You are on page 1of 10

Program Development

HYSTERESIS HEVIA: a new routine to generate


input data for inductors with hysteresis

Orlando P. Hevia Gorostiaga 1483


Santa Fe - Argentina 3000 Santa Fe
Argentina
heviaop@ssdfe.com.ar

1 Introduction

It is a common problem to obtain the full magnetization curve for the saturable inductors
(Type-98 or Type-93). Most of the cases only some points of the curve are readily available as
measurements, or data measured on a similar magnetic material should be used for modelling
the core. Measurements are normally given as RMS quantities, that should be converted into
peak values of current and flux using the SATURA supporting routine of ATP (ref. ATP Rule
Book chapter XIX - G).

Saturable inductors can be represented by hysteresis in ATP, as well, using the routine
HYSDAT to obtain parameters (ref. ATP Rule Book chapter XIX - H ). There is an important
limitation, however: ATP has access to data of only one type of magnetic material (ARMCO
M4 oriented silicon steel). The Rule Book just gives a hint that routine SATURA can be used
to obtain data of other materials, if it is proceeded with care. The hysteresis loop can be
constructed manually by displacing the current values given by SATURA.

The aim of development of the new routine HYSTERESIS HEVIA was to accomplish this
tasks precisely: a) to displace the current values and b) to produce a hysteresis loop with a
power loss equal to a required value. It can also be used to fit the curve to a given residual
flux.

The routine can generate a .PCH file off-line, like all the other auxiliary routines of ATP. But it
can be used to generate hysteresis data for a Type-96 inductor on-line, i.e. when processing the
input data file for a time domain simulation.

2 How the routine works

The routine reads as input the single-valued magnetizing curve of flux versus current, the
power loss arising from the hysteresis, the frequency and the value of required residual flux.

Initially, the program proposes a null displacement. Then calculates the enclosed area for each
segment (the area is approximated by parallelograms, so it can be calculated easily).

EEUG News Feb - May 2000 19


Program Development

This area represents energy. Multiplying it by the frequency results in the power loss for one
cycle. The calculated energy then is compared with the required one. If the result is smaller (in
the first step it will be always the case), displacement is increased by a given step.

The displacement is variable, because the hysteresis loop shows an asymmetry: the
displacement toward to the left is greater than toward to the right. The right displacement is
practically constant, but the left displacement is increasing slowly with the value of flux until a
maximum for the saturation knee. After this point, displacement decreases quickly to be
annulled for the saturation flux.

The variable displacement produces a loop of hysteresis with setbacks. If the calculated power
loss is smaller than required, the process is receded to the value of previous displacement, the
step is divided by two and proceeds with that smaller step. As the applied function is simple,
the process is very rapid and the error in the power can be made very small. The calculation
stops when the relative error is smaller than 10-12 .

The last two points are displaced solely if resulted in a segment with negative slope. The
penultimate point is displaced solely if necessary, but being maintained these points single-
valued. These two points define thus a rectum in the one which there is no hysteresis. In the
routine HYSDAT the penultimate point is defined as saturation point. If this point is displaced,
the area is calculated considering now this new saturation point.

The last point of the flux-current table is not displaced, unless it will be necessary. In this case,
the displacement is effected on the last two points in order to maintain the section saturated
with the same slope. This point does not intervene in the calculation of the area.

The program estimates the maximum value of power loss that can be enclosed by a cycle: the
power loss never will be greater than that of the corresponding extreme flux and current
values, and in base to this, extends the curve displacing the two last points. But it is
responsibility of the user to accept new table.

To obtain a value from equal residual flux to the required, the flux values are multiplied by a
factor, that is determined for each iteration. This can give cause for that the flux values are
removed from the initial data, and it must be controlled by the user.

At the end of the calculation, results are registered in the .DBG file, the auxiliary output of
ATP, a plot is presented on the screen and the routine generates a graphic file in one of the
customary formats of ATP, according to what is specified in the STARTUP file. Finally,
through the $PUNCH command of ATP, a .PCH file is generated that is ready to insert into a
simulation via $INCLUDE.

In certain cases, the routine can not find a displacement and combination of flux that converge
to a solution. In this case, a warning message is added to the ouput file.

