The 3D accelerometer data is proces

The 3D accelerometer data is processed in the front end PCB through a 256-pole low pass FIR filter before it is decimated to 4Hz giving a flat response in the wave processing band.The double integration is performed in the frequency domain, even for the time series output, giving a 3D displacement output used for the wave calculations.A significant amount of development has been invested in optimizing the sensor performances to provide reliable and accurate data when used on any kind of buoy.Mechanical anti-vibration filterWind, waves and current may induce vibrations on the buoy structure.Those vibrations may cause an increased noise being folded back into the wave processing band due to a violation of the Nyquist criteria.To avoid this, a mechanical low pass filter is implemented and has been tuned and tested in the laboratory using a shaker plate(Figure 2).Sweep 95-105Hz DampenedSweep input 95-105HzFigure 2:Effect of the mechanical filter.The blue curve represents the noise induced by the shaker plate.The black curve represents the noise while using the mechanical filter and the red curve represents the ambient noise without any shaker plate excitation.Off-center correctionIt is often impossible or unpractical to install the wave sensor in the buoy’s center of gravity.When the sensor is not positioned in the buoy center, this will result in a displacement error.The MOTUS Wave Buoys use the sensor’s orientation combined with the installation position(x, y, z)relative to the rotational origin and is able to offset compensate by subtracting the added displacement.As a result, the sensor can be installed in a convenient place without sacrificing on the accuracy(Figure 3).Transfer function compensationThe size of the buoy and the mooring design may modify the buoy wave response and the data quality.A user configurable transfer function can be defined to compensate for the buoy behavior and provide maximum system flexibility without sacrificing wave measurement accuracy.Frequency leakage compensationIt is important to reduce frequency leakage coming from the Fast Fourier Transform when implementing an accelerometer-based wave sensor.Even though the frequency leakage is equal in the high and low end, the effect of the leakage will be biased due to the frequency dependent factor 1/(f2)used in the conversion from the acceleration to displacement.In order to keep this leakage minimal, a window function is implemented and reduces this effect.The remaining frequency dependent error is estimated and compensated for by the sensor(Figure 4).