Impact of Wind Power Plant Operation on Voltage Quality Parameters—Example from Poland

Abstract

In Poland, there has been a steady increase in the share of wind-generated energy in the overall energy balance of the country. However, wind power plants belong to unstable energy sources—the amount of generated power is variable in time. Variable generation may cause problems with the quality of energy transmitted in electricity networks. Therefore, this article presents the results of field tests of the impact of the Vestas V90 2 MW wind turbine on selected voltage quality parameters occurring in the grid. Due to the multifaceted influence of the power plant on the parameters of the grid to which it is connected, the article focuses only on four selected parameters: voltage value, zero and reverse asymmetry, and the value of the voltage distortion factor. The research conducted by the authors confirmed the positive influence of the operation of the wind power plant on the quality parameters of the electricity transmitted through the power grid. An increase in the power generated in the tested Vestas V90 power plant resulted in a decrease in the values of both the asymmetry factor (zero and reverse) and the total voltage distortion factor THD_U.

1. Introduction

Wind energy has been used for many years for a variety of applications, from powering mills, floating boats, to supplying electricity to consumers. Due to the increased environmental awareness of the public, efforts are being made to replace conventional sources with renewable sources [1,2]—sources that are self-renewable in the short term, i.e., they do not run out and can be used without fear of one day running out. The most commonly used renewable energy sources in practice are power plants using solar radiation, water, wind, geothermal, and biomass. In Poland, good wind conditions occur in about 40% of the country, with the most favourable conditions in northeastern Poland and in coastal areas. Although wind energy is very clean, it is not stable over time. There are many scientific studies on wind energy. They focus mainly on describing the current status of wind power in a given area [3,4] and trends concerning further development of this type of electricity source [5]. The issue of the technological safety of wind turbines is also a very important subject discussed in these publications [6,7]. The benefits of wind energy, such as environmental sustainability, economic benefits, modularity, scalability, and relative ease of installation, have attracted the attention of the business community and individuals [8]. In Poland, this positive contribution of wind energy to society and industry has already enjoyed strong public support for many years [9]. The reliability of power supply systems for wind farms is also increasingly discussed [10].

The heart of every wind turbine is the generator where mechanical energy is converted into electrical energy. There are two types of generators: asynchronous and synchronous [11]. In power plants with asynchronous generators, there are two converters. One is on the rotor side, whose main task is to control the speed of the generator. The other one is mounted on the grid side and is mainly responsible for the output voltage parameters [12,13].

The issues discussed in the article can be transferred to other European and world countries. However, it should be remembered that the impact of wind power plants on electricity parameters may differ from one location to another. This will be a result of differences in both the construction of a given power system and the diversity of end users. Nevertheless, the observed trends in wind power impact should be analogous to most European electricity networks.

In the case of the power systems of Poland, the main problem is the voltage stability issues [14]. Proper planning and operation of the system requires a thorough analysis of the voltage stability and thus its instability [15]. As the capacity of the wind farms connected to the grid increases, the voltage stability occurring in the power grid decreases [16,17]. The variety of wind turbine designs makes it necessary to determine the inertial response of the power system to the occurrence of a disturbance. The characteristics of the inertial response of wind turbines vary depending on the control methods used in them [18,19,20]. The influence of wind farms on voltage stability is studied in three main aspects: the share of wind energy, the magnitude of disturbances introduced into the grid, and improvement measures. In the medium-voltage power system, the output of grid-connected wind farms is strongly limited by the requirement to maintain voltage stability [17]. This applies both to voltage variations at small and at large disturbances. This paper presents the results of a field study, the influence of switching operations in a Vestas V90 wind power plant with a rated power of 2 MW on selected parameters describing voltage stability and quality.

Current research on wind energy focuses mainly on analytical and simulation processes. Few authors discuss the issues related to the impact of wind power plants on medium-voltage power grids. The field research carried out by the authors was aimed at filling this research gap and confirming or not the information obtained from the mathematical analyses.

2. Materials and Methods

The analysis of the voltage deformation mechanism at the point of connection (PCC) of the load or generator is based on a simplified model of the power system—Figure 1. The diagram consists of the voltage source and the equivalent impedance of the power system. The load or generation is represented by linear and non-linear elements that draw or inject current into the power system. In a steady state, such a system can be considered separately for each harmonic of order h by introducing the source voltage U₀(h), the current I_L(h) and the power system impedance Z_S(h).

