Academia.eduAcademia.edu
All Topics
Engineering
Signal Processing
Active Filtering

Active Filtering

description10 papers
group8 followers
lightbulbAbout this topic
Active filtering is a signal processing technique that utilizes active components, such as operational amplifiers, to enhance or suppress specific frequency components of a signal. It contrasts with passive filtering by providing gain and allowing for more complex filter designs, enabling improved performance in applications like audio processing and communication systems.
lightbulbAbout this topic
Active filtering is a signal processing technique that utilizes active components, such as operational amplifiers, to enhance or suppress specific frequency components of a signal. It contrasts with passive filtering by providing gain and allowing for more complex filter designs, enabling improved performance in applications like audio processing and communication systems.

Key research themes

1. How can active learning strategies be optimized and unified across different domains such as classification and collaborative filtering to efficiently select informative data for labeling?

Active learning (AL) research focuses on designing algorithms and strategies that minimize labeling effort by selecting the most informative data points to label. This is critical due to the high cost and time consumption of obtaining labeled data. A major research thrust has been in creating taxonomies and frameworks for selecting data, extending AL to specialized areas like collaborative filtering recommender systems, and developing unified approaches that generalize across different realistic data settings. These efforts enhance learning efficiency, especially when data distributions are complex, imbalanced, or noisy.

A Survey on Active Learning: State-of-the-Art, Practical Challenges and Research Directions
by Wolfram Schenck

2023, Mathematics

Key finding: This comprehensive survey delineates AL scenarios and query strategies in classification, categorizing query strategies into uncertainty-based and representation-based, emphasizing querying the most informative and... Read more
articleView PaperdownloadDownload
A survey of active learning in collaborative filtering recommender systems
by Mehdi Elahi

2016

Key finding: The paper classifies active learning strategies in collaborative filtering along two orthogonal dimensions: personalization (personalized vs non-personalized item selection) and hybridization (single vs multi-heuristic... Read more
articleView PaperdownloadDownload
Active Learning in Collaborative Filtering Recommender Systems
by Mehdi Elahi

2016

Key finding: This article reinforces the taxonomy of AL strategies by emphasizing the importance of balancing personalization and heuristic hybridization in selecting informative ratings. It offers a systematic comparison of more than 24... Read more
articleView PaperdownloadDownload
SIMILAR: Submodular Information Measures Based Active Learning In Realistic Scenarios
by KrishnaTeja Killamsetty

2022, ArXiv

Key finding: Proposes SIMILAR, a unified active learning framework leveraging submodular information measures to handle various challenging real-world scenarios—imbalanced data, rare classes, data redundancy, and out-of-distribution... Read more
articleView PaperdownloadDownload
Region-Based Active Learning
by Claudio Gentile

2025, International Conference on Artificial Intelligence and Statistics

Key finding: Introduces ORIWAL, a region-based AL framework where the input space is partitioned and a distinct hypothesis is learned per region. By optimally allocating labeling budget regionally using subroutines like IWAL/EIWAL, the... Read more
articleView PaperdownloadDownload
keyboard_arrow_downShow more

2. What are the methodological advances in adaptive filtering and active filtering that enable faster convergence, data selection, and hybrid applications in signal processing and renewable energy systems?

The evolution of adaptive and active filtering methods emphasizes efficient data selection to improve convergence rates and performance, particularly in high-data or complex environments like communications, power quality, and renewable energy integration. Algorithms with data-selective update rules reduce computational cost and enhance robustness by only updating when novel, informative data are present. Simultaneously, new active filter designs and control methods improve harmonic compensation and power quality in power systems, including renewable generation facilities, through advanced control theory, signal decomposition, and system modeling.

