Introduction
There is a common understanding of the necessity of carbon dioxide emissions reduction [
–
]. There are several initiatives across industries, including carbon storage, e-mobility, fossil fuel consumption reduction, etc. focused on CO
2 elimination. Environmental concerns have an enormous impact on modern electric power generation and consumption [
], especially through renewable generation.
The number of dispersed renewable generators is steadily increasing over the past years [
]. Proliferation of renewable into the electricity grids changes its traditional one directional power flow model, from bulk generation to final consumer. There is a long tradition in studies on the impact of renewable generation on the power systems [
]. Various aspects of dispersed generator connection have been standardised [
,
].
Power quality (PQ) [
] is one of the important aspects of renewable generation. Electrical energy is seen as a product on a free market that should fulfil standardised quality criteria [
]. Power generation variations are a new aspect [
] of power quality, which is related to the variations of renewable resources, not present in combustion power plants.
The standard [
] includes power quality requirements for the medium voltage and low voltage grids. It has been widely used and incorporated, sometimes with modification in the grid codes of distribution system operators. There are numerous standards defining disturbance limits in distribution grids [
–
].
This paper presents a case study of power quality aspects of photovoltaic(PV) and doubly feed induction generator (DFIG) integration, which is motivated by increasing concerns about power quality deterioration in the public distribution grids feeding residential, commercial and light industry consumers. Increasing proliferation of distributed generation is seen as a deteriorating factor for PQ.
However, requirements for grid connected micro-generators given in standards mitigate the negative impact on the quality of the power supply. There is no general rule for defining a significant percentage of renewable generation deteriorating PQ. High concentration of dispersed generation along with disturbing loads and week grid causes a remarkable deterioration of power quality. A case oriented perspective is needed.
The daily power demand profiles on the national level and profiles of municipal customers using dispersed generation generally match to decrease energy transmission. Weather conditions may induce an overproduction being problematic for the local distribution grid. Power flow fluctuations increase proportionally to installed renewable generators and are a new aspect of PQ.
Assessment of fluctuations in power generation is not standardised. Moreover, the standard 10 min. window for PQ analysis is insufficient for capturing of power curve variations. Shorter windows are needed. Storage facilities emerge as a necessary component of renewable generation units capable of power fluctuation mitigation.
First, a general characteristic of PQ requirements given in standards is presented along a benchmarking for voltage harmonics limits. Then, the DFIG and PV test sites are briefly introduced, followed by an analysis of measured values. Finally, power curve fluctuations and phenomena exceeding standardised limits are presented.
Requirements for PQ
PQ can be regarded as a part of the electromagnetic compatibility (EMC) [
]. It is understood as the ability of a device or system to operate properly in a given electromagnetic environment. Moreover, no negative impact on this environment and other equipment operating in it should be present. In the case of PQ, the parameters describing an electromagnetic environment are qualitative indices of voltage and current waveforms in electrical power grid [
]. Low frequency disturbances are dominant and are conducted throughout the power grid. Radiated disturbances are less common and are not included in presented considerations.
The most typical disturbing phenomena include outages, interruptions, voltage dips, voltage swells, transients, flicker, asymmetry, power factor variations, and harmonics in voltage and current. Mathematical analysis of waveforms gives energy, active and reactive power, distortion statistics, trends, etc. The Distribution System Operators (DSOs) require specific levels of reactive power, not defined in PQ standard, e.g. in Refs. [
–
].
The IEEE Standard 1159 defines PQ as “the concept of powering and grounding electronic equipment in a manner that is suitable to the operation of that equipment and compatible with premise wiring system and other connected equipment”.
The IEC standards for electromagnetic compatibility section including power quality can be divided into major groups including environment (series 61000-2-
x) [
–
], limits for the levels of disturbance emission into the environment (series 61000-3-
x), and test methods for immunity against distortions(series 61000-4-
x) [
–
].
The immunity and distortion emission limits are given for loads. There are no general standards describing the requirements for low-power generators connected to the grid. The standard in Ref. [
] defines requirements for power factor, levels of harmonic emissions [
], and voltage fluctuations [
], directly calling standards dedicated for loads. It should be stressed that emission test results on disturbances emissions of generators depend on the nature of the grid and connected loads which are fed from the generator under test. In a real network, a PQ study on side is justified.
Generally, the particular requirements for the connection of small generators to the low voltage and medium voltage grids are defined by the regional distribution system operator DSO.
Voltage harmonics limits
Voltage harmonics are one of the important power quality indices [
] mostly because they affect all customers connected to the PCC indicating harmonic content in the voltage. Voltage harmonics are also very common in distribution grids due to large proliferation of nonlinear loads causing them [
].
