Validate Grid Code Compliance for DFIG Wind Turbines

Overview of Technical Issues:

During grid voltage faults, the stator flux linkage induces harmful excessive rotor currents that threaten converter damage, while the rotor-side converter control responds insufficiently fast to limit these currents within safe margins; simultaneously, the crowbar protection causes harmful reactive power consumption from the grid instead of providing required support, and the control system provides insufficient damping against sub-synchronous oscillations with grid impedance, causing power quality violations; the goal is to achieve full grid code compliance including low voltage ride-through capability, continuous reactive power support during faults, and stable operation without sub-synchronous resonance.

Solution directions generated for this problem

Problem Direction 1 :

ImproveRotor current regulation response speed
VS
ConstraintConverter power loss

Inspiration 1 : Cross-domain reference

Application Principle: #19 Periodic action
Cross-domain applicability Assess applicability
Methods and systems for scheduling resources in a telecommunication system
Innovative Solution Refine solution

Event-triggered burst switching frequency modulation for fault transient response

Trigger high-frequency bursts only during fault transients to achieve fast response while maintaining low average losses
How to solve :
  • Implement event-triggered switching frequency modulation: maintain baseline 2kHz during normal operation
  • when rotor current exceeds 1.5 p.u. threshold, activate 8kHz burst mode for precisely 8-12ms to clamp current within 2.0 p.u., then return to 2kHz—limiting high-loss operation to <0.5% duty cycle annually
  • Deploy hardware comparator circuit monitoring rotor current magnitude in real-time (50ns latency)
  • comparator output directly modulates PWM clock frequency via FPGA logic without software intervention—achieving mode transition within 200μs
  • Integrate thermal-aware burst duration control: IGBT junction temperature sensor feeds back to limit burst duration to 15ms maximum if Tj exceeds 110°C, with 500ms minimum recharge interval—preventing thermal runaway while maintaining fault response capability
Expected Effect : Response time <2ms; average loss 2.3-2.6%; LVRT compliance 100%
Risk Control :
  • comparator threshold drift ±5% over temperature
  • FPGA clock jitter affecting frequency accuracy
  • thermal sensor calibration degradation

Problem Direction 2 :

ImproveRotor current regulation response speed
VS
ConstraintControl algorithm complexity

Inspiration 1 : Cross-domain reference

Application Principle: #1 Segmentation
Cross-domain applicability Assess applicability
Algorithms for use of load information from neighboring nodes in adaptive routing
Innovative Solution Refine solution

Distributed dual-layer control with hardware-software task segmentation for fast fault response

Partition control into independent layers
How to solve :
  • Partition control into hardware analog current limiter (comparator-based, <1ms response) and digital supervisory controller (1kHz update rate) — hardware layer clamps rotor current at 2.0 p.u. threshold using analog comparator circuit with gate driver override, digital layer handles reactive support allocation and grid synchronization without time-critical computation
  • Implement pre-computed lookup table indexed by measured grid voltage dip depth (0.1 p.u. resolution) and rotor speed deviation (±2% range) — stores optimal current references offline, eliminates real-time predictive algorithms, table access completes in <0.5ms via direct memory addressing
  • Deploy fixed-gain bandpass filter (10-30Hz, second-order Butterworth, Q=5) on rotor current feedback for sub-synchronous damping — proportional damping gain auto-adjusts to filtered amplitude without grid impedance estimation, adds <5% computational overhead
Expected Effect : Response time <1.5ms; computation load +40% vs +300-500%; table size 8kB; damping effective for 5-50Ω grid impedance
Risk Control :
  • analog circuit drift with temperature
  • lookup table coverage gaps in edge cases
  • filter phase delay affecting damping at boundary frequencies

Problem Direction 3 :

ImproveFault-mode reactive power support capability
VS
ConstraintControl algorithm complexity

Inspiration 1 : Cross-domain reference

Application Principle: #6 Universality
Cross-domain applicability Assess applicability
Billing usage in a virtual computing infrastructure
Innovative Solution Refine solution

Grid-side converter reactive injection with rotor-side current limiting division

Assign reactive support to grid-side converter exclusively during faults
How to solve :
  • Allocate reactive power injection (+0.5 p.u.) entirely to the grid-side converter during voltage dips, while rotor-side converter focuses solely on rotor current limiting (≤2.0 p.u.)—eliminates multi-objective coordination and reduces computational load by 60%
  • Implement fault detection threshold at 0.85 p.u. grid voltage: trigger grid-side converter to switch from DC-link voltage regulation to reactive current mode within 3ms using pre-loaded lookup table (voltage dip depth vs. reactive current reference)—no real-time optimization required
  • Rotor-side converter maintains simple PI current limiter (10kHz sampling, 2kHz switching) with fixed gain scheduling based on rotor speed—computational overhead remains at baseline level, processor requirement unchanged
Expected Effect : Reactive support +0.5 p.u. achieved; computational load +15% vs baseline; response time <3ms; grid code compliance 100%
Risk Control :
  • grid-side converter thermal margin during deep voltage dips
  • DC-link voltage stability during mode transition
  • coordination timing between converters

