Multi-Converter Pulse Pattern Control for Harmonic and Ripple Currents

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In high power converter systems, traditional control methods struggle with harmonic distortion and transient response due to high switching losses and the need for high pulse numbers, which are not feasible in medium-voltage systems, and require additional filters and constraints that diminish harmonic performance.

Innovation Solution

A method for controlling multi-converter systems using model predictive pulse pattern control, which determines optimal switching times and transitions to manage both fundamental and ripple currents without separate filtering, enabling fast and reliable control with reduced harmonic distortion and improved transient response.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If carrier-based pulse-width modulation is used to meet harmonic requirements, then harmonic distortion is reduced, but switching losses increase due to high pulse numbers

Engineering Contradiction:
Improveharmonic distortionVSAvoidswitching losses
Core Design Contradiction:
Object-affected harmful factorsVSLoss of energy

Solution Approach 1:

Optimized pulse patterns (OPPs) are pre-calculated offline to achieve optimal harmonic distortions for a given pulse number. This preliminary optimization allows the system to operate at low switching frequencies while meeting harmonic standards, eliminating the need for high pulse numbers and reducing switching losses.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The invention changes the control approach from continuous carrier-based modulation to discrete optimized pulse patterns with specific switching sequences. This parameter change enables operation at low switching frequencies while maintaining harmonic performance through pre-optimized switching angles and patterns.

Inventive Principle:
Principle #35Parameter changes

2Loss of energy

If optimized pulse patterns are used to reduce switching frequency, then switching losses are reduced, but controller design becomes complicated due to irregular sampling

Engineering Contradiction:
Improveswitching lossesVSAvoidcontroller design
Core Design Contradiction:
Loss of energyVSDevice complexity

Solution Approach 1:

The invention implements a dynamic controller that adapts to the irregular sampling instants caused by OPPs. The model predictive controller dynamically adjusts its operation based on the actual switching times and ripple characteristics, rather than relying on fixed sampling intervals. This dynamic adaptation enables fast transient response without requiring additional filters.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The controller uses feedback from actual current measurements and ripple observations to adjust control decisions in real-time. By incorporating feedback mechanisms that account for the irregular sampling nature of OPPs, the controller achieves fast transient response while maintaining simplicity and avoiding the need for additional filtering components.

Inventive Principle:
Principle #23Feedback

3Object-affected harmful factors

If linear controllers with additional filters are used to handle OPP ripple, then harmonic performance is maintained, but transient response becomes sluggish

Engineering Contradiction:
Improveharmonic distortionVSAvoidtransient response
Core Design Contradiction:
Object-affected harmful factorsVSSpeed

Solution Approach 1:

The invention extracts and separately handles the ripple component through model-based prediction rather than physical filtering. By taking out the ripple effect from the control loop and compensating for it through predictive algorithms, the system maintains harmonic performance without the need for additional filters that would slow down the transient response.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The invention replaces mechanical/physical filters with a computational/model-based approach. Instead of using physical filtering components that introduce delays, the system uses model predictive control algorithms to handle ripple effects computationally, achieving fast transient response while maintaining harmonic performance.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

4Ease of operation

If constraints are added to OPP optimization for smooth angles, then controller compatibility is improved, but harmonic performance diminishes

Engineering Contradiction:
Improvecontroller compatibilityVSAvoidharmonic distortion
Core Design Contradiction:
Ease of operationVSObject-affected harmful factors

Solution Approach 1:

Instead of constraining the OPP optimization to achieve smooth angles, the invention inverts the approach by allowing irregular angle patterns in the optimization and then adapting the controller to handle these patterns. This inversion maintains optimal harmonic performance while achieving controller compatibility through adaptive control algorithms.

Inventive Principle:
Principle #13The other way round (Inversion)

Data Source

PatentEP4304075A1Model predictive pulse pattern control for multi-converter system
Publication Date: 2024.01.10 ABB (SCHWEIZ) AG
  • EP4304075A1 patent drawingFigure 1
  • EP4304075A1 patent drawingFigure 2~3
  • EP4304075A1 patent drawingFigure 4~5

AI summary

A method for controlling an electrical converter system (10) with at least two electrical converters (28a, 28b) comprises: determining a pulse pattern (A1, U1; A2, U2) for each electrical converter (28a, 28b) from a reference converter voltage of the respective converter (28a, 28b), wherein each pulse pattern (A1, U1; A2, U2) comprising switching times (A1;A2) and switching transitions (U1; U2) for the respective electrical converter (28a, 28b); determining a sum current reference ( isum∗) and at least one difference current reference ( idiff∗) produced by the pulse pattern (A1,U1;A2,U2) and determining a measured sum current (isum) and at least one measured difference current (idiff) from measurements in the electrical converter system (10); determining a sum current error (isum,err) and at least one difference current error (idiff,err), wherein the sum current error (isum,err) is determined by subtracting the measured sum current (isum) from the sum current reference ( isum∗), and a difference current error (idiff,err) is determined by subtracting the respective measured difference current (idiff) from the difference current reference ( idiff∗); determining a sum flux modification (Δψsum) by multiplying a gain (L̃t) to the sum current error (isum,err) and at least one difference flux modification (Δψdiff) by multiplying a gain (Ls) to the respective difference current error (idiff,err); mapping the sum flux modification (Δψsum) and the at least one difference flux modification (Δψdiff) to a converter flux modification (Δψ1,abc, Δψ2,abc) for each electrical converter (28a, 28b); modifying the pulse pattern (A1, U1; A2, U2) for each electrical converter (28a, 28b) by moving the switching times, such that the converter flux modification (Δψ1,abc, Δψ2,abc) is compensated by the modified pulse pattern (u1,abc, u2,abc) with the moved switching times; and applying the modified pulse patterns (u1,abc, u2,abc) to the electrical converters (28a, 28b).