pith. sign in

arxiv: 1907.10273 · v1 · pith:MWYM5DMSnew · submitted 2019-07-24 · 📡 eess.SP · cs.SY· eess.SY

Real-time frequency based reduced order modeling of large power grid

Abilash Thakallapelli , Sudipta Ghosh , Sukumar Kamalasadan This is my paper

Pith reviewed 2026-05-24 16:59 UTC · model grok-4.3

classification 📡 eess.SP cs.SYeess.SY
keywords power system reductionfrequency dependent network equivalentrecursive least squaresreal-time simulationelectromagnetic transientz-domain identificationtransient stability analysis
0
0 comments X

The pith

An online recursive least squares algorithm in the z-domain identifies frequency-dependent network equivalents for reduced real-time power grid modeling.

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

The paper presents a method to develop frequency-dependent network equivalents (FDNE) for the external area of large power systems using an online recursive least squares identification algorithm operating in the z-domain. This enables two forms of reduced-order representation: FDNE alone or a hybrid of FDNE with transient stability analysis phasor models. Both forms are tested on real-time digital simulators to support electromagnetic transient studies that would otherwise be computationally infeasible for full-scale grids. A sympathetic reader cares because the approach aims to retain both high-frequency transients and low-frequency electromechanical dynamics while cutting the resources needed for real-time simulation.

Core claim

FDNE can be constructed via an online RLS identification algorithm in the z-domain so that the external area of a power system is represented either solely by the FDNE or by a combination of TSA phasor equivalents and FDNE; both representations are realized on real-time digital simulators while preserving the electromechanical and high-frequency behavior of the original network.

What carries the argument

Frequency Dependent Network Equivalent (FDNE) constructed by online recursive least squares identification in the z-domain.

If this is right

  • Electromagnetic transient studies of large grids become feasible in real time on digital simulators.
  • The external area can be modeled either as pure FDNE or as a TSA-plus-FDNE hybrid depending on the frequencies of interest.
  • Both high-frequency and low-frequency dynamics are retained without requiring a full EMT model of the entire system.
  • The identification step runs online, allowing the equivalent to be updated as system conditions change.

Where Pith is reading between the lines

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

  • If the RLS step remains stable under topology changes, the same identification could support adaptive real-time security assessment.
  • Combining this FDNE with other reduction techniques such as coherency-based aggregation might further lower computational cost.
  • The z-domain formulation could be ported to hardware-in-the-loop testbeds for controller validation.

Load-bearing premise

The external area of the power system can be represented accurately enough by the FDNE identified through online RLS that no significant model mismatch or unmodeled dynamics affect the internal area under study.

What would settle it

A side-by-side comparison on the real-time simulator in which the reduced FDNE or TSA+FDNE model produces voltage or current waveforms that deviate materially from the full-order model during a high-frequency transient or electromechanical swing would falsify the claim.

read the original abstract

Large power systems are complex and real-time modeling of the grid for electromagnetic simulation (EMT) studies is impractical. In general, there are methods that reduce large power system into an equivalent network that requires less computational resource, while preserving electromechanical (low frequency) and high frequency behavior of the original system. This can be achieved by modeling the area not of interest (external area) as a combination of Transient Stability Analysis (TSA) type phasor model equivalent and Frequency Dependent Network Equivalent (FDNE). TSA retains electromechanical behavior, whereas FDNE retains high frequency behavior of the original power system. To this effect, this paper introduces a method of developing FDNE based on an online recursive least squares (RLS) identification algorithm in z-domain, and modeling of reduced power systems as FDNE and as a combination of TSA and FDNE using real-time digital simulators.

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

1 major / 0 minor

Summary. The paper introduces a method of developing Frequency Dependent Network Equivalents (FDNE) based on an online recursive least squares (RLS) identification algorithm in the z-domain. It demonstrates modeling of reduced power systems as standalone FDNE or as a hybrid of Transient Stability Analysis (TSA) phasor equivalents and FDNE, implemented on real-time digital simulators to enable EMT studies of large grids while aiming to retain both low-frequency electromechanical and high-frequency behaviors.