There are different approaches to specifying the maximum allowable levels of voltage harmonics. Usually, the content of voltage harmonics between the second and fiftieth order is expressed as a percentage of the fundamental 50Hz component. The distribution system operator DSO defines the limits in its own grid code. In the presented case of a DFIG wind turbine, the DSO requires that all medium voltage harmonics between 2 and 50 orders should be lower than 1.5% of the fundamental. Moreover, the total harmonic distortion(THD) is required to be lower than 4%. It must be stressed that the limits for the 3rd, 5th, 7th, 11th and 13th harmonic are multiple times lower than the regulations stated in the standards in Refs. [
]. and [
]. Obviously, those tougher regulations given by DSO are harder to fulfil. On the contrary, the other harmonics are allowed to be higher than those given in Refs. [
] and [
]. The fifteenth harmonic can be taken as an example. Accordingly, to the DSO regulation, the wind power plant can cause 1.5% of the fifteenth harmonic distortion. But every customer connected to the grid, even close to the wind generator, should be guaranteed maximally 0.5%. Figure 1 shows the discrepancies in the requirements of various standards.
The THD values for the 10 minute intervals are given in Fig. 2. These values have not exceeded 4%. Occasionally, the 5th harmonic is above the 1.5% limit. THD values and harmonic content are much lower than required in standards (as shown in Fig. 3). The limits given in Refs. [
,
,
] are equal.
DFIG wind turbine test side
Measurement of power variations and PQ disturbances was done at the connection point of a doubly feed induction generator DFIG. The nominal voltage of the generator was
UN=690 V, and the nominal current
IN=415 A. The 2 MW asynchronous generator of the type Vest as V 80 [
] was connected through a transformer to the 15 kV public grids. The switchgear was located on a mast supporting the medium voltage overhead line and linking the wind generator through a 300 meter long ground cable.
The PQ monitoring device was located in the container substation in a standard measurement cell. Voltage probes were attached to a star-connected voltage measurement transformer with a translationkU=150. Current claps were attached to the secondary windings of measurement current transformers kI=16.
It was assumed that the above described wind generator is a very typical example of the middle voltage distribution grid in Europe. Due to its proliferation, it had a significant influence on the power quality in distribution networks.
PV system test side
The research installation with a maximal power output of 15 kW consists of three independent one phase systems constructed in different technologies. The schematic graph is shown in Fig. 4. The photovoltaic research installation was located at the university campus and arranged on the roof of an existing building. The geographical orientation of the roof was accepted and no movable mounting structures were used.
There were three independent 5 kW one phase PV system connected to the grid through SMA Sunny Boy 5 kW inverters. The Monocrystalline and Polycrystalline were oriented south without a transformer while the final one was oriented west with a transformer. The difference in cell technology, converter type and geographical orientation accounted for the differences in switch-on and off times and different power curves.
Wind generator data analysis
The measurement was done over a 3 weeks period. It included days when the wind generator was not producing any energy. It was possible to assess the impact of the wind generator on the PQ indicia (harmonics, THD,Pst, Plt) by benchmarking the results obtained by the stopped and working DFIG. The flicker values were not increased by the operation of the wind turbine, as given in detail in subsection 6.1.
Short-time and long-time flicker
Power line flicker is a phenomenon originated from rapid voltage fluctuation. Those voltage fluctuations can cause the flickering of electric light which is regarded as annoying and disturbing. Moreover, this phenomenon is also dangerous for sensitive electronic equipment. Due to the human eye sensitivity involved in the measurement of voltage fluctuations, a standard was necessary for flicker measurement equipment to guarantee measurement coherence between devices [
].
The short term flicker
Pst was computed over a 10 min interval using momentary data. The long term flicker
Plt was computed using the
Pst values form a two hour window [
]. Wind turbines were usually regarded as a source of flicker [
] with specific flicker assessment procedures [
].
The short time flickers of the considered DFIG generator are shown in Figs. 5 and 6. The one week measurement shows an exceeding of limitPst=0.45.
Table 1 lists the percentage of Pst and Plt values not exceeding the allowed tolerance thresholds set by the local DSO. The average for the whole observation week is presented, followed by flicker values representing the time period with the stopped generator (flicker originating in the grid) and the operating generator (generator contributing to flicker). The differences are small, so the contribution by the DFIG to the flicker values is insignificant.
In this particular case, the flicker values even without generator contribution were higher, then requested by the DSO. As a consequence, the wind turbine operator did not meet the DSO standard even though the wind turbine did not contribute to flicker at all. The DSO limits were not set appropriately. A power quality study at the connection point was recommended as a prerequisite for further regulations. The DFIG was asked to virtually improve flicker values.
DFIG power curves analysis
Figures 7 and 8 present the power generation variation of the DFIG on an arbitrary chosen day. During the course of that day, the generator worked for ca. 9 hours with 50% of the nominal power (1 MW). After a period of 10 hours it was stopped. The generator only occasionally reached its nominal power of 2 MW. The wind power was variable as the weather conditions were. Long term predictions are unreliable. In case of over-generation, the power output must be reduced. Short time power variations (seconds or minutes) are not problematic in wind generation as opposite to PV.
