Superposition of Voltage Disturbances Generated by Direct Current Traction

Abstract

Electric traction vehicles require a power supply with appropriate voltage parameters. In Poland, a direct voltage power supply system has been adopted. This requires the use of traction substations with the conversion of alternating voltage to direct voltage, which generates a number of disturbances related to the quality of electric power. Dynamic changes in the substation load resulting from the movement of traction vehicles are another cause of voltage changes. The article presents an analysis of measurement results carried out simultaneously in systems powering various traction substations and non-traction demand lines. Due to the fact that non-traction demand lines supply both recipients related to the movement of traction vehicles (including railway stations) and other industrial and municipal recipients, the quality of energy is therefore very important.

1. Introduction

Rail transport, especially in Europe, requires a supply of electric energy used to drive traction vehicles (trains). There are several systems that supply electric traction [1,2,3,4,5].

In Poland, rail transport is powered by direct current traction systems. This solution requires the conversion of the alternating voltage of the power system (frequency 50 Hz) to the direct voltage of the traction network. The conversion of alternating voltage to direct voltage takes place at traction substations. Traction transformers are installed at traction substations to reduce the voltage of the supply lines to the level of the input voltage of the AC/DC converter sets. Converters are used to convert voltage. Due to the use of converters, electric traction can be classified as a non-linear recipient of significant power, which is supplied from the power system. The non-linear nature of electric traction affects the quality of electric energy in the power system, mainly in the voltage deformations in the lines supplying traction substations [6,7,8,9,10,11,12,13].

Further disturbances generated by electric traction result from the variable nature of the load on traction substations. Starting in heavy trains, accelerations, especially units with significant unit power, affect changes in the voltage value supplying traction substations. Additionally, voltage disturbances are propagated to the circuits (lines) of their own and non-traction needs. Due to the fact that non-traction lines supply both recipients related to the movement of traction vehicles (including railway stations) and other industrial and municipal recipients, the quality of energy is therefore very important. The supplier, in this case PKP Energetyka, should provide the recipient with energy with appropriate parameters, specified in the regulation [14].

The impact of direct current traction substations was discussed, among others, in publications [15,16]. This article presents the results of further research on the impact of direct current electric traction on the quality of voltage supplying traction substations. The research was extended to the analysis of the superposition of disturbances caused by various traction substations supplying a specific part of the electric traction. Additionally, the results of measurements of parameters characterizing power quality, which were recorded in the line (Un = 110 kV) supplying a traction substation with single-stage energy transformation, were presented.

2. Places and Methods of Measurements

The first stage of the research was to collect the data that will be the starting point for further analysis of disturbances generated by traction substations. Two adjacent traction substations were selected, from which the same part of the electric traction is supplied. Figure 1 shows the electric traction power supply system. Twelve-pulse rectifiers are installed in the substations.

The article contains an analysis of power quality indicators recorded during several measurement cycles. Measurements were carried out in lines supplying traction substations, non-traction needs lines, substation auxiliary needs circuits, and rectifier unit supply circuits. Four PQM 711 analyzers and two Memobox 800 analyzers were used to record the power quality parameters. The PQM-711 analyzer is manufactured in Poland by Sonel, while the Memobox 800 analyzer was manufactured by the Swiss company LEM. The measurements were taken simultaneously at six points, at two traction substations. The analyzer connection points at one of the substations are shown in Figure 2.

The analyzers were connected to the measurement circuits using the current and voltage transformer systems installed at the substation (indirect measurement). Figure 3 shows the method of connecting the PQM—711 m in the measurement fields of the substation’s power supply line and non-traction circuit lines.

The traction substation is supplied from the main supply point by two lines with a nominal voltage 15 kV. Measurements were taken at the traction substation at the following points: A—the line supplying the traction substation, B—the line for non-traction needs, C—supply of rectifier units. The measurements were taken in several cycles, among others, in accordance with the guidelines included in the regulation [14], covering periods of one week of recording. In order to estimate the impact of dynamic changes in the traction substation load on the parameters characterizing the quality of voltage, recordings were also made at short measurement intervals. A significant part of the presented data is expressed in percentages, relating them to nominal values. This is related to the fact that the permissible ranges of voltage changes given in [14,17] are also specified in percentages.

