pith. sign in

arxiv: 2503.15966 · v1 · submitted 2025-03-20 · 📡 eess.SY · cs.SY

Privacy-Preserving Utilization of Distribution System Flexibility for Enhanced TSO-DSO Interoperability: A Novel Machine Learning-Based Optimal Power Flow Approach

Burak Dindar , Can Berk Saner , H\"useyin K. \c{C}akmak , Veit Hagenmeyer This is my paper

Pith reviewed 2026-05-22 23:46 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords privacy-preservingoptimal power flowTSO-DSO interoperabilitymachine learningneural networksdistribution system flexibilitydistributed generatorsmeshed networks
0
0 comments X

The pith

Machine learning models trained only on non-sensitive data let TSOs solve optimal power flow problems using distribution system flexibility while preserving privacy.

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

The paper proposes representing the technical constraints of distribution systems with neural network models trained exclusively on non-sensitive data. This allows the transmission system operator to solve the optimal power flow problem and directly dispatch distributed generators in a single round of communication without receiving private information. The approach uses a novel neural network architecture tailored to represent the feasible region of the distribution systems and is tested on meshed networks with multiple points of common coupling. If correct, this method enables interoperability between TSOs and DSOs without data sharing risks and removes the need for a separate disaggregation step after optimization.

Core claim

The central claim is that neural networks trained on non-sensitive data can accurately represent the feasible region of distribution systems, enabling the TSO to solve the AC-OPF problem at the transmission level while directly determining the dispatch of flexibility-providing units in the distribution system, all without transferring sensitive data and in a single communication round.

What carries the argument

A novel neural network architecture specifically designed to efficiently represent the feasible region of the distribution systems.

If this is right

  • TSO solves the OPF and determines dispatch of DGs directly.
  • Method eliminates the need for an additional disaggregation step.
  • Approach works with various real PQ charts rather than idealized ones.
  • Applicable to meshed DSs with multiple PCCs and varying voltage magnitudes.
  • Privacy is preserved by avoiding transfer of sensitive information.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Such models could potentially be updated periodically without full data exchange if the distribution system topology changes.
  • The method might reduce communication overhead in real-time market operations between TSO and DSO.
  • Extending the NN to handle uncertainty in renewable generation could further enhance its practicality.
  • Integration with existing market mechanisms for flexibility procurement could be explored based on this representation.

Load-bearing premise

Neural networks trained exclusively on non-sensitive data can sufficiently and accurately represent the feasible region of the distribution systems for solving the TSO-level OPF problem.

What would settle it

Running the proposed ML-based OPF on a test case with meshed distribution system and multiple PCCs and comparing the resulting dispatch and power flows against a full AC-OPF solution that has access to all internal data to check for constraint violations or large deviations in objective value.

Figures

Figures reproduced from arXiv: 2503.15966 by Burak Dindar, Can Berk Saner, H\"useyin K. \c{C}akmak, Veit Hagenmeyer.

Figure 1
Figure 1. Figure 1: Schematic representation of the proposed method. [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Data sampling approach using LHS. the rectangles formed by these values fully encapsulate the arbitrary convex polygons. Subsequently, LHS is applied within these bounding rectangles, enabling sampling from the entire arbitrary convex polygon that lies within the bounding rectangle. It is important to note that, as a natural consequence of this approach, some samples are taken from the area between the arb… view at source ↗
Figure 3
Figure 3. Figure 3: Schematic representation of the polytope. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: The architecture of the novel tailored NN. [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: Convex polygon PQ characteristics for DGs in the third DS. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: a) Dataset indicating feasible and infeasible samples. b) Feasible [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: The histogram of the total cost and computational time differences [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
read the original abstract

