Fundamental frequency coupling between HVAC and HVDC lines in the Quebec-

New England multiterminal system - Comparison between field measurements

and EMTDC simulations

Johan Ulleryd Ming Ye Gérard Moreau

ABB Power Systems Hydro-Québec
SE-771 80 Ludvika, Sweden Montréal, Canada

Abstract: Fundamental frequency coupling between HVAC and its terminating impedance, responding to the fundamental
HVDC lines has been investigated through actual field measure- frequency current on the DC line.
ments in the Quebec-New England multiterminal HVDC system. These two effects must be added, thus complicating the
These are the first field measurements, which clearly exhibit the calculation and evaluation of the actual coupling to the DC
fundamental frequency coupling. The measurements covered the sec-
line. The analysis must therefore be performed both with and
tions of interaction located along the Radisson-Des Cantons part
(approximately 1120 km long) of the HVDC transmission line. In without the converters in operation.
these sections, the distances between the AC and DC lines do not In 1990 Hydro-Quebec carried out a series of field meas-
exceed 1,5 km. During the field measurements, the recorded funda- urements of the coupling effects on the Radisson-Des Can-
mental frequency currents in the DC line ranged between 1,05 and tons part of the Quebec-New England HVDC transmission
1,20 A rms. An EMTDC computer model of the coupling phenom- line [4]. The converters were not in operation during the field
enon has been developed and validated based on these measure- measurements and the DC transmission line was directly ex-
ments. This article presents the computer model and the results of posed to electromagnetic coupling from neighboring AC trans-
the validation. mission systems.
To our knowledge, this is the first set of field measure-
Keywords: HVDC, Multi-transmission line theory, EMTDC, AC/
DC line coupling ments of its kind, which clearly exhibits the fundamental fre-
quency coupling between AC and DC lines. Since field meas-
I. INTRODUCTION urements are expensive and time consuming, it is of interest
to develop and validate a computer model that represents
An HVDC line running in parallel with an HVAC line will correctly the coupling effects. The model should be able to
be exposed to a fundamental frequency coupling. The cou- accurately represent detailed geometrical configurations in-
pling effects, if strong enough, have been recognized as an cluding transpositions, different ground resistivities and the
important subject concerning the performance and the design terminating impedances of the lines.
of an HVDC system [1]. Equipment both on the AC and DC In the future, there will be an increased pressure for vari-
sides of the converter station can be affected by this phenom- ous reasons, e.g. economical and ecological, to locate the DC
enon. Induced fundamental frequency current can lead to sat- lines along the same right-of-way as the AC lines. Therefore,
uration of the converter transformer core due to unsymmetri- reliable analyses of more complex configurations, with HVDC
cal magnetization. Saturation leads to an increased noise lev- converters included, are necessary. Some interesting simula-
el and increased heating in the transformer. The commutation tor studies have already been made for AC-DC hybrid trans-
in the converter and the saturation of the transformer may mission systems [2], [3], where the AC and DC lines are lo-
also generate higher order harmonic currents in the AC and cated on the same tower. However, the results presented are
DC networks, which will influence the filter design. not yet confirmed by field measurements.
The superimposed AC current on an HVDC transmission
line is based on two effects: II. DESCRIPTION OF FIELD MEASUREMENTS
• The coupling between AC and DC lines
• The dynamic characteristic of the HVDC converter with The field measurements were carried out on the northern
part of the Quebec-New England Phase II multiterminal
