Predictive Energy Management for Mitigating Load Altering Attacks for Islanded Microgrids Using Battery Energy Storage Systems

Ayobami Olajube; Koto Omiloli; Olugbenga Anubi; Satish Vedula

REVIEW 3 major objections 6 minor 20 references

Predictive Energy Management for Mitigating Load Altering Attacks for Islanded Microgrids Using Battery Energy Storage Systems

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A predictive energy-management layer that dispatches battery storage can keep an islanded microgrid's power balance intact when load measurements are tampered with, and a real-time simulation of a modified IEEE 9-bus system shows the…

desk verdict The control derivations are clean, but the central simulation is undercut by a missing 3600 in the SoC equation that makes the battery appear 3600 times larger than its stated 25 Ah capacity. read the letter →

arxiv 2507.01852 v1 pith:OMHJGFID submitted 2025-07-02 math.OC

Satish Vedula , Koto Omiloli , Ayobami Olajube , Olugbenga Anubi This is my paper

classification math.OC MSC 93C5590C25

keywords load-alteringattacksislandedmicrogridsbatteryenergystoragemodelpredictivecontrolhierarchicalfalsedatainjectionpowertracking

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes an energy management scheme, built on model predictive control, that uses battery storage to counter load-altering attacks in islanded microgrids. The central claim is that the battery absorbs the power mismatch created when false data is injected into load measurements, preventing frequency and power-tracking errors. In a real-time numerical simulation of a modified IEEE 9-bus system, the root-mean-square power-tracking error grows with the attack size when no battery is present, but remains small when the battery is dispatched. The argument matters because islanded microgrids have limited inertia and few resources, so demand-side cyberattacks can drive instability or blackouts; fast storage is one of the few assets that can respond within seconds. The paper extends earlier LAA detection and mitigation work by adding a predictive management layer rather than a new detection scheme.

What carries the argument

The load-bearing mechanism is the model predictive control problem in (27), solved every 1 s over a 5 s horizon, together with the primary controllers of Section 4. The optimization chooses power setpoints for the two generators and the battery subject to the power balance constraint $\sum_i p_i = \sum_j \hat{p}_{L_j}$, the battery state-of-charge update $s_{k+1}=s_k - \frac{T_s}{Q_b v_c} p_i$, ramping limits, and SoC bounds. The attack enters as an additive term $w_L$ on the measured load power; because the MPC treats the corrupted measurement as the demand to track, the battery is dispatched to cover exactly the false increment. The battery cost $C(p_b)=p_b^2$ is chosen so the optimizer prefers battery power for the transient, while generator costs are quadratic in dispatch.

What would settle it

Re-run the same IEEE 9-bus scenario with the corrected SoC update $s_{k+1} = s_k - \frac{T_s}{3600\,Q_b v_c}p_i$ and a 25 Ah battery at 12 kV, and record whether the battery can still hold the power balance during a one-second 50% load-altering spike without violating its SoC limits or its 10 MW limit; also compute the battery's actual energy in watt-hours (25 Ah × 12 kV = 300 kWh) against the energy demanded by the spike.

Watch

Extended reading notes

Core claim

The paper's central discovery is that a battery energy storage system commanded by a receding-horizon optimization can absorb the load alteration caused by an LAA. In the MPC formulation (27), the battery's power is chosen so that the sum of generator and battery powers matches the corrupted load measurement $\hat{p}_L = \frac12 v_c^\top i_L + w_L$, while state-of-charge dynamics enforce a physical limit on stored energy. Simulated on a modified IEEE 9-bus system with two gas-turbine generators and one 10 MW battery, the scheme keeps the power balance constraint satisfied when the measured load is ramped by 50% for one second; without the battery, the tracking error grows monotonically with the percentage alteration. The authors read this as showing that predictive energy management with batteries is a viable mitigation mechanism for static and dynamic LAAs.

Load-bearing premise

The simulation's battery behaves as if it stores 3600 times more energy than a 25 Ah cell actually does, because the state-of-charge update in equation (27) drops the factor of 3600 that converts ampere-hours to ampere-seconds; the demonstrated mitigation may therefore rely on an unrealistically large storage capacity.

