Integration of Renewable Generators in Synthetic Electric Grids for Dynamic Analysis

Fig. 2 shows the block diagram for the governor (drive-train) model WTGT_A. In these block diagrams, the inputs are colored blue, the autocorrection properties are colored green and the outputs are colored yellow. The parameter assignment process of all low level control modules is similar with high level control modules discussed in Section II-B. Various combinations of parameters are extracted and summarized in Table IV. Differentiated by the type of governor response limits, there are two general types of parameter templates, with the total count of 9 and 14 respectively. The range of parameters in these templates are also listed for each type of response limits. The next step is to assign a parameter template to each WTG to match the relative probability that these templates appear in the actual data set.

Fig. 3 shows the block diagram for the stabilizer (pitch-controller) model. Seven parameter templates are extracted from the actual case, with differences in proportional/integral gain of pitch control/compensator as listed in Table V. RThetaMin = -10 and RThetaMax = 10 are considered since 100% of modules in the WECC case have RThetaMin of -10 and RThetaMax of 10. For parameter (ThetaMin, ThetaMax), values are given (0, 17), (-4, 27) and (0, 27) by probabilities of 0.11, 0.17 and 0.72, respectively. Note, this module can only be used with a Type 3 WTG.

Fig. 4 shows the block diagram for the Pref Controller (torque) model. As shown in Table VI, two general types of templates are extracted from actual cases in regard to the control flag with the total amount of 8 and 1 respectively. Last, Twref is set to 60 since 100% of Twref from the WECC case is 60. Different f(Pe) functions are also considered in the parameter template, but the details of each function are omitted in this paper due to limitation of pages. Note, this module can only be used with a Type 3 WTG.

The default data set for this module is used since 100% of the modules used from the studied cases are utilizing the default values. Also, this module can only be used with a Type 3 WTG.

Fig. 6 shows the simulation results for a PV plant with five individual solar generators experiencing a voltage drop event caused by bus fault at the POI. The event starts at 1.00s and is cleared at 1.05s. The mode of operation of the PV plant is plant level V control plus local coordinated V/Q control. At the beginning of the simulation, the plant controller module REPC_B is constantly monitoring the bus at POI (i.e., the regulated bus). At 1.00s, the voltage magnitude at the regulated bus falls below the pre-defined criteria and the voltage control function in REEC_A module is activated. The reactive power output of all the solar generators greatly increased to provide voltage support to the regulated bus. This functionality is vital since during a fault condition the renewable generators should keep connected to reduce the effect of event. Simulation results show this operation mode has impact on the voltage recovery and it’s necessary to incorporate this function in renewables.

Fig. 7 provides the simulation results for a wind plant with two wind turbine generators undergoing an over-frequency event at the POI. The mode of operation at the wind power plant is governor response with down regulation, only (i.e., with no operating reserve for primary frequency response). The event begins at 1.00s, then starts to recover at around 8s and reaches steady-state at around 27s. The plant controller is continuously monitoring the bus at POI. At 1.00s the bus frequency at the regulated bus starts to increase and the frequency control function in REPC_B module is activated. The real power of all the wind generators in the plant start to ramp down and provide downside frequency support to the system. This feature is of great importance in maintaining the power system frequency’s quality and thus it should be included when modeling system dynamics.

Fig. 8 provides the simulation results for a wind plant with two wind turbine generators undergoing a under-frequency event at the POI. The operation mode at the wind power plant is governor response with up and down regulation. The frequency drop event starts at 1.00s, the frequency at POI starts to recover and finally at around 25s return to steady-state. The plant controller is constantly monitoring the regulated bus and the frequency control function is activated when the frequency starts to drop. Then the real power output of the controlled WTGs start to ramp up and provide upward primary frequency response to the system.

