Uncertainty analysis of experimental ship resistance tests

Uncertainty analysis is encountered in most experimental studies. When a problem in engineering modeling inspired by nature, it is created test system by assumptions and idealizations. Mathematical and physical rules are used a lot in the planning of the experimental apparatus and analyzing results. Each test step must define and complete correctly actually most closely to true value. If an error occurs in any step in experiment, the error occurred will adversely affect other results. Errors caused by the measurement system, calculations etc. need to estimate how much will affect the real results (closest to real value). In response to this problem, uncertainty analysis in order to determine the quantity of uncertainties, which is a matter for the purpose of containing both engineering and statistics. Uncertainty analysis can be applied in a different way in each engineering. Because various experiments and calculations are made in the various branches of engineering. There are differences in methods in order to estimate the quantity of uncertainty. The steps that will be applied to each system are standard. First, to determine the resources that may create uncertainty within a test or measurement setup. Then a variety of methods to estimate the uncertainty considered in the dominant sources of uncertainty is applied with engineering experience. Hence the effect on the result of the uncertainty that occurs on the system are analyzed statistically. The average and standard deviation of the test or measurement results are calculated and precision (random) limit is estimated. Later measurement of the uncertainty sources that affect measurement results, the bias (systematic) must estimate as the quantity. By calculating the square root of the sum of the squares of all the uncertainty components are the combined standard uncertainty. Then the results multiplied with the desired reliability coefficient. Finally, the obtained value of the expanded uncertainty. In this study, basic and generally expressed uncertainty analysis. then the uncertainty of the model ship resistance was investigated experimentally calculated. Ship resistance is defined as the forces in calm water and particular, constant speed. This acquired model resistance is used to calculate the value of the ship resistance and effective power with Froude and Reynolds approaches. Because of this feature, ship resistance is an important parameter that is considered in the design stage. If the ship resistance value is not estimated precisely and reliably, the main machine made the wrong choice. This situation significantly adversely affects the production planning and cost. Therefore one of the most reliable way to determine the resistance of the ship are experimental work. However, there are sources of uncertainty occurring in the test system. Sources of uncertainties should be analyzed in terms of the reliability of the test results. Especially the dominant uncertainty must be estimated quantitatively. The procedures used for the analysis of the model resistance results generated by the ITTC. There are procedures about Uncertainty Analysis of Resistance Towing Test which were published in 2002, 2008 and 2014.2008 ITTC procedure was revised at 27th ITTC meeting in 2014. Revised procedure (2014 ITTC) was taken into account in calculations. Then, there are obvious differences between 2002 and 2014 ITTC uncertainty analysis methods. However, as above-mentioned there is no only one method for uncertainty analysis. Purpose of the uncertainty analysis; by combining current academic knowledge and engineering review is to achieve the most accurate results. Therefore, two procedures are analyzed separately and they were applied to the experimental results. Uncertainty components was analyzed and was calculated by two methods. After that, advantages and disadvantages of the procedures were investigated. Two common type of ship are selected to determine the potential uncertainty sources in resistance test. The first type is displacement ship and the other type is a high speed marine vehicle (HSMV). Kriso Container Ship (KCS) was selected for displacement ship. KCS, which is royalty free form, is produced in the laboratory. Other model is selected for experiments performed earlier for HSMV. It is copyrighted form. For uncertainty analysis of KCS model is tested 12 times. On the other hand HSMV model is tested 5 times for uncertainty analysis. Uncertainties are estimated on number of the low, middle and high Froude Number for KCS. They are respectively 0.16, 0.21, 0.26. For HSMV, uncertainties are estimated on two Fr numbers. They are respectively 0.50 and 0.90. There are too many sources of uncertainty in a test system. One by one examination is not possible. However, limited sources of uncertainty can be examined through the existing academic knowledge. In this study, the uncertainties are examined occurred in resistance towing tests. Resistance tests was made in ITU Ata Nutku Towing Tank. Two different methods (2002 and 2014 ITTC) were used and the uncertainty results were estimated. For 0.16, 0.21, 0.26 Fr, expanded uncertainty over total resistance coefficient was found respectively % 3.86, %2.53 and %1.46 according to 2002 ITTC method. Then, according to 2014 ITTC method, this expanded uncertainty was found respectively % 3.13, % 2.10 and % 1.16. In low speed, uncertainty values are relatively high compared to the high-speed due to the resistance measurement system. But uncertainty values can be reduced by improving the current experimental conditions. The results for the HSMV is very good, especially in 2014 ITTC method. On the other hand, in 2002 ITTC method, uncertainty of wetted surface area value was found too high because of model length was smaller than the other models. Due to speed independently of the uncertainty on the CT (about 2% for both speed) less reliability. But the results was analyzed according to the 2014 ITTC method for 0.50 and 0.90 Fr, expanded uncertainty values was found respectively %1.33 and %0.42. At design speed, low percentage uncertainty is very important and quite good results. Uncertainty value will be reduced by improvements in the test system and towing tank conditions. However, uncertainty cannot be reset. Because each system has already uncertainties. If the uncertainty value reduces, the reliability of the results will increase. This applies to every measurement system. Finally, the results obtained from this study were compared with other towing tank results. Results also recommend that make to compare with the computational fluid dynamics.