Editorial extensions

If this is right

Under the simulated conditions, increasing the load alteration from 10% to 50% raises the power-tracking RMSE when no battery is present; adding the battery keeps RMSE near zero across the same range.
A one-second 50% spike in measured load is fully tracked by the generators plus battery, so frequency and voltage regulation hold and no load shedding is needed.
The computational load (1 ms integration, 1 s optimization update) runs on an 8-core real-time target, suggesting the MPC layer is fast enough for practical islanded microgrids.
Because the MPC re-optimizes every second, the framework applies to both static attacks (constant offset) and dynamic attacks (ramped alteration) within the simulation.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The SoC update omits the 3600 s/hour conversion from ampere-hours to ampere-seconds, so the simulated battery stores 3600 times the energy of a 25 Ah cell; correcting this would likely require a much larger battery or a shorter attack to achieve the same result.
The method does not detect the attack; it simply tracks the corrupted measurement. A natural extension is to combine attack detection and reconstruction with the MPC so the battery only covers genuine mismatches, preserving capacity.
The paper's cost functions assume the battery is cheap for fast transients; a different cost, such as one weighting state-of-health or cycle life, would shift how much of the attack the battery absorbs versus how much the generators ramp.
Testing on a larger system or with sustained attacks would reveal the storage size needed, since the simulation only demonstrates a one-second spike.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Quick read for you. The paper builds a hierarchical islanded microgrid model with primary controls (adaptive generator speed/voltage control and sliding-mode load current control) and an MPC-based energy management layer that dispatches a battery to absorb load-altering attacks. The primary-control section is standard but clean: the Lyapunov arguments check out, and Barbalat's lemma is used correctly. The MPC formulation is also standard. What is genuinely new is narrow--combining battery storage with predictive energy management for LAA mitigation and testing it on a real-time simulator. That is an incremental step over [9] and [13], which already use battery-based mitigation and robust MPC for LAAs; the authors acknowledge [13] but overstate their novelty.

The load-bearing problem is equation (27). The continuous battery model in Eq. (23) is \dot{s} = -i_b/(3600 Q_b). With i_b = p_i/v_c, forward Euler gives s_{k+1} = s_k - T_s p_i/(3600 Q_b v_c). Eq. (27) drops the 3600. Unless units are stated that absorb the conversion--they are not--the MPC sees a battery with 3600 times the usable energy of a 25 Ah cell. So the SoC constraints are effectively non-binding, and the dramatic RMSE reduction in Fig. 6 may simply reflect a virtual storage of hundreds of megawatt-hours rather than the stated cell. This is not a minor typo; it is the mechanism behind the central claim.

There are secondary issues. The battery cost C(p_b)=p_b^2 is chosen, in the authors' own words, "to use the battery to negate the LAA scenarios," so the dispatch is built into the objective rather than emerging from a physical or economic model. There is no sensitivity analysis, no error bars, no code or data, and only one attack scenario. The claim of "appropriate battery sizing" is asserted without a sizing study. These alone would make the demonstration suggestive at best.

Credit where due: the control derivations are consistent, the paper is clearly written, and the authors do not hide the ad hoc cost choice. But the central numerical result is not reliable as reported. I would want the authors to fix the unit conversion, rerun, and report what actually happens with a real 25 Ah battery, plus a sensitivity study over attack size and horizon. If the result survives, this becomes a modest but useful extension; if not, it reduces to a control-modeling exercise.

For peer review: I would send it to referees rather than desk-reject, because the modeling is serious and the flaw is identifiable and fixable. But I would not cite the current version.

Referee Report

3 major / 6 minor

Summary. The paper proposes a hierarchical control framework for islanded microgrids: a primary control layer with an adaptive PI/generator controller and a sliding-mode load current controller, plus a battery model with state-of-charge estimation; and a tertiary predictive energy management (MPC) layer that dispatches generators and a battery to balance the measured load under load-altering attacks (LAAs). The framework is tested in a real-time simulation of a modified IEEE 9-bus system with two generators, two loads, and a 10 MW/25 Ah battery. The central claim is that the battery, when included in the MPC, eliminates the power-tracking imbalance caused by the LAA (Fig. 9), and that the RMSE of power tracking remains low as the attack amplitude grows from 10% to 50% (Fig. 6).

