Optimal Voltage Control in Medium Voltage Power Engineering Networks

OPTIMAL VOLTAGE CONTROL IN MEDIUM VOLTAGE POWER DISTRIBUTION NETWORKS Waldemar Szpyra / AGH University of Science and Technology in Cracow Aleksander Kot / AGH University of Science and Technology in Cracow

1. VOLTAGE DEVIATIONS LIMITATIONS Limitations for voltage deviations in power distribution networks are given in the Ordinance of the Minister of Economy dated May 4, 2007 on specific conditions of electric power systems. (Journal of Laws no 93 dated 29 May 2007, item 623) [17], called in short “system regulation”. The Ordinance imposes a duty on power line operators to observe specified quality parameters of supplied power. It stipulates that in a network operating without disturbance, every week 95% out of a set of 10 minute average values of effective input voltage should remain within the following deviation range: • for consumers classified as group I and II connected to a grid:of nominal voltage: Un = 400 kV: +5% /–10%Un Un = 220 i Un = 110 kV: ±10%Un • for consumers classified as group III ÷ V (supplied with network of nominal voltage below 110 kV) – every week 95% out of a set of 10 minute average values of effective input voltage should be in the deviation range ±10% rated voltage. For consumers of group I and II – power quality parameters may, in their entirety or in part, be substituted by other parameters specified in the power sales contract or in a contract for rendering power transmission and distribution services. Failure to observe quality standards of supplied power to consumers from group III, IV and V specified in the system regulations [17], entitles consumers to bonuses and discounts. The discount values are determined according to §37 of the Ordinance of the Minister of Economy dated 2 July 2007 on specific principles governing the calculation tariffs and financial settlements in electric power trading (Journal of laws dated 18 July 2007, no 128, item 895) [16]. Both Ordinances were issued on the grounds of delegation stipulated in the Act on Energy Law [21]. The Acts referred to above do not define the term “grids operating without disturbance”. “Instructions of Transmission System Operation and Maintenance” [4], on the other hand, define disturbance as: “Unplanned automatic or manual shut-down(s) or impossibility to keeping of the expected operating parameters of the components of network assets. The disturbance can take place with or without the damage to the network assets”. A conclusion may be drawn from the definition that energy quality parameters need not be met in systems other than typical/normal systems. Thus, the regulations related to voltage in distribution networks in force today are more liberalised as compared to those binding before the system regulation of May 2007 became effective (regulations on operations of power engineering systems issued before 2007 did not stipulate any restrictions as to “networks operating without disturbance”. Abstract Power flow in elements of the network causes voltage drops in these elements. Therefore, in order to ensure the proper voltage of electric power delivered to consumers it is necessary to regulate voltage in power engineering grids. The article presents voltage requirements in power engineering grids, the impact of regulation on losses in distribution lines and various criteria for optimising voltage regulation. Depending on the adopted criteria, indications for tapping switch settings in transformers and input voltage may differ for various lines or even be quite the opposite.

89

90

Aleksander Kot / AGH University of Science and Technology in Cracow Waldemar L. Szpyra / AGH University of Scie nce and Technology in Cracow

2. DEVIATION AND VOLTAGE DROP BALANCE

The deliberations refer to a MV network supplied by a HV/MV transformer installed in a power distribution substation (DS). A single transformer feeds 6 ÷ 10 medium voltage lines. Each line feeds several to few dozen medium voltage/low voltage (MV/LV) transformer stations, which supply low voltage circuits. An example of the supply system from distribution substation to the consumer connected in point k of the low voltage network with marked deviations and voltage drops is given in Fig. 1.