Due to the transformation of the power system, the effective use of flexibility from the distribution system (DS) is becoming crucial for efficient network management. Leveraging this flexibility requires interoperability among stakeholders, including Transmission System Operators (TSOs) and Distribution System Operators (DSOs). However, data privacy concerns among stakeholders present significant challenges for utilizing this flexibility effectively. To address these challenges, we propose a machine learning (ML)-based method in which the technical constraints of the DSs are represented by ML models trained exclusively on non-sensitive data. Using these models, the TSO can solve the optimal power flow (OPF) problem and directly determine the dispatch of flexibility-providing units (FPUs), in our case, distributed generators (DGs), in a single round of communication. To achieve this, we introduce a novel neural network (NN) architecture specifically designed to efficiently represent the feasible region of the DSs, ensuring computational effectiveness. Furthermore, we incorporate various PQ charts rather than idealized ones, demonstrating that the proposed method is adaptable to a wide range of FPU characteristics. To assess the effectiveness of the proposed method, we benchmark it against the standard AC-OPF on multiple DSs with meshed connections and multiple points of common coupling (PCCs) with varying voltage magnitudes. The numerical results indicate that the proposed method achieves performant results while prioritizing data privacy. Additionally, since this method directly determines the dispatch of FPUs, it eliminates the need for an additional disaggregation step. By representing the DSs technical constraints through ML models trained exclusively on non-sensitive data, the transfer of sensitive information between stakeholders is prevented.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes a machine learning-based optimal power flow method in which a novel neural network architecture, trained exclusively on non-sensitive data, approximates the feasible region of distribution systems. This enables the TSO to solve a TSO-level OPF that directly dispatches flexibility-providing units (DGs) in a single communication round while preserving privacy, and the approach is benchmarked against standard AC-OPF on multiple meshed distribution systems with several PCCs and varying voltage magnitudes. The method also incorporates realistic PQ charts and claims to eliminate the need for a separate disaggregation step.

Significance. If the NN-based feasible-region approximations are sufficiently accurate, the work would provide a practical mechanism for TSO-DSO flexibility coordination that avoids sensitive data exchange and reduces communication rounds. The use of real PQ charts and direct dispatch capability would strengthen applicability to realistic operating conditions.

major comments (2)
  1. [Abstract and §5] Abstract and §5 (Numerical Results): the claim that the method 'achieves performant results' on multiple test systems supplies no quantitative metrics (optimality gaps, feasibility violation rates, out-of-sample error distributions, or ablation studies), leaving the load-bearing assumption that the NN produces a sufficiently tight approximation of the non-convex feasible set for meshed multi-PCC networks unverified.
  2. [§3–4] §3–4 (DS Feasible Region and Novel NN Architecture): the mapping from PCC voltages, injections, and DG set-points in meshed networks with multiple PCCs is topology-dependent and non-convex; no explicit analysis or metric (e.g., Hausdorff distance to the true feasible set or maximum violation rate across voltage ranges) is given to confirm that the chosen training distribution on non-sensitive samples yields an approximation tight enough not to alter the TSO-level optimal dispatch.
minor comments (2)
  1. [Abstract] The phrase 'performant results' is imprecise; replace with concrete indicators (e.g., average cost difference, maximum voltage violation) in both abstract and results section.
  2. [§4] Notation for the NN input/output dimensions and the exact form of the learned constraints inside the TSO OPF should be stated explicitly to allow reproduction.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback and detailed comments. We address each major comment point-by-point below.

read point-by-point responses

  1. Referee: [Abstract and §5] Abstract and §5 (Numerical Results): the claim that the method 'achieves performant results' on multiple test systems supplies no quantitative metrics (optimality gaps, feasibility violation rates, out-of-sample error distributions, or ablation studies), leaving the load-bearing assumption that the NN produces a sufficiently tight approximation of the non-convex feasible set for meshed multi-PCC networks unverified.

    Authors: We agree that explicit quantitative metrics are needed to substantiate the performance claims. In the revised manuscript we will expand Section 5 to report optimality gaps, feasibility violation rates, out-of-sample error distributions, and ablation studies on the meshed test systems, and we will update the abstract to reference these metrics. revision: yes

  2. Referee: [§3–4] §3–4 (DS Feasible Region and Novel NN Architecture): the mapping from PCC voltages, injections, and DG set-points in meshed networks with multiple PCCs is topology-dependent and non-convex; no explicit analysis or metric (e.g., Hausdorff distance to the true feasible set or maximum violation rate across voltage ranges) is given to confirm that the chosen training distribution on non-sensitive samples yields an approximation tight enough not to alter the TSO-level optimal dispatch.