3. Traction Substation as a Non-Linear Receiver

Due to the fact that electric traction requires the use of AC to DC voltage converters, the current in the lines supplying traction substations is deformed (in relation to the sinusoidal course). The degree of current curve deformation is influenced by many factors, e.g., the value of the short-circuit power, the type of converters used, the method of powering the traction substation (single-stage, two-stage transformation), the power and nature of receivers supplied from the non-traction demand lines and auxiliary circuits, and the type of devices used at the substation to reduce current distortion.

Figure 4 shows the current waveforms recorded at traction substation A in (a) circuits supplying rectifier units, (b) non-traction demand lines, and (c) the line supplying the substation.

The current flowing in the line supplying the substation is, therefore, the sum of the currents drawn by all receivers supplied from the substation. Of course, due to the power, traction vehicles have the greatest impact.

One of the effects of the current flowing in the line supplying the substation is voltage distortion in the PCC. As a result of the distorted voltage drop on the impedance of the supply line, which is subtracted from the voltage in the Main Supply Point, the traction substation receives a distorted voltage supplying the traction substation.

As already mentioned, the degree of voltage distortion is influenced by the short-circuit power of the network (impedance of the supply line). The nature of voltage distortion (degree of content of individual harmonics) is also influenced by the type of converter used in the substation.

Figure 5 shows the spectrum of current harmonics and the spectrum of voltage harmonics recorded in the supplying line substation A—measurement point A.

The characteristic current spectrum for the 12-pulse rectifier unit is visible, which results in a similar character to the voltage distortion spectrum. Figure 6 shows the current waveform recorded in the line supplying substation B with 12-pulse rectifiers.

The characteristic feature is the current harmonic spectrum, which translates into the voltage harmonic spectrum—Figure 7.

The voltage distortion value supplying the substation is not only influenced by the electric traction power supply system. The voltage distortion resulting from the nature of the receivers connected to the power system is also very important. Figure 8 shows the changes in the THDU coefficient recorded in the lines supplying substation A and substation B.

The course of changes in the THDU coefficient is similar to the changes in the load of the power system. Additionally, especially in the case of substation B, the impact of direct current electric traction is visible. It should be noted that the voltage distortion is a superposition of disturbances resulting from the impact of non-linear receivers supplied from the power system and the non-linear nature of the traction substation related to the use of voltage converters. Due to the fact that substation A and substation B are adjacent to each other, the courses of the recorded changes in the current distortion (THDI) are similar—Figure 9.

Figure 10 and Figure 11 compare the changes in the THDU and THDI coefficients recorded simultaneously during a week at substation A and substation B. In the case of substation A, the short-circuit power of the supply line, in relation to the power of electric traction vehicles, is much higher than the short-circuit power of the line supplying substation B. This results in greater voltage distortion (at substation B), especially during dynamic load changes.

The lines feeding the traction substations also supply electrical power to the traction substation’s own circuits. The quality of the voltage in the line feeding the traction substation affects the quality of the voltage in the substation’s own needs line.

The voltage curve distortion in the lines supplying the traction substations (substation A and substation B—Un = 15 kV) influences the degree of voltage distortion in the substation’s auxiliary circuits (Un = 0.4 kV)—Figure 12.

Figure 13 shows the THDU correlation at the voltage level U = 15 kV and U = 0.4 kV. The correlation coefficient is: r_THDU = 0.933.

4. Changes in the Average Effective Voltage Value

When analyzing changes in the effective voltage value, the time (measurement interval) used during measurements must be determined. Then, the results obtained must be related to the permissible levels (ranges of changes) adopted in applicable legal acts or standards. For example, in the regulation [14] applicable in Poland, for the III–V connection group, each week, 95% of the set of 10 min average effective values of the supply voltage should be within the deviation range of ±10% of the rated voltage. Figure 14 shows the changes in the average effective voltage value recorded in the line supplying substation A.

Due to the values of the correlation coefficients between the individual voltages, which are r_UL12UL23 = 0.999, r_UL23UL31 = 0.999, and r_UL31UL12 = 0.998, respectively, it is possible to take into account the voltage changes in only one phase for analysis. At the same time, the recorded voltage values in the supply lines of the base have characteristic voltage changes in the power system—Figure 15.

A similar pattern of changes in the average effective voltage value was recorded in the traction substation’s own circuits—Figure 16.