HVDC system, see Fig.1. This part of the system is a bipolar
450 kV overhead transmission line. It is spanning from the
Radisson converter station (2,250 MW), in the James Bay
area, to the Des Cantons converter station (690 MW) near the
US border. The length is 1120 km. A third converter station,
Nicolet (2,138 MW), is located between the two previously
mentioned stations, 90 km north of Des Cantons. At the time
of the field measurements, all converter terminals and their
associated equipment were disconnected. Hence, the trans-
mission line was passive and measurements were not ampli-
fied by resonances with the line terminating impedances.
 Three sections of the HVDC line were exposed to AC line III. RESULTS OF THE FIELD MEASUREMENTS
couplings, as shown in Fig. 1 and detailed in Table 1.
• Section I: Radisson - Némiscau A. Configuration 1
• Section II: Trenche/La Tuque - Trois Rivières/Des Hêtres
• Section III: Bécancour- Nicolet On the DC line, the recordings lasted for about 24 hours
In the first section, starting at about 25 km south of Radis- and indicate that about 1 kV was induced on each pole. A
son station, the DC line runs close to two 735 kV lines (L7062 strong correlation was noticed between these induced voltag-
and L7063) for about 31.5 km. The second section is about es and the positive sequence current on line L2380. This is
94 km long and starts at 800 km south of Radisson: the cou- the AC line running closest to the DC line and for the longest
pling comes mainly from two 230 kV lines (L2370 and L2380). coupling section.
In the last section, immediately north of Nicolet station, the
DC line is exposed for 22.5 km to two 230 kV lines (L2381 B. Configuration 2
and L2382).
The measurements were carried out for two configurations The measurements for DC line configuration 2 lasted only
of the DC line: 90 minutes and showed fundamental frequency currents of
a) Configuration 1: both poles of the DC line were ground- 1.05 and 1.20 A rms for pole 1 and pole 2, respectively.
ed at the southern end (Des Cantons) but left open-circuited The coupling factor in configuration 2, where the DC lines
at Radisson where the pole voltages were measured. are short-circuited to ground in Radisson, is different from
b) Configuration 2: both poles of the DC line were ground- configuration 1. The inductive coupling is dominating and
ed at both ends and the pole currents were measured at Radis- gives a higher value for the current of pole 2, due to its closer
son. location to the AC lines. The tower geometry of the bipolar
A nine-cycle acquisition of the DC line voltages or cur- transmission line and the location of the neighboring AC lines
rents was taken every minute via a digital oscilloscope and are shown in Fig. 1. Considering also the coupling lengths
the data was processed to obtain the first ten harmonics. While and the relative position of the lines given in Table 2, it can
the above DC line measurements were ongoing, voltage and be confirmed that pole 2 is exposed to a stronger inductive
currents of 735 kV line L7063 and of 230 kV lines L2370 coupling compared to pole 1.
and L2380 were being monitored continuously via power The voltage and currents of 735 kV line L7063 and 230 kV
analysers to also obtain the first ten harmonics about every lines L2370 and L2380 are presented in Table 1. The table
seven minutes. 735 kV line L7062 was not monitored as it lists the average magnitudes of the positive-, negative- and
was not in service during that time period and was open- zero-sequence components, calculated from the continuous-
circuited at both ends. Also, since the third coupling section ly monitored ones in configuration 2.
is relatively short, 230 kV lines L2381 and L2382 were not From the hourly records, the average value of the positive
specially monitored and the data came from the hourly records sequence magnitudes for 230 kV lines L2381 and L2382 are
of the system control center. calculated. The negative and zero sequences for these lines
have not been monitored.