    Authors: We acknowledge that direct quantitative measures of approximation quality (Hausdorff distance, maximum violation rates across voltage ranges) are not currently provided. While the AC-OPF benchmarking in Section 5 offers indirect validation, we will add the requested explicit analyses in the revision to confirm that the non-sensitive training distribution produces a sufficiently tight feasible-region approximation. revision: yes

Circularity Check

0 steps flagged

No circularity; empirical ML approximation benchmarked externally

full rationale

The paper proposes an ML-based OPF method using NNs trained on non-sensitive data to approximate DS feasible regions, with direct dispatch of FPUs and benchmarking against standard AC-OPF on meshed multi-PCC systems. No derivation chain, equations, or self-citations are presented that reduce a claimed prediction or result to a fitted parameter or prior self-work by construction. The central claims concern privacy preservation and empirical performance on external benchmarks, which are falsifiable outside the fitted model itself. This matches the default case of a self-contained empirical method without load-bearing circular steps.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The approach depends on the domain assumption that a neural network can faithfully encode the feasible set of a distribution network from non-sensitive data alone; no free parameters or invented entities are explicitly introduced beyond standard NN training.

axioms (1)
  • domain assumption The feasible region of a distribution system can be accurately approximated by a neural network trained exclusively on non-sensitive data.
    This premise underpins both the privacy guarantee and the claim that the TSO can solve OPF directly from the learned model.

pith-pipeline@v0.9.0 · 5859 in / 1294 out tokens · 33197 ms · 2026-05-22T23:46:00.878340+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

46 extracted references · 46 canonical work pages

  1. [1]

    Hierarchical risk assessment of transmission system considering the influence of active distribution network,

    H. Jia, W. Qi, Z. Liu, B. Wang, Y . Zeng, and T. Xu, “Hierarchical risk assessment of transmission system considering the influence of active distribution network,” IEEE Trans. on Power Syst. , vol. 30, no. 2, pp. 1084–1093, 2014

    work page 2014

  2. [2]

    A methodology for improved TSO-DSO coordination in grid operation planning,

    J. Ringelstein, M. V ogt, A. M. Khavari, R. Ciavarella, M. Di Somma, and G. Graditi, “A methodology for improved TSO-DSO coordination in grid operation planning,” Electr. Power Syst. Res., vol. 211, p. 108445, 2022

    work page 2022

  3. [3]

    Basic schemes for TSO-DSO coordination and ancillary services provision,

    H. Gerard, E. Rivero, and D. Six, “Basic schemes for TSO-DSO coordination and ancillary services provision,”SmartNet Deliv. D, vol. 1, p. 12, 2016

    work page 2016

  4. [4]

    A review on TSO-DSO coordination models and solution techniques,

    A. G. Givisiez, K. Petrou, and L. F. Ochoa, “A review on TSO-DSO coordination models and solution techniques,” Electr. Power Syst. Res., vol. 189, p. 106659, 2020

    work page 2020

  5. [5]

    Real-time coordination of integrated transmission and distribution sys- tems: Flexibility modeling and distributed NMPC scheduling,

    X. Dai, Y . Guo, Y . Jiang, C. N. Jones, G. Hug, and V . Hagenmeyer, “Real-time coordination of integrated transmission and distribution sys- tems: Flexibility modeling and distributed NMPC scheduling,” arXiv preprint arXiv:2402.00508, 2024

    work page arXiv 2024

  6. [6]

    Evaluation of TSO–DSO coordination schemes for meshed-to-meshed configurations: Lessons learned from a realistic Swedish case study,

    L. Lind, R. Cossent, and P. Fr ´ıas, “Evaluation of TSO–DSO coordination schemes for meshed-to-meshed configurations: Lessons learned from a realistic Swedish case study,” Sustain. Energy Grids Netw. , vol. 35, p. 101125, 2023

    work page 2023

  7. [7]

    Migliavacca, TSO-DSO Interactions and Ancillary Services in Elec- tricity Transmission and Distribution Networks: Modeling, Analysis and Case-Studies

    G. Migliavacca, TSO-DSO Interactions and Ancillary Services in Elec- tricity Transmission and Distribution Networks: Modeling, Analysis and Case-Studies. Springer, 2019

    work page 2019

  8. [8]

    Ziesemann, F

    C. Ziesemann, F. Gaumnitz, C. S. K ¨ohnen, and I. E. Stoyanova, Chal- lenges and Barriers to the Implementation of TSO-DSO Coordination Concepts: Discussion Paper. Universit ¨atsbibliothek der RWTH Aachen, 2023

    work page 2023

  9. [9]

    A privacy-preserving approach to day-ahead TSO-DSO coordinated stochastic scheduling for energy and reserve,

    M. Habibi, V . Vahidinasab, and M. S. Sepasian, “A privacy-preserving approach to day-ahead TSO-DSO coordinated stochastic scheduling for energy and reserve,” IET Gener. Transm. Distrib. , vol. 16, no. 1, pp. 163–180, 2022

    work page 2022

  10. [10]

