http://www.k-wave.org/forum/topic/high-frequency-modelling-problem Support for the k-Wave MATLAB toolbox en-US Tue, 12 May 2026 22:32:13 +0000 http://bbpress.org/?v=1.0.2 q http://www.k-wave.org/forum/search.php http://www.k-wave.org/forum/topic/high-frequency-modelling-problem#post-3632 Mon, 13 May 2013 15:45:14 +0000 Bradley Treeby 3632@http://www.k-wave.org/forum/ <p>Hi Andrew,</p> <p>What is the size of your domain? Do you truncate it to slightly larger than the size of the transducer? </p> <p>A 2D simulation is certainly a possibility. However, when you perform simulations in 2D, remember you are inherently assuming that your sources, material boundaries, etc., are infinitely extended in the third-dimension (e.g., a 2D point source is actually a cylindrical line source in 3D).</p> <p>Using the C++ or GPU code should decrease the simulation time by around 6-8 times.</p> <p>Brad. </p> http://www.k-wave.org/forum/topic/high-frequency-modelling-problem#post-3619 Fri, 10 May 2013 02:01:39 +0000 ouelletn 3619@http://www.k-wave.org/forum/ <p>I am looking to model a rather unique system, it consists of 4 boundaries(water, copper, steel and liquid steel) totaling a ~3.5cm square and would like to find an approximation of the wave form generated by a single element generating a wave and all elements receiving.<br /> The properties of my transducer are, 24 element, 0.2mm pitch, 0.05mm spacing, 10MHz+/-10% central frequency and 4mm elevation.<br /> I am currently having issues with the size and time of 3d simulations at this frequecy, but was wondering if a 2d simulation would be possible by defining multiple rectangular transmitters/recievers and what errors this may cause.<br /> Also if so should it would it be more efficient on the graphics processor or through a C++ implementation(if uncertain I am willing to try both)<br /> Thanks, Andrew </p>