Skip to main content
Have a personal or library account? Click to login
Hybrid DSTATCOM Design Using Deep Belief Networks for Enhanced Power Quality Assessment Cover

Hybrid DSTATCOM Design Using Deep Belief Networks for Enhanced Power Quality Assessment

Power Electronics and Drives
Volume 11 (2026): Issue 1 (January 2026)
By: ,   and    
Open Access
|Apr 2026
Figures & tables

Figures & Tables

Table 1.

Comparison of different algorithms for analysis of filtering characteristics

Control algorithmYear of publicationRobustnessConvergence capabilityControl flexibilityPower lossCostLC couplingFiltering process
ALMS2018✓×---×Active
KHLMS2019✓✓--------×Active
DRL2020✓✓-------×Active
Delta-Bar-Delta NN2020✓✓------×Active
SLMS2021✓✓-------×Active
ADALINE-LMS2021✓×---×Active
DL2022✓✓-----×Active
DBLN2023✓✓------✓Hybrid

[i] Notation used: ✓= Yes, × = No, -=Low, --=Medium, ---=High

[ii] ALMS, adaptive least mean square; DBLN, deep belief learning network; DL, deep learning; DRL, deep recurrent learning principles are proposed for the purpose of active filtering. No author has attempted to employ such a technique for hybrid filtering. The above literature review has been motivated by the author to suggest a DBLN technique-based LC-supported DSTATCOM to provide the behaviour for better convergence characteristics, better control capability, low power loss, less cost and better shunt compensation.

Figure 1.

Circuit connection of the DBLN based DSTATCOM. DBLN, deep belief learning network; DSTATCOM, distributed static compensator; VSC, voltage source converter.

Figure 2.

Circuit connection of the DBLN-based LC supported. DBLN, deep belief learning network; DSTATCOM, distributed static compensator; VSC, voltage source converter.

Figure 3.

P3W Voltage Source Inverter-based DSTATCOM. DSTATCOM, distributed static compensator.

Figure 4.

Stability analysis of LC filter by using Bode plot.

Figure 5.

Overall DBLN control algorithm of DSTATCOM. DBLN, deep belief learning network; DSTATCOM, distributed static compensator.

Figure 6.

Flow chart for finding the tuned weight to improve the shunt compensation using DBLN technique. DBLN, deep belief learning network.

Figure 7.

Learning mechanism using DBLN for the extraction of reactive part for a-phase. DBLN, deep belief learning network.

Figure 8.

(a) System performance for DSTATCOM based on DBLN, (b). THD of the load current for DSTATCOM based under DBLN, (c). THD of the source current for DSTATCOM based on DBLN, (d). Source side power factor p.f of the phase-a, (e). Load side power factor of the phase-a. DBLN, deep belief learning network; DSTATCOM, distributed static compensator; THD, total harmonic distortions.

Figure 9.

(a) System performance for LC supported DSTATCOM based on DBLN, (b). THD of the load current for LC supported DSTATCOM based under DBLN, (c). THD of the source current for LC supported DSTATCOM based under DBLN, (d). Source side power factor of the phase-a, (e). Load side power factor of the phase-a. DBLN, deep belief learning network; DSTATCOM, distributed static compensator; THD, total harmonic distortions.

Table 2.

Comparative performance evaluation of different types of DSTATCOM.

Performance parameterALMS based DSTATCOMDBLN based DSTATCOMDBLN based LC supported DSTATCOM
is (A), %THD55.88, 4.5655.46, 4.0754.46, 2.07
vs (V), %THD321.4, 2.54321.4, 2.23321, 1.42
il (A), %THD42.96, 18.9651.13, 20.6451.13, 20.64
Power factor0.940.970.99
vdc (V)680671.6605.6

[i] ALMS, adaptive least mean square; DBLN, deep belief learning network; DSTATCOM, distributed static compensator; THD, total harmonic distortion.

Figure 10.

Experimental setup of the DBLN-based LC-supported DSTATCOM. DBLN, deep belief learning network; DSTATCOM, distributed static compensator.

Figure 11.

Experimental waveform of source current, compensator current and load current using DBLN mechanism under (a) constant loading, (b) diversity loading. DBLN, deep belief learning network.

Figure 12.

Experimental results of source current, current of compensator and current of using DBLN controlled LC coupled DSTATCOM for (a). Constant loading, (b) diversity loading. DBLN, deep belief learning network; DSTATCOM, distributed static compensator.

Table 3.

Simulation/experimental parameters for proposed configuration.

SymbolDefinitionValue
vs3-phase source voltage230 V/phase
fsFrequency50 Hz
RsSource resistance0.5 Ω
LsSource inductance2 mH
LfPassive filter inductance0.5 mH
CfPassive filter capacitance10 μF
KprAC proportional controller0.2
CdcCapacitor2,000 μF
KpaDC proportional controller0.01
KiaDC integral controller0.05
vdc (ref)DC link voltage650 V
RcVSC resistance0.25 Ω
LcVSC inductance1.5 mH
KirAC integral controller1.1

[i] VSC, voltage source converter.

DOI: https://doi.org/10.2478/pead-2026-0008 | Journal eISSN: 2543-4292 | Journal ISSN: 2451-0262
Journal RSS Feed
Language: English
Page range: 128 - 141
Submitted on: Dec 20, 2025
Accepted on: Mar 4, 2026
Published on: Apr 17, 2026
Published by: Wroclaw University of Science and Technology
In partnership with: Paradigm Publishing Services
Keywords:
distributed static compensator,
deep belief learning network,
inductor and capacitor coupling,
power quality
Related subjects:
Computer sciences,
Artificial intelligence,
Engineering,
Electrical engineering,
Electronics

© 2026 Pappu Suneetha, Kanna Subba Ramaiah, Nagalamadaka Visali, published by Wroclaw University of Science and Technology
This work is licensed under the Creative Commons Attribution 4.0 License.

Volume 11 (2026): Issue 1 (January 2026)
Previous article
Next article