TABLE 1. SEQUENCE VOLTAGES AND CURRENTS. AVERAGE

VALUES OF THE FIELD MEASUREMENTS FROM
CONFIGURATION 2 OF THE REPORT. DATE: JULY 22,
1990. TIME: 13:38 - 15:04 [4].

Fig. 1. The 1120 km Radisson - Nicolet-Des Cantons ± 450 kV HVDC * Il - phase current, rms value.
system. Ulg - phase to ground voltage, rms value.
 IV. EMTDC COMPUTER MODEL B.1 Simulation for the worst coupling
For each coupling AC line, the simulation is performed in
A. Modeling of the coupling sections two steps. The first one covers the coupling from the posi-
tive-sequence voltage and current, see Fig. 2. A three-phase
The coupling effects are modeled and simulated using the voltage source is connected at each end of the coupling sec-
electromagnetic transients simulation program EMTDC [6], tion and the angle between these two sources is adjusted to
[7]. The selected multi-transmission line model is a Berger- obtain the proper positive-sequence current. Therefore, both
on, single frequency model, tuned to 60 Hz. the capacitive and the inductive couplings related to the pos-
Three AC line sections are considered to have an impact itive sequence are represented simultaneously.
on the fundamental frequency coupling, as shown in Fig. 1. Then, in the second step, the zero-sequence current effects
The EMTDC model, as an approximation, includes these are modeled for each coupling AC line by using a zero-se-
three most dominant sections and neglects the coupling from quence current source. While the simulation is performed for
all other AC lines. Table 2 lists the parameters of the three one AC line, the sources are short-circuited in all the other
sections, as presented by Hydro-Quebec in the technical spec- AC lines to allow the circulation of secondary currents.
ifications of the project [5]. Furthermore, the model includes In the steps above, the sources are adjusted to the average
both AC and DC transpositions, as well as the position (East values recorded during the field measurements for configura-
or West) of the DC line with respect to the AC lines. Outside tion 2, where the DC line was short-circuited at both ends and
the coupling sections, the DC line is extended to its full length, the induced currents on pole 1 and pole 2 were measured.
1,120 km, as a bipolar transmission line. On the other hand, This configuration is considered more interesting than con-
the AC lines are terminated at the end of each coupling sec- figuration 1 because induction issues (presented in the intro-
tion by equivalent impedances representing the AC network. duction of this paper) are more often described in terms of
These impedances vary in time depending of the actual load induced currents. As only the positive-sequence values are
flow and the simulation has to use the appropriate values. available for L2381 and L2382, their zero-sequence currents
were assumed to be 1% (typically used value) of the positive-
TABLE 2. CONFIGURATION PARAMETERS FOR THE THREE sequence currents.
COUPLING SECTIONS.

Fig. 2. Modeling of the positive-sequence voltage and current driving

sources.

These two steps of simulation per AC line will produce 10

values of induced current (2 steps x 5 lines) for each pole of
Note 1: the position of the DC line relative to the AC lines is given by
W (west) or E (east).
the DC line. The worst coupling can be evaluated by two cal-
Note 2: the position of DC pole 1 relative to the center of the tower is culation methods:
given by W (west) or E (east). a) arithmetic sum: the resulting pole current is obtained by
* DC line transposition. an arithmetic sum of the 10 values of induced currents;
b) root-sum-squares (RSS): the resulting pole current is ob-
tained by the square root of the sum of the 10 squared values
B. Modeling of driving sources
of induced currents.
The RSS calculation is usually recommended for telephone
The driving sources for the coupling on the DC line are
interference studies, while the arithmetic sum is used for con-
made of the voltages and currents of the AC lines. These three-
trol and protection studies [1].
phase variables can be represented by the corresponding pos-
itive-, negative- and zero-sequence components. Out of these
B.2 Simulation of simultaneous coupling
components (three for voltage and three for current), only
A final simulation is performed considering the fact that
three are modeled: positive-sequence voltage and current and
the driving sources are phasors. It simultaneously represents
zero-sequence current. Positive-sequence components are im-
the effects of all the positive-sequence voltage and current of
portant because of their magnitudes which are close to one
all the AC lines. Their magnitudes are adjusted. The angles
per-unit. The negative-sequence voltage and current are ne-
between the local voltages and currents are adjusted to the
glected since they are quite small compared to positive-se-
values recorded during the field measurements. However the
quence, while obeying the same coupling mechanisms. The
phase angle between the voltage sources of the various cou-
zero-sequence voltage is also small and cannot couple much
pling sections are obtained from a load flow, corresponding
current on the DC line. Even if the zero-sequence currents
to the light load conditions prevailing during the summer of
are small, their coupling contribution cannot be neglected: as
1990. This load flow indicated that the Radisson (section I)
these currents must return through ground or via overhead
735 kV bus voltage is ahead by 40 degrees of the Nicolet
ground wires.
(section III) 230 kV bus voltage, while there is no phase dif- TABLE 4. 60 HZ CURRENTS RECORDED IN FIELD
ference between Nicolet and Des Hêtres (section II). MEASUREMENTS VERSUS OBTAINED IN EMTDC SIMULATIONS.

