Introduction

In the USA, crude oil is shipped along waterways by means of barges and other vessels, moving between 35 million barrels in 2017 and 93 million barrels in 2018 [1]. As the regulating agency, the United States Coast Guard (USCG) is responsible for monitoring waterways, which involves inspecting and certifying vessels like barges as seaworthy [2]. Due to the hazardous nature of crude oil from previous spills along American waterways, prevention and mitigation of spills is a primary goal of securing marine infrastructure [3, 4]. This is often accomplished through frequent inspections, focused on corrosion, which typically manifests as pitting in local attacks or widespread uniform attack on bulk structures [5], also known as “wastage” in the industry [6,7,8,9]. In 2017, the USCG performed 5758 inspections on vessels classified as barges, with 1518 deficiencies found; that same year 274 major marine casualties occurred involving barges [10].

When a major marine casualty occurs, the National Transportation Safety Board (NTSB) and USCG investigate. The NTSB is an independent Federal agency charged by Congress with investigating every civil aviation accident in the USA and significant accidents in other modes of transportation—railroad, highway, marine, pipeline, and hazardous materials. While mostly known for investigations of aircraft accidents, in the case of marine accidents, the NTSB works with the USCG to investigate those classified as “major marine casualties,” defined as:

  • The loss of six or more lives.

  • The loss of a mechanically propelled vessel of 100 or more gross tons.

  • Property damage initially estimated as $500,000 or more.

  • Serious threat to life, property, or the environment due to hazardous materials.

The accident vessel in the investigation detailed here was an articulated tug and barge (ATB) carrying crude oil. An ATB consists of an independent tugboat (in this case the Buster Bouchard) that connects into a specially shaped stern notch of a cargo-carrying barge (B. No. 255). When connected, the tugboat and barge operate and navigate as a single unit (see Fig. 1). Although an ATB configuration operates as one unit, the tug and the barge each have their own crew.

Fig. 1
figure 1

Overhead view of an ATB configuration (Photograph courtesy of Bouchard Transportation Co., Inc.)

Full size image

On October 20, 2017, at 0430 local time, the crews of the ATB Buster Bouchard/B. No. 255 were preparing to get under way from anchorage to proceed into the Port of Corpus Christi, TX, when an explosion and subsequent fire occurred on the bow of the barge (see Fig. 2). The two crew who were on the bow of the barge were killed in the explosion. Approximately 2000 barrels (84,000 gallons) of crude oil were released either into the water or were consumed in the fire, which was not extinguished until 6.5 h later. The barge sustained over $5 million in damage and was eventually scrapped after the accident.

Fig. 2
figure 2

Barge B. No. 255 bow on fire (Photograph by US Coast Guard)

Full size image

In the morning of the accident, the two crew were on the deck of the barge’s bow to recover the anchor. According to the mate watching from the tug, the barge crew reported the anchor was frequently under “heavy strain” during recovery. During anchor recovery, the mate reported seeing a flash out of his peripheral vision. As he looked forward, he saw blue flames on the bow around the area of the winch, immediately followed by an explosion. Unable to see the barge crew on the bow, the mate sounded the tug’s general alarm and attempted to call the barge crew on the handheld radio, receiving no answer. A second and third explosion followed, causing further damage to the bow of the barge, and a fire erupted in the area of the explosions. The crew of the Buster Bouchard, fearing further explosions, disengaged the tugboat from the damaged barge.

The 468-foot-long barge B. No. 255 was built in 1979 and was extensively modified with the addition of a double hull in 1999. The barge had a total cargo capacity of 188,400 barrels (7.91 million gallons) in its 16 cargo tanks, which included a “forepeak area” in the bow of the barge (see Fig. 3). The forepeak area was divided into four sections and was classified as a void space, meaning it was not intended or designed to hold liquid cargo. Rather, this compartment held equipment such as the anchor chain locker and electrical control box for its winch. The barge crew did not routinely check this area for leaks.