On Fast Converging Data-Selective Adaptive Filtering
by Paulo Diniz

2021, Algorithms

Key finding: This work explores adaptive filtering algorithms—LMS Newton, LMS Quasi-Newton, and Conjugate Gradient—with data selection strategies based on mean squared error thresholds to update coefficients only on informative samples.... Read more
articleView PaperdownloadDownload
Linear Transformation Active Filters
by Anthony Constantinides

2018

Key finding: Presents a systematic design method for active inductorless filters by applying linear transformations to lossless ladder filter port variables. This generalized framework includes existing leapfrog and wave active filter... Read more
articleView PaperdownloadDownload
A proposal for a wind system equipped with a doubly fed induction generator using the Conservative Power Theory for active filtering of harmonics currents
by Adson Moreira

2023, Electric Power Systems Research

Key finding: This study integrates an active filtering capability into doubly fed induction generator (DFIG) wind power systems via Conservative Power Theory-based current decomposition in the grid side converter control. The approach... Read more
articleView PaperdownloadDownload
Harmonic Active Filtering and Impedance-based Stability Analysis in Offshore Wind Power Plants
by Guangya Yang

2023

Key finding: Explores embedding active filtering functions directly into wind turbine grid-side converters to adaptively mitigate harmonics caused by resonances and nonlinearities in offshore wind power plants. It presents system... Read more
articleView PaperdownloadDownload
Active filtering and VAR control of a cascaded H-bridge multi-level StatCom
by Galina Mirzaeva

2023, 2009 International Conference on Power Electronics and Drive Systems (PEDS)

Key finding: Develops a control algorithm for multilevel cascaded H-bridge STATCOMs to simultaneously achieve VAR compensation and active harmonic filtering using predictive current control and symmetrical PWM. The paper introduces a... Read more
articleView PaperdownloadDownload
keyboard_arrow_downShow more

3. How can intelligent information filtering and active learning be combined to enhance automated data selection under constrained labeling, especially in complex input spaces and temporal sequences?

Research in intelligent information filtering and active learning investigates methods to predict user-interest or importance of data based on complex criteria that may depend on future outcomes or contextual variables—termed prospective criteria. Combining predictive modeling, rule extraction, and active sampling strategies can improve filtering accuracy and model interpretability. Furthermore, region-based active learning partitions input spaces to allocate labeling resources efficiently, handling heterogeneous data distributions. The integration of active learning with mechanisms such as face tracking in vision systems exemplifies autonomous data acquisition under minimal supervision.

New Techniques In Intelligent Information Filtering
by Sofus Macskássy

2025

Key finding: Introduces a novel framework for information filtering where user interestingness is a prospective criterion, dependent on future events. It operationalizes such criteria via predictive modeling, enhanced with rule extraction... Read more
articleView PaperdownloadDownload
Region-Based Active Learning
by Claudio Gentile

2025, International Conference on Artificial Intelligence and Statistics

Key finding: Proposes a region-based on-line active learning algorithm (ORIWAL) that partitions the input space into regions, learning distinct hypotheses per region. By optimally distributing labeling effort among these via subroutines... Read more
articleView PaperdownloadDownload
Active Learning in Face Recognition: Using Tracking to Build a Face Model
by Robin Hewitt

2025, 2006 Conference on Computer Vision and Pattern Recognition Workshop (CVPRW'06)

Key finding: Demonstrates an active learning method for autonomous acquisition of face training data using co-training principles combining single-image initialized face detection with tracking to collect diverse, well-localized face... Read more
articleView PaperdownloadDownload
keyboard_arrow_downShow more

Related Topics

  • Power Quality Improvement Using FACTs Devices
  • Renewable energy resources
  • Electrical Machines, Electrical Drives, Power Electronics
  • Power Electronics Converter
  • Renewable Energy
  • Archaeology
  • Electrical Engineering
  • Computer Science

    • All papers in Active Filtering

    Programmable mobile networks
    by Raymond Liao

    2024, Computer Networks

    descriptionView Paperarrow_downwardDownload
    Power Quality Improvement of a Solar Energy Conversion System by a Coordinated Active and LCL Filtering
    by Mohamed Machmoum