3 Off-line data generation

The input data structure of the routine is described as in-line comments in the DCN13.DAT
benchmark file:

20 EEUG News Feb - May 2000


Program Development

BEGIN NEW DATA CASE


C
C 7th of 8 subcases illustrates batch-mode usage of Orlando
C Hevia's supporting program HYSTER. This serves to convert
C a Type-98 element into a Type-96 element. For background,
C see story in the July, 2000, newsletter.
C
C The 3 parameters on the preceding request card govern the screen plot
C that automatically will document hysteresis of the result:
C NOZOOM is a binary switch that indicates whether the entire
C curve or just the hysteresis loop itself is to be plotted.
C Value 1 will plot everything (no zoom) whereas value zero
C will drop the final point, which is outside the loop.
C FSCALE is a scaling factor to waste space above and to the right
C of the 1st and 2nd-quadrant plot. For example, value 1.1
C will waste 10%, providing this much margin.
C DXL2 is the "radius" in screen inches of the squares that mark data
C points. Value zero means that there will be no such marking.
C
C REPLOT to require a new plot with different parameters.
C
C So much for the special-request card. On to reactor. Everything about
C the following is normal Type 98 except for columns 27-74, from which 3
C new floating-point parameters are read. Of course, if currents are in
C amperes and voltage is in volts, power will be in watts and residual
C flux will be in volt seconds:
C
C 345678901234567890123456789012345678901234567890123456
C NOZOOM FSCALE DXL2
HYSTERESIS HEVIA 0 1.1 .04
C <--Losses in W-><--Freq in Hz--><-Residual flux> I
98HEVIA 101519.6 50.0 1370.0 1
C --------------________________
.123028 605.2578
.1488219 781.0518
.1973665 992.4457
.3093331 1088.109
.5095013 1197.105
.8414885 1303.846
1.348816 1388.295
1.965261 1437.115
2.405495 1468.171
2.823462 1485.881
3.30741 1505.801
3.681229 1514.619
3.967086 1521.231
8.539266 1555.819
9999.
$PUNCH
REPLOT 1 1.1 .04
C
C Second subcase: the residual flux is set to 0, indicating that this will be
C a swing unknown.
C
C <--Losses in W-><--Freq in Hz--><-Residual flux> I
98HEVIA 101519.6 50.0 0.0 1
C --------------________________
.123028 605.2578
.1488219 781.0518
.1973665 992.4457
.3093331 1088.109
.5095013 1197.105
.8414885 1303.846
1.348816 1388.295
1.965261 1437.115
2.405495 1468.171
2.823462 1485.881
3.30741 1505.801
3.681229 1514.619
3.967086 1521.231
8.539266 1555.819
9999.

EEUG News Feb - May 2000 21


Program Development

$PUNCH { Show branch cards for the Type-96 hysteretic inductor (now created)
C
C Third subcase: the losses are set to 0, indicating that this will be
C a swing unknown.
C
C <--Losses in W-><--Freq in Hz--><-Residual flux> I
98HEVIA 0.0 50.0 1370.0 1
C --------------________________
.123028 605.2578
.1488219 781.0518
.1973665 992.4457
.3093331 1088.109
.5095013 1197.105
.8414885 1303.846
1.348816 1388.295
1.965261 1437.115
2.405495 1468.171
2.823462 1485.881
3.30741 1505.801
3.681229 1514.619
3.967086 1521.231
8.539266 1555.819
9999.
$PUNCH { Show branch cards for the Type-96 hysteretic inductor (now created)
C
C Fourth subcase: both losses and residual flux are set to 0,
C indicating that both variables will be swing unknown.
C The result is a non hysteretic inductor
C <--Losses in W-><--Freq in Hz--><-Residual flux> I
98HEVIA 0.0 50.0 0.0 1
C --------------________________
.123028 605.2578
.1488219 781.0518
.1973665 992.4457
.3093331 1088.109
.5095013 1197.105
.8414885 1303.846
1.348816 1388.295
1.965261 1437.115
2.405495 1468.171
2.823462 1485.881
3.30741 1505.801
3.681229 1514.619
3.967086 1521.231
8.539266 1555.819
9999.
$PUNCH { Show branch cards for the Type-96 hysteretic inductor (now created)
BLANK card terminates stacked Type-98 elements within HYSTERESIS HEVIA
BEGIN NEW DATA CASE
BLANK

The above data case corresponds to a generator step-up transformer of 825 MVA, 500/21 kV.
Fig. 1 shows the results of of the hysteresis loop measurement. The central curve corresponds
to the values without hysteresis, that is taken as base for the example.