Significance. If the demonstrated mitigation is physically valid, the paper would make a useful contribution: it is, to my knowledge, the first to apply predictive energy management specifically to LAA mitigation, and it includes complete Lyapunov-based derivations of the primary controllers and a real-time simulation setup. The manuscript is clearly written and the mathematical development is mostly consistent. However, the central simulation result rests on two load-bearing issues: a unit error in the battery SoC discretization that effectively enlarges the battery by a factor of 3600, and an explicitly ad hoc choice of the battery cost function that is stated to force the battery to negate the attack. These issues must be resolved before the effectiveness claim can be accepted.

major comments (3)

[Sec. 5, Eq. (27) vs. Sec. 4.3, Eq. (23)] The SoC update in the MPC formulation is inconsistent with the battery model. Equation (23) gives \dot{s} = -i_b/(3600 Q_b), with Q_b in Ahr and time in seconds. Discretizing with i_b = p_i/v_c yields s_{k+1} = s_k - T_s p_i/(3600 Q_b v_c). Equation (27) instead writes s_{k+1} = s_k - T_s p_i/(Q_b v_c), dropping the factor 3600. Unless the units of p_i and v_c are defined to absorb this conversion (they are not), the optimizer sees a battery whose usable energy is 3600 times larger than the rated 25 Ah cell. With v_c ≈ 12 kV and Q_b = 25 Ahr, the virtual battery energy is on the order of 300 MWh rather than the physical ~84 kWh. The SoC bounds therefore become effectively non-binding during the simulated 1 s attack, and the RMSE reduction in Fig. 6 and the success in Fig. 9 may be artifacts of this oversized virtual storage. This must be corrected and the simulations repeated with the proper 3600 factor.
[Sec. 6, battery cost function] The paper states that the battery operating cost is chosen as C(p_b) = p_b^2 'to use the battery to negate the LAA scenarios'. This is an explicit admission that the objective function was tuned to produce the desired mitigation behavior. Since the battery is the only mitigation mechanism in the MPC, the demonstration is at least partly forced by the cost choice rather than being an independent prediction. The authors should either justify the battery cost from a physical degradation model (e.g., a cycle-life or efficiency map) or provide a sensitivity study over the battery cost coefficient, showing that the LAA mitigation persists for a range of reasonable coefficients. Without such support, the central claim that the proposed framework 'shows the effectiveness of the battery' is circular.
[Sec. 6, Fig. 6 and battery sizing] The text claims that 'with an appropriate choice of battery sizing, the error is mitigated', but no sizing analysis is presented. The only battery parameters given are p_b = 10 MW and Q_b = 25 Ahr; the SoC bounds in (27) are never specified, and the initial SoC is not stated. A 10 MW, 25 Ah battery at ~12 kV has only ~84 kWh of stored energy, which is comparable to the 1 s, 50% load spike (~48 MW × 1 s ≈ 13.3 kWh) only if the SoC is near full and the bounds are wide. The lack of a sizing study, along with the unspecified SoC limits, prevents the reader from assessing whether the demonstrated mitigation is physically realistic even after the 3600 error is fixed. Please provide the SoC constraints, initial condition, and a parameter study linking battery energy capacity to attack amplitude and duration.

minor comments (6)

[Sec. 5, Eq. (27)] The notation p_0 = \hat{p}_i is unexplained; it likely means the initial power measurement used as the starting point of the optimization. Please clarify the indexing (e.g., p_{i,0} = \hat{p}_i) and distinguish the scalar measurement from the horizon vector.
[Sec. 6, battery parameters] The SoC limits \underline{s} and \overline{s} used in the simulation are never reported. State these values and the initial SoC, as they directly affect whether the MPC's battery dispatch is feasible in the corrected model.
[Sec. 6, OCV model] The OCV-SoC relation v_oc = 1.071 s + 3.357 kV is given without explaining the series/parallel configuration of the 25 Ah cells. The slope has units of kV per unit SoC, which is an unusual representation for a single Li-ion cell; clarify the module configuration and the base voltage.
[Sec. 4.2, Eq. (16)] The reaching-time expression in Eq. (16) has unbalanced parentheses in the logarithmic argument. The intended formula appears to be \log\left(\frac{(\rho-L)/k}{\|\sigma(0)\| + (\rho-L)/k}\right); please fix the typesetting.
[Sec. 6, Fig. 6] Figure 6 uses different scales on the left and right axes without explicit labels for each axis's scale; please make the two RMSE curves clearly distinguishable and state the units (MW or percentage).
[Sec. 6, real-time claim] The paper claims real-time capability but does not report computation times for the MPC solve. Since the optimization horizon is only 5 s and the step is 1 ms, a per-step solve-time comparison against the 1 s control update would substantiate the real-time claim.

