconds and 10 minutes [4]. These indices identifying small deviations from an ideal or nominal value correspond to normal operation conditions. Additionally, there are dynamic events in the system like voltage dips, interruptions and large-scale blackouts that also have to be covered, but are out of scope of this paper. DEFINITION OF DER PENETRATION The DER penetration level can be expressed as an absolute value (for example the total installed capacity of all DER units, the number of DER units, the total generated power or energy of all units) or as a relative value. The latter (referred to as % DER) is more convenient when comparing different systems as it is, but will introduce some additional methodological questions. To make the results independent of the system size normalization is needed. The absolute value of the power (rated power or generated power) is divided by a reference value. Commonly used definitions are: Ratio of rated DER power to the local fault level (the short-circuit capacity of the system): The disadvantage of this normalization is that the hosting capacity will be a small value. A load or generation size equal to 10% of the fault level will already cause a very large impact of the system. Further many phenomena depend more on actual power flow than on rating, which especially for renewable sources like wind and sun deviates significantly from the actual power output. Two systems with the same relative penetration level may show a completely different performance. Ratio of rated DER power and transformer rating: For a given voltage level the rating of the HV/MV or MV/LV transformer is related to the fault level, so that this normalization is equivalent to the first definition. Ratio of Active Power (power generated by DER and the total power of the local load): When the ratio is equal to %, all loads are supplied by DER. For studies covering a longer period it must be defined which values are taken for both the numerator and the denominator: minimum, maximum or mean values; each decision leading to different penetration levels. CASE STUDIES To assess the impact of DER power flow simulations are carried out for worst-case situations. Assuming a DER operation at maximum generation capacity at minimum load leads to largest reverse power flows with greatest local voltage change, thus identifying possible violations of acceptable limits. To give general recommendations it was decided to perform a thorough analysis of the effects of DER on voltage control in distribution networks depending on DER penetration level, DER generation technology (inverter, synchronous machines, or induction machines) (see [2] for detailed effects), DER generation characteristics (i.e. constant power output for combined-heat-and-power units, intermittent for solar and wind plants), DER spatial location,

network characteristics (load level, feeder design, regional requirements, ), national and regional requirements. Today, DER units do not actively contribute to system control. Therefore only passive distribution networks were considered for the simulations that have been performed in a first step. Several rural, urban and industrial test networks, typical for different European countries, with different sizes and characteristics, have been selected or have been artificially created to perform the simulations. Figure 2 shows the basic structure of an example test network supplying different kind of customers that are equally distributed in typical LV networks with underground cables and overhead lines. Additionally, a distinction was made between MV and LV networks as they have different structures and components; DG resources connected to MV or LV networks have different characteristics in term of reactive power requirements and voltage regulation. 500 m cable, urban network or 5 km ov erhead line, rural network 500 m R=0,156 Ohm/km X=01226 Ohm/km C=235 nf/km 20 kv Sn = 2000 kva Sn = 630 kva Sn = 630 kva Sn = 250 kva 0,4 kv net 1 net 2 net 3 net 4 Figure 2. Example of a test network SIMULATION RESULTS number and type of loads, total installed capacity of the feeder: industry 1200 kw 4 different loads business 800 kw 150 different loads household 400 kw 200 different loads agriculture kw 5 different loads Two examples are presented here to demonstrate the basic results. Example test network As long as there is no active control of distribution networks the injection of active power leads to a voltage rise for customers connected to the distribution feeder. The rise is