22 EEUG News Feb - May 2000


Program Development

Fig. 1
The measured no-load loss is 475 kW (that includes all losses). This value has been measured
at rated voltage and corresponds to three phases. The measured values must be adjusted before
using as input to HYSTERESIS HEVIA.

For one phase the no load loss equals to: P0single = P0 / 3 = 475000 / 3 = 138300 W

The losses caused by the hysteresis was estimated as 57% of the total loss. This 57% was the
average of 3 measurements performed by the owner of the transformer at two different
frequencies and two different fluxes at rated frequency.

Physteresis = 0.57⋅P0single = 78900 W

If this value corresponds to rated voltage and flux: Prated = Physteresis . The rated flux is:

Φ rated = U rated / ( 3 ⋅ π ⋅ f ⋅ 2) = 1299.495 Wb (1)

The saturation flux is 1521.231 Wb (=penultimate value of table, which is the last before the
saturation point as specified Chapter XIX-H of the ATP Rule Book).

The losses of the maximum loop can be approximated by:

Pmaximum = Prated ⋅ (Φmaximum / Φrated)a (2)

The exponent a varies between 1.6 (Steinmetz) to more than 2 for modern materials. Applying
value a=1.6, will result in

Pmaximum = 78900 ⋅ (1521.231 / 1299.495)1.6 = 101519.6 W

The output .lis file contains the results (only the first case is shown):