LV line

LV line Customer load

Customer load

Fig. 1. Deviations and voltage drops in the system from distribution substation to the consumer connected in point k to the low voltage system. Key: R – point of network split; TS - MV/LV transformer station; other designations in text

In the case given in Fig. 1 the voltage deviations δU in point k of the low voltage system may be determined from the deviations and voltage drop datasheet expressed in the following equation: (1) where: δUnn – is the voltage drop in the low voltage system from the TS to point k; ΔUT – drop in voltage in the MV/LV transformer; δUzT – voltage deviation resulting from the position of the tap changer to medium/ low voltage transformer ratio control; ΔUSN – voltage drop in medium voltage network; δUz – deviation voltage in network supply point; Unn – rated voltage of low voltage lines; δUϑ – voltage deviation resulting from the difference between the relations of the transformer rated voltage and the network rated voltage: (2) where: ϑnT – rated MV/LV transformation ratio; ϑnS – the relation of medium network and low network rated voltages; UnG – rated voltage of the MV winding of MV/LV transformer, UnD – rated voltage of the LV winding of MV/LV transformer, USN – rated voltage of MV networks, Unn – rated voltage of LV networks,. Voltage deviation in any point of medium and low voltage system must comply with the range given in the system regulation, i.e. (3) In a normally operating system the maximum voltage deviations occurs in the end of the low voltage line supplied by the most loaded TS distanced from the grid feeding medium voltsage network (usually close to the point of network split). Minimum voltage deviations occur in the case of minimum network load at the beginning the low voltage system, fed by the TS located near DS. To assess the voltage in distribution networks it is necessary to know all the elements of the deviation and voltage drop datasheet (1). Usually a model, reflecting precisely the network parameters from the grid feeding

Optimal Voltage Control in Medium Voltage Power Engineering Networks

MV network to low voltage busbars in TS, is built to analyze a distribution network. Precise models of low voltage networks are not developed. This results from the big number and diversification of low voltage circuits in the system. In practice calculations are made of the power current flow and voltage drop in the MHV network to assess the network operating conditions. Therefore, it is justified to specify the voltage drop limits, i.e. values that allow for the present voltage adjustment system to maintain the deviation level for the consumer within the admissible range. Specifying the admissible voltage drop in the MHV network is possible provided values of certain datasheet components are adopted (1). The most often made assumptions involve: • use of the full admissible voltage range on MV bus bars is possible in the network feeding point, which means that deviation of input voltage δUz may read +10%Un (control practice used by power distributers often restricts the supply voltage deviation in the MHV network in GPZ to δUz = +5%Un – primarily because bigger customers belonging to group III have their own MHV/LV transformer stations) • voltage deviation resulting from the difference between the relations of the transformer rated voltage and the network rated voltage (δUϑ) shall be compensated by relevant setting of tap changer (δUzT) • a voltage drop in the MV/LV transformer (ΔUT) is calculated using the average known transformer load in the circuit and its rated parameters • low voltage networks are designed according to guidelines given in [22] and thus we can assume that voltage drops occurring in these systems (ΔUnn) do not exceed the values given in the last column of Table 1. Table 1. Admissible voltage drops in medium and low voltage lines according to guidelines given in [22] Specification

MV network

Low voltage network

normal

disturbed

Towns supplied by 110 kV/MHV lines located within town borders

2%

4%

4.5%

Towns supplied by MFP located within town borders

8%

10%

(3÷4.5)%

Towns supplied by distanced MFP

8%

13%

(7.5÷10)%

Industrial consumers Supplied from regional grid

8%

13%

(3÷4.5)%

Voltage drop in a MV/LV transformer may be calculated according to the following formulae: (4) where: SN – transformer rated power [kVA]; S – transformer load [kVA]; cosφ – transformer load coefficient (ratio of active to complex transformer demand); ur – active component of transformer’s short-circuit voltage [%]; ux – reactive component of transformer’s short-circuit voltage [%]; (5)

(6)

91

Aleksander Kot / AGH University of Science and Technology in Cracow Waldemar L. Szpyra / AGH University of Science and Technology in Cracow

92

where: Pk – load loss of transformer (cooper losses) [kW]; uk – transformer short circuit voltage [%]. Table 2 presents the voltage drop value for typical MV/LV transformers used in distribution systems in the load function S/SN, and load coefficient of cosφ = 0.9. Table 2. 15.75/0.4 kV transformer voltage drop under varied load Sn

Pk

uk

Transformer load factor S/Sn 0.3

0.4

0.5

0.6

0.7

0.8

0.9

Voltage drop in transformer ΔUT [%]

kVA

kW

[%]