Fig. 3
figure 3

Simplified plan of the Barge B. No. 255 structure

Full size image

Investigation Methods

The on-scene images were photographed with by US Coast Guard personnel. The extracted specimens at the NTSB Materials Laboratory were photographed using a Canon EOS Rebel T3i digital camera. Thickness measurements were taken on a portion of the sectioned bulkhead specimen above the Frame 1 weld using a Krautkramer Branson DME DL ultrasonic thickness gage. The bulkhead section was examined and scanned using two FARO Focus 3D 330 laser scanners in the as-received condition.

This bulkhead was cut using a plasma torch cutter, and then later an emulsion-cooled vertical bandsaw. Later sections were performed using water-emulsion-cooled abrasive cut-off saws. Sectioned specimens produced from the extracted bulkhead were examined using a Keyence VHX-1000 digital microscope, which can produce high-dynamic range and composite high depth-of-field images. A Zeiss Auriga field emission scanning electron microscope (SEM) was used for further microscopic examination. This microscope was equipped with a Thermo Scientific UltraDry NORAN System 7 energy-dispersive X-ray spectroscopy (EDS) used for standardless, semiquantitative analysis.

Several metallographic cross section specimens of the bulkhead and welds were mounted, polished to 0.05 μm grit, and etched using a 2% Nital solution. These specimens were inspected using a Zeiss Axio Observer Z1m inverted microscope.

The hardness was inspected per ASTM E18 and was measured on mounted, polished cross sections in the longitudinal and circumferential directions, using a Wilson RB2000-T hardness tester [11]. The weld cross sections were examined using a LECO AMH43 automated microindentation hardness tester, performed per ASTM E384 [12].

Additional testing was provided by Lehigh Testing Labs in New Castle, DE. This included optical emission spectroscopy with carbon, sulfur, and nitrogen analysis, and mechanical tensile and subsize sample (5 mm × 10 mm) Charpy impact testing per ASTM A370 at 20 °C [13]. Four specimens of the surface corrosion near through-cracks were taken from the forward and aft sides of the bulkhead section. The specimens were inspected using powder X-ray diffraction (XRD) to determine the crystalline compound phases present in the surface corrosion.

As the mechanical testing data were all within typical values for the alloys found in this investigation, they will not be discussed in this manuscript.

Results of the Investigation

The barge had a variety of compartments, as illustrated in Fig. 3. Of note in this investigation were the no. 1 port and starboard cargo tanks, each holding 6700 barrels (281,400 gallons) of crude oil, and the empty forepeak compartments forward. As shown in Figs. 2 and 4, the blasts caused severe damage to the bow of the B. No. 255, primarily to the port side of the forepeak and to the no. 1 port cargo tank, producing a large open hole in the deck.

Fig. 4
figure 4

Explosion and fire damage to port bow of B. No. 255 (Photograph by Coast Guard)

Full size image

During a postaccident on-scene examination of the barge on December 7, 2017, NTSB and USCG personnel inspected an area of the bulkhead between the port side forepeak and the port side first cargo hold (#1P) that exhibited corrosion and horizontally oriented cracks (see Fig. 5). According to records submitted by the classification society, this area had peripheral sections of the bulkhead removed and inserts that were welded into place in April 2008. These welded inserts or patches were placed above and inboard of the original bulkhead. The large middle portion of the bulkhead, which had not been repaired in this manner, exhibited cracks over 33 cm (13 inches) in length (see Figs. 6 and 7). These cracks were oriented horizontally and were located 4.5 m (15 ft.) from the bottom of the tank and about 5 cm (2 inches) above transverse frames welded to the forepeak side of the bulkhead. A 1-meter-by-1.25-meter section of the original bulkhead, including a part of these welded inserts, was torch-cut, removed, and shipped to the NTSB Materials Laboratory for further examination (annotated in Fig. 5).

Fig. 5
figure 5

The port side bulkhead, looking aft, as observed on the oil barge, after the accident (overexposed metallic scaffolding). The dashed section was torch-cut and sent to the NTSB Materials Laboratory

Full size image
Fig. 6
figure 6

Annotated version of the forward face of the bulkhead section, showing noted features such as cracks, corrosion, and welds

Full size image
Fig. 7
figure 7

Annotated version of the aft face of the bulkhead section, showing noted features such as cracks, corrosion, and welds

Full size image

Figure 6 shows the section that was torch-cut from the bulkhead with the different welded pieces diagramed. The sectioned specimen consisted of a large piece of the original bulkhead, which comprised most of the specimen, located toward the center right and bottom areas of the figure. Two pieces of material later welded to the bulkhead prior to the accident were located on the top and inboard side of the section (left side of Fig. 6).

