MEASUREMENT CAMPAIGN AND ASSESSMENT OF THE QUALITY OF SUPPLY IN RES AND DG FACILITIES IN SPAIN

Transcription

1 MEASUREMENT CAMPAIGN AND ASSESSMENT OF THE QUALITY OF SUPPLY IN RES AND DG FACILITIES IN SPAIN Eugenio PEREA*, Eduardo ZABALA*, J. Emilio RODRÍGUEZ*, Asier GIL DE MURO*, Hugo GAGO * * * Fundación LABEIN, Energy Department, Spain ** IBERDROLA Distribución Eléctrica, Customer Support Department, Spain perea@labein.es INTRODUCTION The water fall has the following characteristics: net jumping is 59 m and water stream is 750 l/s. The objective of the DGFACTS project [1], funded by the European Commission, is the application of power electronic systems to improve the Quality of Supply (QoS) of the electrical network, in a high Distributed Energy Resources (DER) penetration scenario. Its main result will consist of the development of a set of power electronic equipment to improve the different aspects covered by the wide electrical QoS concept. In-situ measurements in selected sites of networks with connected Distributed Energy Sources have been performed, in order to assess the existing Power Quality (PQ) levels in partner countries. This paper analyse the results of the PQ measurements performed in a small hydro-plant site in Dima, Spain. Simulations of the small hydro-plant have also been performed in order to validate the models in the steady state and dynamic conditions. The results of simulations obtained from the models of the small hydro-plant have been compared with real measurement data, concretely with the active and reactive power measurements averaged each minute (steady-state condition validation) and with a voltage sag measurement (dynamic condition validation), 30% deep and less than 0.1 s long. Obtained results are good. This work gives a reasonable model of the real DER to be used in bulk simulations. MEASUREMENT SITE DESCRIPTION In the frame of the PQ measurements promoted by the DGFACTS project, Fundación LABEIN and IBERDROLA carried out measurements in two sites, of Spain, during one month in each site: more concretely, in a wind farm of Higueruela (Albacete) and in a small hydro-plant of Dima (Basque Country). Figure 1: Water stream (top), turbine and generator (bottom) The generator is connected in parallel to a capacitor bank to correct the power factor. Both generator and capacitor bank form the generation unit, which is connected to a distribution line at a Point of Common Coupling (PCC) through a 400 kva transformer, with a 13200:380 transformation ratio. The transformer connection is Dyn11. The short circuit apparent power of the PCC equals 29 MVA. The rest of the electrical network and parameters are shown in the following figure. This paper deals with the measurement and simulation results of the small hydro-plant, placed in a rural environment. The measurements were performed in February The small hydro-plant is equipped with a double injector MICHEL-BANKI turbine which directly moves a 350 kw asynchronous generator, with 380 V as terminal phase to phase voltage, 50 Hz as nominal frequency and 600 rpm as rotor speed. The turbine has two injectors: injector1 with 250 l/s of water stream capacity, and injector2 with 500 l/s. Figure 2: One-line diagram of the measurement site

2 POWER QUALITY MONITORING ANALYSIS OF THE PQ MEASUREMENT RESULTS Two measurement equipment were placed. The first one, a Dranetz-BMI Signature System, was placed at the terminals of the generation unit (380 V), exactly between the stator s output and the capacitor bank for power factor correction. This point will be called Low Voltage (LV) side of the transformer. Three phase-to-phase voltages, and three currents were measured. This equipment was provided with: LEM current probes that can measure up to 1500 A, from DC to 5 khz, with a sensitivity of 1 A. Voltage probes that can measure V. The second monitoring equipment, a LEM TOPAS 1000, was placed at the low voltage side of a measurement transformer which monitored the medium voltage side (13.2 kv), between the PCC and the transformer. This point will be called Medium Voltage (MV) side of the transformer. One phaseto-phase voltage, and one current were measured. This equipment was provided with: LEM flexible clip-on current probes that can measure up to 2200 A for sinusoidal signals with a 0.5 % accuracy band from 45 Hz to 3.0 khz. Voltage probes that can measure from 4 to 680 V. The obtained measurement results were analysed attending to their nature, classified in the following groups: Voltage regulation. Voltage sags. Voltage harmonics. Flicker. Voltage unbalance. Transients. Obtained results are available under request in [1]. Voltage regulation It was observed no relevant impact of the DER on the voltage measured in the 13.2 kv line related to the limits stated by the EN50160 standard [3]. The measured mean, maximum and minimum voltage values, each minute, are within the ±10% margin of the nominal voltage, and moreover in the 10 minutes period of time stated by the EN50160 standard. A robust connection exists at the PCC, since the generated power with respect to the short circuit power at that point is approximately 1%, so the generator is not able to regulate network voltage. Voltage sags During the one month measurement period, four voltage sags were registered: Three voltage sags between 10 and 20% depth and duration from 5 to 60 cycles were measured. Another bigger 31% depth sag was captured, but shorter than 5 cycles. This voltage sag was reproduced by simulation. Figure 3: Location of the voltage probes at the LV side The following parameters were configured for the measurement: rms voltage and current, mean, maximum and minimum values each minute. Mean, minimum and maximum active and reactive power and power factor each minute. Total Harmonic Distortion (THD) and individual harmonics calculated each minute. Flicker, concretely Plt and Pst parameters, according to internationally adopted standards [2]. Positive, negative and zero sequence voltages and currents for unbalance calculation each minute. Instantaneous voltages and currents when thresholds were exceeded: 95% and 105% of the rms nominal voltage. Voltage harmonics Harmonics, by definition, occur in the steady state, and are integer multiples of the fundamental frequency. Harmonic distortion is caused by nonlinear devices in the power system, associated with the continuous operation of the load, so its effect is permanent [4]. The relative Voltage Total Harmonic Distortion values, V THD, at the 380 V and 13.2 kv measurement points were below 3%, which is much lower than the maximum allowed by EN50160: 8%. In the same way, individual 3, 5, 7, 9, 11 and 13 th voltage and current harmonics were measured. Individual voltage harmonics were also much lower than the allowed limits. Flicker Loads can exhibit continuous, rapid variations in the current