EEUG News Feb - May 2000 23


Program Development

EMTP begins.
--- 125 cards of disk file read into card cache cells 1 onward.
Alternative Transients Program (ATP), GNU Linux or DOS. All rights reserved by Can/Am user group of Portland, Oregon, USA.
Date (dd-mth-yy) and time of day (hh.mm.ss) = 24-Apr-00 11:52:28 Name of disk plot file is hysthev.pl4
Consult the 860-page ATP Rule Book of the Can/Am EMTP User Group in Portland, Oregon, USA. Source code date is 18 March 2000.
Total size of LABCOM tables = 4939880 INTEGER words. VARDIM List Sizes follow : 6002 10000 10000 340 20000 5000
13140 120000 1450 1900 360 900 36400 255 64800 384 45 254 400000 100000 3000 12000 40000 120 30000 10000 600 126000
--------------------------------------------------+--------------------------------------------------------------------------------
Descriptive interpretation of input data cards. | Input data card images are shown below, all 80 columns, character by character
0 1 2 3 4 5 6 7 8
012345678901234567890123456789012345678901234567890123456789012345678901234567890
--------------------------------------------------+--------------------------------------------------------------------------------
Comment card. NUMDCD = 1. |C data:HYSTHEV.DAT
Marker card preceding new EMTP data case. |BEGIN NEW DATA CASE
Comment card. NUMDCD = 3. |C
Comment card. NUMDCD = 4. |C DATA OBTAINED FROM A MEASUREMENT.
Comment card. NUMDCD = 5. |C THE STEEL IS ARMCO M4
Comment card. NUMDCD = 6. |C CHANGE THE CURRENT TO 8888. !!!
Comment card. NUMDCD = 7. |C 7th of 8 subcases illustrates batch-mode usage of Orlando
Comment card. NUMDCD = 8. |C Hevia's supporting program HYSTER. This serves to convert
Comment card. NUMDCD = 9. |C a Type-98 element into a Type-96 element. For background,
Comment card. NUMDCD = 10. |C see story in the July, 2000, newsletter.
Comment card. NUMDCD = 11. |C
Comment card. NUMDCD = 12. |C The 3 parameters on the preceding request card govern the screen plot
Comment card. NUMDCD = 13. |C that automatically will document hysteresis of the result:
Comment card. NUMDCD = 14. |C NOZOOM is a binary switch that indicates whether the entire
Comment card. NUMDCD = 15. |C curve or just the hysteresis loop itself is to be plotted.
Comment card. NUMDCD = 16. |C Value 1 will plot everything (no zoom) whereas value zero
Comment card. NUMDCD = 17. |C will drop the final point, which is outside the loop.
Comment card. NUMDCD = 18. |C FSCALE is a scaling factor to waste space above and to the right
Comment card. NUMDCD = 19. |C of the 1st and 2nd-quadrant plot. For example, value 1.1
Comment card. NUMDCD = 20. |C will waste 10%, providing this much margin.
Comment card. NUMDCD = 21. |C DXL2 is the "radius" in screen inches of the squares that mark data
Comment card. NUMDCD = 22. |C points. Value zero means that there will be no such marking.
Comment card. NUMDCD = 23. |C
Comment card. NUMDCD = 24. |C REPLOT to require a new plot with different parameters.
Comment card. NUMDCD = 25. |C
Comment card. NUMDCD = 26. |C So much for the special-request card. On to reactor. Everything about
Comment card. NUMDCD = 27. |C the following is normal Type 98 except for columns 27-74, from which 3
Comment card. NUMDCD = 28. |C new floating-point parameters are read. Of course, if currents are in
Comment card. NUMDCD = 29. |C amperes and voltage is in volts, power will be in watts and residual
Comment card. NUMDCD = 30. |C flux will be in volt seconds:
Comment card. NUMDCD = 31. |C
Comment card. NUMDCD = 32. |C 345678901234567890123456789012345678901234567890123456
Comment card. NUMDCD = 33. |C NOZOOM FSCALE DXL2
Request to make Type-96 hysteresis branch cards. |HYSTERESIS HEVIA 0 1.1 .04
Comment card. NUMDCD = 35. |C <--Losses in W-><--Freq in Hz--><-Residual flux> I
Hevia 98. 1.0152E+05 5.0000E+01 1.3700E+03 1 |98HEVIA 101519.6 50.0 1370.0 1
Comment card. NUMDCD = 37. |C --------------________________
(i, psi) point. 1.23028E-01 6.05258E+02 | .123028 605.2578
(i, psi) point. 1.48822E-01 7.81052E+02 | .1488219 781.0518
(i, psi) point. 1.97367E-01 9.92446E+02 | .1973665 992.4457
(i, psi) point. 3.09333E-01 1.08811E+03 | .3093331 1088.109
(i, psi) point. 5.09501E-01 1.19711E+03 | .5095013 1197.105
(i, psi) point. 8.41488E-01 1.30385E+03 | .8414885 1303.846
(i, psi) point. 1.34882E+00 1.38830E+03 | 1.348816 1388.295
(i, psi) point. 1.96526E+00 1.43712E+03 | 1.965261 1437.115
(i, psi) point. 2.40550E+00 1.46817E+03 | 2.405495 1468.171
(i, psi) point. 2.82346E+00 1.48588E+03 | 2.823462 1485.881
(i, psi) point. 3.30741E+00 1.50580E+03 | 3.30741 1505.801
(i, psi) point. 3.68123E+00 1.51462E+03 | 3.681229 1514.619
(i, psi) point. 3.96709E+00 1.52123E+03 | 3.967086 1521.231
(i, psi) point. 8.53927E+00 1.55582E+03 | 8.539266 1555.819
Special termination-of-points card. | 9999.
Request for flushing of punch buffer. |$PUNCH

A listing of 80-column card images now being flushed from punch buffer follows.
===============================================================================
1234567890123456789012345678901234567890123456789012345678901234567890123456789
===============================================================================
C <++++++> Cards punched by support routine on 24-Apr-00 11:52:28 <++++++>
C HYSTERESIS HEVIA 0 1.1 .04
C C <--Losses in W-><--Freq in Hz--><-Residual flux>
C 98HEVIA 101519.6 50.0 1370.0
C C --------------________________
C .123028 605.2578
C .1488219 781.0518
C .1973665 992.4457
C .3093331 1088.109
C .5095013 1197.105
C .8414885 1303.846
C 1.348816 1388.295
C 1.965261 1437.115
C 2.405495 1468.171
C 2.823462 1485.881
C 3.30741 1505.801
C 3.681229 1514.619
C 3.967086 1521.231
C 8.539266 1555.819
C 9999.
96HEVIA HEVIA HYSTER .1488219 790.7990888 1369.999999 1
-3.15626061E+00 -1.53352098E+03
-2.11960930E+00 -1.52459294E+03
-1.44220591E+00 -1.50442434E+03
-9.58897829E-01 -1.48649333E+03
-5.42791755E-01 -1.45504976E+03
-9.20323193E-02 -1.40562050E+03
1.28880636E-01 -1.32011760E+03
1.71138746E-01 -1.21204451E+03
1.73766700E-01 -1.10168827E+03
1.75631950E-01 -1.00483112E+03
1.77072099E-01 -7.90799089E+02
1.78330377E-01 -6.12811233E+02
1.79493923E-01 0.00000000E+00
3.02521923E-01 6.12811233E+02
3.28315823E-01 7.90799089E+02
3.76860423E-01 1.00483112E+03
4.88827023E-01 1.10168827E+03
6.88995223E-01 1.21204451E+03
1.02098242E+00 1.32011760E+03
1.52830992E+00 1.40562050E+03
2.14475492E+00 1.45504976E+03
2.58498892E+00 1.48649333E+03
3.00295592E+00 1.50442434E+03
3.48690392E+00 1.52459294E+03