63

1.20

4.5

1.19

1.59

1.99

2.39

2.79

3.19

3.59

75

1.85

4.5

1.24

1.65

2.06

2.48

2.89

3.31

3.72

100

1.75

4.5

1.19

1.59

1.98

2.38

2.78

3.18

3.58

160

2.25

4.5

1.05

1.41

1.76

2.11

2.47

2.82

3.18

200

3.90

4.5

0.94

1.26

1.57

1.89

2.20

2.52

2.84

250

3.00

4.5

0.99

1.33

1.66

1.99

2.33

2.66

3.00

400

4.25

4.5

0.91

1.21

1.51

1.82

2.12

2.43

2.74

630

6.10

6

1.09

1.46

1.83

2.20

2.58

2.95

3.32

The average peak transformer load in distribution networks reaches 40÷50% of rated power, which means that the average voltage drop in transformers should not exceed 2%. Assuming that: δUz = +10%; δUϑ = –5%; δUzT = +5%; ΔUT = –2%; ΔUnn = 10% and admissible voltage deviation for consumers supplied from low voltage networks, amounting to δUdop = 10%, the voltage drop and deviations (1) show that the maximum voltage drop in MV lines should not exceed:

= 10 – 5 + 5 – 2.5 – 10+10 = 7.5 [%] This means that full range voltage regulations in DS allow for assuring the required low voltage level at the consumer’s end (voltage drop in MV network amounting to circa 7.5%).

3. VOLTAGE REGULATION MEANS Voltage deviation in low voltage networks can be controlled: 1. without investment outlays – using transformer’s regulation capacity, i.e.: a. change of input voltage to the MV network – regulating voltage on MV busbars in DS – by changing the HV/MV transformation ratio operating under load, by ±10% in 8 steps or ±16% in 12 steps b. change of MV/LV transformer ratio control while the transformer is switch-off),– the extent of change depends on the transformer’s year of built and reaches: δUzT = {-5%, 0%, +5%} or δUzT = {–2,5%, 0%, +2,5%, +5%, +7,5%}. 2. investment related – applying additional technical means to reduce the drop in network voltage, i.e.: c. installing condenser batteries to compensate reactive power d. installing condenser’s in series to compensate line reactance e. installing controlling auto transformers in series (buck transformers) f. connecting new circuits to DS taking over delivery to some of the TS g. shortening low voltage circuits by adding new TS. Voltage control options, resulting from application of means mentioned above, are limited because: a. higher input voltage to the medium voltage network is limited by the maximum voltage upward deviations limited by inequality (3), and sometimes by contract conditions with consumers

Optimal Voltage Control in Medium Voltage Power Engineering Networks

93

b. MV/LV transformation ratio is connected with consumer trip off and results in generating costs of the rigging team and is in practice rarely applied (once or twice a year or even less often). Additionally, the increase of rated voltage in low voltage networks in 2003 resulted in growing maladjustment of transformer ratio control and relation of rated voltage in medium and low voltage networks – the value of δUϑ was modified from +0,25% to –5% and as effect up to 5% of the medium/ low voltage transformer regulation range is used for compensating the effects of increased voltage in the low voltage network c. application of additional technical means to reduce voltage drop requires considerable investment outlays, which in practice rarely give the opportunity for return on investment. Every decision related to investments aimed at improving voltage in the network should be preceded by a detailed technical and economic analysis of various solutions to the problem.

4. IMPACT OF VOLTAGE REGULATION ON NETWORK LOSSES The impact of changes in voltage on the network demand is described by the static voltage characteristics of network demand [1, 2, 9]. In the case of minor voltage deviations (±5% Un), changes in network demand are described by coefficients of voltage static characteristics of the network active demand α and reactive demand β. These coefficients show the percentage shift of active and reactive network demand by one percent voltage change. According to [2] the active and reactive network demand on real voltage Ur may be calculated using approximated relationships: (7) (8) where: Pr, Qr – are active and reactive network demand at real voltage Ur respectively; Pn, Qn – active and reactive network demand at nominal voltage respectively; α, β – factor of voltage static characteristics of active and reactive demand respectively; Un – nominal voltage; δU – deviation of supplying voltage: (9) The value of angle factor of voltage static characteristics of active power drawn from the network is given in Table 3 with the value of coefficient of voltage static characteristics of reactive power given in Table 4. Table 3. The value of factor of voltage static characteristics of active demand α for selected types of power distribution systems