    Privacy- preserving power system obfuscation: A bilevel optimization approach,

    T. W. Mak, F. Fioretto, L. Shi, and P. Van Hentenryck, “Privacy- preserving power system obfuscation: A bilevel optimization approach,” IEEE Transactions on Power Systems , vol. 35, no. 2, pp. 1627–1637, 2019

    work page 2019

  11. [11]

    Differential privacy for power grid obfuscation,

    F. Fioretto, T. W. Mak, and P. Van Hentenryck, “Differential privacy for power grid obfuscation,” IEEE Transactions on Smart Grid , vol. 11, no. 2, pp. 1356–1366, 2019

    work page 2019

  12. [12]

    Privacy-preserving obfuscation for distributed power systems,

    T. W. Mak, F. Fioretto, and P. Van Hentenryck, “Privacy-preserving obfuscation for distributed power systems,” Electric Power Systems Research, vol. 189, p. 106718, 2020

    work page 2020

  13. [13]

    Constrained-based differential privacy: Releasing optimal power flow benchmarks privately: Releasing optimal power flow benchmarks privately,

    F. Fioretto and P. Van Hentenryck, “Constrained-based differential privacy: Releasing optimal power flow benchmarks privately: Releasing optimal power flow benchmarks privately,” in Integration of Constraint Programming, Artificial Intelligence, and Operations Research: 15th International Conference, CPAIOR 2018, Delft, The Netherlands, June 26–29, 2018, ...

    work page 2018

  14. [14]

    Ad- vancing distributed AC optimal power flow for integrated transmission- distribution systems,

    X. Dai, J. Zhai, Y . Jiang, Y . Guo, C. N. Jones, and V . Hagenmeyer, “Ad- vancing distributed AC optimal power flow for integrated transmission- distribution systems,” IEEE Transactions on Network Science and En- gineering, 2025

    work page 2025

  15. [15]

    Risk-averse TSO-DSOs coordinated distributed dispatching considering renewable energy and demand response uncertainties,

    T. Jiang, C. Wu, R. Zhang, X. Li, and F. Li, “Risk-averse TSO-DSOs coordinated distributed dispatching considering renewable energy and demand response uncertainties,” Applied Energy , vol. 327, p. 120024, 2022

    work page 2022

  16. [16]

    Ensuring data privacy in AC optimal power flow with a distributed co-simulation framework,

    X. Dai, A. Kocher, J. Kova ˇcevi´c, B. Dindar, Y . Jiang, C. Jones, H. K. C ¸ akmak, and V . Hagenmeyer, “Ensuring data privacy in AC optimal power flow with a distributed co-simulation framework,” Electric Power Systems Research, vol. 235, p. 110710, 2024

    work page 2024

  17. [17]

    Estimating the active and reactive power flexibility area at the TSO-DSO interface,

    J. Silva, J. Sumaili, R. J. Bessa, L. Seca, M. A. Matos, V . Miranda, M. Caujolle, B. Goncer, and M. Sebastian-Viana, “Estimating the active and reactive power flexibility area at the TSO-DSO interface,” IEEE Trans. Power Syst., vol. 33, no. 5, pp. 4741–4750, 2018

    work page 2018

  18. [18]

    TSO–DSO interaction: Active distribution network power chart for TSO ancillary services provision,

    F. Capitanescu, “TSO–DSO interaction: Active distribution network power chart for TSO ancillary services provision,” Electr. Power Syst. Res., vol. 163, pp. 226–230, 2018

    work page 2018

  19. [19]

    Tracing, ranking and valuation of aggregated DER flexibility in active distribution networks,

    A. Churkin, W. Kong, J. N. M. Gutierrez, E. A. M. Cese ˜na, and P. Mancarella, “Tracing, ranking and valuation of aggregated DER flexibility in active distribution networks,” IEEE Trans. Smart Grid , vol. 15, no. 2, pp. 1694–1711, 2023

    work page 2023

  20. [20]

    Aggregate flexibility of virtual power plants with temporal coupling constraints,

    S. Wang and W. Wu, “Aggregate flexibility of virtual power plants with temporal coupling constraints,” IEEE Trans. Smart Grid , vol. 12, no. 6, pp. 5043–5051, 2021

    work page 2021

  21. [21]

    Time-based aggregation of flexibility at the TSO-DSO interconnection point,

    D. A. Contreras and K. Rudion, “Time-based aggregation of flexibility at the TSO-DSO interconnection point,” in 2019 IEEE PES Gen. Meet. IEEE, 2019, pp. 1–5

    work page 2019

  22. [22]

