The real power generated by renewable Distributed Energy Resources (DERs) changes with the cloud cover and the wind. These changes may significantly change the voltages, especially in the distribution circuits. The voltage fluctuations caused by the variable output of the DER may be unacceptable from the standpoint of power quality as well as that of the reliability of other voltage and var controlling devices.
The voltage fluctuations can be compensated either by increasing the generation of kvar, or by reducing the absorption of kvar. To mitigate the voltage fluctuations resulting from the change of reactive power, there must be a sufficient reserve to increase the generation of kvar or to reduce the absorption of kvars.
Following the capability curve, when the kW of the DER falls below its full capacity, the generation of the kvars may exceed the rated kvar value; i.e., even if the initial kvars (under normal weather conditions) are at their nominal value, there still is an additional reserve of kvar for when the kW output drops. It needs to be checked whether this is enough for the compensation of the expected voltage fluctuation. If it is not enough, then the initial kvars should be set below its nominal value.
The voltage fluctuations at any node of the circuit with DER may be caused not only by the variability of the DER, but by also other effects that impact voltage and vars. It means that fluctuations caused by the variability of one group of DERs can be at least partially compensated by another group of DERs. Also it should be taken into account that different methods of compensating the voltage fluctuations affect different aspects of the distribution system, such as power losses, circuit loading, benefits of volt/var optimization, etc.
Some of the issues mentioned above were addressed in  and . This paper analyses the mitigation of voltage fluctuations in the distribution circuits by adjusting the generation or absorption of the DER reactive power based on two different mythologies presented in Section II. The analysis includes an assessment of the methodologies by considering their impact on the compensation of the voltage fluctuations, on power losses, on circuit loading, and on reactive power supply.
Section III presents an illustration of the analysis based on power flow calculations for a simple circuit taking into account the DER capability curves and the load-to-voltage dependencies.
The conclusions are presented in Section IV.
II. Methods of Mitigation of Voltage Fluctuations
Two methods of mitigation of voltage fluctuations due to the DER variability are discussed below:
If the DER is set to maintain a given power factor (PF) to generate reactive power, a drop in active power of the DER will cause a reduction in the generation of reactive power. This will increase the voltage fluctuations even more. But if the DER is set to maintain a given power factor to absorb kvars, a reduction in real power will cause a reduction of the absorption of reactive power, thus acting to reduce the voltage fluctuations. However, in such mode of operations of the DER, the reserve for mitigation of the voltage fluctuations is limited by the ability to reduce the absorbing kvars to zero and cannot be extended into generating kvars. Operating DERs in kvar-absorbing modes increases the need for compensation of the reactive loading of the distribution and transmission circuits. It also may increase the power losses in distribution and transmission, the loading of the critical circuit elements, and the voltage drops in portions of the distribution circuits. Also, in the constant power factor mode of operations, the reactive power of the DER is changed only when the real power of the same DER is changed. Hence, this method does not compensate for the voltage fluctuations caused by the variability of other DERs.
In this mode of operations, the reactive power of DER can change from generating kvars to absorbing and vice versa according to the DER capability curve, following the deviations of voltage. This means that the generating or absorbing kvars can exceed the rated kvars when the kWs are below the rated value. This is a closed-loop control of vars with the voltage as a target. Measures for prevention of "hunting" among several DERs should be considered when setting up this mode of operations. It is assumed here that under the "constant voltage" mode there is not a predefined voltage setpoint that the DER should try to follow. The voltage at the DER terminals (or at another target node) at any time is established based on the current operating conditions of the corresponding distribution circuits and on the operational capabilities of the particular DER . The initial volt/var requirements for the DER under normal conditions can be defined by the controlling entity based on the current operational objectives and the reserve needed to mitigate the expected voltage fluctuations. The objective of the "constant voltage" mode of DER operations considered in this paper is to keep the voltage close to the initial (normal) level by compensating for the voltage deviations from this level by adjusting the reactive power of the DER. In this case, instead of following a predefined voltage setpoint, which in some cases cannot be reached, the "constant voltage" mode of operations will follow setpoints that will be close to the normally established voltage levels at the selected target bus. For instance, such setpoints can be derived as a running average of a given number of previous voltage measurements, or as a running average over a specific time interval.
The illustration of the differences of these two methods is based on two simple circuit diagrams presented Figure 1 and Figure 2.
As seen in the figures, there are two DERs connected either to the secondary voltage buses, or to the primary buses. The following three ratios of the reactance to the resistance of the primary circuits have been considered: 1 to 1; 2 to 1; and 3 to 1. For the constant power factor mode, two settings of the absorbing power factor were considered: 0.95 and 0.9. For the constant voltage mode, the changes of the DERs' kvars were assumed to be available within the capability curve defined by kVA rating of the DER.
The resistance and the reactance of the distribution transformers are assumed to be 1% and 2.5% respectively, and the voltage drop in the secondaries is 2% of the nominal voltage when the DER is out of operations. The CVR-watts factor is 1, and the CVR-vars factor is 4. The load power factor is ~0.9.
In all scenarios, the feeding bus voltage was adjusted to keep the lowest secondary voltage close to the lower voltage limit (for energy conservation).
The voltage fluctuations were considered to be acceptable if their standard deviation did not exceed 0.125% of the nominal voltage.
The following criteria were used to compare the two methodologies:
The results of the analyses are presented in the Table 1 below. The numbers in the table are relative numbers arranged to provide consistency between the assumed circuit parameters, loads, generation, realistic voltage drops and power losses).
The stiffness of the sample circuit is different for different X/R ratios. Therefore, the criteria can be compared only for circuits with the same X/R ratio.
As seen in the table, for X/R = 1, the constant PF method reduces the standard deviation of the voltage fluctuations in comparison with the no-compesation case (e.g., when the DER PF=1), but not enough to meet the stardand deviation criteria assumed here of less than 0.125%. The constant voltage method reduces the voltage fluctuations below the standard deviation of 0.125%.
For X/R = 2, the constant absorbing PF method reduces the standard deviation of the voltage fluctuations below 0.125% when the PF of the DERs is 0.9. When the PF is 0.95, there is not a sufficient reserve for the reduction of the absorbing reactive power to adequately reduce the voltage fluctuations. The constant voltage method reduces the standard deviation of the voltage fluctuation below 0.125% in all cases considered.
Even when the constant PF method reduces the standard deviation of the fluctuations below the assumed criterion, the power losses and the real and reactive loading of the circuits are greater than when the constant voltage method is applied. The losses and the required reactive loading of the circuits are increased due to the greater consuption of the reactive power by the DERs. Also, despite having the same lowest voltage level at the voltage-critical point, the voltages in other nodes are higher in the constant PF method than in the constant voltage case. This is due to a greater voltage drop along the circuit (i.e., a steeper voltage profile along the feeder). Hence, the real and reactive loads are greater in these nodes due to the CVR factors.
For X/R = 3, the constant PF method reduces the standard deviation of the voltage fluctuations below 0.125% when the PF of the DERs is 0.95, except for the standard deviation in Node 2 when the DER is connected to the secondary circuit. The constant voltage method reduces the standard deviation of the voltage fluctuation below 0.125% in all cases considered. As before, the values of other criteria in the constant PF case are higher than in the constant voltage case.
The author would like to thank Mr. Martin Delson for his comments on this paper.