Figure 1.
A schematic showing optimal control of hybrid MTDClinked grids and OWFs.
Figure 1.
A schematic showing optimal control of hybrid MTDClinked grids and OWFs.
Figure 2.
Use of an LQG regulator to generate optimal reference signals for the individual local controllers of the HVDC system and the OWFs: (a) frequency control and (b) voltage control.
Figure 2.
Use of an LQG regulator to generate optimal reference signals for the individual local controllers of the HVDC system and the OWFs: (a) frequency control and (b) voltage control.
Figure 3.
A smallsignal model of HVDClinked grids considering the dynamics of the DC link, the converters, and the inner feedback loops.
Figure 3.
A smallsignal model of HVDClinked grids considering the dynamics of the DC link, the converters, and the inner feedback loops.
Figure 4.
Responses to (left) step and (right) continuous variations in the inverterside and rectifierside loads and in the OWF power generation when the LQG strategy (Case 1) and conventional PIbased strategies (Cases 2 and 3) are applied: (a) inverterside frequency f_{i}, (b) rectifierside frequency f_{r}, (c) DClink voltage V_{dc}, and (d) DC current I_{dc}.
Figure 4.
Responses to (left) step and (right) continuous variations in the inverterside and rectifierside loads and in the OWF power generation when the LQG strategy (Case 1) and conventional PIbased strategies (Cases 2 and 3) are applied: (a) inverterside frequency f_{i}, (b) rectifierside frequency f_{r}, (c) DClink voltage V_{dc}, and (d) DC current I_{dc}.
Figure 5.
Voltage responses to (a) continuous variations in OWF power generation, (b) a communication time delay t_{d}, and, (c) uncertainties in inductance and resistance parameters when a novel SVC strategy (Case 1) and conventional PI and MPCbased SVC strategies (Cases 2 and 3) were applied. In (a), v_{O} and v_{C} are the POC and PCC voltages, respectively, and v_{W}_{1} and v_{W}_{2} are the OWF terminal voltages.
Figure 5.
Voltage responses to (a) continuous variations in OWF power generation, (b) a communication time delay t_{d}, and, (c) uncertainties in inductance and resistance parameters when a novel SVC strategy (Case 1) and conventional PI and MPCbased SVC strategies (Cases 2 and 3) were applied. In (a), v_{O} and v_{C} are the POC and PCC voltages, respectively, and v_{W}_{1} and v_{W}_{2} are the OWF terminal voltages.
Figure 6.
(a) A reconfigurable network with switches and synchronous machinebased and inverterinterfaced generators and (b) the variations in the topology that reduce power loss and restore and shed load.
Figure 6.
(a) A reconfigurable network with switches and synchronous machinebased and inverterinterfaced generators and (b) the variations in the topology that reduce power loss and restore and shed load.
Figure 7.
The (a) conventional and (b) new methods for NR modeling.
Figure 7.
The (a) conventional and (b) new methods for NR modeling.
Figure 8.
The grid frequency and voltage profiles of the new, conventional, and comprehensive models of NR: (a) Case 1 (without NRaided load shedding/restoration) and (b) Case 2 (with NRaided load restoration).
Figure 8.
The grid frequency and voltage profiles of the new, conventional, and comprehensive models of NR: (a) Case 1 (without NRaided load shedding/restoration) and (b) Case 2 (with NRaided load restoration).
Figure 9.
Schematic of a proposed FR strategy for an islanded, reconfigurable MG.
Figure 9.
Schematic of a proposed FR strategy for an islanded, reconfigurable MG.
Figure 10.
A smallsignal model of an islanded reconfigurable MG with supplementary FFCs showing the feedback loops for the inertia response emulation and primary and secondary frequency control.
Figure 10.
A smallsignal model of an islanded reconfigurable MG with supplementary FFCs showing the feedback loops for the inertia response emulation and primary and secondary frequency control.
Figure 11.
(a) Bode plots of the Δf(s)/ΔP_{L}(s) values for the new FR strategy (Case 1) and conventional FR strategies (Cases 2–4); (b) Bode plots of Δf(s)/ΔP_{L}(s) values with the errors in DG parameter estimates; and (c) comparisons of f, V, Σ_{g} P_{Mg}, and Σ_{i} P_{IGi} (from top to bottom) during NRaided load restoration.
Figure 11.
(a) Bode plots of the Δf(s)/ΔP_{L}(s) values for the new FR strategy (Case 1) and conventional FR strategies (Cases 2–4); (b) Bode plots of Δf(s)/ΔP_{L}(s) values with the errors in DG parameter estimates; and (c) comparisons of f, V, Σ_{g} P_{Mg}, and Σ_{i} P_{IGi} (from top to bottom) during NRaided load restoration.
Figure 12.
(a) A flowchart for implementation of the proposed FVCs and (b) comparisons of V_{DG}, V_{Load}, Q_{DG}, and P_{DG} (from top to bottom) between the new proposed VR strategy (Cases 1 and 2) and conventional VR strategies (Cases 3 and 4) for NRaided load restoration.
Figure 12.
(a) A flowchart for implementation of the proposed FVCs and (b) comparisons of V_{DG}, V_{Load}, Q_{DG}, and P_{DG} (from top to bottom) between the new proposed VR strategy (Cases 1 and 2) and conventional VR strategies (Cases 3 and 4) for NRaided load restoration.
Figure 13.
(a) A schematic of a new electrical system inside a commercial building and (b) the experimental setup of an MG including a generator emulator, a battery pack, and a load emulator.
Figure 13.
(a) A schematic of a new electrical system inside a commercial building and (b) the experimental setup of an MG including a generator emulator, a battery pack, and a load emulator.
Figure 14.
A block diagram of the ancillary service provision by buildings including EV batteries.
Figure 14.
A block diagram of the ancillary service provision by buildings including EV batteries.
Figure 15.
Experimental case study results for the (left) conventional and (right) proposed FR strategies: (a) frequency deviations Δf, (b) load demands P_{L}, dispatchable generation P_{G}, and total generation P_{G} + P_{PV}, (c) variations in generator output power ΔP_{G} and in battery input power ΔP_{B}, and (d) dynamic ramping rate of generation dΔP_{G}/dt.
Figure 15.
Experimental case study results for the (left) conventional and (right) proposed FR strategies: (a) frequency deviations Δf, (b) load demands P_{L}, dispatchable generation P_{G}, and total generation P_{G} + P_{PV}, (c) variations in generator output power ΔP_{G} and in battery input power ΔP_{B}, and (d) dynamic ramping rate of generation dΔP_{G}/dt.
Figure 16.
Flowcharts of (a) a conventional, centralized hybrid SE strategy and (b) a proposed, decentralized PHASE strategy, both of which include BD detection, identification, and correction steps.
Figure 16.
Flowcharts of (a) a conventional, centralized hybrid SE strategy and (b) a proposed, decentralized PHASE strategy, both of which include BD detection, identification, and correction steps.
Figure 17.
Comparison of absolute SE errors: (a,b) voltage magnitudes and phase angles for the BD in the SCADA measurements and (c,d) voltage magnitudes and phase angles for the BD in the PMU measurements.
Figure 17.
Comparison of absolute SE errors: (a,b) voltage magnitudes and phase angles for the BD in the SCADA measurements and (c,d) voltage magnitudes and phase angles for the BD in the PMU measurements.
Figure 18.
Variations (a) in the errors of the estimated states and (b) in the objective values, by the number of ADMM iterations for different convergence thresholds ε_{ADMM}.
Figure 18.
Variations (a) in the errors of the estimated states and (b) in the objective values, by the number of ADMM iterations for different convergence thresholds ε_{ADMM}.
Table 1.
List of research challenges in balancing supply and demand.
Table 1.
List of research challenges in balancing supply and demand.
Challenges  Research Questions and Requirements 

