SNAME’s Stability Letter Improvement Project (SLIP) for Passenger Sailing Vessels Jan C. Miles, Captain, Pride of Baltimore II, Baltimore, Maryland, USA Bruce Johnson, Co-Chair SNAME Panel O-49, Annapolis, Maryland, USA John Womack, Co-Chair SNAME Panel O-49, Pittsville, Maryland, USA Iver Franzen, Captain, member SNAME Panel O-49, Annapolis, Maryland, USA

Figure 1, Pride of Baltimore II close hauled – note instrumentation sensors just ahead of the

main mast near the spreaders

ABSTRACT The current required USCG stability letters for passenger sailing vessels provide little useful stability information to the master when operating their vessel in “typical” every day conditions. Sudden squalls, even moderate ones, with their rapid and relatively large increases in wind speed and/or shifts in direction can be of more concern to the master than full fledged storms, both for vessel and passenger safety. This paper will present an overview of SNAME’s ongoing SLIP project to improve operator guidance available for passenger sailing vessels, traditional and otherwise. New approaches to developing operator guidance for non full storm conditions and presenting that guidance in a manner useful to the master will be explored. In addition, the prototype instrumentation and data logging package installed onboard Pride of Baltimore II for her 2005 European cruise will also be discussed including the analysis of two squalls that were encountered.

THE 18th CHESAPEAKE SAILING YACHT SYMPOSIUM ANNAPOLIS, MARYLAND, MARCH 2007

NOTATION Sail symbols starting from aft-forward, a number following a symbol indicates the number of reefs in that sail M = mainsail F = foresail S = staysail J = jib JT = jib-topsail SQT = square-topsail MGT = main-gaff-topsail STG =square top-gallant STDS = studdingsail RT = ring-tail TRYSL = storm-trysail STMJ = storm-jib Symbols used in the MCA based wind heel analysis AZ= vessel righting arm at an assumed KG COG = course over ground from GPS dwhl = derived wind heel lever GPS = Global Positioning System GZ = vessel righting arm at the current KG of the vessel KG = height of vessel center of gravity above the base line LCYC = Large Commercial Yacht Code (MCA) MCA = Maritime and Coast Guard Agency (UK) SMC = Ship Motion Control AB (Sweden) SOG = Speed over ground from GPS WLO = the magnitude of the actual wind heeling lever at 0° which would cause the vessel to heel to the 'down flooding angle' θf or 60° whichever is least θd = the angle at which the 'derived wind heeling' curve intersects the GZ curve. (If θd is less than 15° the vessel will be considered as having insufficient stability for the purpose of the LCYC Code). θf = the 'down-flooding angle' is the angle of heel causing immersion of the lower edge of openings having an aggregate area, in square metres, greater than:- Δ/1500 where Δ = vessels displacement in tonnes INTRODUCTION

In the fall of 2004, Captain Jan C. Miles of the sailing vessel Pride of Baltimore II and the Society of Naval Architects and Marine Engineers (SNAME) started a collaborative study to improve the operator guidance available in USCG Stability Letters for passenger sailing vessels, with a particular focus on traditionally styled and rigged sailing passenger vessels.

“As a sailing master with a long career in traditionally rigged fore & aft rigged sailing vessels, I have explored, through trial, various ways of sailing and uncovered what works better on each and every vessel I have commanded. Yet I cannot today tell you what each of those vessels is likely to

do under any given sail area and wind strength except in general terms. When I first come to my new command I have no introductory information at hand to begin to “understand” what the sailing characteristics of the vessel are except that the vessel has passed the USCG stability regulations. For me, what would be useful is to understand what the actual wind forces would need to be to lean the vessel over to critical angles such as deck edge or bulwark immersion when she is flying her usual sail areas. The current USCG stability letter does say that water tight doors need to be closed and freeing ports opened, but it does not provide any information about how the ship may react to a given wind force, so the existing stability letter offers almost no operator guidance under sail. My interest is to see if it is possible to provide within the USCG Stability Letter format useful introductory vessel stability characteristic information for a new master/mate, or a master/mate new to an unfamiliar vessel, to begin their on the job learning of the nuances of their vessel.”

This collaboration begins with Captain Miles’

observation that the role of the master on a traditionally rigged passenger sailing vessel is to operate the vessel in the most traditional sailing manner possible while remaining safe and on schedule to meet the modern demands of today’s “fixed” itineraries. Sailing Masters face several challenges in meeting this requirement: limited information in USCG Stability Letters, wide diversity of vessel types and seasonal employment. There is only one college level, USCG approved program in the United States that specifically addresses the operation of traditional sailing vessels (Maine Maritime Academy's Associate and Bachelor of Science Degrees). This program of study does not currently address the prediction of a vessel's angle of heel for sail area and wind strength. A master coming aboard a sailing vessel for the first time has little information, other than suggestions from the vessel’s past master(s) (if they can be located), to “understand” how the vessel will respond for different sail area and wind strength combinations except that the vessel has passed the USCG stability regulations (Appendix A-3)

The vessel’s USCG Stability Letter makes no mention

of the vessel’s general stability characteristics accept to direct the master to preserve it; i.e. not to move the ballast, change the vessel physically, to permit water to accumulate in the bilge, and to minimize free-surface affects. All very good advice, but provides no information of the vessel’s likely characteristics under the pressure of sails set during typical operations. The USCG Stability Letter is strictly for very artificial conditions; modest beam winds and all sails sheeted flat. (See Appendix A-3)

This lack of any attempt to provide detailed knowledge of vessel characteristics in conditions other than single criteria is contrary to many other parts of the marine industry. Large cargo ships, whose stability characteristics may vary during a voyage and turnaround, have means to calculate the safe stability zone established by the ships assigned stability letter during loading and unloading procedures. They are also able to refine, in real time, the safe operating zone of vessel stability during the voyage. The United Kingdom, which also certifies auxiliary sail vessels, provides a stability booklet that attempts to describe the affect of wind force on sail area, at least in static equilibrium terms, in order to help the master avoid dangerous down flooding. (Deakin 1990 & Appendix A1)

Internationally speaking, this requirement is recent

and primarily a down flooding avoidance criteria for all sail areas. It does not, however, provide a master with incremental information on their vessel’s actual sailing characteristics under a variety of sail combinations.

The seasonal nature of the US traditional passenger sailing vessel fleet also exacerbates the deficiencies in the USCG stability letter. Approximately 85% of the licensed auxiliary-sail positions are seasonal only. This results in significant turnover on the vessels with some masters/mates serving on different vessels almost each season with little operator guidance for the new vessel. This seasonal nature of employment also means that the previous master is often not available, having long moved on, to provide the new master with their valuable knowledge base on the vessel’s characteristics.