During regular operation, the power factor was kept constant. A change in active power was followed by a change in reactive power level. Modern wind power generators, like DFIG are equipped with power electronics converters, allowing reactive power to change in a wide range. From a PQ perspective, distributed generation can even improve power factor and reactive power management in the distribution grid.
At several time instances, a short time perturbation which lasted for several seconds in power generation was observed as shown in Figs. 9 and 10. There was an underling phase change and small voltage drop due to the switching in the grid. The active power values in three phases oscillated, but the three phase output (the sum) was relatively unaffected as depicted in Fig. 9. At the same time, a perturbation was marking the reactive power generation as demonstrated in Fig. 10. The reactive power was four times higher than outside the perturbation. This case illustrates a short time deterioration of reactive power values and a frequent perturbation in the distribution grid.
Generally, a combination of renewable generation with storage technology is needed to overcome the problem of sudden generation variations. An existing approach is the low voltage right through the capability of wind turbines, as requested in grid codes of DSOs.
PV system data analysis
Electrical power generation in PV systems extensively proliferates into the distribution grid. Reaching more than 10% share in electricity production is feasible and in some aspects advantageous. PV can be seen as a complement to large combustion power plants.
The daily power curve of a PV system on fine days is predictable and similar to each other (Fig. 11). It corresponds to the general power consumption given through the power curve in the transmission system as displayed in Fig. 12 in which the normalised PV power and the transmission system power on Sept. 9, 2016 are shown.
The highest potential of energy generation from photovoltaic panels occurs in the middle of the day and corresponds to the general daily load curve in a power system. The load curve changes in the range of+14% to –21% from the average value on a particular day are depicted in Fig. 13 and listed Table 2. The black solid line represents the average load of the day equal to 1 pu. Assuming that there is a rapid development of installed solar power (e.g. above 10%), an effective tool for alleviation of the midday peak emerges. In addition, siting PV power in the distribution grid immediately feeding consumers reduces transmission power losses.
There is no further detailed presentation of PQ disturbances due to the fact that the limits given in standards [
–
] are not exceeded.
Variable weather conditions deteriorate the smoothness of the power curve (Fig. 12). The afternoon sag is also visible in the reactive power curve as can be seen in Fig. 14. During the start-up and low power output, the generator is not following a predefined power factor. The highest value of reactive power is observed. Similar situation occurs in the late evening, shortly before dusk. With a high power output, the generators are usually operated at PF=1. There is no negative impact on the distribution grid. The negative impact of weather conditions is also visible in the apparent power curve in Fig. 14.
Figures 12, 15 and 16 present the active power curve distortion due to weather conditions on a particular day. The severity of the afternoon deep depends on the averaging window. A typical PQ analysis requires the 10 min averaging window, as defined in Refs. [
,
,
] and this window is applied in commercial PQ monitoring devices. The power curve distortion visible in Fig. 12 is even more distinct by shorter averaging windows including 1 min (Fig. 15) and 1 s (Fig. 16). The power value variations are more frequent and more rapid.
Similar observations are true for reactive power curves given in Figs. 14, 17 and 18 respectively. Substantial rapid power lever variations can be noticed. There are different on and off times for all the phases, which is caused by different panel numbers and technology—mono and polycrystalline. The panels connected to L3 are facing west (maximum delayed) and are connected through a different converter (with a transformer).
Rapid variations in active and reactive power are also visible in power factor plots. Figures 19, 20 and 21 present the 10 min, 1 min and 1 s average respectively. The shorter the window, the more details are visible. The one phase inverters operate at PF=1 when the power is over a predefined threshold. The low power periods in mornings and evenings are problematic.
The total harmonic distortion in voltage and current reaches similar values (Figs. 22 and 23). PV converters operate at low power indicate high current harmonic content. THD values are as high as 20% to 40% (Fig. 23). Due to low RMS current values, it has an insignificant influence on the voltage THD values.
Conclusions
This paper has presented selected topics of PQ assessment in grids with dispersed renewable generation. Power electronic converters are in use in wind and solar systems as well. Power generation from wind and solar energy is primarily influenced by environmental conditions, therefore, prone to short term and long term variations. Those short and long term variations in power are considered as an emerging PQ index. Quantitative treatment of variations is not defined in standards.
Long term averaging, e.g. the 10 min window, attenuates short time variations in power generation curve. The picture for the 1 s window is substantially different. However, it requires more data to be stored and analysed. A detailed analysis of variations is a prerequisite for the implementation of storage facilities. Storage units have to be optimised in terms of generator output short term and long term variations. This problem definitely increases with growing number of distributed renewable generation.
Harmonic distortions are usually related to the operation of converters and other nonlinear devices. However, in the presented cases, there is no much trouble with harmonic levels. Generally, modern converters can operate with low and acceptable harmonic emission and a power factor near one. Troublesome is the operation well below the nominal power.
The high proliferation of dispersed photovoltaic and wind generation can be realised in a manner satisfying the requirements given in PQ standards, as [61000-4-30]. Distortion levels depend on gird parameters and load behaviour as well, so generally a case perspective is required.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(1479KB)