Timeframes  
Forecasting  Longterm datasets with high spatiotemporal resolution should be collected, processed, and utilized to ensure accurate forecasting of VRE generation patterns. Highquality assessment of forecast uncertainty is required to integrate the weatherdependent characteristics of VRE sources into the operational algorithms and analytical tools of grids.

Algorithms and tools  VRE curtailment caused by network congestion can be avoided by improving the accuracy of power flow analyses by enhancing network modeling and nonlinear solvers. It is necessary to capture more details for each service to aggregate the technical details of constraints on individual resources into a single model or algorithm.

Modeling and analysis  Stability constraints should also be represented better, including the limits of stored rotational energy, the frequency regulation reserve, and locational voltage deviations. To establish riskaware power balancing, new optimization methods and new computations should be used to update the deterministic models and tools to the stochastic status, and to further develop existing probabilistic models and tools.

Interactions across regions and sectors  Energy storage systems and controllable loads in lowvoltage networks are costeffective resources for providing essential reliability services to bulk transmission systems. This requires a detailed representation of the complex constraints pertaining to aggregation of distributed energy storage and loads. Coupling of energy sectors must be modeled in sufficient detail (i.e., high resolution) in terms of both flexibility and processspecific constraints, making it possible to include many flexible resources in decarbonized power grids and greatly improve grid operations and economics.

Table 2.
List of research challenges in VRE inverter interfacing.
Table 2.
List of research challenges in VRE inverter interfacing.
Challenges  Research Questions and Requirements 

Desynchronization  How can the change from a synchronous to a nonsynchronous system be seamless while maintaining reliability? Existing synchronous generators and network infrastructures should become more adaptive and cooperative to meet the increased flexibility, stability, and control needs of VREdominated grids.

