Wednesday, August 25, 2004
Residual resistance tables
I have been working the issue of programmatically calculating power curves for ships for slightly more than two years. My previous approach had enough problems, that I set the problem down for about a year. I have picked it back up, and have taken a new tact. Instead of using sparse matrices, I am using complete matrices with all values populated. As The Speed and Power of Ships has graphs that do not cover the space for all ships, I have assumed that they can be extended. The underlying assumption is that residual resistance is a continuous function and has no discontinuities. Given that, it is possible to extend the graphs, by plotting resistance, given the prismatic coefficient and the speed-length ratio, in one direction. The other necessity is to extend the graphs for greater displacement-length ratios, so that we can calculate power for 1860's ironclads. That requires displacement length ratios up to 420. The higher speed-length ratios are needed for fast destroyers that are fairly short. I am only computing for speed-length ratios up to 2.5, but I expect to have to go above 3.0, eventually. I had done enough experiments that it seemed to be possible to get reasonable approximate figures, so I am going forward with the project. A useful aid has been the ability to look at data with line charts, using Excel spreadsheets. That speeds up the process, so I don't have to use paper and end up having to estimate numbers from pencil-drawn graphs.