Types of power distribution systems

Source:

Value of factor of voltage static characteristics of active demand α within hours of:: Morning peak load

Evening peak load

Night load

[2]

0.90÷1.20

1.50÷1.70

1.50÷1.60

[2]

0.60÷0.70

1. 40÷1.60

1. 40÷1.60

Rural networks

[2]

0.50÷0.68

1.50÷1.60

1.50÷1.60

20 kV network of Distribution Company X

[1]

1.20

1. 46

–

15 kV network of Distribution Company Y

[9]

1.15

2.25

0.95

Grid suppluing big towns with small industrial consumers Grid suppluing small towns with small industrial consumers

94

Aleksander Kot / AGH University of Science and Technology in Cracow Waldemar L. Szpyra / AGH University of Science and Technology in Cracow Table 4. The value of angle factor of voltage static characteristics of reactive power β for selected consumers, with voltage deviation in the range ±5%Un Types of power distribution systems

Source:

Value of factor of voltage static characteristics of reactive demand β within hours of: Morning peak load

Evening peak load

Night load

[2]

3.00

2.60

3.10

[2]

3.00

2.60

3.10

Rural networks cosφ ≥ 0.85 0.80 ≤ cosφ < 0.85 0.70 ≤ cosφ < 0.80 cosφ < 0.70

[2]

2.30 2.50 2.80 3.10

2.60

3.10

20 kV network of Distribution Company X

[1]

2.85

4.14

–

15 kV network of Distribution Company Y

[9]

5.95

2.60

2.30

Grid suppluing big towns with small industrial consumers Grid suppluing small towns with small industrial consumers

The impact of supplying voltage changes as well as transformation ratio control on network demand, power and energy losses can be traced on the example of a simple medium voltage circuit comprising of a medium voltage line, MV/LV transformer and low voltage demand. The circuit and equivalent diagram are presented in Fig. 2. Voltage on consumer terminals may be changed by changing the supplying voltage Uz and/or position of the transformer tap changer resulting voltage deviation δUzT. Various combinations of input voltage changes and transformer ratio control are possible. Two cases involving extreme changes are given below: a. changes of input voltage Uz with simultaneous changes position of the transformer tap changer by δUzT, so that the voltage on the consumer terminals Uo remains unchanged b. changes of input voltage Uz with no adjustment of transformation ratio control so resulting in voltage change on the consumer terminals Uo

Fig. 2. Medium voltage circuit and its equivalent diagram

4.1. Adjusting input voltage with simultaneous changes of transformer ratio control Simultaneous changes in input voltage feeding the network and change of transformer ratio control so that voltage in consumer terminals remains unchanged Uo = const – power (and energy) supplied by the network for delivery remains the same. However, the following changes take place: • current in the supply line – inversely proportional to voltage change • power loss of transformer idling – proportional to the square of voltage change value. Changing current causes change a load loss in the circuit – proportional to the square of that change. Loss of energy in the circuit is also subject to change. In this case the direction of loss change depends on: the direction of changed voltage, circuit load and volume of transmitted energy.

Optimal Voltage Control in Medium Voltage Power Engineering Networks

Relative change of energy loss, [%]

Example 1. For the transmission system as in Fig. 2 energy loss was calculated in three variants differing in terms of line input voltage Uz and the location of the transformer tapping switch. Variant „Un”: Uz = Un = 15.0 kV, tapping switch position δUzT = 0% Variant „1.05Un”: Uz = 1.05Un = 15.75 kV, tapping switch position δUzT = – 5%” Variant „0.95Un”: Uz = 0.95Un = 14.25 kV, tapping switch position δUzT = +5%. In the case of such line input voltage and transformer tapping switch positions the output voltage in consumer terminals remains the same for every variant. For each variant calculations were performed using following data: line elementary impedance R0 = 1.227 Ω/km, elementary reactance X0 = 0.398 Ω/km, line length l = 1 km; three values of utilization periods of maximum losses: τ = {1 670; 2 580; 3 560} h/a. (which corresponds to the following utilization periods of peak load: Ts = {3 000; 4 000; 5 000} h/a; transformer load changes range from 25 to 625 kW with power factor cosφ = 0.94; transformer parameters: rated power Sn = 630 kVA, rated transformer ratio ϑn = 15,0/0, 4 kV, load loss Pk = 6.1 kW, no load loss P0 = 0.97 kW, short circuit voltage uk = 6%, idle current i0 = 1%. Calculation results are given in Fig. 3 in graph form showing the relative changes in energy loss in the transmission system in terms of transformer load. The chosen point of reference was the energy loss in the system (corresponding to the given load) calculated at zero voltage deviation (Variant „Un”).