increasing with increasing feeder length and line impedance. As long as the maximum generation is smaller than the minimum load there is no feedback to higher-level networks. Mostly, there are also no limitations within the feeder considering 10-minute average values. The penetration level can be very high in a certain part of the system, but still rather small in the system as a whole. In the network example presented in figure 2 limitations arise in net 4. Mainly at the end of the feeder (node 6) the voltage exceeds the limits. Additionally, the hosting capacity is significantly smaller when DER units are located towards the end of the feeder (figure 3 and 4). As the location of most DER units depends on the availability of the resource and the territory of the owner that mostly is not the DSO there is only a limited choice for the placement of the sites that can be geographically and electrically remote from load centres. 110 % 85 1 2 3 4 5 6 nodes along the feeder DER kw, cosphi = 1, node2 DER kw, cosphi = 0.9, node2 DER kw, cosphi = 0.9, node4 DER kw, cosphi = 0.9, node6 DER kw, cosphi = 1, node6 Figure 3. Voltage along line in net4, maximum load 115 % 110 1 2 3 4 5 6 nodes along the feeder DER kw, cosphi = 1, node2 DER kw, cosphi = 0.9, node2 DER kw, cosphi = 0.9, node4 DER kw, cosphi = 0.9, node6 DER kw, cosphi = 1, node6 Figure 4. Voltage along line in net4, minimum load It has been assumed in the study that the load is constant for a sufficiently long time to reach a steady-state situation so that the voltage at the MV side of the HV/MV transformer equals the nominal voltage. The tap-changer control of this transformer is thus set to the nominal voltage. No line-drop compensation is used. In terms of EN-50160 the voltages resulting from the calculations can be interpreted as the 10-minute rms voltages. If we assume a 106% upper limit for the %-value, the hosting capacity is exceeded for all loading situations for node 6, cos φ = 0.9. For a number of other cases (node 6, cos φ = 1.0 node 4, cos φ = 0.9) the hosting capacity is exceeded only part of the time. Examples of the load shape in case of a working day in winter are presented in figure 5 for the feeder end in net 4. The voltage raise is even higher if the DER units are also feeding in reactive power.

115 110 % 85 00:00 06:00 12:00 18:00 time 00:00 DER kw, cosphi = 0.9, node6 DER kw, cosphi = 1, node6 DER kw, cosphi = 0.9, node4 Figure 5. Voltage shape, working day in winter, net 4, node 6 Voltage control by on-load tap changers may even worse the situation as they increase the secondary voltage in case of low feeder load. Reverse power flows increase the voltage estimation error at target position and lead to a even higher rise in voltage at DER location at dead-end feeders. Turkish distribution network The Turkish distribution system tested here includes one transformer 12.5MVA, 66/10 kv and three radial cables (figure 6). Normal operation of the distribution system is in radial mode and connections at node F with radial feeders 2 and 3 are normally open. The total load of the system is 3.69 MW, 1.76 MVAr; all loads consuming constant power. External Grid A/66kV B/10 kv TRF41470 LATERAL 1 LATERAL 2 LATERAL 3 Etiler 101 Levent 101 Line C_D C/10 kv Line H_K H/10 kv Line M_N M/10 kv Etiler 102 Levent 102 Line D_E D/10 kv Line K_L K/10 kv Line N_P N/10 kv Etiler 103 Levent 103 E/10 kv L/10 kv P/10 kv F/10 kv Voltage 1 0,9 0,99 0,985 0,98 Without DER Syn 0.5 MW Syn1.5 MW Syn 2 MW A C D E F H K L M N P B Bus Name Figure 7. Voltage profile of buses with different DER power output Dynamic Voltage Stability Without DER loads are supplied by the transformer from the grid. In case of an out of service of the line between nodes B and M dynamic simulations showed an immediate voltage drop at node N from 98.4 % to 97.6 %. With the synchronous generator (output power is 1.5 MW), which is equipped with exciter and governor control system, connected to node N the voltage at node N recovers within seconds after the line between node B and M is opened (figure 8). At the same time the connection between node P and F was closed to supply also the other nodes. Generally, DER showed positive effects on voltage stability raising the voltage and reducing voltage dips. 1.000 0.99 Base Case (Without DG) 0.984 p.u. 1.000 0.99 1.5 MW DG 0.9 p.u. 1.5 MW DG 0.988 p.u. Maslak 101 Maslak 102 Maslak 103 0.98 Base Case (Without DG) 0.976 p.u. 0.98 Figure 6. Turkish distribution system Steady State Voltage Stability A synchronous generator with nominal power 4.85 MVA is connected to node N and simulated with 0.5 MW, 1.5 MW, 2 MW power output at power factor 0.98 leading with DER penetration level of 15 %, 40 % and 55 % of total load, respectively. The power output of the generator raises the voltages compared to the base case. The voltage rise is higher due to reactive power support (figure 7). The voltage levels of the nodes in the distribution system also increase with increasing DER penetration level or DER power output power. For higher penetration levels, an overvoltage might occur. 