Journal of Power and Energy Engineering, 2014, 2, 388-394
Published Online April 2014 in SciRes.
How to cite this paper: Johnston, D. (2014) On-Line Independent Tap-Changing of Each Feeder Supplied by a Low Voltage
Distribution Transformer. Journal of Power and Energy Engineering, 2, 388-394.
On-Line Independent Tap-Changing of Each
Feeder Supplied by a Low Voltage
Distribution Transformer
David Johnston
Faculty of Engineering and Environment, Northumbria University, Newcastle upon Tyne, UK
Received Dec emb er 2013
An on-line tap-changing circuit was developed for use with low voltage transformers (10 kV/380 V,
or equivalent), in which the tap positions could be set independently for each low voltage feeder.
This allows for possible variation in loads and distributed generation between different feeders
fed from a given transformer, allowing the line voltages to be kept within limits on all feeders. A
combination of computer simulation and practical experiments was used. A model constructed in
Excel gave preliminary results, which was used to specify a more detailed model in Matlab® S imu-
link. A small-scale 220/380 V distribution network was constructed, with currents limited to 5 A
per phase. Finally, a rotary switch was constructed, suitable for currents up to 500 A, which would
be required for a full-scale low voltage distribution network. The results showed that the voltage
could be kept within limits, even with a large difference in load and distributed generation from
one feeder to another.
Tap-Ch anging ; On-Line; Fe eders; Independent
1. Introduction
Electricity suppliers are required to maintain line voltages within certain limits centered about the nominal vol-
tage. (This is −6%/+10% in the EU, and ±10% is used in many other countries around the world [1] [2].) To en-
able this under changing loading conditions, transformers have multiple tap settings, allowing the output voltage
to be varied in such a way that the supply voltage remains within limits along the full length of each feeder [3].
Medium voltage transformers (33/10 kV, or equivalent) have on-line tap changing, allowing them to respond in
real time to changing loading conditions. However, this means that all the 10 kV feeders from a given transfor-
mer, and all the 380 V feeders fed from them, change their voltages as a single unit, with no independent control
from one feeder to another. Low voltage transformers (10 kV/380 V, or equivalent) have off-line tap changing.
This does not provide real time control; rather it anticipates future loading, until such loading changes, at which
time the taps must be reset. However, it does provide a limited degree of independent control between different
low voltage networks.
D. Johnston
Increasingly, low power distributed generation is being installed on the distribution network. This includes
renewable energy sources (photovoltaic (PV), wind, etc) and combined heat and power (CHP) generators [4].
Utility-owned generators will be installed on the 380 V, 10 kV, 35 kV (or higher) sectionsdepending on output
power, and will be controlled by the network operator [5] [6]. Similar generators may be installed by private
supply companies, and arrangements may be put in place for these to be controlled by the network operators.
Generators used by property owner/occupiers and installed on their premiseson-site generationwill largely be
s mall-scale, and installed on the 380 V network [7]. (A typical exception would be a large generator installed on
industrial premises.) These will be operated by the building occupiers according to their needs and priorities,
and even where an agreement exists with the network operator, the degree of external control will generally be
limited. These generators will reduce the currents in the lines, thus reducing the voltage drop produced by loads,
and in extreme cases they will reverse the current flow, resulting in a voltage rise along the line. In the case of
on-site generation, this will not generally be subjected to network control.
Provision of on-line tap changing on 10 kV/380 V transformers will allow independent real-time control of
different 380 V networks fed from a given 10 kV feeder. This will allow the line voltages to be maintained
within limits for a wider range of loading generating conditions. However, this will not provide independent
control of the feeders supplied by a given 10 kV/380 V transformer, which would be required where the loading
and/or generation varied greatly between feeders. A situation where this might occur is as follows. One or more
feeders supply domestic premises, which have a significant quantity of installed PV systems. Another feeder
supplies small commercial premises (e.g., shops), which have little or no installed generation. During the work-
ing day, the domestic loads are light, and PV generation is high, resulting in excess generation and voltage rises
along these feeders. At the same time, the commercial loads are high, resulting in a voltage drop along the rele-
vant feeder. It may not always be possible to maintain the voltages within limits for all the feeders, with all of
them connected to the same tap of the transformer.
By placing the taps on the output side of the transformer, it is possible for each feeder to be connected to a
different tap. The line voltage along each feeder can then be kept within limits by measuring the line voltage at
the far end of the feeder (and at other points, if necessary) and changing the tap setting accordingly. This can in-
crease the range of loading/generation conditions which can be accommodated on a given low voltage network.
2. Computer Simulation
An Excel based modelling tool and a Matlab® Simulink were used to simulate a low voltage network. The Excel
modelling tool allowed a range of conditions to be modelled in a short time, and a selection of these were used
as the basis for modelling in Simulink.
