Comparison of measurements and DMS600 calculations

Tested equipment with an earth fault current compensation capability and a reactive power compensation capability was modeled in MicroSCADA Pro DMS600 before the measurements. The model of the equipment consisting of a distribution transformer and a wye connected shunt reactor with earthed star point is presented in Figure 32.

Figure 32. The model of a MV-LV substation consisting of a distribution transformer and a wye connected reactor with earthed star point.

Since there is no possibility in DMS600 to model a shunt reactor, a capacitor was coded to DMS600. The nominal reactive power of the capacitor was set to be negative to repre-sent the consumption of reactive power. In order to model the coil’s earth fault current compensation capability an arc suppression coil was coded to the MV-LV substation.

The active power consumed in MV-LV substation 06355 includes the losses of the shunt reactor and the losses of the distribution transformer. The LV-network of the transformer was disconnected so the losses of the transformer were formed merely by the no-load losses of the distribution transformer. The nominal losses of the shunt reactor are 7550 W and no-load losses of the distribution transformer are 405 W, so the losses are mainly generated in the shunt reactor. In order to verify the accuracy of the modeled equipment, calculated and measured reactive and active powers in case of different connection volt-ages are presented in Figure 33 and in Figure 34.

Figure 33. The measured and calculated reactive power values in case of three different voltages affecting over shunt reactor.

Figure 34. The measured and calculated active power values in case of three different voltages affecting over shunt reactors

It can be seen from Figure 33 that the reactive power is accurately calculated in DMS. In Figure 34 it is shown that the calculated active power remarkably differs from the meas-ured active power. The differences between active powers is mainly due to the losses in shunt reactor. In DMS the reactor is modeled with an ideal capacitor, and the losses of the substation are constant no-load losses of the distribution transformer.

6.3.2 Effect of shunt reactors on voltages along feeder

In order to test the accuracy of state estimation, the measured voltages along the feeder are compared to voltages that are calculated based on estimated loadings. Measured pow-ers along the feeder and in the beginning of the feeder can be used in the DMS's load estimation. Usually only measurements from the beginning of the feeder are used because there are no measurements available from the MV-LV substations. Measured voltages are

0

compared to calculated voltages in two different situation. In the first case the loads are estimated based on the measured powers and the voltage in the beginning of feeder. In second case also measured powers from 3 MV-LV substations along the feeder are used in estimation. The measured voltages along the feeder are compared to calculated volt-ages during different voltage levels in substation with and without distributed shunt reac-tors. Results in case the shunt reactors were disconnected are presented in Figure 35.

Results in case shunt reactors were connected are presented in Figure 36. In both figures the voltages of three MV-LV substation are presented in case of three voltage level. Clus-tered column “DMS” represents estimated voltages in case only measured P, Q and V values from substation were used. Clustered column “DMS2” represents calculated volt-ages in case also measured powers from the MV-LV substations were used. The measured powers from the MV-LV substations along the feeder that were used in estimation are presented in Appendix 4. Measured and estimated voltages are presented in Appendix 5.

Estimation tests were done after the real measurements by adding the saved measurement values in each step into DMS-database. The SQL-queries that were used in setting of measurement values are presented in Appendices 6 and 7.

Figure 35. The measured and estimated voltages along inspected feeder in case of dif-ferent tap changer positions. Shunt reactors are not connected.

18 18,5 19 19,5 20 20,5 21

6355 6376 6460 6355 6376 6460 6355 6376 6460

kV

MV-LV substation

DMS DMS2 Measured

Figure 36. The measured and estimated voltages along inspected feeder in case of dif-ferent tap changer positions. Shunt reactors are connected.

Calculated voltages are higher than measured voltages in every measurement points. The differences between measured and calculated voltages is in worse case about 200 V. The use of power measurements along the feeder seem not to make the voltage calculation more accurate in this system. The difference remains that big probably because of the other distribution transformers along the line branches without measurements. Also the fact that the load curves were in somewhat different state at the time the test was done than at the time measurements were done makes some error to the test results. The effect of load curves can be assumed small because there are no loads in inspected feeder that would have large variations in consumption.

However there was no load between the HV-MV substation and the MV-LV substation 06355 so it is expected that the voltage of that line section would be accurately calculated.

In order to test the DMS’s calculation method, the voltage difference occurring in that line section was also calculated with Matlab. The line section was modeled with one pi-section and the receiving end voltage was calculated with the measured sending end volt-age and current. The used Matlab-script is presented in Appendix 8. Line parameters used in calculation are presented in Table 3.

Table 3. The parameters of the first line section.

Length [m] Resistance [Ω/km] Reactance [Ω/km] Susceptance [µS/km]

3040 0,125 0,110 94,20

The measured receiving end voltage and voltage calculated with Matlab and DMS are presented in Figure 37. Only case that is above presented in Figure 35 is presented. Shunt reactors are disconnected.

6355 6376 6460 6355 6376 6460 6355 6376 6460

kV

MV-LV substation

DMS DMS 2 Measured

Figure 37. Voltages in 06355. Measured, Matlab (pi-section) and DMS600

It can be seen in Figure 37 that the results of DMS and Matlab pi-section does not differ from each other much. Therefore the differences between the measured and calculated voltages are caused from the improper calculating parameters rather than improper cal-culation method. The accuracy of voltage measure is ± 0,5 %, the accuracy of reactive power measure is ± 3 %, accuracy of active power measure is ± 1 % and accuracy of current measure is ± 5 % or 1 A [34]. The inaccuracies of measurement values explains part of the differences. One possible reason for differences between calculated and meas-ured voltages might be that the resistance of cable is bigger than resistance used in calcu-lation because there are multiple medium voltage cables close to each other. If the cable warm up it might cause the resistance of the cable to rise. On the basis of ohm’s law the voltage drop is higher in lines with higher resistance.