Adaptive Filter Step Size Adjustment for Harmonic Noise Cancellation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing engine harmonic cancellation systems face instability and noise artifacts when attempting to cancel noise from rotating devices with frequencies that are too close, leading to compromised stability margins and potential divergence of filter algorithms.

Innovation Solution

The system employs an overlap detector to assess the proximity of frequencies and adjust the operation of adaptive filters by modifying parameters such as adaptation step sizes, ensuring stability and effectiveness in noise reduction by decreasing step sizes when frequencies are close, thereby maintaining performance similar to that of a single canceller.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If two adaptive filters are used to cancel noise from both engine and propeller shaft, then the noise cancellation coverage is improved, but the stability margin is compromised when frequencies are close

Engineering Contradiction:
Improvenoise cancellation coverageVSAvoidfilter stability margin
Core Design Contradiction:
Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The system dynamically adjusts the adaptation step sizes of adaptive filters based on the proximity of frequencies being cancelled. When frequencies are close, the system decreases the step sizes to maintain stability, and when frequencies are well-separated, it uses larger step sizes for faster convergence. This dynamic parameter adjustment resolves the contradiction between handling multiple noise sources and maintaining filter stability.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the adaptation step size parameter of the adaptive filters based on the frequency proximity condition. By monitoring the relationship between engine and propeller shaft frequencies, the system modifies the learning rate parameter to prevent divergence when frequencies are close, thus maintaining stability while still providing multi-source noise cancellation.

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If adaptive filters operate with close frequencies, then multi-source noise cancellation is achieved, but filter algorithm divergence occurs leading to noise artifacts

Engineering Contradiction:
Improvemulti-source noise cancellationVSAvoidfilter algorithm stability
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The system implements a feedback mechanism where the frequency proximity between engine and propeller shaft is continuously monitored. Based on this feedback, the adaptation step sizes are adjusted in real-time to prevent filter divergence. This closed-loop control ensures reliable operation when cancelling multiple noise sources with potentially close frequencies.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The adaptation step sizes are made dynamic rather than fixed, allowing the system to respond to changing frequency conditions. When the propeller shaft frequency approaches engine harmonics, the system automatically reduces the step sizes to maintain algorithm stability, preventing the generation of noise artifacts while still achieving effective cancellation.

Inventive Principle:
Principle #15Dynamics

3Speed

If adaptation step sizes are increased for faster convergence, then noise cancellation speed is improved, but system stability is reduced when frequencies are close

Engineering Contradiction:
Improveconvergence speedVSAvoidsystem stability
Core Design Contradiction:
SpeedVSStability of the object's composition

Solution Approach 1:

The adaptation step sizes are dynamically adjusted based on the frequency proximity condition rather than being fixed. The system uses larger step sizes when frequencies are well-separated to achieve fast convergence, and automatically reduces them when frequencies approach each other to maintain stability. This dynamic adjustment resolves the contradiction between convergence speed and stability.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the adaptation step size parameter based on the operational condition (frequency proximity). By monitoring the relationship between input frequencies and adjusting the learning rate accordingly, the system achieves both fast convergence when possible and stable operation when frequencies are close, eliminating the need to choose between speed and stability.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2939233B1Active reduction of harmonic noise from multiple rotating devices
Publication Date: 2016.05.25 BOSE CORP
  • EP2939233B1 patent drawingFigure 1
  • EP2939233B1 patent drawingFigure 2

AI summary

A system and method for reducing noise caused by two or more rotating devices by taking in input signals with frequencies that are related to the rotation rates of the rotating devices, and causing one or more loudspeakers to produce sounds that are at about the same frequencies as the noise and of substantially opposite phase. There is a noise canceller associated with each rotating device. Each noise canceller includes a harmonic frequency computer that computes and provides a harmonic frequency to a harmonic sine wave generator that generates an output sine wave. Each noise canceller also has an adaptive filter to create a noise reduction signal that is used to drive one or more transducers to reduce noise caused by the rotating devices. There is an overlap detector that compares the harmonic frequencies and, based on their proximity, alters the operation of one or more adaptive filters.