The correlation coefficient between the changes in the average 10 min effective voltage value recorded in the line supplying the substation (Un = 15 kV) and the non-traction circuits (Un = 0.4 kV)—Figure 16, is r_U15kVU0.4kV = 0.996. The average 10 min effective values represent only a certain trend of voltage changes, in this case over a period of one week. Figure 17 shows the changes in the average 10 min effective voltage values (U_L1mean) recorded over a week, in one of the phases of the line supplying the traction substation.

The 10 min average effective voltage value varies between 101.82% and 106.68% of the rated voltage of the supply line, assumed as Un = 100%. This means that the measured voltage is within the allowable range of changes.

Due to the adopted 10 min averaging period, when analyzing only the average value, rapid, often very short, voltage changes may be omitted.

Figure 18 shows the changes in the minimum and maximum voltage values (U_L1min, U_L1max) recorded during the week in the line supplying the traction substation.

The voltage changes in Figure 17 and Figure 18 were recorded at the same time.

5. Voltage Swells and Dips

During the measurements, several sudden voltage changes were recorded in the lines supplying traction substations, non-traction lines and the substation’s own circuits. Figure 19 shows the changes in the minimum and maximum effective voltage values that were recorded during the week in all phases of the line supplying the traction substation—substation A.

Table 1. Summary of events recorded in power line—substation A.

Phase L1,L2,L3	<20 ms	20.. < 100 ms	100.. < 500 ms	0.5... < 1 s	1… < 3 ms
Surge > 10.00%	2	38	36		2
Dip > 10.00%
10 < 15%	3	6
10 < 15%		17	2
10 < 15%		13			1
10 < 15%		10	1
Interruption

(UNIPEDE Table) contains a summary of events recorded in power line—substation A.

Similar events were also recorded in the line feeding substation B and are shown in Figure 20 and

Table 2. Summary of events recorded in power line—substation B.

Phase L1,L2,L3	<20 ms	20.. < 100 ms	100.. < 500 ms	0.5... < 1 s	1… < 3 ms
Surge > 10.00%	10	4	12	2	2
Dip > 10.00%
10 < 15%	5	4
10 < 15%	15	6	2
10 < 15%	4	1
10 < 15%		1			1
Interruption

.

It should be emphasized that the duration of some voltage disturbances ranged from a few to several dozen milliseconds. For example, Figure 21 shows the voltage and current oscillograms recorded in the line supplying substation A. A voltage dip (Phase L1) is visible, causing a change in the current in the line.

Figure 22 shows the changes in the effective value of voltage and current corresponding to the disturbance (voltage sag) shown in Figure 21.

Figure 23 shows the current and voltage oscillograms recorded during a disturbance in the line feeding substation B.

6. Dynamic Changes in Substation Load and Supply Voltage Quality

One of the other factors influencing the quality of electric power in traction substation circuits is the nature of the substation load changes resulting from the movement of traction units. Figure 24a shows the changes in current and voltage recorded in the network supplying the traction substation. Figure 24b (enlarged) shows the start-up of a traction unit.

The dynamic increase in the substation load is related to the start-up of the traction unit. This causes an increase in current, a decrease in the RMS voltage supplying the substation, and a deformation of the voltage curve—Figure 25.

The correlation coefficient defining the dependence of changes between THD_U and current at the level of Un = 15 kV is r_{THDU_I} = 0.926—Figure 26

7. Single-Stage Voltage Transformation

One of the ways to reduce the impact of electric traction on the power system is to supply traction substations with a higher voltage (single-stage transformation). Figure 27 shows a diagram of the power supply of a traction substation with single-stage energy transformation.

The methods of powering a traction substation with single-stage energy transformation have been discussed, among others, in publications [18,19,20].

In Poland, such a solution is used, among others, in high-current railways. This is related to the higher power of traction units powered from the substation. For example, the used ED250 Pendolino electric traction units are characterized by a power of 5664 kW [21]. This requires ensuring the supply of electricity with appropriate parameters. Traction units with such significant power also affect the quality of energy on the lines supplying traction substations.

In order to assess the impact of the traction substation (with single-stage energy transformation) on the power system, a number of measurements were taken. The analyzer of electric power quality indicators was connected to the measuring circuits of the line supplying the substation with a voltage of 110 kV (indirect measurement using transformers). The analysis of measurement results was presented, among others, in [22].

The use of single-stage energy transformation (110 kV/15 kV) causes the short-circuit power at the PCC point (the connection point of the traction substation with the power system) to be much higher than the short-circuit power of the line with two-stage energy transformation. Despite the fact that the substation load is much higher than in the previously analyzed cases, the impact on the voltage curve deformation is smaller—Figure 28.