Significance. If the RLS-based FDNE identification proves accurate and stable under realistic boundary conditions, the approach would enable computationally tractable real-time EMT simulation of continental-scale grids by offloading external areas to a combination of phasor and frequency-dependent equivalents.

major comments (1)
  1. [Abstract] Abstract: the central claim that the online RLS algorithm produces an FDNE whose frequency response matches the external area closely enough to preserve internal-area EMT behavior is unsupported by any error metrics, convergence analysis, or comparison to a reference admittance matrix; without these, the weakest assumption (accurate representation without model mismatch) cannot be evaluated.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback. We address the single major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that the online RLS algorithm produces an FDNE whose frequency response matches the external area closely enough to preserve internal-area EMT behavior is unsupported by any error metrics, convergence analysis, or comparison to a reference admittance matrix; without these, the weakest assumption (accurate representation without model mismatch) cannot be evaluated.

    Authors: The manuscript body (Sections III and IV) contains direct comparisons of the identified FDNE admittance against the reference admittance matrix of the external area, with plotted magnitude/phase responses over 0.1 Hz–10 kHz, tabulated maximum absolute errors, and time-domain EMT waveform overlays that confirm preservation of internal-area behavior. RLS convergence is shown via parameter-error trajectories and residual-norm plots for both standalone FDNE and hybrid TSA+FDNE cases. We agree the abstract itself does not cite these quantitative elements and will revise it to include a concise statement referencing the validation metrics and error bounds reported in the results. revision: yes

Circularity Check

0 steps flagged

No circularity: algorithmic RLS identification presented as procedure

full rationale

The paper describes an online recursive least squares (RLS) algorithm in the z-domain to identify an FDNE for the external area, combined with TSA for the internal area. This is framed as a standard system-identification procedure applied to measured or simulated frequency-response data. No equations, derivations, or claims are shown that reduce the identified model to its own fitted parameters by construction, nor are there self-citations invoked as load-bearing uniqueness theorems. The central claim is the feasibility of the real-time implementation on digital simulators, which rests on the external validity of RLS convergence rather than any internal redefinition of inputs as outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The claim rests on the assumption that RLS identification converges to a usable FDNE model for the external area; no free parameters or new entities are named in the abstract.

axioms (1)
  • domain assumption The external power system area behaves as a linear time-invariant system identifiable in the z-domain via RLS
    Required for the online identification algorithm to produce a valid FDNE.

pith-pipeline@v0.9.0 · 5691 in / 1086 out tokens · 21981 ms · 2026-05-24T16:59:28.713676+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

19 extracted references · 19 canonical work pages

  1. [1]

    Aggregation is done in two steps: 1) Network aggregation, 2) Generator and associated controller aggregation

    Aggregation of External area for TSA Initially aggregated TSA model is formulated for retaining electromechanical behavior of the external system. Aggregation is done in two steps: 1) Network aggregation, 2) Generator and associated controller aggregation. a) Network Aggregation For network agg regation first a dmittance matrix ''Y of the external area is...

    work page

  2. [2]

    By tracking input voltage and output current, FDNE is then formulated in z -domain using RLS

    Frequency Dependent Network Equivalent Formulation In this method external area is ene rgized with constant voltage source after short circuiting all voltage sources, and open circuiting all current sources. By tracking input voltage and output current, FDNE is then formulated in z -domain using RLS. Basic principle is as follows. If '' bV is the voltage ...

    work page

  3. [3]

    Here, first, the generator bus voltage is calc ulated as shown in (24 )

    Interfacing TSA type modeling with real-time simulator In this step, input to the TSA block is considered the voltage at the boundary bus, and output from the TSA block is current that is to be injected into the boundary bus. Here, first, the generator bus voltage is calc ulated as shown in (24 ). In this case, generator is model ed in detail to observe t...

    work page

  4. [4]

    study” area using a detailed EMT solution and the “external

    Interfacing FDNE with real time simulator FDNE can be directly implemented since it is computed in z-domain. With process input as bV and process output as bI , and with third order identification, (4) can be written as 1 2 3 1 2 3 ( ) ( 1) ( 2) ( 3) ( 1) ( 2) ( 3) F F F F b b b I k a I k a I k a I k bV k b V k b V k = − − − − − − + − + − + − (27) where F...

    work page

  5. [5]

    Multi-port equivalencing of external systems for simulation of swit ching transients ,

    H. Singh and A. Abur, “Multi-port equivalencing of external systems for simulation of swit ching transients ,” IEEE Trans. Power Delivery, vol. 10, no. 1, pp. 374-382, Jan. 1995

    work page 1995

  6. [6]