24 EEUG News Feb - May 2000


Program Development

3.77097596E+00 1.53352098E+03
3.96708600E+00 1.54021550E+03
8.53926600E+00 1.57523515E+03
9999
=========< End of LUNIT7 punched cards as flushed by $PUNCH request >=======

The .dbg file shows the displacement process for the first case (edited):
Results
--------------------------------- Current --------------------------------- --------- Flux --------
I ORIGINAL CORRECTED LEFT DISP. DISPLACED RIGHT DISP. DISPLACED ORIGINAL CORRECTED
0 0.0000 0.0000 0.1795 0.1795 0.1795 0.1795 0.0000 0.0000
1 0.1230 0.1230 0.3014 0.4244 0.1795 0.3025 605.2578 612.8112
2 0.1488 0.1488 0.3259 0.4747 0.1795 0.3283 781.0518 790.7991
3 0.1974 0.1974 0.3730 0.5704 0.1795 0.3769 992.4457 1004.8311
4 0.3093 0.3093 0.4831 0.7924 0.1795 0.4888 1088.1090 1101.6883
5 0.5095 0.5095 0.6806 1.1901 0.1795 0.6890 1197.1050 1212.0445
6 0.8415 0.8415 0.9704 1.8119 0.1795 1.0210 1303.8460 1320.1176
7 1.3488 1.3488 1.2568 2.6056 0.1795 1.5283 1388.2950 1405.6205
8 1.9653 1.9653 1.4225 3.3877 0.1795 2.1448 1437.1150 1455.0498
9 2.4055 2.4055 1.4466 3.8521 0.1795 2.5850 1468.1710 1486.4933
10 2.8235 2.8235 1.3813 4.2047 0.1795 3.0030 1485.8810 1504.4243
11 3.3074 3.3074 1.1878 4.4952 0.1795 3.4869 1505.8010 1524.5929
12 3.6812 3.6812 0.5250 4.2062 0.0897 3.7710 1514.6190 1533.5210
13 3.9671 3.9671 0.0000 3.9671 0.0000 3.9671 1521.2310 1540.2155
14 8.5393 8.5393 0.0000 8.5393 0.0000 8.5393 1555.8190 1575.2351
Required losses : 101519.6000 W
Calculated losses : 101519.6000 W
Error : 0.3553E-10 %
Saturation current : 3.9671 A
Saturation flux : 1540.2155 Wb
Required residual flux : 1370.0000 Wb
Calculated residual flux : 1370.0000 Wb
Air core inductance : 7.5649 H

Fig. 2 and 3 correspond to the first case. Both figure were obtained in HP-GL format. The
saturated part of the hysteresis loop was considered only on Fig. 3.
.1694 .1733

.1271 .1300
x

x
u

.0847 .0866

Hevia. 1 24-Apr-00 10:01 Hevia. 2 24-Apr-00 10:01


l

Peak current = 4.36379E+00 Peak current = 9.39319E+00


.0424 .0433
Peak flux = 1.69424E+03 Peak flux = 1.73276E+03
F

0.0 0.0
.0436 .1309 .2182 .3055 .3927 .0939 .2818 .4697 .6575 .8454
Current x 10** 1 Current x 10** 1

Fig. 2 Fig. 3

4 On-line data generation

The data for the Type 96 inductor with hysteresis can be generated internally by the ATP. In
this case, the data should be in the format as shown by the following example:

