FYP disaster ahead

It's 3am and I'm still up trying to figure out my project. In retrospect, I think I may have made some bad decisions but unfortunately, the benefits of hindsight always come too late.

Initially I had hoped to do modeling of acoustic emissions on board of a jackup engine room using data from my previous intern company at Keppel FELS. At the time it seemed like a good idea but when I got back to Melaka, I was worried about how I would approach the problem as noise and acoustics was never taught in my syllabus. Even though I did some reading in the holidays regarding the problem, I wasn't confident that the project would be doable within the timeframe given.

Thus I switched to design of an offshore floating wind structure. At the time it was already almost deadline for submission of final year project title so I didn't really understand the whole contract, as it were, before signing on the dotted line. This single most important decision was made in haste and with the title secured, the next task of gathering information on the problem came next.

I settled on using ANSYS as the software to perform my analysis. Specifically, the ANSYS website gave a glowing review of their ASAS offshore module which seemed to do everything that I required so I thought the whole thing was just a matter of putting together the model and running it through the CAE platform.

Wrong approach. Turns out ANSYS ASAS has practically ZERO available information on how to actually use the module. No supplied tutorials (although ANSYS bundled extensive tutorials for their CFD and Structural modules) and zero guide online on how to use ASAS effectively killed the software. Using it was a nightmare, because when you started the ASAS solver, all there is was just a window prompting you to input your data file. I tried converting a model from ANSYS APDL to ASAS (using a cryptic macro hidden deep within ANSYS's help system) but it didn't get me very far. However, ANSYS happily promotes their 3 day training course for using ASAS, and I suspect it will cost a pretty penny to attend.

I briefly flirted with the idea of setting up an analysis using pipe59 element as it came together with wave loading capabilities built into the element. However it was unsuitable for my design which was a suction caisson.

Next up I tried using ANSYS AQWA for modelling, and it's a more promising approach with the new AQWAWB interface very useful in setting up a 3D model. However, I ran into problems with trying to "cut the model at the waterline using a plane", as one of the requirements for importing a model to be analyzed. I have absolutely no idea what that line means, and I tried cutting the model but I end up important 2 separate bodies into the software, whereas it required 1 body with 2 portions. So that's another dead end.

The next step is to try ANSYS V12.1, with a more complete integration of AQWA into the Workbench, supposedly this will alleviate the problem of passing models into AQWA properly. I suspect however, that this will be far from easy, and even after that, AQWAWB is only capable of doing a surface diffraction analysis, the output of which would be a time-domain pressure reading which has to then be imported into Mechanical-Structural so that I can perform the analysis on the force impact on the structure.

The irony of it all is that my original idea of doing an acoustic analysis turns out to be as easy as setting up Acoustic modeling as the option in the Workbench then specifying the domain to be meshed and finally inputting the noise source parameters. Of course, interpretation of the data would be harder, but at least that's 75% of the project done right there.

The main problem faced by the project is how to accurately perform wave loading on my model. ASAS is capable of both PM and JONSWAP wave loads, but to replicate the loads in standard ANSYS is damn near impossible (even with pipe59), at least with my understanding, while AQWA and ASAS might as well be a black box due to the lack of any proper documentation on how to use the software. I suspect both of them rely heavily on using hand-massaged text files as input, and I salute the engineers dealing with these two softwares. It reminds me of the days when I use to create websites by hand using nothing more than Textpad, imagination, and hitting lots of Ctrl+R keys.