Maybe the relatively few accidents on record, considering the number of departures, passenger miles and sailing vessels, ranging from 16th to 20th century sailing rigs and hull forms are a testament to the nature of the certified sailing vessel system and the considerable professionalism of the officers. But with the recent advances in Naval Architecture now available to analyze and predict a sailing vessel’s characteristics, there may now be an opportunity to provide better assistance to the sailing master so as to head off avoidable incidents or accidents. This is particularly important when under sail in weather conditions that could experience rapid, though not necessarily major, changes such as summer squalls.” OVERVIEW OF PROPOSED SOLUTIONS

Based on the above comments and discussions with sailing masters, naval architects, and others in the traditional sailing vessel community, any improvement in stability analysis and operator guidance for traditional passenger sailing vessels will come in three areas; (1) new analysis methods to create the technical stability information required, (2) new methods to convey that technical information in a manner useable by the master,

and (3) new methods in training the master in how to use the new stability guidance.

New Stability Analysis Methods

The new stability analysis methods being developed will include both new calculation tools and new assumptions on weather parameters such as potential wind gust magnitudes and durations that could be experienced. First and foremost, any analysis methods must be able to handle less than full storm conditions. Sudden squalls, even moderate ones such as those that occur during normal summer weather, with their rapid and relatively large increases in wind speed and/or shifts in direction can be of more concern to the master than full fledged storms, both for vessel and passenger safety.

This deviation from the traditional one size fits all stability criteria approach is needed for several reasons. One, full storms generally build over a significant period of time which allows for sail to be reduced ahead of time. Second, recent onboard data recorded on Pride II during two squalls demonstrate that several current assumptions concerning wind gusts may be significantly deficient. Gust magnitudes in full storm conditions are currently estimated to be about 40% over the nominal wind speed levels, yet our data showed temporary gust magnitudes of 100% occurred during the squall events (see below for details). Lastly, well over 95% of the time US passenger sailing vessels are underway in good weather only. These vessels will not get underway in storms and will often return to port well in advance of known threatening conditions. It is the summer squalls and other similar events that are unpredictable, that can catch a vessel with too many sails up.

The new stability analysis methods must also be able to handle a combination of weather parameters to be able to determine the true worst case condition for an individual vessel. The principal parameters envisioned are; the magnitude of the base wind speed increase, the time in which the base wind speed increases (gradual/rapid), the magnitude of the gusts levels, the magnitude of the wind shifts, and the speed at which the wind shifts occur (gradual/rapid). An example of a possible combination would be a moderate dry squall that also included a rapid and large wind shift.

This example also leads into the next requirement for the new stability analysis methods; level of risk determination to key vessel points. In the example given, it is highly unlikely that any capable vessel would capsize, but it could experience heel angles large enough to put passengers and crew in danger or at a greater risk of down flooding. By analyzing the risks to key points, information can be developed to assist the master in sail selection to minimize passenger/crew risks, yet still maximize the sailing experience. Key points analyzed should include

passenger comfort heel angles, deck edge submergence, bulwark submergence, down flooding angles, knockdown forces and maximum allowable heel angle levels under the stability rules which apply to a particular vessel. The potential loads on the rig should also be noted for consideration by the master when making sail selections.

The last requirement for the new stability analysis methods is the ability to include the dynamic effects of wind gusts. Traditional sailing vessels, with their heavy masts and keels, have a large roll moment of inertia. This results in different responses to wind gusts depending on the speed and duration of the gust’s increase and decrease. PROTOTYPE REAL-TIME SAILING PERFORMANCE LOGGING for PRIDE OF BALTIMORE II (2005)

A prototype real-time sailing performance logging system (Figure 2) was installed onboard the Pride of Baltimore II, a 91 foot waterline fore and aft rigged topsail schooner, (Gillmer 1989) for her 2005 summer European cruise. This system recorded the following sailing performance parameters. - Vessel’s course and speed recorded from the ship’s

GPS unit. - Apparent wind speed and direction recorded from the

ship’s wind sensor mounted just ahead of the main mast above the radar ring. (See figures 1 and 4)

- Vessel’s pitch and heel angles recorded from an SMC 2 axis heel and pitch sensor temporarily located onboard.

- Sails set and major sail heading were recorded by manual notes.

Note: Several previous full scale experiments have been described in CSYS proceedings (Deakin 1991,Grant, et al. 1997 & 2001.)

Figure 2 - Prototype Logging System

The data recorded by this prototype logging system was translated into an Excel file for analysis. Figure 3 is an example of the Excel spreadsheet with the raw data plotted. The review of this data leads to several important conclusions that will need to be incorporated into future data logging systems.

Figure 3 Time histories from September 5th, 2005

First the apparent wind speed and direction will need to be corrected for the motion sensor readings of both the heel and pitching of the vessel as is done in the B&G ”Deckman” performance measuring system used aboard high performance racing yachts such as the Volvo 70 class. This is apparent when one observes that the Sept. 5th heel angle data leads the uncorrected wind speed data by several seconds (Figures 9 and 10 following).

The authors also have considerable concern about

time shifts caused by differing sensor time constants (lag and lead). It is obvious from the apparent wind speed and direction corrections done on the spreadsheet that the analysis is overestimating the effects of the mast mounted wind sensors. This may be caused by likely (but not currently known) difference in time constants between the anemometer accelerations and the hull rotational accelerations as measured by the SMC motion sensor. The authors are curious how the B&G Deckman system handles this problem, especially as it applies to the systems installed on the Volvo 70s.

In June, 2006, the SNAME T&R Committee funded the installation of port and starboard Young wind sensors as shown in the Figure 4.

Figure 4 - Pride’s Autohelm Wind Sensor (center) and Port and Starboard Young Marine Junior Sensors

Capt Miles immediately noticed that the old centerline

sensor in the center of Figure 4, gave wind velocities up to 1/3 higher than the windward port and starboard sensors when close hauled. This shows that there is an acceleration of the wind speed between the masts (see Figure 1) causing the anemometer located in the slot between the masts near the radar dome on the main mast to read too high. (Most tall ships fly pennants at the mast head which interferes with locating an anemometer there.) Therefore, columns were added in the spreadsheet (Figure 6) to correct the measured wind speed by a factor to take into account the slot (venturi) effect.

For 2005, the raw wind data error is assumed to be

20%, which equals the correction factor of 0.8 used in the example analysis. This one size correction factor is

obviously incorrect for all headings and wind velocities. The project needs new 2007 data guidance on correcting the old data for our forensic analysis of the dismasting. Pride II will now be able to get a better measure of apparent relatively undisturbed wind from the Young wind sensors during the 2007 season.