    2024, Periodica Polytechnica Electrical Engineering and Computer Science

    Power converters play an essential role in Photovoltaic (PV) system to maximize the power transfer to the electrical grid. However, the generated harmonics in the grids due to these power converters and nonlinear loads are considered one... more
    descriptionView Paperarrow_downwardDownload
    Harmonic Mitigation Methods in Large Offshore Wind Power Plants
    by Bo Hesselbæk

    2024

    Various sources of harmonic problems in large wind power plants (WPPs) and optimized harmonic mitigation methods are presented in this paper. The harmonic problems such as sources of harmonic emission and amplification as well as harmonic... more
    descriptionView Paperarrow_downwardDownload
    Harmonic Mitigation Methods in Large Offshore Wind Power Plants
    by Bo Hesselbæk

    2024

    Various sources of harmonic problems in large wind power plants (WPPs) and optimized harmonic mitigation methods are presented in this paper. The harmonic problems such as sources of harmonic emission and amplification as well as harmonic... more
    descriptionView Paperarrow_downwardDownload
    A proposal for a wind system equipped with a doubly fed induction generator using the Conservative Power Theory for active filtering of harmonics currents
    by Adson Moreira

    2023, Electric Power Systems Research

    This study describes a strategy that allows the addition of active filtering functionality to a doubly fed induction generator (DFIG) wind power system. The active filtering is performed through an algorithm that uses the mathematical... more
    Fig. 18. Electric current in the grid before the injection of active power by the wind system and without active filtering (a) and harmonic spectrum (b). Fig. 17. Reactive (a) and active power produced by DFIG (b).
    Fig. 8. Block diagrams illustrating the DFIG rotor current control loops.
    Fig. 11. Open-loop frequency response of the filter L plant cascaded with the designed PI controller.
    DC link filter and capacitor parameters.
    Parameters of the wind power system studied.
    Parameters of non-linear load.
    Fig. 3. Schematic diagram for the stator magnetic flux estimator. Given a generic polyphase circuit of m conductors with or without return conductor under periodic operation condition. The CPT authors define detailed mathematical formulation of CPT that can be found in Refs. [16] and [30]. However, the concepts necessary for the applica- tion of CPT in this study are defined [35].
    Fig. 4. Schematic diagram of the dq current control loop for the GSC.
    Fig. 2. DFIG current control loop (a) and electric current transform diagram of the rotor from abc to dq and rotational velocity used for decoupling and feed-forward compensation (b).
    Fig. 20. Electric current in the grid with active power supply by the wind system and with active filtering (a) and frequency spectrum (b).
    Fig. 7. Voided current measurement (a) and control loop GSC with the active filtering function (b). Thus, we add the direct and quadrature axis components of the residual current with the dq components of the active and reactive power control of the GSC, in order to generate the references for the control system in synchronous coordinates. According to the results that will be presented in the simulation results section this operation allows to add the active filtering function to the wind power system with DFIG. P, is the active phase power; W,, is the phase reactive power. It is shown by the Tellegen theorem that Relations (35)-(38) re- present conservative quantities for any systems, independent of voltage
    The concept of expanded plant consists of calculating an equivalent transfer function that allows designing the controller gains given by Gep(s): Fig. 10. Block diagrams illustrating the current control loop of the L filter.
    Fig. 15. Void current calculated in CPT.
    Fig. 16. Voltage in the direct current link (a) and electrical current in the DFIG stator (b).
    Fig. 9. Bode diagram for the rotor current control loop.
    Fig. 1. Schematic diagram proposed for a wind system with active filtering function.
    Fig. 12. Block diagram of DC link voltage control: complete (a) expanded plant (b). The wind system shown in Fig. 1 was simulated using components
    Collective instantaneous active power and collective instantaneous reactive energy:
    Fig. 13. Open-loop frequency response of the cascade-augmented plant with the designed PI controller.
    Fig. 14. Electrical current measured in the set of loads (a) harmonic spectrum (b). The electric current shown in Fig. 18-a is distorted, because of the electric current drained by the set of non-linear loads. The THD of the electric current is 16.85% and the grid current harmonic spectrum contains components of 5th, 7th, 11th and 13th orders, in Fig. 18-b.
    Fig. 19. Electric current in the grid with active power supply by the wind system and without active filtering (a) and harmonic spectrum (b).
    Fig. 21. Electric current in the grid without active power from the wind system and with active filtering (a) and frequency spectrum (b)
    descriptionView Paperarrow_downwardDownload
    Design and Optimization of Hybrid Renewable Energy System (2MWH/D) for Sustainable and Economical Power Supply at Jec Jabalpur
    by BHUPENDRA SINGH THAKUR