Two horizontal L-shaped frames had been welded across the original bulkhead and later the inboard welded insert. Pieces of gusset plate remnants had been welded to the frames on the inboard insert corner. The frame sections were labeled Frame 1 (upper frame section) and Frame 2 (lower frame section). Above the frames, there were two horizontally oriented cracks, highlighted in light blue in Fig. 6. Toward the port (or outboard) side of these cracks were additional areas of through-corrosion, denoted in green in Fig. 6.

For the purposes of this report, the horizontal through-cracks on the specimen were labeled Crack 1 and Crack 2, located above Frame 1 and Frame 2, respectively. Both cracks exhibited similar morphologies, with most of the crack lengths oriented horizontally. The outboard-most part of each crack was oriented 45° in an upward direction, toward the port side. The inboard-most part of each crack was oriented 45° in the opposite downward direction, toward the starboard side. The inboard-most part of each crack turned toward vertical when approaching the vertical weld between the inboard insert and the original bulkhead. The horizontal length of Crack 1 was 35 cm (13.75 inches), with the angled outboard segment measuring 8.9 cm (3.5 inches). The horizontal length of Crack 2 was 34 cm (13.5 inches) long and the outboard segment measured 5.7 cm (2.25 inches).

Figure 7 shows the aft face of the sectioned bulkhead specimen, as received, with the location of the welds, cracks, and other through-corrosion annotated. The aft face would have faced the #1P cargo hold compartment, which typically contained crude oil during operation of the barge prior to the accident. The horizontal and vertical welds between the inserts and the original bulkhead (as well as between the inserts) are annotated yellow in Fig. 7. The frame welds are annotated in blue, noting that these fillet welds did not exhibit any through-wall penetration to the aft side.

Stray horizontal welds were present on the aft and forward sides of the bulkhead and one of the frames—they exhibited multiple lobes, consistent with multiple passes performed when welded. Their surfaces exhibited circular-shaped steps, consistent with having been torch-cut.

Figure 8 shows a closer view of Crack 2 from the forward and aft sides, respectively. This crack and Crack 1 exhibited comparable lengths and morphologies. While oriented horizontal, the cracks themselves exhibited a rounded sawtooth shape.

Fig. 8
figure 8

Crack 1 above Frame 1, viewed from the aft side, as received

Full size image

The additional through-corrosion was present on the outboard side, manifest as holes in the center of thinned material above Frame 1, but as small cracks above Frame 2. The lengths of these additional cracks above Frame 2 varied between 0.63 and 1.9 cm (0.25 and 0.75 inches). Figure 9 shows closer views of these smaller cracks above Frame 2, viewed from the forward side.

Fig. 9
figure 9

Closer view of the two inboard-most cracks above Frame 2

Full size image

The sectioned pieces around Crack 1 and Frame 1 are shown in Fig. 10, arranged as oriented prior to sectioning. The inserts, frames, original bulkhead, and welds are labeled in Fig. 10. Figure 11 shows a cross section view through the horizontal weld in Fig. 10. The insert thicknesses were greater than that of the bulkhead. The original bulkhead thinned as one moved away from the welds. Both the horizontal and vertical welds themselves were thicker than the joined plates. The cross sections of the original bulkhead in this area revealed that while the aft surface was relatively flat, the forward surface exhibited a pock-marked and pit-shaped morphology (Figs. 9, 10, and 12).