From preliminary data recorded on the Pride of Baltimore II, large intense gusts with a rapid speed increase/decrease only caused a modest increase in the base heel angles. The force of the wind gusts is absorbed in the vessel’s rig before the wind gusts could overcome the vessel’s roll inertia. Preliminary data also showed that smaller, longer lasting gusts with a slow speed increase/decrease caused significant increases in the vessel’s heel angle as predicted by current equilibrium theory.

Figure 5 Indicated Apparent Wind Corrections for roll and pitch

In September 2006, further analysis of the time series graphs suggests that the mast rotation under gust loading absorbs enough of the wind gust energy to make the observed anemometer readings much smoother than would be expected given the roll and pitch response time traces of the vessel. In order to compensate the indicated wind speed for the effect of rolling and pitching at an assumed elevation of the sensor at 19 meters above the roll axis, a previous roll only analysis was modified (as shown in the spreadsheet printouts in Figure 6) and it now treats the apparent wind as a vector. The measured wind speed vector is divided into transverse and longitudinal components; the pitch rate correction is added to the longitudinal wind speed component and the roll rate correction to the transverse wind speed component. The resulting vector components are recalculated as a corrected wind velocity vector. The roll rate and pitch rate are currently calculated from backward and forward differencing of the heel and pitch time traces so that additional smoothing is not deemed necessary.

What is still unknown is how the anemometer actually

responds to the rapid changes in roll and pitch and what degree of overcorrection has been applied. This problem

is readily apparent from the large vector based corrections to the apparent wind direction as shown in Figure 6. To do this correctly, one may need a 3-D sonic wind sensor and a

fast response gyro based motion sensor, perhaps both with identical low pass filters to insure a common time constant.

Figure 6: September 5th 2005, First Squall data extract.

-

Figure 7 Apparent wind directions, raw and corrected for mast motions and GPS COG.

The following figure 8 shows that although the amplitudes of the heel and pitch motions are greatly different (see figure 3), the pitch and roll rates and the corresponding velocity corrections at an anemometer height of 19 meters are both significant.

Figure 8 Time series expansion of the anemometer velocity correction data

The second lesson proved that when using the GPS unit used for data acquisition, the speed and course time averaging period must be set to a low value, generally around 1 second. The higher value of one minute averaging in use on September 5th 2005 masked critical peak values and created a SOG and COG time lag in the data. Therefore, in the detailed squall analysis below, the time series of the apparent wind has been shifted by 10 seconds to line up the peak squall corrected velocities with the greatest change in apparent wind direction.

The third lesson indicated that the location of wind sensors is critical in getting accurate wind data. The 2005 prototype logging system used the ship’s wind sensor located forward of the mainmast just above the crosstrees, see Figure 4. This led to incorrect estimates of apparent wind as discussed previously. A port and starboard or a centerline sensor well clear of any sail and mast effects is crucial to minimizing sensor error corrections.

Description of two squall events onboard Pride by Captain Jan Miles. Narrative of the sailing situation leading up to the squalls:

The sailing scenario of the day the squalls were

experienced was one of performing as well as possible in the context of a race with other modern and traditionally rigged sailing vessels (so called tall ships). The fleet was a small one of about 8 vessels. Rigs varied from full rigged ships to Bermudian rigged ketches.

On the day of the squalls, PRIDE was sailing with mainsail, foresail, forestaysail, jib, jib-topsail and reefed square-foretopsail. The wind was about 40 degrees apparent off the starboard bow. With the wind from the SSW PRIDE was close-hauled heading SExS and sailing at speeds of 6-8 knots. Standard wind strength of the apparent wind was measured at 20-25 knots with periods of greater and lesser strength; up to 30 and down to 15 (prior to the slot correction).

Drawing 1, Pride II Sailplan Narrative of the first wet squall (see Figure 9)

The source of the first squall (wet squall) became

visible to windward some length of time before PRIDE was affected by it. There was not a great deal of vertical dimension to the cloud formation, but there was no horizon visible under due to rain under the cloud. The question of how to handle the rain cloud was to first determine if it would pass by ahead or not. With time it became evident the cloud and rain formation was not going to pass by. Continuous scrutiny of the rain under the cloud showed no sign of slanting rain. There were also no gusty or changing wind directions as PRIDE got closer to the squall.

I decided we would sail on into the rain. As PRIDE

entered the rain (time about 1:29) the angle of heel increased with an increase of wind. There was no gusting or changing direction but there was a feeling of reduced temperature, probably due to feeling the wet of the rain. As the forces increased we steered PRIDE closer to the wind (see apparent wind curve in Figure 7), reducing the wind angle and slowing PRIDE down. Observed wind strength was 40-45 knots (see Figure 3) before the slot correction. After it became evident the initial onset of rain and increased wind was not going to reduce, the jib topsail was taken in. After the wind dissipated, PRIDE’s jib topsail was reset. When the dissipation occurred, wind strength was less than the standard wind speed and had backed toward the south from southwest. This wind strength reduction coincided with the rain dissipating and the horizon becoming visible. After about a half hour, the wind increased to the standard wind speed and veered back toward the SW as shown in Figures 3 and 8.

Gust magnitudes and anemometer effects (illustrated in figure 9): Note from the comparison of apparent anemometer wind speed and the roll and pitch compensated apparent wind the vessel absorbs much of the short term gust forces and averages the apparent wind

speed. The 30 degree roll (Time 1:34:30) appears to have been caused by three or four short gusts of up to 46 knots when corrected for mast motions.

Figure 9 Time series data expansion during the First (Wet) Squall

Narrative of the second (dry) squall (see Figures 10a & 10b)

The second squall (dry squall) provided no warning. I was aware of a rain cloud to the starboard beam (windward side) that was tracking past as PRIDE was sailing on with an apparent wind of around 40 degrees. This cloud had no particular verticality and also had falling rain under it, heavy enough that I could not see the far horizon through the rain. Again, there was no discernable angularity to the rain as it fell. This cloud/rain formation was about 2-3 miles off. Because it was at the beam and passing aft, I dismissed this formation. I was more interested in another cloud/rain formation ahead. I did not want to sail into that one without first reducing sail and was very interested in finding out if it would pass by the bow before we got close to it. As I studied the cloud/rain formation ahead, the wind suddenly increased (Time 2:26:15). Apparent wind angle remained the same but at an observed (uncorrected) strength of more than 40 knots. There was no rain and no noticeable decrease in air temperature. I steered PRIDE closer to the wind and waited for her to begin to reduce her

angle of heel as the apparent wind angle reduced. Moments later the angle of heel had not reduced. I checked the wind meter and I recall finding it indicating 50-60 degrees apparent at more than 40 knots. I turned the helm some more and looked forward to observe the results. Moments later the bowsprit broke and all three headsails flew to leeward. Angle of heel immediately reduced and weather helm became evident.