    2023, International Journal of Current Research and Review

    A hybrid renewable energy system may be used to reduce dependency on either conventional energy or renewable system. Optimization of hybrid renewable energy systems looks into the process of selecting the best components and its sizing... more
    Figure 5. Monthly Load Profile of JEC Jabalpur
    Figure 7. Cost Curve of Wind Turbine (NW/100) Those factors will determine the maximum level of radiation. Other factors, such as eight above sea level, water vapour or pollutants in the atmosphere, and cloud cover, decrease the radiation level below the maximum possible. Solar radiation does not experience the same type of turbulence that wind does, but there can be variations over the short term. Most often, these are related to the passage of clouds. The initial installation cost of photovoltaic arrays are taken approximate as $5000 and $3000, respectively. The lifetime of the PV arrays are taken as 25 years and no tracking system is included in the PV [5]. approximate as $5000 and $3000, respectively.
    Figure 6.Cost Curve of Solar PV System Solar energy is the most promising of the renewable energy sources in view of its apparen unlimited potential. The sun radiates its energy a the rate of about 3.8 x 1023 kW per second. Mos of this energy is transmitted radially as electromagnetic radiation which comes to abou 1.5kW/m’ at the boundary of the atmosphere. After traversing the atmosphere, a square metre of the earth's surface can receive as much as 1kW of solar power, averaging to about 0.5 over all hours of daylight. The solar radiation resource is fundamentally determined by the location on the earth’s surface, the date, and the time of day.
    The high capital cost has been a major barrier to widespread use by utilities. However, the costs of both PV and wind systems have decreased substantially in recent years and further reductions over the next 10 years are expected. Renewable energy systems such as PV, solar therma electricity such as Dish-Stirling systems, and WT are appropriate solar and wind technologies tha can be considered for electric power generation a the distribution system level. Other renewable energy technologies, such as the solar centra receiver, hydro-electric generation, geothermal, eceiver, hydro-electric generation, geothermal, A monthly average wind dataset for Jabalpur ware collected from Environment Jabalpur climate. This is an average of last ten year and indicates that annual average wind speed and shown in Figure 2. From the above given data, wind speed probability function and average hourly wind speed throughout the year is shown in Figure 2. The autocorrelation factor (randomness in wind speed) is found to be 0.85. Average wind speed in the summer season is slightly higher than the winter season as shown in Figure 4. Wind turbines generate electric power by extracting kinetic energy from the wind to drive a turbine connected to an ac generator. The power in wind available to a turbine is proportional to the cube of the wind