The harmonic voltage value of order h at the point of connection U_pcc(h) can be described by the equation [21,22]:

$$U_{pcc}\left(h\right)=U_0\left(h\right)\pm I_L\left(h\right)·Z_S\left(h\right)$$

(1)

The voltage at the point of connection comprises the sum of the background distortion U₀(h) (caused by the non-linear loads present in the power system) and the distortion caused by the analysed load or generator I_L(h)·Z_S(h). Depending on the phase shift of a given current harmonic I_L(h), the voltage drop across the system impedance can have a positive or negative sign. This situation is shown in Figure 2.

Figure 2a corresponds to the situation when |U_pcc(h)| > |U₀(h)|. In this case, the sub-connection of a load or generator leads to an increase in voltage distortion at the point of connection (PCC). Figure 2b illustrates the situation where |U_pcc(h)| < |U₀(h)| and the connection of a load or generator leads to a reduction of the voltage distortion at the PCC.

The value of the total voltage distortion unit (THD_U) can be determined from the general formula [23]:

$$TH{D}_U=\frac{\sqrt{{{\displaystyle \sum }}_{h=2}^{\infty }U{\left(h\right)}^2}}{U\left(1\right)}100\%$$

(2)

where: U(1)—rms voltage of the fundamental harmonic, U(h)—rms voltage of the h-th harmonic, h—order of the harmonic.

In the case under consideration, the value of the THD_Upcc coefficient at the connection point, taking into account Equation (1), is described by Equation (3):

$$\left\{\begin{array}{l}TH{D}_{Upcc}=\frac{\sqrt{{{\displaystyle \sum }}_{h=2}^{\infty }{\left[U_0\left(h\right)\pm I_L\left(h\right)·Z_S\left(h\right)\right]}^2}}{U_0\left(1\right)-I_L\left(1\right)·Z_S\left(1\right)}100\%forload\hfill \\ TH{D}_{Upcc}=\frac{\sqrt{{{\displaystyle \sum }}_{h=2}^{\infty }{\left[U_0\left(h\right)\pm I_L\left(h\right)·Z_S\left(h\right)\right]}^2}}{U_0\left(1\right)+I_L\left(1\right)·Z_S\left(1\right)}100\%forgeneration\hfill \end{array}\right.$$

(3)

For practical analyses, the value of the impedance Z_S(h) is difficult to obtain. Therefore, it is more common, when describing system parameters, to use the short-circuit power S_k^″, which can be related to the system impedance according to the following relations:

$$Z_S\left(1\right)=\frac{1,1·U_n^2}{S_k^{″}}{Z}_S\left(h\right)=w_h·Z_S\left(1\right)$$

(4)

in which w_h is the coefficient of impedance gain—together with the harmonic order, described by the following equation [24]:

$$w_h=\sqrt{h}+jh$$

(5)

Given Equations (4) and (5), Equation (3) takes the form:

$$\left\{\begin{array}{l}TH{D}_{Upcc}=\frac{\sqrt{{{\displaystyle \sum }}_{h=2}^{\infty }{\left[U_0\left(h\right)\pm I_L\left(h\right)·w_h·\frac{1,1·U_n^2}{S_k^{″}}\right]}^2}}{U_0\left(1\right)-I_L\left(1\right)·\frac{1,1·U_n^2}{S_k^{″}}}100\%forload\hfill \\ TH{D}_{Upcc}=\frac{\sqrt{{{\displaystyle \sum }}_{h=2}^{\infty }{\left[U_0\left(h\right)\pm I_L\left(h\right)·w_h·\frac{1,1·U_n^2}{S_k^{″}}\right]}^2}}{U_0\left(1\right)+I_L\left(1\right)·\frac{1,1·U_n^2}{S_k^{″}}}100\%forgeneration\hfill \end{array}\right.$$

(6)

When the generator is connected to the power grid, the value of the denominator of Equation (6) will increase. At the same time, in modern wind power plant designs, there should be a very low deformation of the generated current, which will cause a small increase in the value of the denominator of Equation (6). Therefore, a reduction in the value of the connection point voltage distortion coefficient THD_Upcc should be expected with wind power plant operation. In addition, the THD_Upcc value will be influenced by the short-circuit power S_k^″, which should also increase in value due to the connection of the source [25].