The voltage waveform distortion in the line supplying the traction substation (with a rated voltage of Un = 110 kV) is influenced by changes in the substation load and other non-linear receivers supplied from the power system. Direct supply of the traction substation from a high voltage line (110 kV) ensures the greater quality and reliability of the power supply. Analyzing the changes in voltage (U_max, U_min) and current I_max, only one significant voltage dip was recorded—Figure 29.

One of the effects of voltage dips in the line supplying the traction substation is a sudden increase in the flicker values P_st and P_lt—Figure 30.

The recorded rapid voltage changes occur sporadically (several changes within a week). As a result, no annoying flickering phenomenon was observed.

8. Discussion

In Poland, direct current electric traction is one of the largest recipients supplied from the power system. In the case of receivers with power of several MW with dynamically changing load or non-linear character, parameters characterizing power supply conditions are very important. A very important parameter is the value of the short-circuit power of the line supplying traction substations. The short-circuit power at the Main Supply Point (MSP_A) of the traction substation (substation A) is S_CC = 497 MVA (at the supply line voltage level Un = 15 kV). The substation is supplied by a 500 m cable line. In the case of the second of the tested substations (substation B), the short-circuit power at the Main Supply Point (MSP_B) is S_CC = 196 MVA (at the supply line voltage level Un = 15kV). The substation is supplied by an overhead line 1210 m long.

One of the disturbances recorded in the lines supplying substation A and substation B are sudden voltage dips and swells. Figure 31a shows examples of such disturbances recorded simultaneously in the lines supplying the substation in phase L1. Figure 31b (enlarged) shows dips and swells that were recorded simultaneously in the line supplying substation A and substation B. Disturbances of a similar nature occurred in the other phases.

One of the effects of voltage dips and surges is the phenomenon of light flicker. The phenomenon of light flicker is characterized by the flicker coefficients, short-term P_st and long-term P_lt. Figure 32 shows the changes in the minimum U_min, maximum U_max voltage values and the short-term flicker index P_st of the light recorded during the week in the lines supplying traction substations.

A sharp increase in the P_lt coefficient was recorded several times during the week. At substation A, the long-term flicker index P_lt was exceeded in phase L1. The 95% value of P_lt is 1.816—

Table 3. Summary of parameters recorded during the week in the line supplying substation A. Blue—Maximum value above limit value. Red—95%-value above limit value.

Parameter			Maximum Value			95%-Value
	Unit	En50160.gwd	L1	L2	L3	L1	L2	L3
Voltage variations		8670.00V
Maximum 100/95% Minimum 100/95%	% [Un] % [Un]	+10/+10 −15/−10	8.22 −4.99	5.28 −8.22	7.49 −6.33	7.82 −4.99	3.78 −3.78	6.97 −6.33
Interrruption < 1%	Number of	100	1	1	1
Events	Number of	100	28	24	28
Harmonics
23. Harm.	% [Un]	1.50	0.71	0.64	0.70	0.55	0.49	0.54
Flicker Plt	Plt	1.00	3.301	3.330	3.293	1.816	0.799	0.856
Unbalance		2.00	0.48			0.40
Signallingvoltages	% [Un]		No
frequency		50 Hz
Maximum 100/95% Minimum 100/95%	% %	+4/+1 −6/−1	0.60 4.00			0.20 −0.20

.

Substation A has significantly better power supply conditions than substation B. The short-circuit power of the line supplying substation A is S_CCA = 497 MVA, while the short-circuit power of the line supplying substation B is S_CCA = 196MVA. The ratio of short-circuit power is: S_CCA/S_CCB = 497 MVA/196 MVA = 2.535. As a result, the impact of direct current traction on power quality is greater. Similarly to substation A, the parameter that exceeded the permissible values specified in the regulation was the flicker index. The long-term flicker index was exceeded in phases, L1, L2, and L3. The 95% value P_lt was: P_lt1 = 3.05, P_lt1 = 2.73, P_lt1 = 2.41, respectively—

Table 4. Summary of parameters recorded during the week in the line supplying substation B. Blue—Maximum value above limit value. Red—95%-value above limit value.