    Solar Resource web site. The annual average solar radiation fot this area is 5.43 kWh/m’/d. Figure 1 shows the solar radiation profile over a one-year period. Solar and wind energy systems are characterizec by relatively high capital costs, low operation anc maintenance (O&M) costs, and zero fuel costs. (Source: http://eosweb.larc.nasa.gov/sse) web site. The annual average solar radiation for Table 1- Solar and Wind data of JEC Jabalpur The solar resource was used for a site in Jabalpur Engineering College, Jabalpur at a location of 23°.10’N latitude and 79°.59’E longitude. Solar radiation data for this region was obtained from the NASA Surface Meteorology and Solar Energy The solar resource was used for a site in Jabalpur
    and large wind farms, are normally connected to the grid at the sub transmission or transmission level because of the higher power capacities of these systems. Presently, PV and WT technologies appear to be the most viable candidates for integration into power distribution systems. Sunlight is converted directly to direct current (dc) electricity by PV cells made of semiconductor materials such as silicon. The cells are wired in series-parallel combinations to form modules or panels. The dc is converted to alternating current (ac) by an inverter or power conditioning unit. speed. Thus, wind speed variations will cause variations in the wind turbine power output.
    performing the sensitivity analysis over a large Klectrical Load The load profile is based on a_ hypothetical Jabalpur Engineering College, Jabalpur (MP). Figure 1 illustrates this profile. A small base load of 15 KW occurs throughout the day and night. Small peaks of 30 KW occur in the morning and at noon, while the majority of the load occurs in the evening. This evening load, with a peak load of 60 KW, would likely include compact fluorescent lighting and a radio. The total daily load averages 2 Mega watt-hours per day. HOMER can perform a sensitivity analysis by accepting multiple values for a particular input variable such as the average load. This analysis determines how changes in the input variable affect the performance of the system and the relative ranking of different systems. By range of load sizes, the study simulates a range of load types from a single home to a large
    Figure 9. Optimization Results of Hybrid Energy System for Jabalpur Engineering College, Jabalpur
    Option II - System architecture
    Option III - System architecture
    Figure 8. Block Diagram of a Conventional PV—Wind-_Diesel-Battery System For the purpose of energy storage, lead acid batteries are included in this analysis for economic considerations. The use of hydrogen as another possible storage alternative is not currently economically viable given the prohibitive costs of electrolyzers and fuel cells and the low efficiency resulting from the electricity-hydrogen—electricity conversion cycle. In this case, commercially available battery models, such as Surrette 6CS25P model (6 V, 1156 Ah, 9645 kWh), were used in the scheme. Cost of one battery was considered The efficiencies of the inverter and rectifier were assumed to be 90% and 85% respectively for all sizes Considered. The simulations were done for each system switching the power between the inverter and the generator. Both devices were not allowed to operate in parallel. Initial and replacement cost for the converter is 700$, with no cost for operation and maintenance.
    descriptionView Paperarrow_downwardDownload
    A proposal for a wind system equipped with a doubly fed induction generator using the Conservative Power Theory for active filtering of harmonics currents
    by Ramon Souza

