Real-Time Wave Force Estimation Using Unscented Kalman Filter
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating the excitation force of waves on wave energy systems are inadequate, particularly for non-linear systems, as they either provide delayed estimates or are computationally heavy, and existing solutions do not accurately determine wave force across various frequency ranges in real-time.
Innovation Solution
A method using a fragrance-free Kalman filter that incorporates models of radiation, drag, and nonlinear dynamics to estimate the excitation force based on kinematic measurements and force applied by the conversion machine, suitable for all wave energy systems, including non-linear ones, allowing real-time determination and control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a bank of independent harmonic oscillators and a Luenberger observer is used to estimate wave force, then the estimation can be performed, but the estimates are significantly delayed (out of phase) relative to the true excitation force
Solution Approach 1:
The patent changes the fundamental parameters of the estimation approach by using a state-space model with state variables representing wave force and its derivative, rather than relying on harmonic oscillator frequencies. This parameter transformation enables real-time estimation without phase delay while maintaining accuracy across varying wave conditions
Solution Approach 2:
The patent implements a dynamic state-space model that adapts to changing wave conditions in real-time, rather than using fixed-frequency harmonic oscillators. The state variables and transition matrices allow the system to dynamically track the true wave force without phase lag, even as wave spectrum varies over time
2Adaptability or versatility
If a very large number of frequencies are considered to work under realistic varying wave spectrum conditions, then the method can be effective, but the computation becomes very demanding
Solution Approach 1:
The patent transforms the problem from analyzing many frequency components to tracking a small number of state variables (wave force and its derivative) through a state-space model. This parameter reduction maintains adaptability to varying wave spectra while dramatically reducing computational complexity from O(n) frequencies to O(1) state variables
Solution Approach 2:
The patent extracts only the essential dynamic characteristics of wave force into state variables, separating the critical information (wave force magnitude and rate of change) from the redundant frequency spectrum details. This extraction enables real-time processing while capturing the essential behavior across all wave frequencies
3Speed
If an extended Kalman filter is used to reconstruct the spectrum considering wave excitation force as a single sinusoid, then real-time estimation is achieved, but the method can only be effective for waves within a very narrow frequency band
Solution Approach 1:
The patent creates a universal state-space model that functions across all wave frequency ranges by using state variables that capture the essential dynamics without frequency-specific assumptions. The model adapts to any wave spectrum through its transition matrices, making it multi-functional for both narrow and broad frequency ranges simultaneously
Solution Approach 2:
The patent uses a dynamic state-space representation that automatically adapts to changing wave frequencies through its time-varying transition matrices, rather than being constrained to a fixed narrow band. This dynamic approach maintains real-time performance while covering the full operational frequency range
4Measurement precision
If pressure measurements from sensors distributed over the entire surface are integrated to calculate forces, then all forces applied by the swell can be calculated, but this is an expensive and unrobust solution
Solution Approach 1:
The patent extracts the wave force information from the complex pressure field by using a state-space model that processes only a few key measurements (position, velocity, acceleration of the floating structure). This extraction eliminates the need for expensive distributed pressure sensor arrays while maintaining accurate force calculation
Solution Approach 2:
The patent introduces a state-space model as an intermediary between the measurable kinematic parameters (position, velocity, acceleration) and the unmeasurable wave force. This intermediary model enables accurate force estimation without direct pressure measurements, avoiding the complexity and cost of distributed sensor arrays
Data Source
AI summary
The present invention consists of a method for establishing in real time the forces applied by the swell incident on a movable means of a wave energy system. The method is based on the construction of a model of the radiation force which is applied to the movable means, a model of the drag force which is applied to the movable means and a non-linear model of the dynamics of the wave energy system. The invention uses only measurements of the kinematics of the float (position, speed and, where applicable, acceleration) and the force applied by the conversion machine, which measurements are normally available on a wave energy system because they are used for controlling and monitoring it. These models, these measurements and an unscented Kalman filter are used to establish the excitation force applied by the swell incident on the movable means.