0.97 0.97 1s 10s 20s 1s 10s 20s Figure 8. Dynamic voltage at bus N without and with DER GENERAL OUTCOMES Even for one phenomenon the hosting capacity is not a fixed value: it depends on various local parameters like structure and degree of automation of the network, type, generation characteristic and spatial location of the DER unit, kind and behaviour of load, climate parameters (in case of wind or solar power), different regional requirements. As it is not possible to quantify the impact of some dependencies to allow comparisons between different national and regional systems values for the hosting capacity should

be handled with care. The hosting capacity may by increased by technical means, including: Reconfiguration or upgrading of the network (upgrading conductors and transformer, installation of new lines or substations), Connection of DER units at higher voltage levels or closer to the feeder, Reduction of primary substation voltage, Operation of generator at leading power factor, Temporarily constraining of DER generation according to network loading conditions. In any case the solutions imply additional costs in term of capital investment and ongoing operating costs. Vice versa, it is always possible to increase the hosting capacity with additional investments (figure 9 and figure 10). For that reason, the integration of DG into distribution systems becomes an economic issue rather than a technical one. Who is paying for the measures is an important further discussion, but is out of the scope of this paper. As it is generally much more cost-effective to implement monitoring, communications and control facilities than it is to reinforce the network to accommodate the same level of DER penetration active management methods will be the future way to overcome barriers for DER integration and will therefore erase all limits for the hosting capacity. FUTURE RESEARCH Preliminary results have been shown in this paper. Analysis and simulations will continue considering more active distribution networks. Further there are strong interacts of DER penetration with operation, control, or stability. All these topics will also be handled within the EU-DEEP project. Publicly accessible reports on the various topics will be available for download at www.eu-deep.org. ACKNOWLEDGEMENT This work is funded by the European Commission as part of the EU-DEEP project under grant SES6-CT-2003-503516. REFERENCES performance index increasing additional investments limit penetration level (% DER) Figure 9. Performance index for increasing investments [1] www.eu-deep.org [2] V. Knayazkin, T. Ackermann, 2003, "Interaction between the distributed generation and the distribution network: operation, control and stability aspects", CIRED 2003, paper 4-40 [3] T. Bopp, A. Shafiu, I. Cobelo, I. Chilvers, N. Jenkins, G. Strbac, H. Li, P. Crossley, 2003, "Commercial and technical integration of distributed generation into distribution networks", CIRED 2003, paper R4-45. [4] M.H.J. Bollen, M. Häger, C. Schwaegerl, 2005, "Quantifying voltage variations on a time scale between 3 seconds and 10 minutes", this conference. hosting capacity additional investments Figure 10: Hosting capacity as a function of the investment. ACTIVE DISTRIBUTION NETWORKS AS A MEANS TO INCREASE DER PENETRATION Active management techniques enable the DSO to maximise the use of the existing circuits by taking full advantage of generator dispatch, control of transformer taps and any reactive compensation devices in an integrated manner [3]. It provides real time network monitoring and control at key network nodes combined with state estimation and real time modelling for critical issues. Coordinated voltage control communicates with generator controls (i.e. controlling power factor of synchronous generators and inverters), loads and network devices, such as reactive compensators (preferably at the point of DER connection to the network), voltage regulators, and on-load tap changers.