Circularity Check

1 steps flagged · score 6.0 of 10

Battery-mitigation 'validation' is partly forced by the cost function chosen to make the battery negate LAA; the rest of the derivation is largely self-contained.

fitted input called prediction [Section 6, 'Real-Time Numerical Simulation' (battery cost choice and Fig. 9 discussion); see also Eq. (27)]
"The operating cost for the battery is chosen as C(pb) = p2 b to use the battery to negate the LAA scenarios. ... Fig. 6 shows ... with an appropriate choice of battery sizing, the error is mitigated. ... This shows the effectiveness of the predictive energy management framework and the battery in tackling LAAs."

The MPC in Eq. (27) minimizes the sum of operating costs Ci(pi), and the battery dispatch is the minimizer of that objective. By explicitly choosing C(pb)=pb^2 'to use the battery to negate the LAA scenarios,' the authors make battery usage the intended objective rather than an independently predicted outcome. The subsequent RMSE comparison and Fig. 9 therefore demonstrate that the optimizer does what it was tuned to do: the battery is dispatched to absorb the altered load because the cost function was selected for exactly that purpose. No battery-health or economic cost is derived; the choice is an input tailored to the target result. Thus the central 'effectiveness' claim is partially forced by construction.

full rationale

The paper's physical modeling (generator, load, battery, primary control) and the stability proofs are self-contained and do not reduce to the claimed conclusion. The MPC problem in Eq. (27) is fully stated, so the self-citation to [19] is not load-bearing. The main circular element is the battery operating cost: C(pb)=pb^2 is chosen 'to use the battery to negate the LAA scenarios,' so the demonstrated mitigation in Fig. 9 is in part a restatement of the design objective rather than an independent validation. Separately, Eq. (27) writes s_{k+1}=s_k - Ts/(Qb vc) pi while Eq. (23) has dot{s}=-ib/(3600Qb), omitting the 3600 s/h conversion; as written, this gives the simulated battery 3600 times the usable energy of the stated 25 Ah cell. That is a serious modeling/unit inconsistency and a physical-plausibility concern, but it is a correctness issue rather than a circular reduction, so it does not further raise the circularity score. Overall, the central claim is partially forced by an ad hoc objective choice, but the underlying model and control derivation retain independent content.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The central demonstration depends on several hand-chosen simulation parameters, particularly the battery cost function and battery size, plus an incorrect SoC conversion factor that artificially enlarges the battery's energy. No new physical entities are introduced.

free parameters (7)

Battery operating cost coefficient = 1 (cost C(p_b) = p_b^2)
Chosen ad hoc in Sec. 6 'to use the battery to negate the LAA scenarios'; the MPC dispatch and the demonstrated mitigation depend on this choice.
Generator cost coefficients = a=0.4, b=5.5, c=500 and a=0.6, b=5.3, c=400
Chosen in Sec. 6 to represent efficiency, but not derived from data in the paper.
Prediction horizon = h=5 s
Chosen in Sec. 6 with no sensitivity analysis.
Battery capacity = Q_b = 25 Ah
Chosen for the simulation; no battery-sizing method is provided.
OCV-SoC linear fit coefficients = beta1=1.071 kV, beta2=3.357 kV
Linear curve-fit to OCV-SoC data from [18], stated in Sec. 6.
Attack size and duration = 50% of load, 1 s
Chosen for the single test scenario; no parameter sweep is performed.
Power and ramping limits = 5-95% rated, battery ramping at 95%
Chosen in Sec. 6 to define the feasible dispatch region.