The research was carried out in a system supplying a Vestas V90 type wind power plant with a nominal capacity of 2 MW. The power plant uses an asynchronous generator driven by a three-bladed wind turbine with power output controlled by changing the blade angle (pitch). The wind turbine is connected to a medium voltage (15 kV) overhead power line supplied from a 110/15 kV substation. The distance from the connection point to the 110/15 kV substation is approximately 5.3 km. The short-circuit capacity at the power plant connection point is approximately 52 MVA (short-circuit factor 26).

The Vestas V90 2.0 MW wind turbine is a pitch-regulated upwind turbine with active yaw and a three-blade rotor—

Table 1. Rotor data.

Diameter	90 m
Swept Area	6362 m2
Rotational Speed Static, Rotor	14.9 rpm
Speed, Dynamic Operation Range	9.6–17.0 rpm
Rotational Direction	Clockwise (front view)
Orientation	Upwind
Tilt	6°
Hub Coning	2°
Number of Blades	3
Aerodynamic Brakes	Full feathering

. The turbine has a rotor diameter of 90 m with a generator rated at 2.0 MW. The turbine utilises a microprocessor pitch control system and variable speed feature. With these features, the wind turbine is able to operate the rotor at variable speed, helping to maintain the output at or near rated power.

The turbine is equipped with a 90-m rotor consisting of three blades. Based on the prevailing wind conditions, the blades are continuously positioned to help optimise the pitch angle.

The energy input from the wind to the turbine is adjusted by pitching the blades according to the control strategy. The pitch system also works as the primary brake system by pitching the blades out of the wind. This causes the rotor to idle.

The yaw system is designed to keep the turbine upwind. The nacelle is mounted on the yaw plate, which is bolted to the turbine tower. Asynchronous yaw motors with brakes enable the nacelle to rotate on top of the tower. The turbine controller receives information about the wind direction from the wind sensor. Automatic yawing is deactivated when the mean wind speed is below 3 m/s.

The generator is a three-phase asynchronous Doubly Fed Induction Generator (DFIG) with a wound rotor that is connected to the Vestas Converter System (VCS) via a slip-ring system—

Table 2. Generator data.

Type Description	Asynchronous Doubly Fed Induction Generator (DFIG) with wound rotor, slip rings and VCS
Rated Power (PN)	2.0 MW
Rated Apparent Power	2.08 MVA (Cosφ = 0.96)
Frequency	50 Hz
Voltage, Generator	690 Vac
Voltage, Converter	480 Vac
Number of Poles	4
Winding Type (Stator/Rotor)	Random/Form
Winding Connection, Stator	Star/Delta
Rated Efficiency (Generator only)	>97%
Power Factor (cos)	0.96 ind–0.98 cap
Overspeed Limit According to IEC (2 min)	2900 rpm
Vibration Level	≤1.8 mm/s
Weight	Approximately 7500 kg

. The generator is an air-to-air cooled generator with an internal and external cooling circuit.

The generator has four poles. The generator is wound with form windings in both the rotor and stator. The stator is connected in Star at low power and Delta at high power. The rotor is connected in Star and is insulated from the shaft.

The transformer is located in a separate locked room in the nacelle. The transformer is a two-winding, three-phase, dry-type transformer—

Table 3. Transformer data.

Type Description	Dry-type cast resin
Primary Voltage	15.0 kV
Rated Power	2100 kVA
Secondary Voltage 1	690 V
Rated Power 1 at 690 V	1900 kVA
Secondary Voltage 2	480 V
Rated Power 2 at 480 V	200 kVA
Vector Group	Dyn5
Frequency	50 Hz
HV-Tappings	±2 × 2.5% off-circuit

. The windings are Delta-connected on the high-voltage side. The low-voltage windings have a voltage of 690 V and a tapping at 480 V and are Star-connected. The 690 V and 480 V systems in the nacelle are TN-systems, which means the star point is connected to earth.