Modeling and analysis  How can inverter design and the functionality of VREdominated power grids be optimized? It is necessary to implement various forms of VRE inverter control that reference local conditions, parameter settings, and controller tuning, and to ensure that these are incorporated with gridlevel control. It is necessary to understand the limitations of existing simulation models and tools for power electronic converters and to develop new models and tools that support the planning and interconnection of VREdominated grids. Existing, positivesequence fundamentalfrequency simulation tools must be updated to represent faster controllers with advanced functionality and limiting conditions, focusing on accurate VRE inverter controller representation. Analytical simulation models that are needed to handle the characteristics of inverterbased generators, storage, and loads with realworld, complex, nonlinear operational characteristics. The fidelity of generic and manufacturerspecific EMTlevel simulation models and tools for largescale network studies must be enhanced.

Frequency and voltage stabilities  Which parameters best indicate the supplyanddemand balance and how the grid frequency should be controlled? It is necessary to understand the effects of system harmonics and subsynchronous oscillations from VRE inverters on grids and to design mitigation method, such as shaping of the inverter harmonic impedances at specific frequencies.

Shortcircuit currents  Traditional protection schemes, designed for synchronous generators should be replaced by new protection schemes for a wide variety of VRE inverters. New methods are needed for the restoration and black start of VREdominated grids.

Behindthemeter units  
Table 3.
Features of the proposed and conventional strategies.
Table 3.
Features of the proposed and conventional strategies.
HVDC Control Strategies in Figure 4  DCLink Voltage  Control Target Grids  Secondary Frequency Control 

Proposed  Case 1  timevarying  bothside grids  LQG 
Conventional  Case 2  timevarying  bothside grids  PI 
Case 3  fixed  inverterside grid  PI 
Table 4.
Comparisons of the results for the step and continuous response tests.
Table 4.
Comparisons of the results for the step and continuous response tests.
Frequency Deviations  Case 1  Case 2  Case 3 

Individual  Total  Individual  Total  Individual  Total 

Figure 4 (left)  Δf_{i}_{max} [Hz]  1.08  3.16  2.22  4.44  0.86  4.04 
Δf_{r}_{max} [Hz]  1.08  2.22  3.18 
Figure 4 (right)  Δf_{i,rms} [Hz]  0.25  0.53  0.44  1.24  0.10  1.13 
Δf_{r,rms} [Hz]  0.28  0.80  1.03 
Table 5.
Features of the proposed and conventional strategies.
Table 5.
Features of the proposed and conventional strategies.
SVC Strategies  Reference Voltages 

v_{C}^{ref}  v_{O}^{ref}  v_{Wk}^{ref} (or Q_{Wk}^{ref}) 

Prop.  Case 1  LQG  timevarying  timevarying  timevarying 
Conv.  Case 2  PI  1 pu  Δv_{O}^{ref} = 0 pu   
Case 3  NoSVC  1 pu    ΔQ_{Wk}^{ref} = 0 pu, ∀k 
Case 4  MPC  timevarying  timevarying  timevarying 
Table 6.
Comparison of continuous response test results.
Table 6.
Comparison of continuous response test results.
Maximum Variations in Figure 5a  $\Delta {v}_{O}{}_{\mathrm{max}}(\times {10}^{5})$  $\Delta {v}_{C}{}_{\mathrm{max}}(\times {10}^{5})$  Δv_{W}_{1}_{max}  Δv_{W}_{2}_{max} 

Case 1 [pu]  6.4840  4.3595  0.0363  0.0354 
Case 2 [pu]  18.1581  5.8330  0.0610  0.0614 
Case 3 [pu]  348.8569  5.8436  0.0478  0.0474 
Table 7.
Features of the proposed and conventional FR strategies.
Table 7.
Features of the proposed and conventional FR strategies.
FR Strategies  SFC, PFC, and IRE Gains 

Proposed  Case 1  Set as default values in [64] 
Conventional  Case 2  Set as default values in [64] 
Case 3  Increasing SFC gains (P_{f} = 3 and I_{f} = 6) 
Case 4  Increasing PFC and inertia gains (m = 0.30, n = 0.05, and K = 15) 
Table 8.
Comparisons for the continuous load variations.
Table 8.
Comparisons for the continuous load variations.
Comparison Factors in Figure 11c  Proposed (Case 1)  Conventional 