BEGIN NEW DATA CASE


0.00001 0.10
1000000 1 1 1 1
C
MODELS
C INTEGRAL OF NODE VOLTAGE=FLUX

EEUG News Feb - May 2000 25


Program Development

INPUT VOLT {V(HEVIA)}


MODEL FLUJO
INPUT VOLT
VAR FI
EXEC
IF T=0.0 THEN INTEGRAL(VOLT):= 0.0 ENDIF
FI:=INTEGRAL(VOLT)
ENDEXEC
ENDMODEL
USE FLUJO AS FLUJO
INPUT VOLT:=VOLT
ENDUSE
RECORD
FLUJO.FI AS FLUJO
ENDRECORD
ENDMODELS
BLANK
98HEVIA HEVIA HYSTER 101519.6 50.0 1370.0 3
C --------------________________
.123028 605.2578
.1488219 781.0518
.1973665 992.4457
.3093331 1088.109
.5095013 1197.105
.8414885 1303.846
1.348816 1388.295
1.965261 1437.115
2.405495 1468.171
2.823462 1485.881
3.30741 1505.801
3.681229 1514.619
3.967086 1521.231
8.539266 1555.819
9999.
BLANK END OF BRANCHES
GENEA HEVIA MEASURING 4
BLANK END OF SWITCHES
C
C VOLTAGE FOR NAXIMUM FLUX
C 2*PI*50.0*1555.819= 488774.954
C ------__----------__________----------__________----------__________---------
14GENEA 488774.954 50. 0.0 -1.
C
BLANK END OF SOURCES
GENEA
BLANK END OF OUTPUT
BEGIN NEW DATA CASE
BLANK

The MODELS part was added to obtain the flux by integrating the voltage.

The ATP results are shown next (edited):


EMTP begins.
--- 61 cards of disk file read into card cache cells 1 onward.
Alternative Transients Program (ATP), GNU Linux or DOS. All rights reserved by Can/Am user group of Portland, Oregon, USA.
Date (dd-mth-yy) and time of day (hh.mm.ss) = 24-Apr-00 12:17:50 Name of disk plot file is HYSTHEVS.pl4
Consult the 860-page ATP Rule Book of the Can/Am EMTP User Group in Portland, Oregon, USA. Source code date is 18 March 2000.
Total size of LABCOM tables = 4939880 INTEGER words. VARDIM List Sizes follow : 6002 10000 10000 340 20000 5000
13140 120000 1450 1900 360 900 36400 255 64800 384 45 254 400000 100000 3000 12000 40000 120 30000 10000 600 126000
--------------------------------------------------+--------------------------------------------------------------------------------
Descriptive interpretation of input data cards. | Input data card images are shown below, all 80 columns, character by character
0 1 2 3 4 5 6 7 8
012345678901234567890123456789012345678901234567890123456789012345678901234567890
--------------------------------------------------+--------------------------------------------------------------------------------
Comment card. NUMDCD = 1. |C data:HYSTHEVS.DAT
Marker card preceding new EMTP data case. |BEGIN NEW DATA CASE
Misc. data. 1.000E-05 1.000E-01 0.000E+00 |0.00001 0.10
Misc. data. ***** 1 1 1 0 0 0 1 0 0 | 1000000 1 1 1 1
Comment card. NUMDCD = 5. |C
No TACS section. Begin MODELS section. |MODELS
Comment card. NUMDCD = 7. |C INTEGRAL DE LA TENSION DE NUDO = FLUJO (SI HAY L...)
MODELS data card. | INPUT VOLT {V(HEVIA)}
MODELS data card. |MODEL FLUJO
MODELS data card. | INPUT VOLT
MODELS data card. | VAR FI
MODELS data card. | EXEC
MODELS data card. | IF T=0.0 THEN INTEGRAL(VOLT):= 0.0 ENDIF
MODELS data card. | FI:=INTEGRAL(VOLT)
MODELS data card. | ENDEXEC