The tests were carried out at the point of connection of the wind power plant to the 15 kV power grid—Figure 3. The quality parameters of electricity were recorded during switching operations carried out in the tested power plant. These activities consisted in recording the operation of the generator and then switching off the wind power plant and recording the parameters of the system without an operating source (the so-called background test), after which the power plant was started up. In order to eliminate errors resulting from momentary variations of the power grid parameters, these activities were repeated three times in the same time intervals. The tests were carried out when the wind speed allowed the power output to exceed 75% of the rated power. This allows the system response to the switching actions occurring in the power plant (generating the highest possible power gradients) to be correctly investigated. The measurement was conducted with an interval of 200 ms.

A SONEL PQM-701 portable power quality analyser was used to record changes in electrical quantities occurring during the operation of a wind turbine. The analyser registers network parameters in accordance with class A of the EN 61000-4-30 standard. It is a programmable device that measures, calculates, and stores in its memory the parameters of single- or three-phase power networks.

One of the power quality parameters determined was the voltage asymmetry coefficients. The first of them was the coefficient of opposite voltage asymmetry (voltage asymmetry) calculated from the relation [26]:

$$\begin{array}{l}{\underset{\_}{k}}_{U2}=\frac{{\underset{\_}{U}}_2}{{\underset{\_}{U}}_1}=k_{U2}{e}^{j{\phi }_{U2}}\hfill \\ k_{U2}=\left|{\underset{\_}{k}}_{U2}\right|=\frac{U_2}{U_1}\hfill \\ k_{U2\%}=k_{U2}·100\%\hfill \end{array}$$

(7)

and the coefficient for zero voltage asymmetry (voltage imbalance), described by the equation

$$\begin{array}{l}{\underset{\_}{k}}_{U0}=\frac{{\underset{\_}{U}}_0}{{\underset{\_}{U}}_1}=k_{U2}{e}^{j{\phi }_{U0}}\hfill \\ k_{U0}=\left|{\underset{\_}{k}}_{U0}\right|=\frac{U_0}{U_1}\hfill \\ k_{U0\%}=k_{U0}·100\%\hfill \end{array}$$

(8)

where: U₀, U₁, U₂—symmetrical components of voltages; k_U0, k_U2—composite asymmetry coefficients of phase voltages; k_U0, k_U2—modules of asymmetry coefficients of phase voltages; k_U0%, k_U2%—percentage asymmetry coefficients of phase voltages.

The obtained measurement data set was statistically analysed using the following indicators:

• arithmetic mean

$$\overline{x}=\frac{1}{n}{\displaystyle \sum _{i=1}^n}{x}_i$$

(9)

• median

$$Me=\left\{\begin{array}{ll}{x}_{\left(n+1\right)/2}\hfill & foroddn\hfill \\ \frac{1}{2}\left(x_{n/2}+x_{\left(n+2\right)/2}\right)\hfill & forevenn\hfill \end{array}\right.$$

(10)

• standard deviation

$$s=\sqrt{\frac{1}{n}{\displaystyle \sum }_{i=1}^n{\left(x_i-\overline{x}\right)}^2}$$

(11)

• striae

$$R=x_{max}-x_{min}$$

(12)

• coefficient of variation

$$V=\frac{s}{\overline{x}}100\%$$

(13)

where: x_i—individual values of the random variable, n—sample size, x_{(n + 1)/2}—value of the element with number (n + 1)/2 in a non-descriptively ordered set of values of the random variable, x_min—smallest value of the random variable, x_max—largest value of the random variable.

When investigating the impact of wind power plant operation on the voltage quality in the power grid, the authors decided to analyse the dependence of power quality indicators on the power plant capacity. Correlation analysis was used to analyse the significance of the interdependence of the mentioned quantities. The analysis started with the preparation of scatter diagrams (correlation diagrams) illustrating the relations between individual power quality indicators and the power plant power. A straight line determined by linear regression (distance-weighted least squares method) was fitted to the plot points.

