... <obtained for an infinite fluid). Among the first Chung (1977) published real comparisons of experimental added masses and wave damping coefficients with theoretical coefficients computed by the Frank's method, showing that agreements are much better for cIrcular cross sectIons, and better for heavIng than for swaying motions. In Matsumoto (1989), a good deal of realistIc results are presented, which are referred to the applIcatIon of MorIson formulation and to direct lab avaIlability of the pendulum test machinery. In Matsumoto (1989) only square cylinders have been considered, here general- Ized to the case oT both polygonal and circ- ular cylinders, where the hydrodynamIc forces on cIrcular cylInders have stronger tendency to dIrectly depend on the characteristics of each vortex shedded from the body (see Tab.l). The dIrect experience of opportune pendulum tests treating single and multiple cylInder patterns has allowed reaching the aimed and reqUIred knowledge. PHYSICAL PROBLEM INTRODUCTION AND FORMULATION The physIcal problem development essentially and basIcally consists in determining those parameters which do really contribute to the CONFIGURATION PREDICTED SQUARE CYLINDERS SINGLE 1 - 15 CIRCULAR CYLINDERS SINGLE 1 - 50 POLYGONAL SINGLE 1 - 70 configuration of the system, and which do influence its behavior. The field of study is now wider than in RomagnolI et al., (1986, 1989, 1990), where the object of the research has been only the ,practical behavior of the ice masses. A clear separation has to be done between the mass and the drag effects of the total hydrodynamIC forces upon the oscIllat- ing body systems, In order to evaluate both the partial and the single contrIbutIons. The physical problem consIsts in studYIng and in analyzing the hydrodynamic forces actIng upon arbitrarily oscillating bodie~ (in general on groups of them), In order to forecast and to check the real phenomena occurrIng in the sea during the operational stages. A practIcal, safe and reliable calculation procedure comes to be reqUIred, in order to determIne all the time dependent velocity distributIons around the oscillating bodIes and - as saId before - the hydrodynamic forces actIng on them. The range of applicability of the theoretIcal possible solutions has, anyway, to be always validated by the performance of experlment~l tests (e.g. the pendulum test), at least WIth reference to the case of deals of perIodIcal oscillations of chosen body assemblements. The resulting forces are computed by applYIng CONFIGURATION PREDICTED MEASURED 1 - 5 COUPLED 1 - 50 1 - 10 1 - 8 COUPLED 1 - 150 1 - 20 1 - 10 COUPLED 1 - 200 1 - 25 Tab. 1: Errors (expressed in percent) and discrepancies from the theoretical values of predicted and measured (by the pendulum test) time dependent hydrodynamic forces acting on bodies oscillating in water 'data from literature]. 85 the MorIson formula, assuming that the mass and the drag coefficients keep constant, with the correct tIme dependent relative veloc1t1es and acceleration values. The optimlzed prediction procedure can allow also the d1rect computatlon of how the time dependent hydrodynam1c forces actlng on the chosen bodies are influenced by the problem physical parameters. THEORETICAL BACKGROUND AND DEVELOPMENT The theory upon which the study has been performed includes the following assumptions: the hydrodynamic forces wh1ch solicit the oscillating bodies are in good accordance wlth the Morison formula, the drag and mass coeffIcients keeping constant during the oscillation cycles; the drag and the mass coefficients applied 1n the Morison formula can be assumed to be the same as 1n the range of very large KC (Keulegan-Carpenter) numbers (Matsumoto, 1989 ; Morison et al., 1950), the veloc1ty distribution in the wake generated by arbitrary body oscillations is computed by lntegrating the induced velocities by impuls- ive inflow velocity to the body for each of the chosen time steps; the induced wake speed field due to any impulsive body - motion v v o t velocity (also 1n case of real shock) and the corresponding propagation in time and in space can be obtained by solving the linear- 1zed boundary-layer equation(s) in free and turbulent unsteady wake (Matsumoto, 1989). Ind1cating w1th the expression "unsteady and turbulent wake theory" as 1n Matsumoto (1989) the prediction procedure of the wake velocity due to arbitrary body oscillations, the aim is pursued also in the present paper of determining the induced velocitIes due to any arbitrary inflow velocit1es with respect to the considered bodies. Moreover a detailed analysis is performed of the slngle parameter influence in a project oriented way, as far as the design of the offshore structure comes to be concerned. The inflow velocity is assumed to be due to body motions, and to the effects of the wake velocities generated in the past h1story. and also to their propagation 1n time and space, (Matsumoto, 1989, where the 1st approxlmation of considering the induced velocity inslde an unsteady turbulent wake behind a body as due to the inflow steady velocity gIven at zero time and after, has been here overcome by the practical introduction of the saId character- ization inside the mathematical linearizatIon v V(s) ~ o t Fig. 1: Step functioned inflow velocity to a body. and consequent induced wake velocity (v.v(s).V.t in formulas). 86 procedure of the two dimensional boundary layer equation in free turbulent unsteady wake). The result of the said linearization being expressed as: 8v 8v 2 2 2 2 + V K ( 8 v I 8x + 8 v I 8y (1) 8t 8x (8 part1al derivative operator; V = inflow veloclty; v = induced velocity; K = virtual kinematic viscosity; and {x, y} =ccoordinate system which moves wlth the inflow velocity), the integrat10n yields the following result for the induced wake velocity, computed as due to an apt (opportunely defined) step- functioned inflow veloCity in the form (see Fig. 1>: 2 (x-Vt)(x-Vt)+ y ] 14K t) e vIs) = c (2) (4 IT K t) (t = elapsed time after the referred -inflow velocity is glven; c = numerical constant Of course the group of the point sources must be given for {x=O; y=O} for t ~ 0, with a known law. At the positive infinite limit, vIs) tends to a value given...