212 



THEORY OF SEAKEEPING 



\'isually a swell when the wind exceeds certain strength 

 (say, force 5 on the Beaufort scale). Yet a swell often 

 can be distinguished bj^ its period on the wave record. 

 Howe\'er, the observer on board usually does not have 

 enough time and facilities to read and analyze the re- 

 corder tape. The best solution would be to record on a 

 magnetic tape, provided a portable analyzer is on hand. 

 Should this prove impossible, a tabulating attachment to 

 the tape recorder would pro\'ide a typed talile which an 

 observer can scan for a rough estimate of longer periods. 



A ship-borne wave recorder will give the sea spectrum 

 at a moving point. No method for evaluating the sea 

 directional spectrum from a moving ship is available at 

 present, and, as a result, all analyses have to depend on 

 an assumed directional distribution within each identifi- 

 able system. ^^ The predominating direction of each, 

 for instance the wind-sea and one or two swells, can be 

 visually estimated during daylight hours. This estimate 

 is practically impossible at night, and in a strong wind 

 is difficult for swells even in the day time. Roll (1952) 

 described a method of evaluating the direction, wave 

 length, and A^elocity of propagation of a swell In' re- 

 peatedly photographing the radar screen. This arrange- 

 ment is recommended on ships instrumented for ex- 

 tensive observations. Without it night observations 

 are not recommended. 



It is necessary to continue to look for ways of record- 

 ing directional wave spectra from a moving ship. Stereo- 

 photography is an extremely tedious method which is 

 suitable only for short waves. It appears impossible 

 to include a sufficient number of long waves in a photo- 

 graph in order to make a statistically valid record. 



In most of the previous sea tests, observation periods 

 of 2 or .3 n^in duration were used. Such recordings are 

 of little use in view of the properties of irregular seas. 

 Kent and Kempf commented on the "beat" cycles of 

 ship pitching with periods from 30 to 80 sec. The author 

 observed, however, a longer C3'cle containing a number 

 of such beats. These appeared to have half-hour pe- 

 riods. One becomes aware of these long cycles when 

 trying to photograph outstanding events (for example, 

 spectacular shipping of water), or trying to record 

 slamming. 



A 12-min observational run is definitely too short. 

 This has been demonstrated by the \'ariations of sig- 

 nificant wave heights determined by Cartwright's test 

 (also Tucker, 1-1957). A record length of 300 waves 

 has occasionally been recommended. Generally, a 

 broad wind-sea spectrum may be satisfactorily evaluated 

 by a shorter record, and a narrow swell spectrum re- 

 quires a longer one. It appears to the author that test 

 runs of an hour's duration may be satisfactory under 

 average conditions. Trial programs must be limited 

 therefore to a small number of runs. 



In the past the usual equipment on extensively in- 

 strumented observational ships consisted of gyros for 



'6 In special sea tests the directional sea spectrum can he ob- 

 tained from the information on wave slopes telemetered from a 

 suitably instrumented buoy. 



pitching and rolling and accelerometers for heaving. 

 Not having gj'roscopic instruments, Schnadel (1936) 

 and Horn (1936) estimated pitch angles from accelerom- 

 eter readings. Cartwright (1957, NSMB Symposium) 

 presented a comprehensi\'e spectral analysis based en- 

 tirely on accelerometer readings. On the SS Nissei 

 Mam the yawing-angle record was obtained from the 

 ship's gyro compass. However, this is not satisfactory 

 because compass repeaters usually operate in steps of 

 about half a degree. A continuous record is necessary. 



The author believes that in the future all six modes of 

 ship motion and the rudder angle must be recorded simul- 

 taneously along with the waves. Four groups of ac- 

 celerometer installations are visualized for the motions — 

 at the bow, at the stern, and two not far from amid- 

 ships. Of the latter two, one should be located very 

 low in the engine room and the other as high as possible 

 on the superstructiu'e. The accelerometers are to 

 measure vertical and lateral accelerations at each station. 

 Surging accelerations also should be measured at the 

 two stations amid.ships. All accelerometers should be 

 duplicated to guard against loss of records caused by 

 failures. If it is preferred to use gyros for recording 

 angles, the electrical output of their sensing pick-ups 

 must be identical with that of accelerometers to permit 

 interchangeability of electronic equipment. 



The central recording station should consist of a 

 battery of identical amplifiers (including a reserve 

 stand-by) and devices for adding and subtracting read- 

 ings in order to obtain records of translational and 

 rotational motions. All recorders should be identical 

 and spare recorders must be pro\-ided. It must be 

 possible to switch inputs instantly to a spare amplifier 

 and a spare recorder. It is extremely annoying to lose a 

 record of a valuable event in a storm because of failure 

 of some delicate recorder detail or, worse yet, because of 

 the need to change the recording paper. 



Attention should be gi\'en to shortening the warm-up 

 period for electronic equipment. It appears to be an 

 ironical law of Nature that many interesting events occur 

 unexpectedly outside of the scheduled recording periods. 

 These could be recorded if the equipment could be put 

 into' operation instantly. The only alternative is to 

 have the electronic equipment sufficiently rugged to stay 

 in operation continuously for the duration of an ocean 

 crossing; i.e., up to 3 weeks. 



Recording seven modes of motion (including rudder 

 movements) and using recording periods of an hour or 

 more make it a foregone conclusion that magnetic tapes 

 will be used. In addition to the timing signal, two to 

 four channels per tape may be expected, so that several 

 recorders will be needed. The wave record must be 

 contained on each tape. As a result, it will be possible 

 to evaluate by means of a cross-spectral analj'sis the am- 

 plitude and phase of each ship motion in relation to the 

 wave. If the analyzing equipment is provided, analysis 

 of an hour's recording probably can be completed and 

 presented in the form of several spectrum plots within 

 an hour. Thus, observers on board can appraise a ship's 



