Fig 1.
The HL Hunley as it would have appeared in attack position on the evening of February 17, 1864.
Image courtesy Michael Crisafulli of The Vernian Era. More renderings and details of the construction of the Hunley can be found at http://www.vernianera.com/Hunley/.
Fig 2.
Photograph of the scale Hunley model, nicknamed the CSS Tiny.
[a] threaded attachment for spar [b] access port (2 total, one each at bow and stern) to fill and empty the ballast tanks, can be sealed with threaded insert [c] Rings (3 on model) for carrying the vessel and attaching lines [d] Gasket-sealed panel for interior access [e] Data ports (2 on model) for gauges [f] Bulkhead fittings (4 on model) for gauge wires.
Fig 3.
Diagram of the Hunley and charge, illustrating the steps of the analyses using relative equivalency (not to scale).
Analysis was performed both starting with known black powder equivalencies and calculating risk to the crew, and also by starting with the minimum internal pressure required to cause fatality and calculating necessary TNT equivalency.
Fig 4.
Representative curves of initial blast wave from black powder charges of two sizes.
The open circle indicates the peak of the wave, and the grey circle indicates one time constant of decay. The 1 kg charge was at a range of 1.8 m with a time constant of 265 μsec. The 490 g charge was at a range of 0.8 m with a time constant of 408 μsec.
Fig 5.
Scaled time constant as a function of scaled distance.
The data for black powder show the power law trend consistent with other known explosive types.
Fig 6.
Representative waveform showing transmission into the scale model.
This test had a shock tube orientation perpendicular to the side of the scale model. [a] Waveform in water [b] Waveform inside the boat hull.
Fig 7.
Blast transmission into the model hull of 1 kg charge on a spar.
The circle in the lower panel indicates the selected point of peak pressure (defined as peak pressure achieved with <2 ms local rise time).
Fig 8.
Waveforms transmitted through the 1.6 cm (5/8”) thick mild steel plate.
Shock tube waveforms have been overlaid with the ‘external’ incident pressure Friedlander curves produced by these shock tubes in this configuration without the reflection back into the tube from the steel plate. Gauge 1 in the 140 mil tests was dislodged during the test and shows some spike-shaped anomalies between 2–5 ms.
Fig 9.
Ratio of peak pressure propagating through the wall [24, 47].
Dashed lines show 95% confidence intervals. The vertical line indicates the βs value calculated for the Hunley explosion. The calculated transmitted impulse is shown in grey for reference; axes selected for consistency with previous works [22, 24, 26, 47, 48].
Fig 10.
Risks of injury and fatality for a range of TNT relative equivalencies and transmission levels.
Grey rectangles indicate the range of physically reasonable values (RE [0.24, 0.46]; transmission [6.3%, 10.5%]) (a) Risk of pulmonary fatality [50] (b) Risk of pulmonary injury [50] (c) Risk of fatality from traumatic brain injury [36]. The exposure to the Hunley crew (closed circle) was calculated using a 8.4% transmission level and RE = 0.43, the median value found in the literature (see Methods). This exposure is a low estimate because 8.4% was the transmission occurring at the low experimental overpressure ratio of 3.4, and rate of transmission would continue to increase up to the Ps/P0 = 68 expected of the full-sized explosion (value of 68 justified in the Discussion).