Fifty kinds of marine diesel fuel oils are prepared by blending atomospheric bottoms with distillates in various proportions, and their rates of carbon-flower formation are compared with each other after 24 hous' running in a pre-combustion type diesel engine. A cylindrical spacer is inserted into the pre-combustion chamber to accelerate the carbon formation around its injection nozzle by reducing the effective volume of the chamber.
The rates of carbon formation in these fuel oils are numerized by counting the frequencies of remarkable changes in exhaust gas temperature, which are regarded to be correspondent to the rise and fall of carbon flowers. The longer is the interval (Y), the less the carbon formation.
Seven variables representing the characters of these fuel oils are selected as follows: carbon residue, specific gravity, viscosity, sulfur content, aniline point, diesel index, and spot test grade (sedimental matter) .
Among the coefficients of correlation between y-value and each variable, only that for the combination between y-value and carbon residue is significant. The dispersions of y-values observed in this simple correlation is also suggested to be attributed to the influences of other variables than carbon residue, and then multiple regression analyses are followed over the 64 linear equations constituted with all combinations of the seven variables.
The coefficients of multiple correlation in all regression equations become larger in accordance with the numbers of variables composing each equation, but some of variables are found to be insignificant as for their factors after t-value testing for each variable.
Finally five linear regression equations are left to be reasonable as the representations for the carbon formation tendency in these fuel oils tested, and the y-values calculated by means of these equations agree well with those observed.
The limiting y-value that would be appropriate to judge the safety in actual performance of such fuel oils as used in this work is proposed to be about 85 minutes in referring to the published data in ship tests, and the safety grade for any one of these fuel oils is predicted by comparing the y-value calculated through one of these equations with the limiting y-value.

View full abstract

In this paper some calculating methods of the forced vibration of marine diesel engine shaftings are discussed and the calculating results are compared with measured one. They are summarized as follows;
(1) Two direct calculating methods are proposed, which provide not only amplitudes for any point of vibrating system or frequency but also their phase angles.
One of them is to solve a group of independent uncoupled equations, which are derived by transforming the physical co-ordinates into normal one with the modal matrix. The modal matrix is given by solving the undamped free vibrating equation.
Another one is to calculate steady-state amplitudes and phases, by applying the mechanical impedance method to the multi-degree freedom equation. Both methods are applicable to coupled or uncoupled torsional-axial vibrating equation. But the coupled forced vibrating equation that contains two kinds of bending motions, is so complex for numerical analysis that another simplified equation, which excludes these elements, is introduced.
(2) The estimating method of axial vibrating and damping energy is introduced and the axial amplitude is found by equating the vibrating energy to damping one. For the coupled torsional-axial vibration, the axial amplitude is estimated from coupled torsional-axial mode, which is calculated by solving the undamped vibrating equation. With this so-called energy method, one can calculate only the resonance amplitude.
(3) Numerical calculations by above three methods for three different shaftings are compared with measured data on board and their agreements are fairy good.

View full abstract