    Time domain modeling of external systems for electromagnetic transients programs,

    A. Abur and H. Singh, “Time domain modeling of external systems for electromagnetic transients programs, ” IEEE Trans. Power Syst , vol. 8, no. 2, pp. 671-679, May 1993

    work page 1993

  7. [7]

    Interfacing convolution based linear models to an electromagnetic transients program ,

    B. Gustavsen and O. Mo, “Interfacing convolution based linear models to an electromagnetic transients program ,” in Proc. Int. Conf. Power Systems Transients, Lyon, France, Jun. 4-7, 2007, pp. 4-7

    work page 2007

  8. [8]

    Rational approximation of frequency domain responses by vector fitting,

    B. Gustavsen and A. Semlyen, “ Rational approximation of frequency domain responses by vector fitting,” IEEE Trans. Power Del, vol. 14, no. 3, pp. 1052-1061, Jul. 1999

    work page 1999

  9. [9]

    Frequency domain modeling of external systems in an elec tro-magnetic transients program,

    A.Semlyen and M. R. Iravani , “Frequency domain modeling of external systems in an elec tro-magnetic transients program, ” IEEE Trans. Power Syst., vol. 8, no. 2, pp. 527-533, May 1993

    work page 1993

  10. [10]

    Development and analysis of applicability of a hybrid transient simulation platform combining TSA and EMT elements,

    Y. Zhang et al., “Development and analysis of applicability of a hybrid transient simulation platform combining TSA and EMT elements,” IEEE Trans. Power Syst., vol. 28, no. 1, pp. 357-366, Feb. 2013

    work page 2013

  11. [11]

    Available: http://nptel.ac.in/courses/Webcourse- contents/IIT-Kanpur/power-system/chapter_3/3_6.html

    Power System Analysis. Available: http://nptel.ac.in/courses/Webcourse- contents/IIT-Kanpur/power-system/chapter_3/3_6.html

    work page

  12. [12]

    Inertial and slow coherency aggregation algorithms for power system dynamic model reduction,

    J. Chow, P. Accari, and W. Price, “Inertial and slow coherency aggregation algorithms for power system dynamic model reduction,” IEEE Trans. Power Syst., vol. 10, no. 2 , pp. 680 -685, May 1995

    work page 1995

  13. [13]

    K. J. Åström and B. Wittenmark , Adaptive control . Reading, MA: Addison-Wesley Publishing Company, 1995

    work page 1995

  14. [14]

    J. H. Chow, Power system coherency and model reduction. New York, NY, USA: Springer, 2013

    work page 2013

  15. [15]

    The localness of electromechanica l oscillations in power systems,

    S. Ghosh and N. Senroy, “ The localness of electromechanica l oscillations in power systems,” Int. J. Elect. Power Energy. Syst., vol. 42, no. 1, pp. 306-313, Nov. 2012

    work page 2012

  16. [16]

    Doubly Fed Induction Generator (DFIG) -based Wind Farm Control Framework for Primary Frequency and Inertial Response Application

    S. Ghosh, S. Kamalasadan, N. Senroy and J. Enslin, “Doubly Fed Induction Generator (DFIG) -based Wind Farm Control Framework for Primary Frequency and Inertial Response Application”, IEEE Trans. Power Syst., no. 99, pp. 1-11, Jun. 2015

    work page 2015

  17. [17]

    A PID Controller for Real -Time DC Motor Speed Control using the C505C Microcontr oller

    S. Kamalasadan and A. Hande, “A PID Controller for Real -Time DC Motor Speed Control using the C505C Microcontr oller”, CAINE, pp. 1- 8, 2004

    work page 2004

  18. [18]

    System Equivalent for Real Time Digital Simulator

    X. Lin, “System Equivalent for Real Time Digital Simulator ”, Ph.D dissertation, Dept Elect Eng. , Univ . of Manitoba, Winnipeg, Manitoba, Jan. 2011. [Online]. Available: http://hdl.handle.net/1993/4375

    work page 2011

  19. [19]

    Multiport frequency -dependent network equivalencing based on simulated time -domain responses,

    A. Ubolli and B. Gustavsen, “Multiport frequency -dependent network equivalencing based on simulated time -domain responses,” IEEE Trans. Power Del., vol. 27, no. 2, pp. 648-657, Apr. 2012

    work page 2012