2.1. Excel Based Model
The Excel modelling tool used a simple resistive network to model a set of low voltage feeders. Each feeder is
divided into sections, with the length and cable type specified by the user, thus determining the resistance of
each section. Loads and distributed generators are located at nodes between sections, with the power of each
specified by the user. The difference between these is the net power. The resistance of each (net) load is calcu-
lated based on the rated voltage (R = V2/P). Thus, generators are represented as negative resistances. (Modelling
these as nonlinear sources in combination with resistive loads and lines would require iterative solutions.) This
results in the voltages being approximate. However, the modeling tool allows the network parameters (loads,
line resistances, etc.) to be changed quickly, so that a wide range of operating conditions can be investigated in a
short time. An appropriate selection of these can then be analysed further in Simulink. The first and second se-
ries of models were for a fixed tap, and a common on-line tap for all feeders respectively. The results for a
common on-line tap are shown in Figure 1. As seen, the voltage on the feeder with distributed generation ex-
ceeds the maximum limit, while the voltage for the heavily loaded feeder is below the minimum limit. Therefore,
there is no common tap setting, for which the voltages on all the lines will be within the limits. By changing the
tap settings for each feeder separately, it is possible to keep the voltages on both feeders within limits along their
entire length, as shown in Figure 2.
2.2. Simulink Model
A Simulink model was used for more detailed and accurate analysis of the scenarios selected using the Excel
D. Johnston
Figure 1. Line voltages along feeders connected to a com-
mon tap (102.5% = 225.5 V).
Figure 2. Line voltages along feeders connected to separate
taps (100% = 220 V; 105% = 231 V).
modeling tool. A circuit diagram of the model is shown in Figure 3. For clarity, a single feeder and a single
phase are shown, and the feeder is reduced to two line sections and two nodal loads. These are shown here as re-
sistances, but in general resistive/inductive elements are used. The tap-changing transformer is represented by
two transformers (identified as “Main” and “Tap”), with their inputs connected in parallel and the outputs con-
nected in series. The main transformer has outputs at 0 V and the minimum of the tapping range. The tap trans-
former has five outputs covering the tapping range from minimum to maximum voltage. Each of these tap out-
puts is connected to a contactor, which is connected to the near end of the feeder. A control block closes one of
these contactors, connecting the feeder to the required tap. The control block receives a voltage measurement
signal from the far end of the feeder. If the far-end voltage is below the minimum limit, the control block in-
creases the tap setting by one step. If the far-end voltage is above the maximum limit, the control block decreas-
es the tap setting by one step.
0 1 2 34 5 6
Volt age
Feeder with DGHeavily loaded feeder
012 34 5 6
Volt age
D. Johnston
Figure 3. Simulink model of OLTC transformer and a feeder (one phase).
A number of scenarios were tested, each with a large loading/generation differential between feeders. In one
example, 300 kW of net loads were place on the nodes along one feeder, and up to 240 kW of net generation on
each of the remaining feeders. With independent tap setting, the far-end voltage on the heavily loaded feeder
was 208 V, and the far-end voltage on the feeder with the largest level of distributed generation was 240 V. Both
of these voltages are within the limits.
3. Practical Evaluation
Two experimental arrangements were used, which were based on the limitations of the power available in the
laboratorya reduced version of a transformer/feeder network, and a rotary switch capable of passing the cur-
rents typical of a low voltage network.
3.1. Reduced Power Low Voltage Network
This network was a practical implementation based on the Simulink model. An Elgar Electronics Smart WaveTM
power generator was used to supply the Chinese standard voltage of 220/380 V from the (United Kingdom)
230/400 V supply. The current limit of this generator is 5 A per phase, and this set the limit for the rest of the
equipment/circuits used.
The main transformer was made up from two toroidal transformers with dual 55 V outputs. These were con-
nected in series to give 220 V (nominal), which was reduced to 210 V when supplied at 220 V. The tap trans-
former was a multi-tap transformer with outputs of 0 - 6 - 12 - 18 - 24 V (nominal), giving a range of 23 V when
supplied by 220 V. By connecting this in series with the toroidal transformers, a tapping range from 210 to 233
V was available. This arrangement was replicated for each phase.
D. Johnston
The contactors were implemented by three-pole relays, which simultaneously changed the tap settings for all
three phases of a given feeder. A block of five such relays was connected to each feeder. The output stage of the
control block was a digital decoder, thus ensuring that only one relay in each block was closed at any given time,
to avoid shorting of the outputs of the multi-tap transformer.
Because of the 5 A per phase limit, the resistances of the loads were higher than those of a real network. Load
resistances of 220 Ω to 2.2 kΩ were used, giving a maximum of 1 A at each node. (Not all nodes drew this
maximum current.) In order to produce voltage drops similar to those on real feeders, in the presence of smaller
currents, larger line resistances were used. The total resistance per phase of each feeder was up to 20 Ω, made up
of sections of 1 to 4 Ω. The resistors used have inductances giving X/R ratios similar to those for real cables and
loads. A single generator (driven by a D.C. motor) was used to represent the larger number of generators that
would generally be present on a real system.