    Complex power flexibility evaluation using energy arbitrage between transmission and distribution,

    R. Vijay and P. Mathuria, “Complex power flexibility evaluation using energy arbitrage between transmission and distribution,” Electr. Power Syst. Res., vol. 203, p. 107641, 2022

    work page 2022

  23. [23]

    Reduced and aggregated distribution grid representations approximated by polyhedral sets,

    P. Fortenbacher and T. Demiray, “Reduced and aggregated distribution grid representations approximated by polyhedral sets,” Int. J. Electr. Power Energy Syst., vol. 117, p. 105668, 2020

    work page 2020

  24. [24]

    A novel two-stage TSO–DSO coordination approach for managing congestion and voltages,

    M. Usman, M. I. Alizadeh, F. Capitanescu, I.-I. Avramidis, and A. G. Madureira, “A novel two-stage TSO–DSO coordination approach for managing congestion and voltages,” Int. J. Electr. Power Energy Syst. , vol. 147, p. 108887, 2023

    work page 2023

  25. [25]

    Monetarization of the feasible operation region of active distribution grids based on a cost-optimal flexibility disaggregation,

    M. Sarstedt and L. Hofmann, “Monetarization of the feasible operation region of active distribution grids based on a cost-optimal flexibility disaggregation,” IEEE Access, vol. 10, pp. 5402–5415, 2022

    work page 2022

  26. [26]

    Leveraging two-stage adaptive robust optimization for power flexibility aggregation,

    X. Chen and N. Li, “Leveraging two-stage adaptive robust optimization for power flexibility aggregation,” IEEE Trans. Smart Grid , vol. 12, no. 5, pp. 3954–3965, 2021

    work page 2021

  27. [27]

    Aggregate modeling of distribution systems for multi-period OPF,

    E. Polymeneas and S. Meliopoulos, “Aggregate modeling of distribution systems for multi-period OPF,” in Power Syst. Comput. Conf. (PSCC) . IEEE, 2016, pp. 1–8

    work page 2016

  28. [28]

    Coordinated vertical provision of flexibility from distribution systems,

    H. Fr ¨uh, S. M ¨uller, D. Contreras, K. Rudion, A. von Haken, and B. Sur- mann, “Coordinated vertical provision of flexibility from distribution systems,” IEEE Trans. Power Syst., vol. 38, no. 2, pp. 1834–1844, 2022

    work page 2022

  29. [29]

    Characterizing the reserve provision capability area of active distribu- tion networks: A linear robust optimization method,

    M. Kalantar-Neyestanaki, F. Sossan, M. Bozorg, and R. Cherkaoui, “Characterizing the reserve provision capability area of active distribu- tion networks: A linear robust optimization method,” IEEE Trans. Smart Grid, vol. 11, no. 3, pp. 2464–2475, 2019

    work page 2019

  30. [30]

    Modelling and characterisation of flexibility from distributed energy resources,

    S. Riaz and P. Mancarella, “Modelling and characterisation of flexibility from distributed energy resources,” IEEE Trans. Power Syst. , vol. 37, no. 1, pp. 38–50, 2021

    work page 2021

  31. [31]

    Estimating the robust PQ capability of a technical virtual power plant under uncertainties,

    Z. Tan, H. Zhong, Q. Xia, C. Kang, X. S. Wang, and H. Tang, “Estimating the robust PQ capability of a technical virtual power plant under uncertainties,” IEEE Trans. Power Syst., vol. 35, no. 6, pp. 4285– 4296, 2020

    work page 2020

  32. [32]

    Optimal control of wind farms for coordinated TSO-DSO reactive power management,

    D. S. Stock, F. Sala, A. Berizzi, and L. Hofmann, “Optimal control of wind farms for coordinated TSO-DSO reactive power management,” Energies, vol. 11, no. 1, p. 173, 2018

    work page 2018

  33. [33]

    V oltage- dependent PQ reserve capacity evaluation at TSO-DSO interface con- sidering uncertainties of DGs and FLs,

    Y . Xu, L. Yao, T. Pu, S. Liao, F. Cheng, Y . Li, and X. Wang, “V oltage- dependent PQ reserve capacity evaluation at TSO-DSO interface con- sidering uncertainties of DGs and FLs,” CSEE Journal of Power and Energy Systems, 2022

    work page 2022

  34. [34]