    2023, Electric Power Systems Research

    This study describes a strategy that allows the addition of active filtering functionality to a doubly fed induction generator (DFIG) wind power system. The active filtering is performed through an algorithm that uses the mathematical... more
    Fig. 21. Electric current in the grid without active power from the wind system and with active filtering (a) and frequency spectrum (b)
    Fig. 18. Electric current in the grid before the injection of active power by the wind system and without active filtering (a) and harmonic spectrum (b). Fig. 17. Reactive (a) and active power produced by DFIG (b).
    Fig. 19. Electric current in the grid with active power supply by the wind system and without active filtering (a) and harmonic spectrum (b).
    Parameters of non-linear load.
    Fig. 16. Voltage in the direct current link (a) and electrical current in the DFIG stator (b).
    Fig. 7. Voided current measurement (a) and control loop GSC with the active filtering function (b). Thus, we add the direct and quadrature axis components of the residual current with the dq components of the active and reactive power control of the GSC, in order to generate the references for the control system in synchronous coordinates. According to the results that will be presented in the simulation results section this operation allows to add the active filtering function to the wind power system with DFIG. P, is the active phase power; W,, is the phase reactive power. It is shown by the Tellegen theorem that Relations (35)-(38) re- present conservative quantities for any systems, independent of voltage
    The concept of expanded plant consists of calculating an equivalent transfer function that allows designing the controller gains given by Gep(s): Fig. 10. Block diagrams illustrating the current control loop of the L filter.
    DC link filter and capacitor parameters.
    Fig. 8. Block diagrams illustrating the DFIG rotor current control loops.
    Fig. 11. Open-loop frequency response of the filter L plant cascaded with the designed PI controller.
    Fig. 9. Bode diagram for the rotor current control loop.
    Parameters of the wind power system studied.
    Fig. 4. Schematic diagram of the dq current control loop for the GSC.
    Fig. 12. Block diagram of DC link voltage control: complete (a) expanded plant (b). The wind system shown in Fig. 1 was simulated using components
    Fig. 14. Electrical current measured in the set of loads (a) harmonic spectrum (b). The electric current shown in Fig. 18-a is distorted, because of the electric current drained by the set of non-linear loads. The THD of the electric current is 16.85% and the grid current harmonic spectrum contains components of 5th, 7th, 11th and 13th orders, in Fig. 18-b.
    Fig. 13. Open-loop frequency response of the cascade-augmented plant with the designed PI controller.
    Fig. 2. DFIG current control loop (a) and electric current transform diagram of the rotor from abc to dq and rotational velocity used for decoupling and feed-forward compensation (b).
    Fig. 1. Schematic diagram proposed for a wind system with active filtering function.
    Fig. 20. Electric current in the grid with active power supply by the wind system and with active filtering (a) and frequency spectrum (b).
    Collective instantaneous active power and collective instantaneous reactive energy:
    Fig. 3. Schematic diagram for the stator magnetic flux estimator. Given a generic polyphase circuit of m conductors with or without return conductor under periodic operation condition. The CPT authors define detailed mathematical formulation of CPT that can be found in Refs. [16] and [30]. However, the concepts necessary for the applica- tion of CPT in this study are defined [35].
    Fig. 15. Void current calculated in CPT.
    descriptionView Paperarrow_downwardDownload
    Estimation of Energy Storage and Its Feasibility Analysis
    by MOHAMMAD ARIF

    2023, Energy Storage - Technologies and Applications

    descriptionView Paperarrow_downwardDownload
    ENGI 990A EJECE formmated publication
    by M. Tariq Iqbal

    2023

    St. Lewis is one of the isolated communities situated in Newfoundland and Labrador (NL). The easternmost province of Canada includes over seven thousand small islands with scattered populations. It brings various challenges to the... more
    descriptionView Paperarrow_downwardDownload
    Harmonic Active Filtering and Impedance-based Stability Analysis in Offshore Wind Power Plants
    by Guangya Yang