26 EEUG News Feb - May 2000


Program Development

MODELS data card. |ENDMODEL


MODELS data card. | USE FLUJO AS FLUJO
MODELS data card. | INPUT VOLT:=VOLT
MODELS data card. | ENDUSE
MODELS data card. | RECORD
MODELS data card. | FLUJO.FI AS FLUJO
MODELS data card. | ENDRECORD
MODELS data card. |ENDMODELS
Optional blank card follows last of MODELS data. |BLANK
Hevia 98. 1.0152E+05 5.0000E+01 1.3700E+03 3 |98HEVIA HEVIA HYSTER 101519.6 50.0 1370.0 3
Comment card. NUMDCD = 26. |C --------------________________
(i, psi) point. 1.23028E-01 6.05258E+02 | .123028 605.2578
(i, psi) point. 1.48822E-01 7.81052E+02 | .1488219 781.0518
(i, psi) point. 1.97367E-01 9.92446E+02 | .1973665 992.4457
(i, psi) point. 3.09333E-01 1.08811E+03 | .3093331 1088.109
(i, psi) point. 5.09501E-01 1.19711E+03 | .5095013 1197.105
(i, psi) point. 8.41488E-01 1.30385E+03 | .8414885 1303.846
(i, psi) point. 1.34882E+00 1.38830E+03 | 1.348816 1388.295
(i, psi) point. 1.96526E+00 1.43712E+03 | 1.965261 1437.115
(i, psi) point. 2.40550E+00 1.46817E+03 | 2.405495 1468.171
(i, psi) point. 2.82346E+00 1.48588E+03 | 2.823462 1485.881
(i, psi) point. 3.30741E+00 1.50580E+03 | 3.30741 1505.801
(i, psi) point. 3.68123E+00 1.51462E+03 | 3.681229 1514.619
(i, psi) point. 3.96709E+00 1.52123E+03 | 3.967086 1521.231
(i, psi) point. 8.53927E+00 1.55582E+03 | 8.539266 1555.819
Special termination-of-points card. | 9999.
Blank card ending branches. IBR, NTOT = 0 2 |BLANK TERMINAN RAMAS
Permanently-closed switch used for metering. | GENEA HEVIA MEASURING 4
Blank card ending switches. KSWTCH = 1. |BLANK TERMINA INTERRUPTORES (AMPERIMETROS)
High res. added by EMTP across Type-99 elem. 1 | 0HEVIA .1E+11
Comment card. NUMDCD = 45. |C ------__----------__________
Comment card. NUMDCD = 46. |C TENSION NOMINAL
Comment card. NUMDCD = 47. |C 14GENEA 408248.291 50. 0.0 -1.0
Comment card. NUMDCD = 48. |C
Comment card. NUMDCD = 49. |C TENSION PARA MAXIMO FLUJO
Comment card. NUMDCD = 50. |C 2*PI*50.0*1555.819= 488774.954
Comment card. NUMDCD = 51. |C ------__----------__________----------__________----------__________----------
Source. 4.89E+05 5.00E+01 0.00E+00 -1.00E+00 |14GENEA 488774.954 50. 0.0 -1.
Comment card. NUMDCD = 53. |C
Comment card. NUMDCD = 54. |C TENSION PARA SATURACION
Comment card. NUMDCD = 55. |C 2*PI*50.0*1521.231= 477908.813
Comment card. NUMDCD = 56. |C
Comment card. NUMDCD = 57. |C 14GENEA 477908.813 50. 0.0 -1.
Comment card. NUMDCD = 58. |C 14GENEA 800000.000 50. 0.0 -1.
Blank card ends electric network sources. |BLANK TERMINAN FUENTES

List of input elements that are connected to each node. Only the physical connections of multi-phase lines are shown (capacitive
and inductive coupling are ignored). Repeated entries indicate parallel connections. Switches are included, although sources
(including rotating machinery) are omitted -- except that U.M. usage produces extra, internally-defined nodes "UMXXXX".
--------------+------------------------------
From bus name | Names of all adjacent busses.
--------------+------------------------------
HEVIA |TERRA *TERRA *GENEA *
GENEA |HEVIA *
TERRA |HEVIA *HEVIA *
--------------+------------------------------
Nonlinear resistances, time-varying resistances, and type-94 MODELS components are ignored during the phasor solution. Nonlinear
inductors are linearized.