assumptions (5)

domain assumption Electrical dynamics are fast compared to mechanical dynamics, so the electrical torque can be approximated as Te = v_cr^T i / omega_e (Assumption 1)
Invoked in Sec. 4.1 to simplify the generator model; if the assumption fails, the PI speed controller design changes.
domain assumption All currents and voltages in the microgrid are measurable
Assumed in Sec. 4.4 before defining the active power measurements; measurement noise is not modeled.
domain assumption Battery parameters (rb, rp, cp) are known
Remark 3 in Sec. 4.3; parameter estimation is only mentioned as a possible extension.
standard math MPC feasibility and stability follow from convex optimization literature (Boyd and Vandenberghe [20])
Invoked in Sec. 5 to argue that the polytope constraint set is well-studied; no stability proof specific to this paper is given.
domain assumption Power balance can be enforced using the attacked load measurement p_hat_L as the required total power
Equation (26) defines balance using p_hat_L that includes the false data wL; no attack detection or robust reformulation is used, so the MPC is not attack-resilient in a feedback sense.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Predictive Energy Management for Mitigating Load Altering Attacks for Islanded Microgrids Using Battery Energy Storage Systems." pith.science (2026). https://pith.science/paper/OMHJGFID

@misc{pith2026250701852,
  author       = {Pith},
  title        = {Pith review of: Predictive Energy Management for Mitigating Load Altering Attacks for Islanded Microgrids Using Battery Energy Storage Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OMHJGFID}},
  note         = {Machine review of arXiv:2507.01852}
}

read the original abstract

An increasing number of smart devices controlling loads opens a potential pathway for false data attacks which could alter the loads. The presence of energy storage with its ability to quickly respond to discrepancies in loads offers a promising solution for security by preventing further instabilities and potential blackouts. This paper proposes a control methodology for secure predictive energy management that uses batteries to mitigate the impact of load-altering attacks. To that extent, we develop a microgrid model along with the primary control for microgrid. The developed models and the optimization algorithm are validated through a real-time numerical simulation of a modified IEEE 9 bus system involving a battery as one of the energizing sources. The results show the effectiveness of the battery in countering the load alterations.

Figures

Figures reproduced from arXiv: 2507.01852 by the authors.

**Figure 2.** AC microgrid consisting of an AC generator, DC battery connected [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Numerical simulation: real-time target setup and implementation [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗

Figures from the paper (6 more)

**Figure 4.** Figure 4: Modified IEEE 9 bus system used as a numerical example to test the [PITH_FULL_IMAGE:figures/full_fig_p012_4.png]

**Figure 5.** Figure 5: Tracking performance of the gas turbine speed (frequency) and the [PITH_FULL_IMAGE:figures/full_fig_p013_5.png]

**Figure 6.** Figure 6: Root mean square power tracking error without battery on the left [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]

**Figure 7.** Figure 7: Load alteration [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]

**Figure 8.** Figure 8: Power sharing under load altering without the presence of battery [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]

**Figure 9.** Figure 9: Power sharing under load altering with the presence of battery [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

20 extracted references · 20 canonical work pages

[9]

Load altering attacks detection, reconstruction and mitigation for cyber-security in smart grids with battery energy storage systems,

G. Rinaldi, M. Cucuzzella, P. P. Menon, A. Ferrara, and C. Edwards, “Load altering attacks detection, reconstruction and mitigation for cyber-security in smart grids with battery energy storage systems,” in 2022 European Control Conference (ECC), 2022, pp. 1541–1547

work page 2022
[13]

Deep learning detection and robust mpc mitigation for ev-based load-altering attacks on wind-integrated power grids,

A. Abazari, M. M. Soleymani, M. Ghafouri, D. Jafarigiv, R. Atallah, and C. Assi, “Deep learning detection and robust mpc mitigation for ev-based load-altering attacks on wind-integrated power grids,” IEEE Transactions on Industrial Cyber-Physical Systems , vol. 2, pp. 244–263, 2024. 16

work page 2024
[1]

Hierarchical control of droop-controlled ac and dc microgrids—a general approach toward standardization,

J. M. Guerrero, J. C. Vasquez, J. Matas, L. G. de Vicuna, and M. Castilla, “Hierarchical control of droop-controlled ac and dc microgrids—a general approach toward standardization,” IEEE Transactions on Industrial Elec- tronics, vol. 58, no. 1, pp. 158–172, 2011

work page 2011
[2]