I steered immediately to leeward and called for the

mainsail to be dropped. This was done quickly. By the time the mainsail was down the squall was dissipated and felt like it had dropped to a velocity below the pre squall wind speed. PRIDE was not responding to my effort to turn away from the wind. Instead she was now all but stopped in the water and continuing to swing westward toward the now reduced wind strength, putting the sea swell on her port bow. As PRIDE went head-to-wind to a heading of west by north the foremast began to fall (around time 2:27).

Figure 10a Time series data expansion for second (dry) squall

Figure 10b further expansion of Second (Dry) Squall with apparent wind time shifted to correlate peaks

The first indication of the impending fall of the foremast was very loud cracking and popping sounds coming from the foremast at a level close to the deck. The noise was so loud as to draw the acute attention of everyone on deck. Being the person farthest aft I could see everyone focus on the noise just as I was focused on the noise. I was standing to port having been involved with trying to get the ship to head to the left and away from the wind and sea direction. As I looked forward to the noise coming from the foremast I noted that its shrouds were

going slack and taught alternately port and starboard to the rhythm of the rolling caused by the sea on the port side of the vessel. After a couple of cycles the mast tilted aft and started to fall quickly. During these oscillations everyone on deck that I could see was rooted to the deck looking toward the noise and the mast. Everyone on deck could see what direction the mast was falling and some moved to safe locations and some did not need to move. I moved to starboard as the mast fell aft and swung down past the mainmast close to the port side of the mainmast. The

foremast landed somewhat softly as compared to the mainmast when it fell, apparently due to fetching up on the spring stay that links the tops of the lower-masts. The mainmast broke half way up. It wobbled port and starboard a bit and fell with a loud thud and vibration just to starboard of the helm. It had crushed through the full height of the transom bulwark.

Throughout the dismasting there were no screaming or loud orders, hence no distraction to the focus of everyone on deck. I believe this lack of panic or additional orders being shouted contributed to the injury free disaster. No one on deck was unaware of where the spars were falling to. However, it remains a mystery as to how no one was struck by the mast shrouds as they fell with the masts. Certainly none of us aboard were watching out for the shrouds. In the moment of silence after all the spars fell, I could see all of us standing frozen with the scene of disaster laid out before us.

As I was formulating what the next step might be, I

noted some of the crew beginning to move as if they were going to try to deal with the disaster independently. I immediately interrupted those intensions with loud orders to muster all hands on deck and amidships. This redirected the attention of all hands from individual “best intentions” towards working within the normal command and control system of coordinated direction stemming from the captain and mates. Once all hands were located and mustered and no injury was revealed, the recovery process began.

Gust magnitudes and anemometer effects (illustrated in figures 10a & 10b): It appears that Pride got hit with almost no warning by three very short duration gusts (Time 2:26:20) over 40 knots (light blue lines) which correspond to the sharp rise in heel angles (red line). The gusts were spaced so that Pride returned to the pre gust heel angle and then was hit by an even more sudden and higher gust speed (time 2:26:40). During this second gust response, the bowsprit iron most likely broke allowing the bowsprit to lift and the forward sails to go over the side (see Appendix B for description of old and new irons.).

- Forensic comparison of the two squall events:

The first “wet” squall build up of forces can be seen as gradual unlike in the second “dry” squall. This extra time provided the opportunity to reduce forces through steering PRIDE closer to the apparent wind, reducing the apparent wind angle and thus reducing the forces of the squall on the sails and rigging. Steering closer also slowed the vessel down which contributed to reducing forces. The wet squall according to the data logging reached a similar intensity to the dry (second) squall but no rigging failure occurred most likely to the combination of reduced angle and slower vessel speed. The jib-topsail was struck right after the greatest heeling angle was experienced.

In the second (dry) squall the forces built up quickly. The first action taken was the same as in the first squall, to steer closer to the wind, but this action came after the first gust struck. Vessel speed was high and took time to bleed off after the helm was adjusted to come closer to the wind. However the wind was veering and the vessel speed did not reduce quickly. In fact vessel speed did not begin to reduce until the second even higher gust was experienced, which corresponds with the failure of the bowsprit.

Captain Miles notes the irony of the results after his description of observed weather risk levels. Both squalls exhibited similar signs and were deemed to be of moderate intensity. The wet squall was dead ahead verses the dry squall being off to the side. The wet squall was sailed into. The resulting experience provided the desire to avoid doing so again. The dry squall was off to the side and Captain Miles observed it to be of little risk, yet it produced much more rapid increase in wind speeds experienced as well significant rapid veering. This irony demonstrates the sailors’ inability to forecast squall wind characteristics with any assurance. CURRENT SLIP RESEARCH PROJECTS Stability Calculation Spreadsheet

The first step in developing a spreadsheet model of the wind heeling analysis was to convert the Pride II lines as drawn in a traditional lines plan to a table of offsets for use in the MaxSurf CAD program. The table of offsets was faired in MaxSurf to produce a 3-D surface (Figure 11)

Figure 11, 3D MaxSurf representation of Pride II

This surface was then analyzed in the Hydromax

stability analysis program to produce tabular curves of form and a tabular set of righting arm curves for an assumed center of gravity (AG) of 10 feet above the baseline.

Figure 12: Effect of KG on GZ for Pride II

These tables were imported into an Excel spreadsheet

(Figures 12, 13 and 14a) and used as a lookup table based on displacement for the AZ table in building the GZ curve as a function of KG. The effect of the small variations in KG is illustrated in Figure 12. The corrected GZ curve is calculated by specifying a delta KG between the assumed pole height of 10 ft. and the assumed KG based on current conditions (12.45 feet, in this example, see Figure 13)). The British MCA LCYC (Large Commercial Yacht Code) criteria (see Appendix A) were set up for Pride II. The WLO (see notation for definition) curve is set to have the same value as the corrected GZ curve at 60 degrees (Pride minimum safe downflooding heel angle is 24 degrees). This WLO criteria value is then calculated at zero degrees heel by dividing Pride’s GZ60 by the Cos^1.3 at 60 degrees value of 0.41 giving WLO0 = 4.5 ft. The LCYC derived wind heel lever, dwhl, is set at one half the WLO value corresponding to a sustained wind speed increase factor of 1.414 (twice the wind pressure which is proportional to the square of the wind speed) in the upright condition.

