AwaSys can generate regular, irregular and solitary waves of high quality.
Regular waves can be generated using linear theory, complete second order theory, cnoidal wave theory with shallow water wave generation or ad-hoc unified generation based on stream function wave theory input. The linear theory results in spurious free higher harmonics to be generated. The result is a wave without constant form which is very visible for shallow water waves (h/L small) and/or for steep waves (H/L large). The second order theory eliminates the spurious second harmonic and can be a very large improvement over 1st order theory. The approximate stream function theory is a method to eliminate super harmonics of higher order. That theory is recommended for generation of regular waves in intermediate or shallow water when using piston mode generators. The method based on cnoidal wave theory is useful in highly nonlinear waves in shallow water where convergence of the stream function is difficult to achieve.
Solitary waves, single waves, N-waves as well as user defined long wave profiles can be generated for simulation of tsunami waves with or without leading through.
AwaSys comes with 13 predefined irregular wave spectra (Pierson-Moskowitz (ITTC-81), Bretschneider (ISSC), Bretschneider-Mitsuyasu, Modified Bretschneider-Mitsuyasu, JONSWAP, Generalised Γ, TMA, FRF, Top-Hat, Gaussian Swell, Torsethaugen, Ochi-Hubble 1P, Ochi-Hubble 2P). On request, more predefined spectra can be added. Additionally, user-defined spectra and surface elevation time series can be generated. Moreover, composite sea states consisting of two or more of above waves may be generated (for example sum of two irregular wave spectra to generate a combined swell and wind sea). Also focused irregular waves (freak waves) can be generated based on all of above spectra and also based on Gaussian Wave Packets.
The following wave synthesis methods for irregular waves are implemented:
- The Random Phase Method
- The Random Complex Spectrum Method
- The White Noise Filtering Method
- Ad-hoc unified generation based on Boussinesq numerical model input
The Random Phase Method is a deterministic method simulating random waves in the frequency domain by assigning random phases to each frequency component. Subsequent use of the InvFFT-algorithm provides the time domain representation of the wave train. The method will in theory provide a spectrum 100% in agreement with the target after one FFT block length.
The White Noise Filtering Method is a non-deterministic method, which simulates random waves in the time domain by means of digital filtering of white noise. Filters for each paddle are generated in accordance with the specified directional energy spectrum. By using the white noise method the natural variability of the wave spectrum is simulated which convergence against the prescribed spectrum for increasing sample duration. The white noise filtering method provides an energy spectrum with no discontinuities. This is important to reproduce natural wave grouping and could also be important for the response of structures with tuned eigen frequencies which might be missed in the discrete spectrum generated by InvFFT methods. Moreover, white noise generation can simulate long time series without repetition of signal. The steering signal is generated in real time and the preparation time for long time series is thus significantly reduced compared to the InvFFT based methods.
Nonlinear interaction between the individual wave components in the wave trains give rise to the so-called bounded sub- and superharmonics which are of second order. In both wave synthesis techniques, a correct reproduction of the sub harmonics (group bounded long waves) is possible. Moreover, the random phase method includes correct reproduction of the bound superharmonics. The 2nd order generation is based on the complete method derived by Schäffer and Steenberg (2003) which has been derived without any narrow spectrum assumptions and includes corrections for evanescent mode interactions.
The ad-hoc unified generation makes it possible to generate waves with higher nonlinearities than the 2nd order theory. The input for the method is calculated by a Boussinesq wave model that is delivered with AwaSys. The numerical model is thus started automatically by AwaSys and is invisible to the user.The numerical model is GPU accelerated to reduce test preparation time. The method leads to similar improvements for generation of nonlinear irregular waves as the approximative stream function wavemaker method does for nonlinear regular waves.
To avoid damage to the wave maker due to excessive velocities, accelerations and strokes is applied in real time a special soft clipping algorithm to the control signals. Warnings are given to the user in case clipping has occurred during a test.