3 magnitude that can cause voltage variations often referred to as flicker. The term flicker is derived from the impact of the voltage fluctuation on lamps such that they are perceived to flicker by the human eye. Voltage fluctuation is an electromagnetic phenomenon consisting in systematic variations of the voltage envelope or series of random voltage changes, their magnitude not usually exceeding the voltage ranges specified by EN Then, flicker is an undesirable result of the voltage fluctuation in some loads [4]. The EN50160 standard gives only a limit (at MV and LV levels) for the long term flicker severity Plt (95 percentile). This is considered to be a more significant parameter when describing the supply voltage [1]. During the measurements, the Plt value at 13.2 kv and also at 380 V was always below 1, under normal operation conditions of the generator. Voltage unbalance The unbalance of a three phase supply voltage consists of a loss of symmetry of the phase voltage vectors (magnitude and/or angle). Practically, the unbalance, Uu, of the supply voltage is defined by the negative sequence component, Ui, related to the positive sequence component, Ud. Taking into account the EN50160 standard, under normal operation conditions, during each period of one week, 95% of the 10 minutes mean rms values of the negative phase sequence component of the supply voltage shall be within the range from 0 to 2% of the positive phase sequence component. In some areas with partly single phase or two phase connected customers installations, unbalances up to about 3 % at threephase supply terminals occur [1]. During this measurement campaign, the negative sequence was far from the 2% maximum limit for the negative sequence stated by EN50160 standard. The zero sequence was also measured and it was also very low. However, it must be stated that the current injected by the generator was very asymmetrical, which does not affect the voltage because of the robustness of the PCC. Transients The term transient has long been used in the analysis of power system variations to denote an event that is undesirable but momentary in nature. The notion of damped oscillatory transient due to a RLC (resistance, inductance and capacitance) network is probably what most power engineers think of when they hear the word transient. Transients are usually dispatched within a few cycles. Transients are usually associated with changes in the system such as switching capacitor banks [4]. Lots of transients occurred at LV, but they were meaningless, due to the high number of generator start, connection and disconnection sequences. On the other hand, these data were not available for the MV measurement point (only one phaseto-phase voltage was measured). SIMULATION MODEL The simulation software used was the Alternative Transient Program-ElectroMagnetic Transient Program (ATP-EMTP). The generator model As the water stream in the turbine was not measured and the control of the injectors was not known, a simple electrical model of the generation asynchronous unit was created. The model of the turbine was avoided, and the inertia constant of the rotor modified accordingly (13 kg m 2 ). The asynchronous machine model of the ATP-EMTP software package was used as the induction generator. The model includes the three phase-to-phase capacitor bank in parallel with the generator. The value of each capacitor was 0.004F. Following this, it was placed the 13200:380 transformer. The network model The generation unit is connected to an infinite busbar by means of the 13200:380 transformer. A 1.6 MW, 0.5 MVAr load is also connected to this infinite busbar. The short circuit power of the infinite busbar was Scc=111 MVA, because of an adequate inductance. Steady state simulation The steady state validation of the asynchronous generator used in the small hydro-plant of Dima was made from the mechanic torque (Tm), calculated by equation (1), knowing that ω=600 rpm. Calculations were made for a limited number of points because each Tm value corresponds with a simulation of about 5 s, until the generator was stabilized. P meas = T m Applying the obtained mechanical torque from equation (1) to the ATP-EMTP model, the obtained result was compared with the measured one. Dynamic simulation In order to force this dynamic condition, a 30% depth voltage sag, shorter than 0.1 s, previously measured, was introduced in the MV distribution line, near the 13200:380 transformer, by means of the "pointlist" function of ATP-EMTP. Complete voltage measurements (three phase-to-phase voltages) were performed at the LV side of the transformer, when at the MV side only one phase-to-phase voltage could be measured. Then, the MV phase-to-phase voltages were extrapolated from the LV measurement results, considering the type and connections of the transformer. Later, one of the extrapolated phase-to-phase voltage was compared with the measured one at MV with good results (figure 4). The pointlist function was fed with points of the extrapolated three phase-to-phase voltages at the MV side of the ω (1)