Dynamic load altering attacks in smart grid,

S. Amini, H. Mohsenian-Rad, and F. Pasqualetti, “Dynamic load altering attacks in smart grid,” in 2015 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT) , 2015, pp. 1–5. 15

work page 2015
[3]

Distributed resilient interval observers for bounded-error lti systems subject to false data injection at- tacks,

M. Khajenejad, S. Brown, and S. Mart ´ ınez, “Distributed resilient interval observers for bounded-error lti systems subject to false data injection at- tacks,” in 2023 American Control Conference (ACC), 2023, pp. 3502–3507

work page 2023
[4]

Multi-model resilient observer under false data injection attacks,

O. M. Anubi, C. Konstantinou, C. A. Wong, and S. Vedula, “Multi-model resilient observer under false data injection attacks,” in 2020 IEEE Con- ference on Control Technology and Applications (CCTA) , 2020, pp. 1–8

work page 2020
[5]

Resilient control of dc microgrids against fdi attacks on communication links,

Q. Liu and H. Zhang, “Resilient control of dc microgrids against fdi attacks on communication links,” in 2023 62nd IEEE Conference on Decision and Control (CDC), 2023, pp. 5574–5579

work page 2023
[6]

Distributed internet-based load altering attacks against smart power grids,

A.-H. Mohsenian-Rad and A. Leon-Garcia, “Distributed internet-based load altering attacks against smart power grids,” IEEE Transactions on Smart Grid , vol. 2, no. 4, pp. 667–674, 2011

work page 2011

Show all 20 references

[7]

Dynamic load alter- ing attacks against power system stability: Attack models and protection schemes,

S. Amini, F. Pasqualetti, and H. Mohsenian-Rad, “Dynamic load alter- ing attacks against power system stability: Attack models and protection schemes,” IEEE Transactions on Smart Grid , vol. 9, no. 4, pp. 2862–2872, 2018

2018
[8]

Detecting dynamic load altering attacks: A data-driven time-frequency analysis,

S. Amini, F. Pasqualetti, and H. MohsenianRad, “Detecting dynamic load altering attacks: A data-driven time-frequency analysis,” in 2015 IEEE International Conference on Smart Grid Communications (SmartGrid- Comm), 2015, pp. 503–508

2015
[10]

On the feasibility of load-changing attacks in power systems during the covid-19 pandemic,

J. Ospina, X. Liu, C. Konstantinou, and Y. Dvorkin, “On the feasibility of load-changing attacks in power systems during the covid-19 pandemic,” IEEE Access, vol. 9, pp. 2545–2563, 2021

2021
[11]

Adversarial dynamic load- altering cyberattacks against peak shaving using residential electric water heaters,

E.-N. S. Youssef, F. Labeau, and M. Kassouf, “Adversarial dynamic load- altering cyberattacks against peak shaving using residential electric water heaters,” IEEE Transactions on Smart Grid , vol. 15, no. 2, pp. 2073–2088, 2024

2024
[12]

Mitigating load-altering attacks against power grids using cyber-resilient economic dis- patch,

Z. Chu, S. Lakshminarayana, B. Chaudhuri, and F. Teng, “Mitigating load-altering attacks against power grids using cyber-resilient economic dis- patch,” IEEE Transactions on Smart Grid , vol. 14, no. 4, pp. 3164–3175, 2023

2023
[14]

Optimal dynamic load-altering attacks against power systems,

V. Katewa and F. Pasqualetti, “Optimal dynamic load-altering attacks against power systems,” in 2021 American Control Conference (ACC) , 2021, pp. 4568–4573

2021
[15]

Modeling of grid-forming and grid-following inverters for dynamic simulation of large-scale distribution systems,

W. Du, F. K. Tuffner, K. P. Schneider, R. H. Lasseter, J. Xie, Z. Chen, and B. Bhattarai, “Modeling of grid-forming and grid-following inverters for dynamic simulation of large-scale distribution systems,”IEEE Transactions on Power Delivery , vol. 36, no. 4, pp. 2035–2045, 2020

2020
[16]