The next stage of the correlation analysis was to determine the values of indicators defining the examined interdependencies and to carry out statistical tests on their basis concerning the existence (or not) of significant relationships between the quantities. As a measure of the interdependence of phenomena, the Bravais–Pearson linear correlation coefficient (r) was selected, the value of which can be calculated based on the following formula:

$$r=\frac{cov(y,x)}{s_y·s_x}=\frac{{{\displaystyle \sum }}_{i=1}^n\left[\left(x_i-\overline{x}\right)\left(y_i-\overline{y}\right)\right]}{\sqrt{{{\displaystyle \sum }}_{i=1}^n{\left(x_i-\overline{x}\right)}^2·{{\displaystyle \sum }}_{i=1}^n{\left(y_i-\overline{y}\right)}^2}}$$

(14)

where: cov(y, x)—covariance of variables y and x, x_i, y_i—individual values of a random variable, respectively explaining (x) and explained (y), s_x, s_y—standard deviations of respectively explaining (x) and explained (y) variable, $\overline{x},\overline{y}$—arithmetic means of the explanatory variable (x) and the explained variable (y), respectively, n—sample size.

On the basis of the values of the calculated indicators, statistical tests (t-test) were carried out for the existence of significant relationships between the quantities, for which a significance level of α = 0.05 was adopted. In the case of tests for the existence of relationships between the variables:

• A positive test result (a “+” sign) means that there are no grounds to reject the hypothesis of independence of the tested variables at the given significance level α;
• A negative test result (“−” sign) means that the hypothesis of independence of the tested variables at the given significance level α should be rejected.

3. Results

As a result of the conducted field research on the quality parameters of electricity generated in the Vestas V-90 wind power plant with a capacity of 2 MW, the waveforms of the variability of these parameters during the switching operations carried out in the power plant were recorded. The registered course of the variability of the power generated in the investigated system is presented in Figure 4.

Figure 5 and Figure 6 show the waveforms of the phase and line voltages corresponding to the values of the active power generated at a given moment in the wind turbine under study.

The variation curves of the values of the voltage asymmetry coefficients are shown in Figure 7—zero asymmetry coefficient (k_U0) and Figure 8—opposite asymmetry coefficient (k_U2).

The determined course of variation of the value of the total voltage distortion factor during three times of stopping and starting the tested wind power plant is presented in Figure 9.

The results of the statistical analysis of the quantities characterising the voltage quality at the point of connection of the analysed wind power plant to the power grid are summarised in

Table 4. Results of statistical analysis of quantities characterising voltage quality at the point of connection of the analysed wind power plant to the power grid.

Size	Unit	$\overline{\mathit{x}}$	Me	s	xmin	xmax	R	V
UL1	kV	8.830	8.835	0.078	8.674	8.965	0.291	0.892
UL2	kV	9.855	9.857	0.087	9.678	9.997	0.318	0.890
UL3	kV	8.783	8.793	0.085	8.608	8.919	0.311	0.975
UL1-2	kV	15.775	15.787	0.148	15.482	16.007	0.525	0.939
UL2-3	kV	15.888	15.896	0.143	15.605	16.112	0.506	0.905
UL3-1	kV	15.837	15.845	0.144	15.551	16.060	0.508	0.909
kU0	%	7.859	7.850	0.092	7.650	8.140	0.490	1.180
kU2	%	0.416	0.410	0.030	0.330	0.530	0.200	7.231
THDUL1	%	1.124	1.120	0.075	0.960	1.290	0.330	6.749
THDUL2	%	1.083	1.080	0.074	0.930	1.240	0.310	6.867
THDUL3	%	1.178	1.160	0.076	1.030	1.370	0.340	6.484
PL1	kW	420.348	559.446	265.756	−20.018	681.728	701.747	63.222
PL2	kW	479.254	637.063	301.787	−18.699	777.342	796.041	62.970
PL3	kW	430.951	573.585	273.605	−20.678	703.075	723.753	63.488
PIII	kW	1330.554	1769.900	841.138	−58.727	2161.960	2220.687	63.217

.

Scatter plots showing the effect of power plant output on power quality indicator values are shown in Figure 10, Figure 11, Figure 12, Figure 13 and Figure 14. Due to the repeatability of the obtained waveforms in individual phases, the parameters of only one of the phases were analysed.

The values of calculated correlation coefficients and results of statistical tests for individual power quality indicators depending on the capacity of the wind power plant analysed are presented in

Table 5. Results of the statistical analysis of the dependence of individual power quality indicators on the three-phase power (P_III) of the power plant under analysis.

