Delft Series Resistance Equations - Variety

Delft resistance equations

The dimensions of the Dehler 33 are:

Length water line 9.123 m
Width 2.412 m
Draft canoe body 0.41 m
Volume c. b. 3.785 m^3
The Dehler 33 is the type that was used for the "Dyna" sail dynamometer by Karsten Hochkirch

The Delft regression coefficients are from HISWA 1998, Keuning / Sonnenberg
and from Lorient 2008, Keuning / Katgert

The fact that the residual resistance can become negative at FN=0.1 appeared already with the very first Delft-model, the Sysser 1. For me it is a sign, that the Delft-friction line is not a realistic estimate of the frictional resistance of the hull. The true wave resistance is larger than the calculated residual resistance. The equations of 2008 avoid this problem, because the lowest FN-number of the table is 0.15.

You guys live around the globe and can ask questions 24 hours. I need to sleep and work in between.
Regards
Uli

 

Several published curves from Delft of residual resistance based on experimental results show the residiual resistance at almost zero for Fn less than 0.2. I have to wonder if it is smaller than the combination of experimental uncertainty and uncertainty of the ITTC 1957 curve.

It would be interesting to know if any of the individual test results showed negative residual resistance when the ITTC 1957 curve was used.

Also of interest would be whether the parameter sets of any of the Delft series models result in negative residual resistance at low Froude numbers.

 

Your contributions to this discussion are much appreciated.

 

As published in Larsson/Eliasson book, Delft series starts from Fn=0.125. I just checked RR for D32 and it is positive at Fn=0.125.

Actually there would be no surprise if one gets negative results at lower Fn, models of Delft series are quite small (1.6...2.0m) and scale effect would be significant. But again, with version of series we use never experienced that.

 

Just a note: if one is using coefficients from Fossati book p.24, pls look at digits they have - some coefficients are rounded up and have only one-three digits. This will significantly effect the result.

 

Alik, which Delft series formula do you use? The equation included in Principles of Yacht Design, Third Edition by Larsson and Eliasson, 2007 is the 1998 equation.

 

I am using formula from 1994 or 2000 edition. My spreadsheet is old

 

The Third edition of Principles of Yacht Design has coefficients listed starting at Fn = 0.10 in 0.05 increments. Presumably earlier editions had different versions.

 

I also remember that first published series (1981?) starts from 0.125.

Wonder how one can get reliable results for lower small speeds for 1.6-2.0m models.

 

Extension to the first post in this thread:

Another regression formula was presented by Keuning and Sonnenberg at the 1998 HISWA symposium. It's available at http://www.hiswasymposium.com/asset...ion of the hydrodynamic forces on a saili.pdf This uses the same form of the equation as the 1996 version, but with different coefficients and covers a Froude number range from 0.10 to 0.60 in 0.05 increments. This is the equation included in Principles of Yacht Design, Third Edition by Larsson and Eliasson, 2007.

 

OK, I am looking at this paper now. See table 4 on p.113 - the coefficients are rounded up again. Engineering practice requires at least 4 digits in intermediate calculations, in come coefficients there is only one digit. This might have significant effect when doing calculations for low Fn.

 

Given that the coefficeints are determined by regression adding more digits may not improve accuracy. Perhaps that is why Delft went to 4 digits after the decimal point.

The number of digits needed depends on the final accuracy required and the details of the calculation.

 

Reply to post #17

As indicated in my last post, if you take the parameter set of Delft model #1 as published in the 1998 HISWA paper and also the coefficients of that paper, the calculation gives a negative residual resistance at FN=0.1
I assume that Prof. Keuning is aware of this discrepancy and omitted therefore the speed range below FN=0.15 in his new equations in 2008. Possibly the team at Delft suspected some laminar flow at low speeds considering the small size of the models.
I did some boundary layer calculations (CFD) for the canoe body and it turned out, that the Delft-friction line is definitely too high. The Grigson line fits more closely the CFD-results. Laminar flow is not an issue.
Regards
Uli

 

Once those coefficients are multiplied they should keep 4 significant digits, not 4 digits after decimal point. This can be source of inaccuracy at small Fn.

 

Arbitrially adding digits, even if they are generated by a computer program, doesn't alway increase accuracy.

I don't expect the resistance formulas based on regressions of the DSYHS data to be accurate to .01% as implied by four significant digits. If not that accurate then no reason to insist on 4 significant digit accuracy in the final answer.

Uli has probably explained the unusual results at low Froude numbers.