Note that providing downward regulation for an over-frequency event is always possible no matter whether or not the renewable resources are operating at the maximum power point. However, only renewable generators that are operating at a sub-optimal power point (i.e., with additional energy headroom to sustain the response) can provide the upward frequency regulation. Operating renewable power plants at a sub-optimal condition to provide upward frequency response have economic consequences while this ability provides a vital reliability benefit [26]. Both aspects are considered in this work. About 15$\%$ of all renewable generators are set to be operating at a sub-optimal condition, which is the same as we have summarized from the actual WECC cases.

After testing the plant-level or converter-level control functionalities of the renewable generators integrated 2000-bus case, this part will focus on demonstrating and validating the dynamic response of the full system. As illustrations, a N-2 contingency event and a bus fault event are considered in the simulation. Case details for both the original 2000-bus case (“Original Case”) and the renewable generators integrated case (“Renewable Case”) are given in Table. VII. The 109 renewable generators are assigned with suitable dynamic models and parameters in the Renewable case, while only power flow data for these generators is available in the Original Case.

The first example is a N-2 contingency event. Two large generators with total real power output of 2589 MW are disconnected from the system at 1.00 s. The simulated bus frequencies and voltages collected from the Renewable Case are displayed in Fig. 9 in light-grey lines. Some representative values are plotted in solid color lines. We observe that the bus frequencies (voltages) start to recover soon after the contingency event and settle down within 20s.

The second example is a bus fault event. A three-phase fault is applied to a 115kV bus at 2.00s and cleared at 2.05s. Simulation results for frequencies of all buses are plotted using light-grey lines in Fig. 10. Similarly, the solid color lines represent some representative bus frequency values. The fast dynamics in bus frequencies are eliminated within seconds after the fault is cleared as shown in the results. The bus voltages drop immediately when the fault applies and they begin to recover at the fault clearing time. The post-fault voltages amplitudes of all buses are approaching to their pre-faut values.

Reference [9] considers three transient stability metrics for N-1 contingencies: the minimum damping ratio of generator rotor angle ($M_{r}$), the minimum and maximum bus frequencies after the last contingency event ($M_{f}$) and the minimum ratio of bus voltage nadir to pre-contingency value ($M_{v}$). The transient stability requirements of these metrics are $M_{r}$ larger than 3$\%$, $M_{f}$ between [59.5, 60.5] Hz and $M_{v}$ larger than 75$\%$.

Fig. 12 depicts the calculations of these metrics. In the boxplot, the blue box indicates the results of a generator outage event and the green box indicates the results of a bus fault event. The upper and lower extreme lines shows the maximum and minimum value. A box is created from the first quartile to the third quartile of each metric and the solid maroon line which goes through the box indicates the median. These results verify that the Renewable Case meets the transient stability requirements after being subjected to selected contingencies.

Simulation results for a N-2 and a N-1 contingency event are given as examples. The results for the N-2 event (loss of two large generators at 1.00s) are given in Fig. 13. The frequencies and voltages for all buses are plotted with light-grey lines and selected illustrative signals are highlighted in solid color lines. We can see that bus frequencies and voltages start to vary and oscillate when the N-2 event happens at 1.00s, but the oscillations are damped out within 30s. The post-contingency frequencies and voltages are at acceptable levels for this severe and rare event.

Fig. 14 displays the simulation results for a N-1 contingency event. A balanced three-phase fault is applied to a 161kV bus at 1.00s and then cleared at 1.05s. The light-grey area indicates the results for all bus frequencies. Similarly, solid color lines denotes signals that are representative. After the fault is cleared, the frequency variations decline gradually and the system enters an equilibrium condition. We notice that this renewable generators integrated case has satisfactory dynamic responses and meets transient stability requirements. Also, this case performs to have satisfactory system responses and transient simulation results for a severe N-2 event.

In general, the proposed approach can be used to model renewable generators in large-scale network models. The generated case can be used in many power system studies. One application is to use this renewable generation integrated case as the base for interactive simulations developed in [27] and evaluating the participant’s performance using metrics described in [28].