Reflections on Volt/Var Control by Multiple Smart Inverters of DERs

Reflections on Volt/Var Control by Multiple Smart Inverters of DERs

Nokhum Markushevich

The performance of the Volt-var control function by a single smart inverter was discussed in [1]. In this article, the performance of the function by multiple smart inverters is addressed.

The sample circuit used for the analysis is presented in  Figure 1.

Figure 1. Sample circuit

The volt-var curve for all inverters of the sample circuit is presented in Figure 2.  It is, basically, a CVR oriented curve. The simulation of the volt-var control is based on the recursive formula as presented in [1]. 

Figure 2. Volt-var curve for all inverters in the sample circuit

First, we simulated autonomous volt-var control by all inverters for a sunny day, when all DERs generate 100% of their kW capabilities.  The relationships between the local grid sensitivities and the droop values of the corresponding volt-var curves are presented Figure 3. The droop values for different DERs are different because the available kvars are different.  As seen in the figure, the grid sensitivities are smaller than the droop values for all nodes. It means that a convergence of the volt-var control can be expected [1]. The resultant voltages for consecutive iteration are presented in  Figure 4.

Figure 3. Grid sensitivities and droop values.

Figure 4. Convergence of the volt-var control

The volt-var control simulation was performed for different bus voltages. The results are presented in Figure 5. As seen in the figure, the target voltage is reached for the most of the nodes under the lowest bus voltage. It means that the distribution system operator (DSO) should keep the bus voltages in accordance with the abilities of the autonomously controlled smart inverters. This, in turn, implies some information support for the DSO from the DER’s controllers.  If the DSO cannot keep the voltage required for this particular circuit, the voltages in some nodes will deviate from the target. In this example, we estimate the voltage quality by the kW-weighted deviation of the voltage from the target. The smaller this index, the closer is the voltage to the target.

Figure 5. Secondary voltages for different bus voltages

In Figure 6, two components of the voltage quality index are presented: the component for the deviation above the upper target and the deviation below the lower target voltage.  As seen in the figure, the “above” voltage quality index is increasing with the increase of the bus voltage. However, the total absorbing DER’s kvar limit is not reached. It is possible that there still is a potential for the smart inverter to improve the overall voltage. The autonomous volt-var control does not recognize voltage deviations in remote nodes. To utilize this potential, an involvement of the DSO in the overall volt-var control is needed.

Figure 6. kvar parameters of DERs and voltage quality indices.

We will address such an involvement based on the simulation of the volt-var control for the cloudy conditions. Under these conditions, the DER’s generate about 10% of their kW capabilities, and the available kvars are much greater than in the previous case. That is why the droop values become smaller, and they can be smaller than the grid sensitivities in some nodes.  As seen in Figure 7, this happened for nodes 1207, 1206, and 1210. This means that the volt-var simulation may not converge.

Figure 7. Grid sensitivities and droop values for cloudy conditions

As seen in Figure 8, the simulation diverges in all nodes, although the grid sensitivity is greater than the droop in three nodes only. The instability of the “guilty” nodes destabilizes the entire circuit.

Figure 8. Volt-var simulation does not converge

In order to increase the droop values in the critical nodes we reduced the available kvars of the volt-var curve for these inverters by 50% (see Figure 9).

Figure 9. Volt-var curve for nodes 1206, 1207 and 1210.

The adjusted relationships between the grid sensitivities and the droop values are presented in Figure 10. As seen in the figure, the droop values are bigger than the grid sensitivities in all nodes.  The convergency of the simulation under these conditions can be seen in Figure 11.

Figure 10. Adjusted relationships between the grid sensitivity and the droop values.

Figure 11. Volt-var simulation converges

The process converges, but the voltage quality is not satisfactory, as can be seen in  Figure 12 and Figure 13. It is also seen in Figure 13 that the DERs’ total kvars by far did not reach the available limits.

Figure 12. Secondary voltages under cloudy conditions and autonomous volt-var control by smart inverters.

Figure 13. DER’s kvars and voltage quality indices under cloudy conditions and autonomous volt-var control

Trying to keep the voltage at the bus on a low level for the CVR purpose, e.g., at 1.01pu, resulted in secondary voltage deviations below the lower border of the voltage tolerance in nodes 1209 and 1210.  There are no voltages above the upper border of the tolerance.  At the same time, there are unused generation kvars of some DERs.