Parameter			Maximum Value			95%-Value
	Unit	En50160.gwd	L1	L2	L3	L1	L2	L3
Voltage variations		8670.00V
Maximum 100/95% Minimum 100/95%	% [Un] % [Un]	+10/+10 −15/−10	6.03 1.19	5.43 0.49	4.49 −0.28	5.68 1.19	5.00 0.49	4.13 −0.28
Interrruption < 1%	Number of	100	0	0	0
Events	Number of	100	47	41	43
Harmonics
5. Harm.	6.00	1.50	1.85	1.83	1.80	1.68	1.66	1.61
Flicker Plt	Plt	1.00	4.431	4.267	34.385	3.058	2.731	2.412
Unbalance		2.00	0.38			0.38
Signallingvoltages	% [Un]		No
frequency		50 Hz
Maximum 100/95% Minimum 100/95%	% %	+4/+1 −6/−1	0.60 −0.40			0.20 −0.20

.

Figure 33 shows the changes in the flicker coefficients P_st and P_lt recorded during the week. The maximum values and 95% values are marked in red and blue in Table 4.

Figure 34 and Figure 35 present the statistical analysis of the flicker indices corresponding to the curves in Figure 33.

Statistical analysis shows that the 95% value of the flicker coefficient P_lt = 3.058 is greater than the 95% value of the flicker coefficient P_st = 0.308. This is contrary to the definition of flicker indices. The long-term flicker coefficient is determined from the following formula:

$$P_{lt}=\sqrt[3]{{\sum }_{i=1}^{12}{\displaystyle \frac{P_{st}^3}{12}}}$$

(1)

In the case of sudden voltage disturbances (recorded in the line supplying the traction substation) lasting from milliseconds to several seconds (voltage dips and swells, Figure 32), there is a significant increase in the short-term flicker index P_st (P_{st_max}=7.114). With a sliding “measuring window”, this causes a disturbance in determining the long-term flicker index P_lt.

In the case of a traction substation with single-stage energy transformation, the impact of dynamic load changes on the power system is much smaller. Figure 36 shows changes in the values of the voltage harmonic distortion factor THDU and current THDI.

In the case of substations supplied from 15 kV lines, the influence of voltage disturbances (voltage dips and swells) on the occurrence of the flicker phenomenon is much greater than in lines with single-stage energy transformation. With single-stage energy transformation, the short-circuit power of the 110 kV line is high enough to ensure that the flicker indices do not exceed the permissible values. Figure 37 compares the changes in the effective voltage value and the short-term flicker index P_st.

The use of substations with single-stage energy transformation significantly reduced the impact of direct current traction on the power system. However, this did not provide adequate power supply conditions for traction vehicles with significant unit power. Direct current traction with a voltage of 3 kV significantly limits the development of high-speed railways. Providing power supply with appropriate parameters is only one of the elements that make up the infrastructure related to rapid rail transport. It is worth mentioning the appropriate railway infrastructure [23,24] and the connection between the vehicle and traction [25,26].

9. Conclusions

Estimating the impact of individual traction substations on the power system requires taking into account many factors. The value of the short-circuit power of the supply network (in relation to the rated power of the traction substation) has a significant impact on the size of the disturbances generated by traction. In most cases, traction substations in Poland are supplied from medium voltage lines with a rated voltage of U_n = 15 kV. The introduction of standardization of the rated voltages of transmission and distribution lines in the power system led to a change in the supply voltage from U_n = 30 kV to U_n = 15 kV of many traction substations. This resulted in the significant deterioration of the power supply conditions of the traction system. Especially with long supply lines (between Main Supply Points—110/15 kV power stations and traction substations) the substation power supply parameters deteriorate—substation B.

The most common disturbance in the power system is voltage sags, which constitute about 95% of all disturbances. This statistic was also confirmed by the analysis of recorded indicators characterizing the quality of energy presented in the article.

In order to improve the power supply conditions of electric traction, the electric traction power supply with single-stage voltage transformation 110/3 kV is being introduced. This ensures the stabilization of the supply voltage levels in a significant range of the traction substation load. The scope of the power supply system modernization is closely related to the planned maximum speed, the rolling stock used and the traffic density.

In the author’s previous publications, the influence of individual substations on the power system was estimated. The novelty presented in the current article is the assessment of the simultaneous influence of several traction substations on the power supply system and the influence on the quality of power in circuits supplied directly from the traction transformer. The collected measurement data, selected fragments of which are presented in the article, will be helpful in modeling studies involving the assessment of the influence of the traction substation power supply conditions on the magnitude of disturbances generated by direct current electric traction.