Although the criteria uses the term “wind gust”, the velocity increase must be sustained sufficiently long for the vessel to settle out at the new equilibrium condition. The authors of this paper were concerned that temporary wind gusts much greater than the 40% used by the criteria are frequently encountered in squalls. In fact wind speed increases on the order of factors of 100% to 500% have been recorded in short duration squalls. Frequently, however, weather cell gusts in isolated squalls are more like several hammer blows rather than a hydraulic ram which forces the vessel over and holds it there until the

wind gust subsides. Pride’s experience is that short duration wind gusts (see figures 9 and 10a) heel the vessel to much less than the equilibrium condition. For example, the quasi-equilibrium heel angle for a wind speed of 41 knots corresponding to the first large gust in both squalls is greater than 60 degrees whereas the maximum actual heel angles were around 33 degrees. (Compare to Deakin, 1991.)

Figure 13, Pride II GZ curve for KG=12.45’ segmented to heel angle ranges relative to the criteria described in this paper. (Figure 16)

In the analysis spreadsheet (figure 14a), the yellow

boxes are the user inputs. On the left side are inputs for the vessel displacement, vertical center of gravity and wind speed data. Individual wind parameters are entered on the lower left of figure 14a. This worksheet lets one choose which sails including the number of reefs in the sails by putting a sail combination number in the yellow cell in Figure 14b. So far, 24 sail combinations have been tabulated, but only sail combination 3, used on Sept 5th, has been analyzed. For PRIDE the sail combinations are defined in Figures 14b, 16 & 17 as: ALL LOWER SAILS: 4LS or M+F+S+J Mainsail (M), Foresail (F), Fore-staysail (S), Jib (J) ALL PLAIN SAILS: APS or 4LS+SQT Mainsail (M), Foresail (F), Square-fore-topsail (SQT), Fore-staysail (S), Jib (J). FULL SAILS: FS or 4LS+SQT+STG+JT+MGT or APS+MGT+JT+STG This represents All Plain Sails plus main gaff-topsail (MGT), Jib-topsail (JT) and Square top-gallant-sail (STG).

Figure 14a – Basic Pride II Stability Analysis Page Extract

Figure 14b Basic Pride II Stability Analysis Page Extract

If any of these sails are reefed, then M1 represents the

main with the 1st reef, M2 represents the mainsail with a 2nd reef and so on, Thus: FS represents full sails, i.e. everything up except studding sails and main ringtail 4LS+SQT1+JT represents all lower sails plus a single

reefed Square-fore-topsail and the Jib Topsail (the combination for Sept 5th data)

M1+F+S+J+SQT1 represents single reefed Main and single reefed Square-fore-topsail with the Foresail, Fore-staysail, and Jib.

4LS+SQT1 represents all lower sails plus a single reefed Square-fore-topsail.

M1+F+S represents a single reefed Main plus the Foresail and Fore-staysail.

M2+F1+S1 represents a double reefed Main and single reefed Foresail and Fore-staysail. And so on.

For each sail combination, the heeling moment contribution from each sail is calculated, summed and wind heeling arms for various apparent wind speeds can then be calculated along with the percent contribution to the total heeling moment. The dynamic pressure used for the wind heeling arms (1/2*ρ*V2 with V in knots) is based on the air temperature and density and the knots conversion appropriate to the measurement of density.

The dynamic pressure times the assumed side force coefficient of 1.2, including both lift and drag for a close hauled sail, gives the wind loading heel coefficient of 0,00405 The appropriate value of the assumed side force coefficient should allow matching of the predicted heel angle with the recorded heel angle. It is assumed that this value will vary with the apparent wind.

Using this inputted data and the vessel’s hydrostatic

and sail information available in lookup tables, the vessels pertinent righting arm and wind heeling arm curves are plotted in figure 14 along with the British MCA criteria (see Appendix A). From these curves, the expected steady state heel angles, gust heel angles, and available reserve stability can be determined. Table 2 is a comparison of the predicted and observed average heel angles for data obtained during close hauled winds on September 5, 2005 using a single value of 1.20 for the side force coefficient. This gives a value of 0.004 for the wind loading heel coefficient which was used by Martin in his wind pressure formula (see Skene, 2001, p 92). The agreement is reasonably good over the measured wind speed range used with this sail combination. Note that the sail combination in use on September 5th was well within the MCA LCYC safety parameters.

Since this analysis is for a single sail combination and

heading (close hauled), additional work is needed to determine other sail combinations and headings. For the next season, the apparent wind direction, wind speed and sail sheeting conditions will be systematically varied for various sail combinations to find a best fit assumed side force coefficient(s).

Table 2: Predicted vs. Observed Wind Heel angles

Development of Wind Gust Characteristics

One of the key wind parameters needed is the characteristics of wind gusts that a vessel may experience. As previously discussed, while gust characteristics for full storms have been investigated, little is currently known about wind gusts during squall conditions. The characteristics and probabilities of occurrence that will need to be determined are the magnitude and duration of the gusts over the base steady state wind speeds, the rate at which those gust speeds escalate (rapid/gradual), the magnitude of wind directional shifts, and the speed at which those shifts occur (rapid/gradual/swirling).

An additional wind gust characteristic that needs to be investigated is “mircobursts at sea”. These phenomena, while well known for their effects on airplanes, are not well known for sailing vessels. And since they occur with such unpredictability and unknown intensity, they are of particular danger to sailing vessels. The original Pride of Baltimore was likely a victim of being in close proximity to the water impact of a sustained mircoburst. In her case eye witness survivor accounts clearly show that there was 100% cloud cover at the time of the “white squall” gust that capsized her. Her eye witness accounts also clearly describe a sky that did not depict any concentrated darkness suggesting heaviness of potential rain. They had not experienced rain for some hours before capsize. The survivors gave a description of a “white wall” of water (spume) advancing on them from windward. This appears to actually be the leading edge of the gust front lifting sea water into the air.

The last wind gust parameter needed is the likelihood (or risk) that a given wind gust will occur. This will require the development of predictive methods for squalls and other wind gust events based on common weather data that can be determined by gathering much data at sea by vessels equipped with Doppler Weather Radar. New Stability Guidance Presentation Methods

With the new stability analysis methods, reams of highly technical data on a vessel’s characteristics will now be known to the naval architect. This technical data though, is basically indecipherable to most sailing vessel masters. GM, LCB, GZ, or CLP, while the language of naval architects, is going to require a translation to be useful by the master. It is the presentation method, i.e. “Stability Letter” that should serve as that translator between the naval architect and master.

The basic requirements for any presentation method are relatively simple. 1. Be written to provide stability guidance, not to dictate

the boat’s operation. 2. Present the safe sailing conditions clearly, both

visually and written.