V. SIMULATION RESULTS

A. Worst coupling, comparison with field measurements

The results of the EMTDC simulation, according to the

worst coupling method, are shown in Table 3. The dominat-
ing contribution of fundamental frequency coupling corre- B. Simultaneous coupling, comparison with field
sponds to positive sequence values in coupling section II. It measurements
implies that the distance between AC and DC lines and the
line length of paralleling are dominating factors. Besides, the The simultaneous coupling, with a specific load flow con-
higher resistivity of earth may result in a stronger capacitive dition, seems to give a good approximation of the reality even
coupling. if it does not include the zero sequence coupling. It seems
Section II will also contribute most to the coupling from that conservative results are obtained with only the positive
the zero sequence driving sources. It can be seen that the cor- sequence generated in the driving sources. A further investi-
responding magnitudes for this section will even be higher gation taking into account zero sequence has not been done.
than the positive sequence magnitudes of sections I and III. Depending on the phase displacement between the sequenc-
The zero sequence current was typically assumed to be 1 es, inclusion of the zero sequence effects, could help in bring-
% for lines L2381 and L2382. Even if a higher value was ing the simulation results closer to the field measurements.
assumed for this current, such as 2-3 %, it would not affect Typically, zero sequence effects tend to add to the positive
significantly the final result. sequence effects, the zero sequence being affected by such
factors as ground resistivity and line configuration.
TABLE 3. AMPLITUDE OF INDUCED CURRENTS IN THE DC LINE
AT THE RADISSON TERMINAL (A rms). EMTDC SIMULATION,
C. Simultaneous coupling, comparison with worst coupling
WORST COUPLING.

Taking into account the phase displacement of the various

positive sequences reduces the induced current from 1.5 - 1.9
A rms, see Table 3, to 1.4 - 1.6 A rms, see Table 4.