Grid forming inverter modeling, control, and applications,

D. B. Rathnayake, M. Akrami, C. Phurailatpam, S. P. Me, S. Hadavi, G. Jayasinghe, S. Zabihi, and B. Bahrani, “Grid forming inverter modeling, control, and applications,” IEEE Access, vol. 9, pp. 114 781–114 807, 2021

2021
[17]

W. E. Dixon, A. Behal, D. M. Dawson, and S. P. Nagarkatti, Nonlinear control of engineering systems: a Lyapunov-based approach . Springer Sci- ence & Business Media, 2003

2003
[18]

An open-circuit-voltage model of lithium- ion batteries for effective incremental capacity analysis,

C. Weng, J. Sun, and H. Peng, “An open-circuit-voltage model of lithium- ion batteries for effective incremental capacity analysis,” ser. Dynamic Sys- tems and Control Conference, 10 2013

2013
[19]

Model and load predictive control for design and energy management of shipboard power systems,

M. M. Bijaieh, S. Vedula, and O. M. Anubi, “Model and load predictive control for design and energy management of shipboard power systems,” in 2021 IEEE Conference on Control Technology and Applications (CCTA) , 2021, pp. 607–612

2021
[20]

Boyd and L

S. Boyd and L. Vandenberghe, Convex Optimization . Cambridge Univer- sity Press, 2004. 17

2004

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [9]

Load altering attacks detection, reconstruction and mitigation for cyber-security in smart grids with battery energy storage systems,

G. Rinaldi, M. Cucuzzella, P. P. Menon, A. Ferrara, and C. Edwards, “Load altering attacks detection, reconstruction and mitigation for cyber-security in smart grids with battery energy storage systems,” in 2022 European Control Conference (ECC), 2022, pp. 1541–1547

work page 2022

[2] [13]

Deep learning detection and robust mpc mitigation for ev-based load-altering attacks on wind-integrated power grids,

A. Abazari, M. M. Soleymani, M. Ghafouri, D. Jafarigiv, R. Atallah, and C. Assi, “Deep learning detection and robust mpc mitigation for ev-based load-altering attacks on wind-integrated power grids,” IEEE Transactions on Industrial Cyber-Physical Systems , vol. 2, pp. 244–263, 2024. 16

work page 2024

[3] [1]

Hierarchical control of droop-controlled ac and dc microgrids—a general approach toward standardization,

J. M. Guerrero, J. C. Vasquez, J. Matas, L. G. de Vicuna, and M. Castilla, “Hierarchical control of droop-controlled ac and dc microgrids—a general approach toward standardization,” IEEE Transactions on Industrial Elec- tronics, vol. 58, no. 1, pp. 158–172, 2011

work page 2011

[4] [2]

Dynamic load altering attacks in smart grid,

S. Amini, H. Mohsenian-Rad, and F. Pasqualetti, “Dynamic load altering attacks in smart grid,” in 2015 IEEE Power & Energy Society Innovative Smart Grid Technologies Conference (ISGT) , 2015, pp. 1–5. 15

work page 2015

[5] [3]

Distributed resilient interval observers for bounded-error lti systems subject to false data injection at- tacks,

M. Khajenejad, S. Brown, and S. Mart ´ ınez, “Distributed resilient interval observers for bounded-error lti systems subject to false data injection at- tacks,” in 2023 American Control Conference (ACC), 2023, pp. 3502–3507

work page 2023

[6] [4]

Multi-model resilient observer under false data injection attacks,

O. M. Anubi, C. Konstantinou, C. A. Wong, and S. Vedula, “Multi-model resilient observer under false data injection attacks,” in 2020 IEEE Con- ference on Control Technology and Applications (CCTA) , 2020, pp. 1–8

work page 2020

[7] [5]

Resilient control of dc microgrids against fdi attacks on communication links,

Q. Liu and H. Zhang, “Resilient control of dc microgrids against fdi attacks on communication links,” in 2023 62nd IEEE Conference on Decision and Control (CDC), 2023, pp. 5574–5579

work page 2023

[8] [6]

Distributed internet-based load altering attacks against smart power grids,

A.-H. Mohsenian-Rad and A. Leon-Garcia, “Distributed internet-based load altering attacks against smart power grids,” IEEE Transactions on Smart Grid , vol. 2, no. 4, pp. 667–674, 2011

work page 2011