The unused generation kvars of a DER j after the completion of the autonomous control can be defined as follows:

(Unused kvars)j = (Current kvar generation limit)j – (current kvar generation)j   (1)

However, not all unused kvars can be utilized to raise voltages in the critical points. The additionally available kvars can be limited by nodal voltages due to the nodal grid sensitivity to their own or to a remote DER’s kvar additions. Additional kvars generated by a DER changes the voltages not only in its own node, but in other nodes due to the nodal grid voltage sensitivity to the kvars of remote DERs. The grid voltage sensitivities to the remote DERs for our example are presented in Figure 14.

Figure 14. Nodal grid sensitivities to DER’s kvars

Hence, the additional kvars from a particular DER can be limited by the room for voltage increase in any node due to the nodal grid sensitivities to its own or remote DER.

The room for additional voltage increase in node i is defined by the difference between the upper tolerance limit V3 and current voltage in this node Vi after the autonomous control:

ΔVi= V3 – Vi   (2)

The resultant additional dispatchable kvars from DER j (Δkvarj) is defined as:

Δkvarj = Min {ΔVi/(Grid sensitivity)I,j  ,  (Unused kvars)j}  (3)

 The operating conditions after the completion of the autonomous volt-var control are presented in Table 1. As seen in the Table, the voltages in nodes 1209 and 1210 are below V2 by 0.14% and 1.46% respectively. There are no unused kvars of the DERs in these nodes. However, there are additional dispatchable kvars of other DERs. The DSO may issue a broadcast request to the DER operators to use maximally available kvars to mitigate the low voltage in the critical nodes. If the dispatchable kvars of all DERs are added to raise voltage in the critical node 1210, the total initial effect can be expected to be +2.04%. The results of such exercise are presented in Table 2. 

Table 1. Results of autonomous volt-var control for the bus voltage 1.01pu.

Table 2. Results of a centralized addition of all dispatchable kvars.

[1] A negative number means that the voltage exceeded the upper limit

As seen in Table 2, the voltage deviation from the tolerance was eliminated for node 1209, and it was reduced to 0.17% for node 1210. However, the voltages in a number of other nodes exceeded the upper border V3 due to the cumulative effect of the additional dispatchable kvars.  Note that the total effect on the voltage in node 1210 was smaller that the originally expected 2.04%. It was 1.29%.  It could be due to responses of other operating parameters to the introduced changes, e.g., load dependencies on voltage, as well as due to the performance of the iteration process.

Another attempt to increase the voltage in node 1210 was based on addition of dispatchable kvars of selected DERs. The selection was based on the most “influential” DERs in reference to voltage in node 1210. As follows from Table 1, these are DERs in nodes 1205, 1206, and 1207. The initially expected cumulative effect of these additions on the voltage in node 1210 is 1.54%. The results of this exercise are presented in Table 3. As seen in the Table, the voltage in node 1210 was increased by 1.05%, and there were no voltage deviations above the upper limit V3 in other nodes. The total (below and above) voltage quality index was significantly better in comparison with the autonomous control. As also seen in the table, not much of additional dispatchable kvars of other DERs were left after these additions of kvars in the selected nodes.

Table 3. Results of a centralized addition of selected dispatchable kvars.

Conclusions.

1. Simulations of volt-var control by a number smart DER inverters based on proposed volt-var curve was performed for a 13-node balanced circuit model. 

2. Based on these simulations, the following conclusions are suggested:

   1. The stability of autonomous volt-var control depends on the relationships between the droop of the volt-var curve and the grid voltage sensitivity to DERs contribution of kvars.

   2. Instability of volt-var control by a subset of DERs may lead to instability in the entire closely connected circuit.

   3. The relationships between the drop of the volt-var curve and the grid voltage sensitivity may be changing in near real-time due to the change of the DERs’ available kvars and/or change of the circuit connectivity.

   4. The settings of the volt-var curves of some inverters may need to be change to keep up with the changes of the available DER kvars and grid sensitivities.

   5. Not all information needed for the updates of the volt-var curve is locally available to the DER’s operators. 

   6. Autonomous control may not utilize all available resources for optimal volt-var control in the relevant distribution circuit.

   7. Involvement of the DSO (DMS) may be needed either to mitigate violations of operational limits, or to optimize the overall operating conditions.

   8. The DSO (DMS) decision-making process involves information about multiple distributed variables and an analysis of their effect on a number of operational parameters, such as voltages, currents, losses, and demand.

3. This centralized analysis is a near real-time optimization process, which requires significant information support and analytical capabilities.  Therefore, the distribution system planning should include an assessment of the incremental benefits and cost of the centralized volt-var control.

Reference.

1. N. Markushevich, Reflections on the Topic of Volt/Var Control by Smart Inverters of DERs. Available: https://www.energycentral.com/c/cp/reflections-topic-voltvar-control-smart-inverters-ders