by Capt. Kirk Greiner
Numerous cases involving groundings or collisions of ocean going vessels require reconstruction of the trackline of each vessel involved to analyze the cause.
Data Sources
The source of information which is used to establish the trackline(s) is the course recorder, the fathometer and the engine bell log. Ancillary to these sources are ships handling characteristics, speed curves, navigable water characteristics, accurate correlation of times, and current or tide information. Where reconstruction uses fathometer records, current water depths are needed.
Time Correlation
In reconstructing a trackline, it is necessary to use a common time so that speed changes, headings of the vessel(s), and fathometer readings can be coordinated to a common time. It is not really important that the times used be correct, but they must have a commonality. Thus, immediately upon the occurrence of an accident, a check of the time on various recording devices and the log books must be made. Where two vessels are involved, correlation between the times on each vessel are also needed.
Although I have written about this need before, I have yet to do a reconstruction where the times have been checked after an accident. Although testimony will indicate that times were corrected immediately before the vessel sailed, or at noon the previous day, there is almost inevitably an error between the times on various instruments.
Data Analysis
Reconstruction which I have done include the EXXON VALDEZ grounding, the ARCO ANCHORAGE grounding, the MING WINTER grounding in the Columbia River, the collision between the CHEVRON LOUISIANA and the PROSPERITY, the collision between the C. E. DANT and the AEGEAN SEA, the collision between the WAYWAY and the ORIENTAL MARINER and others which I cannot off hand recall. While each has had its own peculiarities, common among them has been the collection and correlation of data, the interpretation of that data and the plotting of the result.
In a recent collision, several factors not usually present made a conventional reconstruction impossible. Two tankers collided at a bend in the channel, one upbound and the other downbound. Each claimed the other had crossed into their side of the river. One had no recording devices in operation. The other had a course recorder, bell logger and fathometer in operation.
Because the vessel with the course recorder in operation was navigating a river, the heading changed continuously. It was thought that the manual reconstruction of the trace line using headings from the course recorder would be impossible, or inaccurate at the least.
Computer Assisted Reconstruction
It was decided to digitize the information and create a software program to draw a poly line to be used as an overlay on the Corps of Engineers River Condition Survey Drawing. The rudder angles used up to the time of collision and the gradual changes in heading were considered small enough so that the tendency of the vessel to crab through the water or slow down in a turn would be negligible. These tendencies were therefore ignored.
The course recorder trace was placed on a digitizing table and calibrated. The program created for this was developed so that the digitized information from the course recorder could be verified immediately. After verification, the course recorder was traced using a "puck" which transferred the data into the computer.
The program would then print out a table of headings at set time intervals. Intervals of any increment could be chosen. I selected each tenth of a minute. In order to create this, I did not have to chose and record points from the course recorder every tenth of a minute. Where the heading was constant, or where its rate of change was constant, points could be selected every 30 or 60 seconds or even at greater time intervals. The computer would assume a straight line between the points and provide a reading at intermediate chosen intervals.
One of the problems encountered was the calibration of the heading. Recorders can have several errors in them. One is mechanical slack in the mechanism at the top and the bottom of the trace (at the end of each quadrant). This would be indicated if the trace did not reach either the top or the bottom of the chart when shifting quadrants. Another error which can be visually identified is an offset of the graph either up or down on the recorder. In order to compensate for these errors, when calibrating the graph on the digital table, it was assumed that the bottom of the trace was 90o and the top 0o (for the first quadrant). The actual calibration lines on the recorder graph were ignored.
In addition to the table, the program would create a trackline on a plotter. Where this trace was to be overlaid on a chart or drawing such as a Corps of Engineers Condition Survey of the channel, the scale of the drawing and the speed of the vessel have to be inputted into the program.
Reconstructions Using Fathometer
The only remaining means of reconstructing the track line was to use the fathometer record. A different software program was developed for this.
A depth recorder trace has no time scale on it. Times are periodically indicated by the operator who creates a vertical line on the trace chart and manually writes the time alongside of the line. Like the course recorder, the fathometer graph was placed on the digitizing table and calibrated. Time calibration used the manually entered times when the mate would place a vertical line on the graph using the fathometer scribe. It was recognized that when using two lines to determine the time interval, a time error of up to one minute could occur due to the fact that the time was recorded to the nearest minute. Depth calibrations used the scale on the fathometer graph.
An overlay was then made on a clear piece of cellulose using Letraline colored tape to outline the channel. Then, starting at the last navigational point on the river which was logged abeam, time segments were placed on the overlay. These were based on the speed of the vessel that was constant for this time span.
The water depths from the table of the depths printed out by the computer were entered into a spreadsheet. Corrections were made for the draft of the vessel and the depth of the water. If the fathometer transducer is located aft, a correction for drag may also have to be made. The result is the depth of water from datum at noted times.
This information can now be written on the overlay at the edge of the channel. At each time interval for which there was a depth, the areas across the width of the channel where these depths occurred were circled.
When this was done, the overlay was examined to determine if there was a line of soundings through which a line could be drawn to represent the track line. If not, the overlay was moved up or down river to accommodate time errors until a "match" between sounding and water depths was found which resulted in a reasonable track line.
This method will only work if the bottom is not smooth and of constant depth. It is one of the methods relied on by the Coast Guard years ago when it manned weather stations in the middle of the ocean. These stations were beyond the range of LORAN A (LORAN C had not come into existence) and there were no satellites. Thus, when the skies were overcast for prolonged periods and navigational sights could not be taken, soundings were the only method of determining the ship's position.
The methodology of using state of the art computer technology to determine track lines of vessels will improve reconstructions and make the results highly reliable.
Home ! Search Page ! Master index ! Table of Contents for This Issue