Figure 4 shows the results for a lightly loaded feeder with a 200 W generator at the far end, and a heavily
loaded feeder. The results are in general agreement with those of the Simulink model, and the voltages on both
feeders are within limits.
3.2. High Current Rotary Switch
A rotary switch was constructed with a current rating typical of currents in a low voltage network. On-line tap
changers used at higher voltages use break-before-make contacts, in order to avoid shorting between taps.
However, the resultant interruption of current can cause high voltage transients, where inductive components are
present in the system. One solution is to place resistive elements between the contacts of the tap changer, such
that they only come into circuit during a transition between tap settings. This method was used in the design of
the rotary switch. A cross section is shown in Figure 5. A cam, embedded in the rotor, connects to one of the
contacts placed around the rim, thus connecting to the corresponding transformer tap. This cam is connected
along the axis to a slip ring, which is connected via a contact to the relevant feeder. The cam and contacts are
made of phosphor bronze to reduce friction. Graphite sections are placed either side of the rotary cam. During a
tap transition, as the rotor is rotating, one of the graphite sections makes contact with the next contact, thus pro-
viding a continuous conductive path. This avoids voltage transients due to inductive components, while also
providing sufficient resistance to avoid shorting of adjacent taps. A stepper motor was used to rotate the rotor, in
response to overvoltage or undervoltage signa ls.
Figure 4. Line voltages for a low-power network with a single ge-
nerator on one feeder.
Feeder with DGHeavily loaded feeder
D. Johnston
Figure 5. Cross section of the
rotary switch, showing cam and
contact s.
A 10 V, 300 A D.C. power supply was used to evaluate the conductivity of the switch. Temperature measure-
ments showed that there was a moderate increase above ambient, but not sufficient to cause overheating. In or-
der to test the performance during switching, a 230 V A.C. supply and an inductive load of 2 mH. Resistive
heating of the graphite was negligible and voltage transients were less than 150% of the steady-state voltage.
4. Conclusions
The tap-changing system was tested under a range of conditions, focusing on those where there is a large load-
ing/generation differential between feeders. Independent tap setting was able to keep the voltages on all feeders
within limits, under operating conditions in which a fixed or common tap would result in voltages exceeding the
limits. This allows a greater range of loads and particularly distributed generation to be installed. It may also re-
duce or avoid the need for alternative network support measures, such as battery storage. Electric vehicles (EVs)
have the potential to act as such storage, but in high density residential areas, home charging may not be possi-
ble, and EVs would not be available for network support. In such cases, independent tap changing could provide
the degree of support necessary to allow the use of in-si tu generation. Transformers for independent tap setting
would require multiple taps on the secondary coils. The cost involved could effectively be reduced by installing
such coils in new transformers, to replace older units, as part of the maintenance/upgrade of the low voltage
The analysis in this work considered loads and generators distributed along each feeder, which resulted in
voltages falling or rising steadily along each feeder. However, if a large load or generator is placed part-way
along a feeder, the resulting minimum or maximum voltage would be located at this point. An additional voltage
measurement would be needed at this point, to ensure that the voltage is within limits along the entire length of
the feeder. The control system could be designed to allow additional voltage measurement inputs after installa-
tion, to allow for large loads or generators being installed on the network at a later time.
Acknowledgem ent s
The author would like to thank Edward Bentley, Tianxiang Jiang and Pasist Suwanapingcarl, whose previous
research contributed to this work.
[1] EN 50160European Committee for Electrotechnical Standardisation.
[2] GB/T 12325 (2003) Admissible Deviation of Supply Voltage.
[3] Weedy, B.M. (1994) Electric Power Systems. John Wiley & Sons.
[4] Aly, G.E.M., El-Zeftawy, A.A., El-Hefnawy, A. A. and Eraky, S.A. (1999) Technical Assessment of Electric Utility-
Interactive Residential Photovoltaic Systems. Electric Power Systems Research, 51, 175-185 .
[5] Milanović, J.V . (2001) Characteristics of Voltage Sags in Radial Networks with Dynamic Loads and Embedded Gene-
rators. Proceedings of the IEEE Power Technology Conference, P ort o.
[6] Acker mann, T., Andersson, G. and Söd er, L. (20 01) Distributed Generatio n: A Definition. Electric Power Systems Re-
D. Johnston
search, 57, 95-204 . 0010 1 -8
[7] El-Khattam, W. and Salama, M.M.A. (2004) Distributed Generation Technologies, Definitions and Benefits. Electric
Power Systems Research, 71 , 119-128 .