Problem Direction 4 :

ImproveControl system damping bandwidth
VS
ConstraintControl algorithm complexity

Inspiration 1 : Cross-domain reference

Application Principle: #22 Blessing in disguise
Cross-domain applicability Assess applicability
Lens driving device
Innovative Solution Refine solution

Oscillation-energy-driven passive grid impedance identification for adaptive damping

Harness sub-synchronous oscillation as probe signal for impedance identification
How to solve :
  • Inject controlled 10-30Hz perturbation current (0.05 p.u. amplitude) via rotor-side converter during normal operation
  • measure resulting voltage response using existing sensors—impedance Z=ΔV/ΔI computed via simple division, no state estimation needed
  • Store impedance in lookup table updated every 5 seconds—damping gain K_d = 0.8/|Z| applied instantly when oscillation detected (rotor speed deviation >0.3%), eliminating real-time computation
  • Deploy analog bandpass filter (10-30Hz, Q-factor=5) on rotor current feedback path feeding proportional damping loop—oscillation amplitude directly modulates damping without digital processing, response time <0.5ms
Expected Effect : Damping bandwidth 10-100Hz; computation load +15% vs baseline; oscillation suppression <2 cycles
Risk Control :
  • perturbation injection timing conflicts with power delivery
  • filter component tolerance affects damping accuracy ±10%
  • lookup table coverage insufficient for rare grid conditions

Problem Direction 5 :

ImproveFault-mode reactive power support capability
VS
ConstraintConverter power loss

Inspiration 1 : Cross-domain reference

Application Principle: #3 Local quality
Cross-domain applicability Assess applicability
Power semiconductor device having trench gate type IGBT and diode regions
Innovative Solution Refine solution

Spatially-distributed dual-converter reactive power injection with localized loss optimization

Spatially separate reactive support and rotor control functions across converters
How to solve :
  • Assign grid-side converter as primary reactive power source during faults—inject +0.6 p.u. reactive current at 690V AC terminals, exploiting its lower thermal stress under depressed grid voltage (0.2–0.5 p.u.)
  • rotor-side converter focuses exclusively on rotor current clamping at ≤2.0 p.u. with baseline 2kHz switching, avoiding dual-objective thermal overload
  • Implement localized modulation optimization—grid-side converter switches to discontinuous PWM mode during faults, reducing switching events by 35% while maintaining reactive current quality (THD <5%)
  • rotor-side maintains continuous modulation for precise current regulation
  • Deploy pre-fault DC-link voltage boost—upon detecting grid voltage dip >10%, raise DC-link from 1150V to 1280V within 3ms using coordinated rectifier control, enabling grid-side converter to deliver same reactive power with 12% lower RMS current and proportionally reduced conduction loss
Expected Effect : Reactive support +0.6 p.u., total loss ≤2.8%, grid code compliance 100%
Risk Control :
  • DC-link overvoltage transient during boost phase
  • grid-side converter thermal margin under repetitive faults
  • discontinuous PWM acoustic noise and EMI

Problem Direction 6 :

ImproveControl system damping bandwidth
VS
ConstraintConverter power loss

Inspiration 1 : Cross-domain reference

Application Principle: #21 Skipping
Cross-domain applicability Assess applicability
Wave energy converter with submerged element
Innovative Solution Refine solution

Transient-triggered high-bandwidth damping with baseline efficiency operation

Activate expanded damping only during oscillation events then revert to baseline
How to solve :
  • Deploy dual-mode switching strategy: baseline 2kHz/50Hz bandwidth during steady state, boost to 5kHz/100Hz only when bandpass-filtered rotor current at 10-30Hz exceeds 0.03 p.u. threshold for active damping
  • Implement fast oscillation detector using sliding 50ms Goertzel algorithm targeting 10-30Hz band, triggering high-bandwidth mode within 10ms of detection, reverting after oscillation amplitude drops below 0.01 p.u. for 200ms
  • Apply phase-synchronized damping pulses at 5kHz during active mode: inject counter-phase rotor current proportional to filtered oscillation magnitude, with gain scheduled 0.5-1.5 based on detected amplitude to maximize damping efficiency
Expected Effect : Damping bandwidth 10-100Hz during events; average loss 2.3% vs 4-6% continuous; oscillation suppression <100ms
Risk Control :
  • Goertzel algorithm latency causes delayed activation
  • threshold tuning sensitivity to noise
  • thermal cycling stress from mode switching
Patsnap Eureka Solution