    2023

    Nowadays, to eliminate harmonics injected by the wind turbines in offshore wind power plants there is a need to install passive filters. Moreover, the passive filters are not adaptive to harmonic profile changes due to topology changes,... more
    Figure 12 Impedance profile, seen at the point of connection, which varies depending on array cable length.
    Figure 5 Bode plot of the open-loop transfer function.
    ‘igure 15 Fourier decomposition of the normalized voltage at the point 0 connection as well as command signal in the controller (Case 3, cable length: 24km).
    Figure 8 Voltage waveform measured at the point of connection.
    Firstly, the proposed AF was tested/simulated on a system presented in Figure 11. For example, 31.5km MV cable capacitance is causing a resonance at the POC close to the 11" harmonic (see to Figure 12). Thus, the NF resonance frequency is set to eliminate the 11" harmonic. It can be seen based on the open-loop Bode magnitude plot that the controller can selectively attenuate the 11" harmonic component (see Figure 18). It is important to note that the NF response shown in the Bode plot represents the ideal condition (- 100dB). In reality, a limited attenuation can be caused by the nonidealities, e.g. digital implementation, sampling resolution. Figure 19 Harmonic spectrum of the point of connection voltage with and without active filtering (Case 1, cable length: 50km).
    Figure 18 Bode plot of wind turbine converter control including the notch filter. Where, @, is the resonance frequency and ¢ defines the attenuation of the NF. The open-loop control system with the NF is shown below in Figure 18.
    Figure 17 Butterworth notch filter placement in the wind turbine control chain. Figure 16 Lowpass Butterworth filter placement in the wind turbine control chain.
    Figure 7 Current waveform showing wind turbine connection measured at the point of connection.
    Figure 22 Nyquist diagram of the open-loop control with the notch filter. From the Bode plot (Figure 18) and the Nyquist diagram (Figure 22) the stability of the system with the NF is verified with GM of 6 dB and PM of 70°. Therefore, the robust AF solution is provided which effectively reduces the harmonic distortion at the POC. It is worth to emphasize that no additional passive filters across the WPP electrical infrastructure are needed. In practice, the AF can be digitally implemented after the A/D signal conversion and before the modulation block inside the WT converter control loop.
    The simplified diagram of the investigated offshore WPP is presented in Figure 1. It comprises four identical Type-4 WTs. The WTs are connected to the offshore MV array cable system via a step-up transformer. The POC is defined at the LV side of the offshore transformer (i.e. MV _ level). Furthermore, the WPP is connected to the onshore grid at the point of common coupling (PCC) via long HVAC cable. The WT model is simplified for the harmonic simulation purposes, i.e. the DC link circuit is represented as a constant DC voltage source. The modelled part of the WPP is marked with red dotted line in Figure 1. Figure 1 Simplified diagram of the investigated offshore wind power plant. A. System Parameters
    In Figure 9, the active power level is rising in steps from 6MW to 24MW as the number of connected WTs is increasing. Whereas, the reactive power level is kept always at zero MVAr during the whole simulation to maintain unity power factor at the POC. Figure 7 shows increasing current injection to the POC. The first few waveform cycles are neglected to omit the transient behaviour at the beginning of the simulation. Figure 9 Active and reactive power changes at the point of connection while new wind turbines are connected.
    Based on the results it can be seen that the implementation of LFs and NFs in WT control chain is reducing the harmonic distortion level in the WPP offshore electrical system. Table 4 Active filtering effectiveness at the point of connection (wind power plant study).
    Figure 3 Current controller in the synchronous reference frame. To control the output power, active power reference (Pret) and reactive power reference (Qrer) can be rewritten from Equation (2) with setting the V, signal to 0 in the phase locked loop (PLL).
    Figure 4 Wind turbine grid-tied converter overall control strategy. The entire WT control scheme (including control blocks, modulation block, passive components, converter, measurements) is shown in Figure 4.
    Figure 21 Harmonic spectrum of the point of connection voltage with and without active filtering (Case 3, cable length: 24km). After inserting the filters (i.e. LF and NF) in the control chain, the simulations were run again. From Figure 19, Figure 20, and Figure 21 it can be seen that NF is capable of effectively and selectively reducing the harmonic content at the POC. In all three cases, the NF is adjusted and successfully eliminating the major harmonic components as presented in Table 3.
    Figure 13 Fourier decomposition of the normalized voltage at the point of connection as well as command signal in the controller (Case 1, cable length: 50km).
    Figure 2 Bode plot of the wind turbine LCL passive filter. It can be seen from the Bode plot in Figure 2 that the resonance frequency is at 6.68x10? rad/second. Additional damping resistance allows to reduce the resonance gain. Moreover, the attenuation of the LCL-filter is seen to become gradually higher for higher order harmonics.
    Figure 20 Harmonic spectrum of the point of connection voltage with and without active filtering (Case2, cable length: 31.5km).
    From the open-loop and closed-loop frequency responses the following characteristics of the current controller are [8], [9] (i) the open loop gain margin (GM) > 6dB, (ii) phase margin (PM) >45°, (iii) closed-loop bandwidth > 0:25 kHz.