Case 2  Case 3  Case 4 

Δf_{pk}  [Hz]  0.134  0.825  0.716  0.613 
Δf_{rms}  [Hz]  0.026  0.183  0.159  0.163 
ΔP_{M,rms}  [pu]  0.163  0.130  0.147  0.118 
ΔP_{IG,rms}  [pu]  0.160  0.126  0.145  0.142 
Table 9.
Features of the proposed and conventional strategies.
Table 9.
Features of the proposed and conventional strategies.
VR Strategy  Description 

Proposed  Case 1  No uncertainties in the parameter estimates 
Case 2  30% uncertainties in the parameter estimates 
Conventional  Case 3  PIbased output feedback loop 
Case 4  Robust state feedback loop 
Table 10.
Comparisons for the continuous load variations.
Table 10.
Comparisons for the continuous load variations.
Comparison Factors  Proposed  Conventional 

Case 1  Case 2  Case 3  Case 4 

ΔV_{rms,avg}  [×10^{−3} pu]  1.564  1.816  6.684  3.808 
ΔV_{pk,max}  [×10^{−2} pu]  0.962  1.163  2.741  2.418 
Σ_{i} ΔQ_{SG}_{i,rms}  [pu]  0.118  0.137  0.111  0.131 
Σ_{k} ΔQ_{IG}_{k,rms}  [pu]  0.082  0.099  0.075  0.092 
Table 11.
Comparison of the AMAEs for the corrupted SCADA data.
Table 11.
Comparison of the AMAEs for the corrupted SCADA data.
AMAE [×10^{−3}]  Proposed  Conventional 

DPHASE  DSE w/o BDP  DSELNRT  RDSE 

V  θ  V  θ  V  θ  V  θ 

14bus  0.05  0.21  1.04  1.01  0.75  0.69  0.62  0.69 
118bus  0.19  0.24  3.84  4.02  1.75  2.92  1.99  3.52 
1062bus  0.17  0.18  8.99  9.57  3.08  3.23  4.37  8.52 
Table 12.
Comparison of the AMAEs for the corrupted PMU data.
Table 12.
Comparison of the AMAEs for the corrupted PMU data.
AMAE [×10^{−3}]  Proposed  Conventional 

DPHASE  DSE w/o BDP  DSELNRT  RDSE 

V  θ  V  θ  V  θ  V  θ 

14bus  1.20  1.97  66.8  27.9  55.4  22.5  65.5  25.3 
118bus  7.2  5.9  84.5  56.8  63.3  46.6  77.0  54.7 
1062bus  13.1  14.5  126.6  93.3  107.8  92.4  108.6  91.4 
Table 13.
Buses in the five subareas in the IEEE 118bus network.
Table 13.
Buses in the five subareas in the IEEE 118bus network.
SubAreas  Buses ^{†} 

1  1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 117 
2  21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 70, 71, 72, 73, 74, 75, 113, 114, 115 
3  33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59 
4  68, 69, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 116, 118 
5  60, 61, 62, 63, 64, 65, 66, 67, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112 
Table 14.
List of research challenges in energy and capacity adequacy.
Table 14.
List of research challenges in energy and capacity adequacy.
Challenges  Research Questions and Requirements 

Adequacy planning tools  
Metrics and calculation methodologies  Use of several current metrics, including LOLP, LOLE, and lossofload hours (LOLH), and development of new metrics that consider societal needs and load and storage flexibilities. Use of different reliability levels, such as an event occurring once in 10 years, as the common LOLP target and two events per year as a lower reliability target. Improving the calculation of adequacy benefits afforded when neighboring regions are interfaced and larger geographical areas are connected via transmission lines with limited capacities.

Contributions of emerging technologies  Improving representations of demandside flexibility, energy storage, sector coupling, grid limitation, and expansion cost, to better predict the investments required for grids with high VRE proportions and accurately determine the optimal VRE mix.

Modeling and data  Transmission networks and distributed generators should be modeled in a manner that accurately captures their contributions to adequacy. The contribution of demand and storage to adequacy must be determined accurately using models of realworld operating characteristics, including responses to electricity prices. New models must capture how sector coupling among electricity, transportation, heating, and natural gas affects adequacy. More highquality longer time frame (10+ years) data on supply and demand power are required to ensure precise and robust calculations of adequacy.

Table 15.
List of research challenges in electricity market design.
Table 15.
List of research challenges in electricity market design.
Challenges  Research Questions and Requirements 

Changes in electricity markets  Enhance market and regulatory frameworks to ensure that demand becomes more responsive to price in a manner that resolves potential market challenges, including price volatility, revenue imbalance, resource deficiency, and reduced flexibility.

Electricity prices and investment signals  Establish wellorganized markets that incentivize longterm investment in an optimal mix of energy resources by accurately valuing those resources and the required attributes of future electricity systems with 100% renewables. Design marketbased approaches that incentivize flexible use of energy resources guided by the tradeoff between cost and reliability and between the valuations of various market participants.