    TSO- DSO interaction: Privacy-preserving optimal power flow with distributed generators using a machine learning-based approach,

    B. Dindar, C. B. Saner, H. K. C ¸ akmak, and V . Hagenmeyer, “TSO- DSO interaction: Privacy-preserving optimal power flow with distributed generators using a machine learning-based approach,” in IEEE PES 15th Asia-Pacific Power Energy Eng. Conf. (APPEEC). IEEE, 2023, pp. 1–6

    work page 2023

  35. [35]

    A machine learning-based privacy-preserving approach to incorporate distributed generators in ac optimal power flow,

    B. Dindar, C. B. Saner, D. Y . Polat, H. K. C ¸ akmak, and V . Hagenmeyer, “A machine learning-based privacy-preserving approach to incorporate distributed generators in ac optimal power flow,” in 2024 IEEE PES Innovative Smart Grid Technologies Europe (ISGT EUROPE) . IEEE, 2024, pp. 1–6

    work page 2024

  36. [36]

    Towards smarter grids: developing TSO and DSO roles and interactions for the benefit of con- sumers,

    ENTSO-E, “Towards smarter grids: developing TSO and DSO roles and interactions for the benefit of con- sumers,” Availabe: https://eepublicdownloads.entsoe.eu/clean- documents/Publications/Position%20papers%20and%20reports /150303 ENTSOE Position Paper TSO-DSO interaction.pdf, 2015

    work page 2015

  37. [37]

    Improvements to and limitations of Latin hypercube sampling,

    D. Huntington and C. Lyrintzis, “Improvements to and limitations of Latin hypercube sampling,” Probabilistic engineering mechanics , vol. 13, no. 4, pp. 245–253, 1998

    work page 1998

  38. [38]

    D. A. C. Schneider, Estimation of flexibility potentials in active distri- bution networks. BoD–Books on Demand, 2021, vol. 34

    work page 2021

  39. [39]

    On feasibility and flexibility operating regions of virtual power plants and TSO/DSO interfaces,

    S. Riaz and P. Mancarella, “On feasibility and flexibility operating regions of virtual power plants and TSO/DSO interfaces,” in 2019 IEEE Milan PowerTech. IEEE, 2019, pp. 1–6

    work page 2019

  40. [40]

    Computing the feasible operating region of active distribution networks: Comparison and validation of random sampling and optimal power flow based methods,

    D. A. Contreras and K. Rudion, “Computing the feasible operating region of active distribution networks: Comparison and validation of random sampling and optimal power flow based methods,” IET Gener. Transm. Distrib., vol. 15, no. 10, pp. 1600–1612, 2021

    work page 2021

  41. [41]

    Estimating B-coefficients of power loss formula considering volatile power injections: an enhanced least square approach,

    B. Liu, F. Liu, W. Wei, and J. Wang, “Estimating B-coefficients of power loss formula considering volatile power injections: an enhanced least square approach,”IET Gener. Transm. Distrib., vol. 12, no. 12, pp. 2854– 2860, 2018

    work page 2018

  42. [42]

    MAT- POWER: Steady-state operations, planning, and analysis tools for power systems research and education,

    R. D. Zimmerman, C. E. Murillo-S ´anchez, and R. J. Thomas, “MAT- POWER: Steady-state operations, planning, and analysis tools for power systems research and education,”IEEE Trans. Power Syst., vol. 26, no. 1, pp. 12–19, 2011

    work page 2011

  43. [43]

    Knitro: An integrated package for nonlinear optimization,

    R. H. Byrd, J. Nocedal, and R. A. Waltz, “Knitro: An integrated package for nonlinear optimization,” Large-scale nonlinear optimization, pp. 35– 59, 2006

    work page 2006

  44. [44]

    TensorFlow: Large-scale machine learning on heterogeneous systems,

    M. Abadi et al. , “TensorFlow: Large-scale machine learning on heterogeneous systems,” 2015, software available from tensorflow.org. [Online]. Available: https://www.tensorflow.org/

    work page 2015

  45. [45]

    Chollet et al., “Keras,” https://keras.io, 2015

    F. Chollet et al., “Keras,” https://keras.io, 2015

    work page 2015

  46. [46]

    A review on evaluation metrics for data classification evaluations,

    M. Hossin and M. N. Sulaiman, “A review on evaluation metrics for data classification evaluations,” International journal of data mining & knowledge management process, vol. 5, no. 2, p. 1, 2015

    work page 2015