Transformer demand factor, [%Sn]

Fig. 3. A relative change of energy loss as a function of transformer demand factor in the case of simultaneous changes supplying voltage and transformation ratio control

The graphs show that when the transformer is underloaded an increase in input voltage concurrent with the same relative increase of transformer ratio cause energy loss in the system. The loss increases with the decrease in transformer load and the shortening of time intervals of peak power consumption. For example in result of growing input voltage and transformation ratio by 5% and transformer load of So = 30% Sn and time values for peak power consumption Ts = 3 000 h/a, energy loss grows by less than 8.5% and in time value Ts = 5 000 h/a circa 6.5%. Relative loss changes diminish with growing transformer load (when the transformer load exceeds a specific value the direction of change switches to the opposite sign, i.e. losses decrease with growing voltage). On the basis of graphs in Fig. 3 we can state that in the case of adjustments involving simultaneous changes of input voltage and transformer ratio, a change in energy loss direction in the system depends above all from the system load and time intervals of peak power consumption. In the case of small load and short time intervals of peak power input voltage should be decreased and simultaneously the transformer ratio reduced in order to reduce losses. In contrast, with big loads and long time intervals of peak power consumption, input voltage and transformer ratio should be increased. On the other hand, the comparison of calculation results for two line lengths indicates that losses decrease with falling transformer loads – resulting from the impact of bigger voltage drop on transformer idle loss.

95

Aleksander Kot / AGH University of Science and Technology in Cracow Waldemar L. Szpyra / AGH University of Science and Technology in Cracow

96

Fig. 4 shows the range of load power factor cosφ depending on the value of factor of voltage static characteristics of the active power α, when load loss of power in the circuit fall with the growing voltage. The curves represent three values of factor of voltage static characteristics of reactive power β. The curves are constructed so that the intersection point of the straight line, representing load power factor value cosφ, with the straight line representing the value of factor of voltage static characteristics of the active power α, lies above the curve representing the value of factor of voltage static characteristics of the reactive power β, then load power loss in the circuit diminishes with the growing voltage.

1.00

Power factor cosφ

0.95 0.90

6.0 2.6 2.3

0.85 0.80 0.50

0.55 0.60 0.65 0.70. 0.75 0.80 0.85 0.90 0.95 Factor of voltage static characteristics of active power α

1.00

Fig. 4. Range of power factor cosφ in terms of factor of voltage static characteristics of active power α, when load loss of active power decreases with the growing voltage

Table 3 indicates that the factor of voltage static characteristics of the active power α is less than one in principle only during the morning peak load when the value of load power factor is low (power factor values cosφ in MV networks are given in Table 5). In practice this means that situations of network load loss decrease while supplied voltage grows occur very rarely. Table 5. Power factor cos� in medium voltage network in various seasons, days and hours [15] Kind of network

Season (daytime)

Network supplying big town

Network supplying rural areas

Power factor cosφ Before noon

In the evening

At night

Winter (workday)

0.86

0.89

0.77

Summer (workday)

0.74÷0.80

0.74÷0.80

0.63

Winter (workday)

0.50÷0.70

0.98

0.98

Summer (workday)

0.52÷0.67

0.78÷0.98

0.90÷0.98

Summer (Sunday)

0.88

0.98

0.78÷0.93

Generally, input voltage growth accompanies growth of load losses in the circuit. Only in the cases of high power factors cosφ, in that time of the day when the factor of voltage static characteristics of the active power α < 1, the load losses may decrease with the growing input voltage. When the factor α ≥ 1, load losses always grows together with growing supplying voltage (because of the factor α is always bigger than 1). As the factor of voltage static characteristics of the active power is always bigger than zero, growing supplying voltage will always be accompanied by the growth of active power consumption. In most cases (except for consumers requiring a fixed amount of energy for their technological process) the amount of energy used by the consumers also grows.

