Wot only can the computer be programmed to read 

 the film and type out results but it also can be 

 programmed to analyze the original data including 

 normalization, linearization and individual 

 instrument calibration correction. This greatly 

 enhances an effort to allow simplicity (and 

 thereby reliability) of systems placed in the 

 ocean. The result of this principle is that the 

 computer becomes truly a functionary part of the 

 transducer. In the analysis of the current meter 

 film, the computer memory can store the charac- 

 teristics of every instrument and the calibration 

 curves of the Savonius rotor. In addition, 

 because the CRT is under computer control, the 

 computer can present the raw data and the com- 

 puted results in graph form for immediate visual 

 presentation. 



APPLICATIONS OF COMPUTER CAPABILITY 



Additional judgment making capabilities are 

 built into the computer program to prevent 

 grinding out meaningless data. For example, if 

 there is a break in the reference channel, or too 

 many channels, these facts are typed out. These 

 algorithms or rules for rejecting or accepting 

 results can be elaborated manyfold to include, 

 for example, that a direction of 121° is 7 ones 

 and its absence in some pattern signifies unre- 

 solved bits. The question of unresolved bits 

 and the measure of confidence one can assign to 

 any given reading has already been approached in 

 a simple and direct way with the computer. This 

 is done by printing out the level of the least 

 significant digit which was a zero in the 7-level 

 binary vane or compass reading. The fifth place 

 numbers in Fig. 5 preceding the 3 place compass 

 and vane readings are these values . A zero occurs 

 only when all bits are resolved. Thus, a 7 indi- 

 cates the highest level (3°) of confidence in 

 the value. Obviously, any other digit down to 1 

 could be equally precise since all channels can 

 legitimately be on, so this is only a qualitative 

 indication. 



The fact that up to 600 scans per inch of 

 film may be made leads to an evaluation of the 

 recording idiosyncracies and thence to improved 

 reading judgment capability. These results are 

 also fed back to evaluate the performance of the 

 current meters themselves. In fact, a program 

 has been written to test and calibrate every 

 instrument using a short length of test film 

 such as shown in Fig. 2. This program gives the 

 percent each channel is on (theoretically 50$) , 

 the amount of noise on the film (i.e., images 

 that are not intentionally recorded data), the 

 density or width of each channel, the mean posi- 

 tion of each channel, the maximum drift of each 

 channel, channel width variations if they occur 

 and the variance of the reference channel along 

 the film. 



Via the CRT plotter, a variety of data pre- 

 sentations can be made automatically by the com- 

 puter. Some of these have been programmed 

 including trajectory summations, compass rose 



current diagrams and a chart of the ocean area 

 under investigation with superimposed animated 

 current vectors . The last presentation may even 

 be done isometrically to give a 3 dimensional 

 view. Standard plots such as power spectra are 

 easily handled. 



A PERFORMANCE ANALYSIS OF THE CURRENT METER 



Some of these data presentations can be used 

 for an analysis of the instrument performance 

 itself as well as for a study of the ocean. An 

 example is a histogram presentation of the cur- 

 rent meter compass and vane indications which 

 was used to evaluate several mooring and instru- 

 mentation methods tested during an experiment 

 in Vineyard Sound off Woods Hole, Massachusetts. 

 Five current meters were moored in an area of 

 nearly uniform 65-foot depth where a fairly 

 strong tidal current running parallel to the 

 shoreline reversed approximately every 6 hours. 

 The arrangements were as follows: (l) a standard 

 mooring using an 800 pound cast iron block anchor 

 connected with polypropylene line to a current 

 meter (serial no. G-I85) and a subsurface 

 inflated polyvinylchloride float; all parts 

 shown in Fig. 6, (2) a rigid mooring (Fig. 7) 

 consisting of a 20-foot length of iron pipe with 

 an 800 pound anchor block on an iron plate with 

 k pipe legs on the bottom to the top of which 

 was bolted a current meter (serial no. 3°2), 

 (3) a current meter (serial no. G-190) with a 

 fiberglass fin attached (Fig. 8) moored in an 

 otherwise standard fashion, (k) a current meter 

 (serial no. G-l8l) with grease packed ball-bearing 

 swivels above and below it on an otherwise stan- 

 dard mooring and (5) a current meter (serial no. 

 299) with the recording film advance 10 times 

 that of (l) . 



Histograms of the currents for meters G-I85 

 and G-I90 are shown in Figs. 9 a nd 10 respec- 

 tively. Current direction is represented along 

 the abscissa in 128 increments from 000° to 

 357° an d the number of times each direction 

 reading occurred is represented along the 



Fig. 6. Mooring gear for current meter mooring 

 comparison experiment . 



81+ 