Sinusoidal steady-state phasor solution, branch by branch. All flows are away from a bus, and the real part, magnitude, or "P"
is printed above the imaginary part, the angle, or "Q". The first solution frequency = 5.00000000E+01 Hertz.
Bus K Phasor node voltage Phasor branch current Power flow Power loss
Bus M Rectangular Polar Rectangular Polar P and Q P and Q

HEVIA 488774.954 488774.954 .488774954E-4 .488774954E-4 11.945047782885 11.945047782885


0.0 0.0 0.0 0.0 0.0 0.0

TERRA 0.0 0.0 -.488774954E-4 .488774954E-4 0.0


0.0 0.0 0.0 -180.0000000 0.0

HEVIA 488774.954 488774.954 0.0 .44791619154764 0.0 0.0


0.0 0.0 -.4479161915476 -90.0000000 109465.10795978 109465.1079598

TERRA 0.0 0.0 0.0 .44791619154764 0.0


0.0 0.0 .44791619154764 90.0000000 0.0
Total network loss P-loss by summing injections = 1.194504778289E+01

Output for steady-state phasor switch currents.


Node-K Node-M I-real I-imag I-magn Degrees Power Reactive
GENEA HEVIA 4.88774954E-05 -4.47916192E-01 4.47916194E-01 -89.9937 1.19450478E+01 1.09465108E+05

Solution at nodes with known voltage. Nodes that are shorted together by switches are shown as a group of names, with the printed
result applying to the composite group. The entry "MVA" is SQRT( P**2 + Q**2 ) in units of power, while "P.F." is the
associated power factor.
Node Source node voltage Injected source current Injected source power
name Rectangular Polar Rectangular Polar P and Q MVA and P.F.

GENEA
HEVIA 488774.954 488774.954 .488774954E-4 .44791619421444 11.945047782885 109465.10861151
0.0 0.0 -.4479161915476 -89.9937478 109465.10795978 0.0001091

Card of names for time-step loop output. | GENEA


Blank card ending requests for output variables. |BLANK TERMINAN PLOTEOS

Column headings for the 6 EMTP output variables follow. These are divided among the 5 possible classes as follows ....
First 3 output variables are electric-network voltage differences (upper voltage minus lower voltage);
Next 2 output variables are branch currents (flowing from the upper node to the lower node);
Next 1 output variables belong to MODELS (with "MODELS" an internally-added upper name of pair).
Branch power consumption (power flow, if a switch) is treated like a branch voltage for this grouping (1 variables);
Branch energy consumption (energy flow, if a switch) is treated like a branch current for this grouping (1 variables).
Step Time HEVIA GENEA GENEA GENEA HEVIA MODELS
TERRA HEVIA HEVIA TERRA FLUJO
*** Phasor I(0) = 4.8877495E-05 Switch "GENEA " to "HEVIA " closed in the steady-state.
0 0.0 488774.954 23.8900956 488774.954 -.3342E-20 -.1092E-17 0.0
% % % % % % Final time step, PLTFIL dumps plot data to ".PL4" disk file.
Done dumping plot points to C-like disk file.
10000 0.1 488774.954 87756.0239 488774.954 9841.44696 .179493923 -.29612E-8

Input data exhausted! CIMAGE will pass back innocuous blank card.
Input card pointer NUMDCD = 1. Exit "CIMAGE" with blank card image.
Blank card terminating all plot cards. |BLANK card build by CIMAGE because no more input exists.
.....

EEUG News Feb - May 2000 27


Program Development

The plot of flux - current curve is shown on Fig. 4.

Fig. 4

The zero frequency component of the Fourier spectrum of the power, with a fundamental
frequency of 100 Hz, turns out to be 101535.9 W, very near to the value of 101519.6 W
obtained by formula (2).

The error for rated voltage can be reduced by readjustment of exponent a. Doing so, the
power for the maximum loop varies in the same sense that the exponent, and as a consequence,
the losses for rated voltage. Through comparisons with this exponent (in reality, with the
power of losses for maximum cycle), the value of calculated hysteresis losses can be made
practically equal to that of the measurements. For the maximum loop the error is generally
smaller.

It is worth to clarify that for the examples shown above the mistakes are small. In most of the
cases where no real hysteresis data is available, this small mistake is more than acceptable. The
error that is committed upon accepting an allotment of losses between eddy currents and
hysteresis certainly much greater.

5 Acknowledgement

Author is thankful for the collaboration of Dr. W. Scott Meyer, who implemented a primitive
version of the routine in the source code of ATP and for the measurement of the hysteresis
loop performed by Ing. Eduardo Lalla.

28 EEUG News Feb - May 2000

You might also like