Optimal Voltage Control in Medium Voltage Power Engineering Networks

4.2. Adjusting input voltage with no change of the transformer ratio control Whereas the change of input voltage is not accompanied by a change of the transformation ratio, voltage changes in consumer terminals. The relative changes by δU in the supplying voltage supplying the circuit will cause almost the same relative voltage changes on the transformer terminals and consumer terminals. As a result of above voltage change the next changes will take place: • of the consumed active and reactive power supplied by the network – in compliance with the voltage static characteristics of the consumed power • of energy consumed from the network • power loss of transformer idling. Changes in the delivered power consumed result in changes of the current in the circuit, and thus the load loss. The direction changes of total power loss in the circuit depend on the power factor, ratio of transformer load and time of the day.

Relative change of energy loss, [%]

Example 2. Similarly as in example 1, calculations were made for the transmission system of power changes and energy consumed from the network, power and energy losses in the system with rated transformation ratio for three values of supplying voltage (the same as in example 1). Other parameters used in the calculations were the same as in example 1. Fig. 5 shows relative changes in energy loss against energy loss at rated voltage.

Transformer demand factor, [%Sn]

Fig. 5. A relative change of energy loss as a function of transformer demand factor in the case of changes supplying voltage and fixed transformation ratio

Fig. 5 indicates that an increase of input voltage by 5% causes growing power and energy loss of over 10%, whereas a 5% voltage drop leads to a nearly 10% reduction of power and energy loss.

4.3. Conclusions The deliberations presented above indicate that: 1. Voltage regulation in distribution networks have an impact on both power and energy loss in the network and power and energy consumption from that system, and thus on company costs and revenues. 2. In extreme cases voltage regulation, reducing power and energy loss, may result in decreasing the amount of energy used by consumers and thus reduce revenues for transmission charges. We should emphasise that the example selected well depicts the nature of changes in progress. In real network circuits supplying a bigger number of stations, operating under varied loads and time of peak consumption, the situation is not as clear. In order to determine the input voltage level and MHV/LV transformer ratio setting that is the most appropriate, in terms of loss, MHV/LV transformer ratio requires optimising calculations.

97

Aleksander Kot / AGH University of Science and Technology in Cracow Waldemar L. Szpyra / AGH University of Science and Technology in Cracow

98

5. OPTIMAL VOLTAGE CONTROL The key objective of voltage regulation is assuring voltage deviation in every point of the medium and low voltage networks within the admissible range. In the process of achieving this objective we can also optimise the voltage level in the network. The solution for optimising voltage regulation comes down to finding voltage values for MV busbars in DS feeding the network and all MV/LV transformation ratio setting values, where the target function of a specified quality control criterion reaches the optimum and is concurrently compliant with limitations resulting from admissible voltage deviations and technical capacity to effect the control (e.g. transformation ratio control range). The number of targeted voltage values for MHV busbars depends on the number of time zones per day The setting of MV/LV transformer tapping switches is the same for all time zones in the analysed period. Calculations are performed for the period of a year or separately for particular seasons, e.g. autumn/ winter and spring/summer. The solution for optimum voltage control for networks feeding n TS in time T, comprising r time intervals, is the vector determining all MV/LV transformer tapping switches settings in the analysed network, containing information on the relevant level of input voltage in DS where the target function of adopted optimum criterion reaches an extreme value. (10) where: δUzTi – voltage deviation connected with the position of the MV/LV transformer tapping switch in i station; δUzp – oltage deviation in MV busbars in DS connected with the position of the medium to low voltage 110 kV/MV transformer tapping switch in time interval p. If we assumed optimisation for a period of one year broken down to hours, the number of input voltage levels in DS to be identified would be huge and amount to r = 8 760. Thus, the vector (10) would be very long and the problem difficult to solve. Therefore, the need for decomposition. Decomposition of the problem involves a breakdown of the hourly sets to a small number of subsets called zones, where a fixed input voltage level in DS is assumed. This corresponds to agreeing on network load intervals for which input voltage to DS remains constant. Usually several (4–6) such zones are agreed. In this situation the solution vector takes the following form (11) where s – the number of time zones, s