3. Provide some means for conveying the risks for various sail combinations.

4. Be comprehensible by crews with little or no formal training. In summary, the goal for any presentation method is to

provide the captain with practical stability guidance and a way to gauge the risks based on sails set, weather, and other factors, and let them run their commands. Color Coded Safe Sail Combination Matrixes: Downflooding Risk Assessment

Several methods are currently being proposed to meet these requirements. The first method is to use color coded matrixes similar to ones being developed for the small commercial fishing vessel fleet. Interestingly, sailing and fishing vessels share similar requirements for presenting stability guidance to their respective crews. It was in fact Captain Miles seeing a SNAME developed prototype color

coded loading matrix for fishing vessels, see Figure 15, that lead him to follow up with SNAME Panel O-49, Small Working Vessel Operations and Safety, with his initial proposal for the stability letter improvement program (SLIP).

These risk based loading matrixes, particularly the

color versions, have been readily accepted by the crews because of the many advantages they offer. First the color gives a very quick intuitive indication of the current risk of capsize or downflooding for any conceivable loading condition. Second, the matrixes allow the crew to plan ahead to ensure adequate stability. With all of the loading conditions on a single sheet, the crew can literally plot their trip on the load matrix and adjust loading, ballast, or fuel levels to suit expected weather conditions. These attributes of the color coded loading matrixes allow a fishing vessel crew to maximize fishing effort (income) while safely operating their vessel.

Figure 15 - Weather Dependent Safe Loading Matrix (Womack 2002)

The similarity between sailing vessels and fishing

vessels is the goal to maximize the sailing experience for the passenger (income) while maintaining safe operations. If a similar safe sail handling matrix for sailing vessels was available, this goal could be advanced.

The format developed for Pride II may be usable for vessels with complex rigs such as those with multiple reefs or mixed square and fore/aft rigged sails. For these vessels, a possible prototype matrix based on the MCA downflooding criteria is illustrated in Figure16. This figure is for close hauled sailing only and additional diagrams will be needed for beam reaching and broad reaching.

Figure 16 displays nearly all the “gears” of choice to pick from in a rising or decreasing wind circumstance while open water sailing. The assumption in the gear order is that PRIDE is sailing upwind at her most efficient with every expectation she will be continuing to sail upwind for the foreseeable future. The sequence is established with respect to weather helm (see reduction between #1 & #2) followed by strength of the lighter spars and associated rigging. For example, the reduction between #2 & #3 is to reduce the foretopmast strain by reefing the square-foretopsail while leaving the Jib-Topsail flying from the foretopmast. This preserves for as long as possible the “fore-triangle” for overall power to pull PRIDE ahead. Eventually conditions become too strong for even the jib, because of the design of the traditionally rigged supporting structure (i.e. deadeye and lanyards which continue to stretch as load increases).

Figure 16 – Close-Hauled Downflooding Avoidance Diagram for Pride II

For the matrix format in Figure 16, the color coding is used to indicate the level of risk present. The basic colors are proposed to indicate the following risk levels. Green - Indicates a angle of heel below deck edge

submergence for Pride II Yellow - Indicates the deck edge is likely to be submerged

but the heel angle remains in the safe zone established by the MCA code (below 24 degrees of heel for Pride II)

Orange – Indicates a heel angle greater than the MCA code limit but less than 29 degrees for Pride II which corresponds to bulwark immersion.

Red - Indicates an angle of heel corresponding to between 84% and 100% of GZ max for Pride II and results in a high level of weather related risk from downflooding through hatches

Black – Indicates sailing at heel angle at or beyond GZ max (must be avoided to prevent serious risk to the vessel)

For all color zones, the position in each color band

also provides an additional indication of the risk. For example if winds and waves are steady and there is little risk of squalls, operate at the top of the green band to maximize the passengers sailing experience. However, if

the winds are gusty or there is a significant risk of squalls, operating down in the middle or lower portion of the green band would be advisable if passengers are onboard. For crew only though, the master could operate near the top of the green band to maximize speed as the crew would be better able to handle reducing sail in gusts. Color Coded Safe Sail Combination Matrixes: Passenger Comfort Risk Assessment

Using the Day Sail Combinations in Figure 14b, Figure 17 was created which represents a possible operator guideline for passenger comfort. It presents suggested limiting heel angles for various sail combinations as a function of wind velocity when close hauled. Additional diagrams will be needed for beam reaching and broad reaching.

“Comfort” sailing parameters are established recognizing crew effort and passenger experience reward rather than detailed weather conditions. Day sails are often 3 hours or less. Unless weather prevents a short day sail from occurring, sail combination is chosen based on the reason for the day sail. For public relations more sail is better than less because of photographic or TV news interests.

Figure 17 Close hauled passenger comfort risk assessment for Pride II

The comforts heel limits depend on the client populations, but based on experience with Pride II,

Green = Heel less than 8 degrees Yellow = Heel less than 12 degrees but more than 8

degrees and so on as the legend of Figure 17 indicates.

When day sailing does not involve public relations concerns there are two other client populations that help the master decide on the combination to sail with. For sailing amateurs/enthusiasts there is reason to set combination #17 (see Figure 17 for sail definitions) because it is as full a sail plan that can be set without resorting to the “kites” which add significant sail stowing work for the crew at the end of the day. For those day sails that are corporate based, especially if there will be food and drink dispensed, #18 is the most efficient for providing a sail that is not engine assisted yet not spill or tip over the guests or refreshments (guests often come dressed in street cloths to impress their corporate host). PRIDE will tack to windward on smooth water with all of these sail combinations, even in very light air. Combination #18 also reduces the work of the crew who often are stowing sail well into the evening after a day that starts at 0800.

Combinations #19 & #20 are “pinch hit” solutions to

weather conditions that are just a little bit threatening to the sense and sensibilities of non pure sailing interests or maybe even for those wilder days when purists are paying for the thrill of sailing aboard a traditional vessel. The drawback to #20 is not using the square-topsail in either

full or reefed condition. Most of those day sailors interested in having a sail aboard PRIDE are drawn by the square-rigged aspects of her rig. It must also be noted that day-sailing sail combinations are most often chosen before getting underway. Thus, in the interests of time, even with a full crew aboard (due to Pride’s full ocean service crew size of 12, day sails can often be sailed with half crew and thus provide regular time off), reefing takes too long to make fine adjustments after getting underway. Therefore, any post getting underway adjustments usually take the form of gross single or multiple sail reduction (or set) in the interests of time. Conclusions/Summary

Our investigation to date has shown that useful operator guidance for passenger carrying sailing vessels can be generated from a combination of knowledge of vessel stability characteristics, a reasonably accurate sail plan, and a limited amount of full scale data on wind heel as a function of wind velocity and direction. Using the Large Commercial Yacht Code (UK) as a starting point for downflooding risk analysis, operator guidance in terms of expected heel angle for various close hauled apparent wind speeds was generated for various sail combinations. The same analysis with different passenger comfort criteria was also generated for day sailing combination commonly used on Pride of Baltimore II. The full scale data will be expanded during 2007 to include beam reaching and stern quartering winds in the analysis.