VI. CONCLUSIONS

Fundamental frequency coupling between HVAC and

HVDC lines has been investigated through actual field meas-
urements, for the Quebec-New England multiterminal HVDC
In Table 4, the sum of the effects from the positive sequence system. An EMTDC model of the coupling phenomenon has
and the zero sequence driving sources are calculated (arith- been developed and validated based on these measurements.
metically and RSS) and compared with the field measure- The following conclusions can be made:
ments. The superimposed 60 Hz currents in the DC lines have mag-
When the arithmetical summing method is used, the simu- nitudes ranging from 1.05 to 1.20 A rms, according to the
lation gives magnitudes of induced currents that are twice the field measurements. In EMTDC simulations, the coupling
ones observed during the field measurements. This method currents vary between 1.0 A to 2.5 A rms, depending on the
will always lead to the biggest overestimation because it does calculation method.
not allow for any cancellation effects. The overestimation will The method that arithmetically adds the contribution from
generally increase with the number of coupling sections. each AC line section, in most cases, results in an overestima-
The RSS calculation corresponds very well with the field tion of the magnitude of superimposed fundamental frequen-
measurements. This can be a quite appropriate method spe- cy current (1,0 A compared to 2,5 A), but can give the upper
cially when there are many coupling sections, but one has to limit of the estimated worst coupling current.
remember that it could in some occasions lead to underesti- The RSS method will, in the case of multiple coupling sec-
mations. tions, generate a closer approximation to reality compared to
the arithmetical value. It sets the lower limit of the worst cou-
pling current.
Conservative results and close matching with field results
(1.4 A and 1.6 A), can be obtained by using the simultaneous
coupling method based on the appropriate load flow condi-
tion of the AC system. The load flow will determine the phase
relationships between the driving sources. It is recommended
to assume peak load conditions in the neighboring AC system [6] D. A. Woodford, “Validation of digital simulation of DC links”, IEEE
in order to obtain the worst condition possible. Trans. Power Apparatus and Systems, Vol. PAS-102, No. 6, June,
1983, pp. 1616-1623.
The arithmetic and RSS calculation methods can be used [7] Omprakash Nayak, Garth Irwin, Arthur Neufeld: “GUI Enchances
to identify the range of the worst coupling current for multi- Electromagnetic Transients Simulation Tools” , IEEE Computer
ple driving sections. The concept of simultaneous coupling Applications in Power , Vol.8, No. 1, January 1995.
gives conservative results and is recommended for analysis
of more complex configurations. The drawback of this meth- VIII. BIOGRAPHIES
od is the special concern regarding the terminating imped-
ances and the loadflow conditions of the neighboring AC lines, Johan Ulleryd graduated from Chalmers University of Technology,
Gothenburg, Sweden, in 1993 with an M. Sc. degree in E.E. In 1994 he
which can be difficult to assess in some studies. joined ABB Ludvika, Sweden, as a trainee engineer, which included a half
The coupling effects between the HVDC and the AC lines year stay at Manitoba HVDC Research Centre, Winnipeg, Canada. From
being adequately represented, the EMTDC model can be con- 1995 he is working as a system engineer at ABB Power Systems AB, HVDC
fidently expanded to include the dynamic characteristics of division. He is non-member of IEEE.
Ming Ye received his B.Sc. in E.E. from Jiaotong University, Xian, in
the HVDC converters. 1982, and M. Sc in E.E. from China University of Mining Technology,
Beijing, in 1986. He obtained his Ph.D. degree in 1994 from Uppsala
VII. REFERENCES University, Sweden. In 1995, he joined ABB Power System AB and is
currently working as a system engineer with the design of insulation
[1] E. V. Larsen, R. A. Walling, and C. J. Bridenbaugh, “Parallel AC/DC coordination for HVDC transmission system. His research area and interests
transmission lines steady-state induction issues”, IEEE Transaction on include protection of HVDC power transmission system from various
Power Delivery, Vol. 4, No. 1, January, 1989, pp. 667-673. transient overvoltages and EMC of electric equipment in low voltage power
[2] K. Ramesh, T Adhikari, S.C. Kapoor, D.P. Gupta. “ Parallel operation installations. He is non-member of IEEE.
of AC and DC lines running on the same tower.”, Cigré, Paper 14-06, Gérard Moreau received his B.Sc. in E.E. from Laval University,
1988. Quebec, Canada in 1971. He worked for six years for Shawinigan
[3] R. Verdolin, A.M. Gole, E.Kuffel, N. Diseko, B. Biweski, “Induced Engineering Ltd, a consulting firm in Montreal, as a control and protection
overvoltages on an AC-DC hybrid transmission system”, IEEE engineer in various commissioning projects locally and overseas. He joined
Transactions on Power Delivery, Vol. 10, No. 3, July 1995. the Hydro-Quebec field tests department in 1979 where he has been
[4] J.-G. Boisclair,“Mesures relatives au couplage des lignes de puissance responsible of the acceptance tests and system integration of various SVC’s
c.a. sur la ligne c.c. Radisson-Des Cantons”, Internal Hydro-Quebec and HVDC transmission systems. He also conducted field evaluation of
report SEET no 92096, 1993. capacitive and inductive current switching schemes, of disconnect switching
[5] “Quebec-New England Phase II multi-terminal HVDC system.”, transients and of very fast transients in GIS switchgear. He is member of
Technical Specification, Hydro-Quebec,1986, Chapter F. the IEEE and of the Order of Engineers of Quebec.