Size	Linear Correlation Coefficient r	Critical Level of Significance α	Test Result α > 0.05
UL1	0.856	0.000	-
UL2	0.847	0.000	-
UL3	0.879	0.000	-
UL1-2	0.872	0.000	-
UL2-3	0.863	0.000	-
UL3-1	0.858	0.000	-
kU0	−0.657	0.000	-
kU2	−0.745	0.000	-
THDUL1	−0.686	0.000	-
THDUL2	−0.754	0.000	-
THDUL3	−0.803	0.000	-

. According to the values given in Table 5, the error of type I (α) for all analysed values is 0. The error of type II was not analysed by the authors, due to its minor significance for the conducted research.

4. Discussion

When a distributed, unstable energy source has a small share of power compared to conventional sources, the effect of interference caused by the operation of that source can be very small. However, if this share is large, as is most often the case in rural networks, the risk of disturbance increases significantly. In articles, we can find proposals for various technical solutions to monitor transients [27,28]. Analytical methods for detecting power quality disturbances in networks coupled to a wind turbine are also given [29]. Om Prakash Mahela [30] proposed an algorithm using the multiresolution capability of the discrete wavelet transform. Muljadi [31] in his work addressed the impact of reactive power compensation of wind farms on the parameters of the electricity generated in the farm. However, all the abovementioned works deal with theoretical analyses, and their results were not supported by field tests. Shalukho [32] presented the results of research conducted on a laboratory model of a wind power plant. He showed that the modelled power plant negatively affects the supply voltage distortion. However, this is not consistent with the results of tests carried out on a real wind power plant. The research conducted by the authors confirmed, in comparison with the results of the research conducted on a 30 MW wind farm [33], the positive influence of the operation of the wind power plant on the quality parameters of the electricity transmitted through the power grid. As shown in Figure 5 and Figure 6, according to the theory, an increase in the generated power causes an increase in the voltage value occurring at the power plant connection point. More precisely, this relationship can be observed in Figure 10 and Figure 11, which show the relationship between the power generated in the power plant under study and the voltage values in the grid. The set of registered values was approximated by a straight line, with the Bravis–Pearson linear correlation coefficient close to unity (r > 0.85). This means a close relationship between the amount of power generated at the source and the voltage value occurring at the power plant connection point. Since the determined line in Figure 10 and Figure 11 is increasing, the relationship between voltage and power is directly proportional. It is greater for the phase-to-phase voltage (Figure 11)—the angle of the straight line concerning the axis of ordinates is greater. However, due to the fact that the power plant is equipped with power gradient control systems, the voltage fluctuations caused by switching on and off the power plant do not exceed 2.5% of the nominal network voltage. This is in line with the law in force in Poland. It is worth emphasising that the operation of the power plant reduces both the value of the zero and reverse voltage asymmetry factor as well as the value of the voltage distortion factor. These relations can be observed both on the recorded time courses of these quantities (Figure 7, Figure 8 and Figure 9), as well as on the diagrams presenting the results of the statistical analysis of the obtained test results (Figure 12, Figure 13 and Figure 14). The relationship between these quantities can also be approximated by a straight line while maintaining a high Bravais–Pearson linear correlation coefficient (above 0.6). The determined straight line is decreasing, which indicates that the relationship between the power generated by the power plant under study and the coefficients of voltage asymmetry and distortion is inversely proportional (as the power increases, the values of the coefficients decrease). The slope of the straight line concerning the axis of ordinates is the smallest (the dependence on the generated power is then the smallest) for the value of the opposite voltage asymmetry. It can therefore be concluded that, in the case of the THD_U coefficient, the field tests have unambiguously confirmed the conclusions drawn from the theoretical analysis carried out by the authors.

5. Conclusions

The increase in global demand for electricity forces the continuous development of renewable energy, including distributed energy using wind power plants for generation. A growing share of electronic devices, sensitive to supply voltage parameters, in the group of receivers causes power quality tests to be more and more commonly performed. More and more theoretical works on the influence of power plants on the parameters of the power grid can be found in the literature. However, as shown in this paper, the results obtained under laboratory conditions are not always reflected in the operation of real generation systems. Field tests carried out by the authors on a real wind power plant connected to a medium-voltage grid confirmed (according to the tests carried out on wind power plants connected to a high-voltage grid) a positive impact of the source on the quality parameters of electric energy transmitted through the grid. An increase in the power generated in the tested Vestas V90 power plant resulted in a decrease in the values of both the asymmetry factor (zero and reverse) and the total voltage distortion factor THD_U.