Once this first phase of the SLIP project is completed, another traditional sailing vessel will be analyzed on a similar spreadsheet and the wind heel predictions will be checked with full scale measurements, insuring that the analysis goes beyond “making the data fit the crime”.

In addition, the project clarified several problems in full scale data experiments that will need to be addressed in future work.

1. It is very important to insure the wind anemometer operates in an area where flow restrictions caused by the sail interactions are minimized.

2. For time domain analysis, the time constants for the various instruments should be adjusted to be as close as possible.

3. The full scale data appeared to validate the cos^1.3 power used by the LCYC to correct the heeling force as a function of roll angle as opposed to the Cos^2 function used by the USCG regulations.

4. Much more probabilistic data is needed for predicting the magnitude and duration of small scale gusts generated by both wet and dry wind squalls. This data is being collected by land based weather radars, but almost no sea based data is available.

Acknowledgements

The authors wish to thank the SNAME Technical and Research (T&R) Steering Committee for sponsoring this research and Formation Systems for the kind loan of their MaxSurf Software used in this paper. Thanks also to SMC Corporation for the loan of the motion sensor and to Tri-Coastal Marine for supplying Figures 20 and 21. The authors also wish to thank those who reviewed this paper; especially Barry Deakin and Roger Long. They also acknowledge the support of the Pride of Baltimore II Foundation for letting us use Pride II as a vessel of opportunity for the experiments. References Chappelle, Howard, 1973. The American Fishing Schooners 1825-1935, WW Norton Co, NY Deakin, Barry, 1991 Model Test Techniques Developed to Investigate the Wind Heeling Characteristics of Sailing Vessels and their Response to Gusts, 10th Chesapeake Sailing Yacht Symposium, Deakin, Barry, 1990 The Development of Stability Standards for UK Sailing Vessels, Transactions of RINA, 1991 Gillmer, T, 1989 The Design and Construction of the Second Pride of Baltimore, 9th Chesapeake Sailing Yacht Symposium Grant H. and Stephens, O. J. 1997, On Test Measurement in Full Scale Sailing Test Programs, 13th Chesapeake Sailing Yacht Symposium. Grant. H. et. al, 2001 Schooner Brilliant Sail Coefficients and Speed Polars, 15th Chesapeake Sailing Yacht Symposium Skene, N. 2001 Elements of Yacht Design, Sheridan House, 2001 edition. Womack, J., 2002, Small Commercial Fishing Vessel Stability Analysis, Where are We Now Where are We going, Proceedings of the 6th International Ship Stability Workshop, Webb Institute, 14-16 October 2002 and Marine Technology, Vol 49, No 4,SNAME, October 2003. Appendix A. Brief summary/review of current sailing vessel stability criteria and guidance methods. 1 U.K. LARGE COMMERCIAL YACHT CODE (MCA, UK)

Large is 24 metres and over in load line length and the Code of Practice applies to yachts which are in commercial use for sport or pleasure, do not carry cargo and do not carry more than 12 passengers. 11.2.2 Sailing vessels 11.2.2.1 Monohulls .1 Curves of statical stability (GZ curves) for at least the Loaded Departure with 100% consumables and the Loaded Arrival with 10% consumables should be produced. .2 The GZ curves required by .1 should have a positive range of not less than 90o. For vessels of more than 45m, a range of less than 90o may be considered but may be subject to agreed operational criteria. .3 In addition to the requirements of .2, the angle of steady heel should be greater than 15 degrees (see Figure18). The angle of steady heel is obtained from the intersection of a "derived wind heeling lever" curve with the GZ curve required by .1. In the figure:- 'dwhl' = the "derived wind heeling lever" at any angle θ° = 0.5 x WLO x Cos¹.³θ where WLO = GZf / Cos¹.³θf

Figure 18, Large Commercial Sailing Vessel Code (UK) Noting that:- WLO is the magnitude of the actual wind heeling lever at 0° which would cause the vessel to heel to the 'down flooding angle' θf or 60° whichever is least. GZf is the lever of the vessel's GZ at the down flooding angle (θf) or 60° whichever is least. θd - is the angle at which the 'derived wind heeling' curve intersects the GZ curve. (If θd is less than 15° the vessel will be considered as having insufficient stability for the purpose of the Code). θf - the 'down-flooding angle' is the angle of heel causing immersion of the lower edge of openings having an aggregate area, in square metres, greater than:- Δ/1500 where Δ = vessels displacement in tonnes

All regularly used openings for access and for ventilation should be considered when determining the

downflooding angle. No opening regardless of size which may lead to progressive flooding should be immersed at an angle of heel of less than 40°. Air pipes to tanks can, however, be disregarded.

If as a result of immersion of openings in a superstructure a vessel cannot meet the required standard those superstructure openings may be ignored and the openings in the weather deck used instead to determine θf. In such cases the GZ curve should be derived without the benefit of the buoyancy of the superstructure.

It might be noted that provided the vessel complies with the requirements of 11.2.2.1.1, 11.2.2.1.2 and 11.2.2.1.3 and is sailed with an angle of heel which is no greater than the 'derived angle of heel', it should be capable of withstanding a wind gust equal to 1.4 times the actual wind velocity (i.e. twice the actual wind pressure) without immersing the 'down flooding openings', or heeling to an angle greater than 60°. (The value of 1.4 used for gusts is not an MCA figure. It was derived from work conducted by meteorologists and oceanographers: "Spectra and Gust Factors for Gale Force Marine Winds” by Smith & Chandler, Boundary-Layer Meteorology vol 40, pp 393-406, 1987, Reidel Publishing Company) 2. European STIX stability criteria. http://rorcrating.com/stix/stix.htm a. Provides minimum stability levels for different sea conditions. b. Not designed for traditional sailing vessels. 3. USCG CFR 46 http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?c=ecfr&sid=47f7010c7c768 § 171.055 Intact stability requirements for a monohull sailing vessel or a monohull auxiliary sailing vessel. (a) Except as specified in paragraph (b) of this section, each monohull sailing vessel and auxiliary sailing vessel must be shown by design calculations to meet the stability requirements in this section. (b) Additional or different stability requirements may be needed for a vessel of unusual form, proportion, or rig. The additional requirements, if needed, will be prescribed by the Commandant. (c) Each vessel must have positive righting arms in each condition of loading and operation from— (1) 0 to at least 70 degrees of heel for service on protected or partially protected waters and (2) 0 to at least 90 degrees of heel for service on exposed waters. (d) Each vessel must be designed to satisfy the following equations:

(1) For a vessel in service on protected or partially protected waters

where— X=1.0 long tons/sq. ft. (10.9 metric tons/sq. meter). Y=1.1 long tons/sq. ft. (12.0 metric tons/sq. meter). Z=1.25 long tons/sq. ft. (13.7 metric tons/sq. meter). (2) For a vessel on exposed waters— HZA, HZB, and HZC are calculated in the manner specified in paragraph (e) or (f) of this section.

where X=1.5 long tons/sq. ft. (16.4 metric tons/sq. meter). Y=1.7 long tons/sq. ft. (18.6 metric tons/sq. meter). Z=1.9 long tons/sq. ft. (20.8 metric tons/sq. meter). A=the projected lateral area or silhouette in square feet (meters) of the portion of the vessel above the waterline computed with all sail set and trimmed flat. Sail overlap areas need not be included except parachute type spinnakers which are to be added regardless of overlap. H=the vertical distance in feet (meters) from the center of A to the center of the underwater lateral area or approximately to the one-half draft point. W=the displacement of the vessel in long (metric) tons. (e) Except as provided in paragraph (f) of this section, HZA, HZB, and HZC must be determined as follows for each condition of loading and operation: (1) Plot the righting arm curve on Graphs 171.055 (b), (c), and (d) or (e). (2) If the angle at which the maximum righting arm occurs is less than 35 degrees, the righting arm curve must be truncated as shown on Graph 171.055(a). (3) Plot an assumed heeling arm curve on Graph 171.055(b) that satisfies the following conditions: (i) The assumed heeling arm curve must be defined by the equation—HZ=HZA cos^ 2 (T) where— HZ=heeling arm. HZA=heeling arm at 0 degrees of heel.

T=angle of heel. (ii) The first intercept shown on Graph 171.055(b) must occur at the angle of heel corresponding to the angle at which deck edge immersion first occurs. (4) Plot an assumed heeling arm curve on Graph 171.055(c) that satisfies the following conditions: (i) The assumed heeling arm curve must be defined by the equation— HZ=HZB cos ^2 (T) where— HZ=heeling arm. HZB=heeling arm at 0 degrees of heel. T=angle of heel. (ii) The area under the assumed heeling arm curve between 0 degrees and the downflooding angle or 60 degrees, whichever is less, must be equal to the area under the righting arm curve between the same limiting angles. (5) Plot an assumed heeling arm curve on Graph 171.055 (d) or (e) that satisfies the following conditions: (i) The assumed heeling arm curve must be defined by— HZ=HZC cos 2 (T) where— HZ=heeling arm. HZC=heeling arm at 0 degrees of heel. T=angle of heel. (ii) The area under the assumed heeling arm curve between the angles of 0 and 90 degrees must be equal to the area under the righting arm curve between 0 degrees and— (A) 90 degrees if the righting arms are positive to an angle less than or equal to 90 degrees; or (B) The largest angle corresponding to a positive righting arm but no more than 120 degrees if the righting arms are positive to an angle greater than 90 degrees. (6) The values of HZA, HZB, and HZC are read directly from Graphs 171.055 (b), (c), and (d) or (e). (f) For the purpose of this section, the downflooding angle means the static angle from the intersection of the vessel's centerline and waterline in calm water to the first opening that cannot be rapidly closed watertight. (g) HZB and, if the righting arms are positive to an angle of 90 degrees or greater, HZC may be computed from the following equation:

where— I=the area under the righting arm curve to— (1) the downflooding angle or 60 degrees, whichever is less, when computing HZB; or (2) the largest angle corresponding to a positive righting arm or 90 degrees, whichever is greater, but no greater than 120 degrees when computing HZC.

T=the downflooding angle or 60 degrees, whichever is less, when computing HZB or 90 degrees when computing HZC. § 171.057 Intact stability requirements for a sailing catamaran. (a) A sailing vessel that operates on protected waters must be designed to satisfy the following equation:

Where— B=the distance between hull centerlines in meters (feet). As=the maximum sail area in square meters (square feet). Hc=the height of the center of effort of the sail area above the deck, in meters (feet). W=the total displacement of the vessel, in kilograms (pounds). X=4.88 kilograms/square meter (1.0 pounds/square foot). (b) A sailing vessel that operates on partially protected or exposed waters must be designed to satisfy the following equation:

Where— B=the distance between hull centerlines in meters (feet). As=the maximum sail area in square meters (square feet). Hc=the height of the center of effort of the sail area above the deck, in meters (feet). W=the total displacement of the vessel, in kilograms (pounds). X=7.32 kilograms/square meter (1.5 pounds/square foot). Appendix B: Comparison of Old and New Bobstay Iron fittings: Bobstays are the rigging spanning between the stem at the waterline and the outer end of the bowsprit (Figure 19), rigged so as to prevent the bowsprit from being pulled upward through the forces of the forestays when the jibs and stay sail are set. Aboard PRIDE II the connection of the bobstays to the bowsprit is in way of metal fittings fashioned to fit around the bowsprit and receive the upper dead-eye metal band via a bolt passing through metal tangs in the bowsprit iron.

Figure 19 Bobstay irons on bowsprit attached to bobstays. The original bowsprit iron design was of a discontinuous ring with bent-out tangs, as can be seen in Figure 20 below.

Figure 20: Old bobstay iron (actually now steel) which fractured

Figure 20 illustrates the concentration of loading

under strain. There was no early sign of impending failure after 16 years of sailing. The failure occurred in year 17 in an area of high stress concentration.

The re-designed bobstay iron is a full un-broken ring

band that encircles the bowsprit as can be seen in Figure 21.

Figure 21: New bobstay iron for the bowsprit

The new bobstay iron comes as a result of stress analysis of the first iron. The improvements were in making the iron a full uninterrupted ring including lengthening the fore & aft dimension plus doubling the ring thickness were the bobstay attaches. What is interesting about this redesign is the similarities it has to bowsprit irons of the Atlantic Cod Fishing Schooners circa 1890 to 1920 (see Chappelle, 1973, pages 375-380, ).

Traditionally speaking, circa 1812 Baltimore

Schooners would have used rope stropping to make this attachment. Modern wire rope is a full strength substitute and fully adequate to withstand PRIDE’s squall experience. Substitution by a non traditional metal banding was driven by the high level of maintenance required by wire rope.