4 13200:380 transformer. Results in the dynamic conditions Concerning the dynamic condition, the sag of figure 4 was forced at the MV side of the transformer in the model by means of the "pointlist" function of ATP-EMTP. Measured and simulated voltages and currents at the LV side of the transformer were compared, to study the reaction of the generator to this voltage sag. Figure 4: Voltage sag at the MV side, both measured and extrapolated from LV side measurements Figure 6 shows the measured and simulated phase-to-neutral voltage at the LV side of the transformer of phase C (380 V nominal phase-to-phase voltage). Rather good agreement is obtained, considering that the applied voltage at the MV side of the transformer in the model is an extrapolation of the measured voltage at the LV side, what is compared now. OBTAINED RESULTS. CORRELATION Results in the steady state conditions With the measurement results as reference data and the mentioned premises for the model of the generation unit, the network and the PCC, simulations in steady state and dynamic conditions were performed. In steady state condition simulations, the active and reactive powers were compared with the measurements with good agreement. Figure 5 shows the correlation between the measurement and simulation results for active power. Blurred graphs (lower power graphs) represent the measured active powers at each phase, being the continuous black graph their sum, thus the total measured active power. The grey line with black points represents the simulated total power, obtained from the mechanical torque. The offset of phase B with respect to phases A and C in the measured active power, is a consequence of the already mentioned current unbalance. Figure 6: Measured and simulated phase to neutral voltage in Phase C at the LV side of the transformer Figure 7 shows the measured and simulated current of phase C. Disagreements in the result are due to two facts appeared because of the lack of the turbine model: The inertia constant of the rotor was modified accordingly, trying to reflect the turbine system inertia. The inertia constant of the rotor introduced in the model was 13 kg m 2, a big one for a 350 kw of nominal power machine. The time response derived from the control of the two water injectors to the turbine control, water stock dependent, was not known. The time response of the model was then simpler than in the reality. Figure 5: Measured and simulated active power Figure 7: Measured and simulated current in Phase C at the LV side of the transformer

5 In the simulation, the rotor accelerates more than in the real case and consequently provides a bigger current. CONCLUSIONS In the frame of the EU funded DGFACTS project a measurement and simulation work was performed by Fundación LABEIN and IBERDROLA in Spain. Measurements were done with the objective of characterizing DER sites and to provide valuable information for simulation purposes. Simulations were performed to obtain reasonable models of DER, in order to continue investigating potential problems of these systems in distribution networks. Measurements in a small hydro-plant in Dima were done during one month. Several parameters were measured, processed and analysed, such as rms voltage and current, active and reactive power, total harmonic distortion and individual harmonics, flicker, voltage and current unbalance in steady state conditions, instantaneous voltages and currents when concrete thresholds were exceeded... There were not detected special problems originated by the DER facility in the distribution network. Current unbalance was remarkable as a problem, but only to the plant owner, since it did not affect to the PCC due to its high short circuit power. Concerning simulation, some of the measured results were used for a simple electric model validation in the steady state and dynamic regimes. The model was developed in ATP- EMTP. In steady state conditions, the mechanical torque was analytically calculated from the active power measurements, and then introduced in the generator model. Later on, the simulation was executed, and the active and reactive power simulation results were correlated with good agreement with the measurement results. In dynamic conditions, a voltage sag, previously extrapolated from measurements at the LV side of the transformer, was introduced in the distribution line at the MV level. Obtained voltages and currents at the LV side from the simulation were compared with those measured with rather good agreement, taking into account the simplicity of the electric model. FURTHER WORK electrical lines of different parameters (indicative resistive against inductive ratio of the distribution system), several loads of different type (linear or non-linear) and composition (combinations of different power factors) is now available for bulk simulation of scenarios with high DER penetration. The objective is, then, to identify potential problems in the distribution network due to an important interconnection of DER attending to different criteria (dispersed apparent power generated against short circuit power of the line), and to evaluate the financial losses that could occur due to those problems. Actually, simulations of the so-called DGFACTS devices in the available distribution network are being done. The purpose is to analyse if they can help solve the identified potential problems in the distribution network, and, if so, what would then be the economical impact. REFERENCES [1] [2] EN :2000(IEC :2000), Limits - Limitation of voltage changes, voltage fluctuations and flicker in public low-voltage supply systems - Equipment with rated current <= 75 A and subject to conditional connection, 2000 [3] EN50160, Voltage characteristics of electricity supplied by public distribution systems, 1996 [4] Roger C. Duncan, Mark F. McGranaghan, H. Wayne Beaty, 1996, Electrical Power Systems Quality, McGraw-Hill ACKNOWLEDGMENT The authors are grateful for the assistance to Mr. Iñaki Gorriño (CH IBAIAK), owner of the Small Hydro Plant. He was aware of every contingency occurred during the measurement period. The authors also want to thank Mr. Iñigo Cobelo (Fundación LABEIN), for his help in the modelling, simulation and reporting work. In the frame of the DGFACTS project, from the measurement and simulation work performed, some generation units like this small hydro-plant have been validated. Merging all these models, a representative distribution network composed by: asynchronous generators, synchronous generators and generation units based on photovoltaics,