Fig. 10
figure 10

Additional bandsaw sections of the upper left portion of the bulkhead section

Full size image
Fig. 11
figure 11

View of the bandsaw cut cross section across the horizontal weld in Fig. 10

Full size image
Fig. 12
figure 12

View of the lower half of Crack 1, showing an area that thinned to a small point

Full size image

The crack above Frame 1 is highlighted in Fig. 12, showing an area where the metal thinned to a sharp edge. Locally, this thinned edge was random and tortuous in orientation and shape. Multiple areas exhibited this local thinning along the length of the crack. The forward side of the bulkhead adjacent to the crack exhibited the pock-marked surface condition common over the rest of the original bulkhead. There were also grooves of additional thinning perpendicular to the crack. The surface material near the crack was consistent with surface corrosion as well as residue from crude oil.

Figure 13 shows a typical view of the thinned edge of Crack 1. In the thinnest crack edges, the bulkhead material tapered to approximately 20 μm. The edge itself exhibited rounded, ball-shaped features, consistent with corrosion of the underlying metal. Much of the forward and aft bulkhead faces adjacent to the crack edge exhibited pit-like surface features. All the surfaces examined exhibited characteristics and artifacts (including charging in the electron microscope) consistent with surface corrosion.

Fig. 13
figure 13

Secondary electron (SE) micrograph of the thinnest portion of the Crack 1 (upper face)

Full size image

On thicker sections of the crack edge, the surface corrosion-product layer had spalled, revealing the underlying metal surface. Where spalled in the manner described above, the fractured surface oxide, consistent with iron oxide, typically exhibited an elongated faceted morphology. These features were consistent with a columnar grain structure of the surface oxide (Fig. 14). These features, mirrored on the fractured regions of the oxide layer, were consistent with the underlying grain structure exposed from spalling of the outer layer.

Fig. 14
figure 14

SE micrograph of a higher-magnification view of Fig. 13

Full size image

Some portions of the thicker crack edges exhibited localized plastic deformation near the tip. In areas where the fracture surface was not damaged, dimpled rupture could be observed, as typified in Fig. 15. These fracture features were consistent with overstress fracture. The surface areas near these fracture features exhibited regions of cracked and spalled surface oxide. On these thinnest areas of the crack edge, no other features were present that would be consistent with features typical of a fracture surface (such as dimpled rupture, fatigue striations, or cleavage facets).

Fig. 15
figure 15

SE micrograph of dimpled rupture on the fracture surface of a thicker area of the crack

Full size image

One of the thinnest areas of the Crack 1 edge was cross-sectioned and cold epoxy-mounted for metallographic examination. While the aft surface was flat, the forward face exhibited a pitted surface consistent with features seen throughout the bulkhead forward face. The cross section of the surface oxide layer was approximately 75 μm thick and was present uniformly over the forward and aft surfaces of the crack edge. This oxide layer was present, with the same thickness, over the tip of the crack. These features of the oxide film present on the crack were consistent with corrosion of this region, and this oxide thickness was consistent in size and morphology across multiple cross sections of the bulkhead.

The corrosion product was present primarily outside the metal/oxide interface (or surface), with no indications of intergranular corrosion or attack of the metal near or away from the metal/oxide interface. However, the oxide closest to the metal/oxide interface exhibited a darker contrast than the rest of the surface scale. This was common on both the aft and forward sides (Fig. 16). As shown in Fig. 16, the oxide scale exhibited a columnar grain morphology, consistent with the shape of the fractured oxide scale observed previously. The forward oxide layer was found to be consistent with a region of elevated sulfur under the oxide along the metal/oxide interface (darker area in Fig. 16).

Fig. 16
figure 16

Bright field (BF) optical micrograph montage of a cross section through the crack tip. (~10X, etched 2% Nital)

Full size image

This crack edge cross section was also examined using optical metallography. Figure 17 shows a montage of cross section images, demonstrating the degree of thinning that occurred closer to the crack edge. Figure 18 shows closer views of the crack tip, showing an adherent, uniform surface oxide layer present over all the surfaces examined, consistent with having been present prior to the accident. This oxide scale was consistent with the oxide scale found away from the crack tip on both the forward and aft sides (as typified in Fig. 16). The oxide surface scale on all these figures was consistent in thickness and microstructural features, consistent with uniform and prolonged corrosion over this region of the bulkhead.

Fig. 17
figure 17

BF optical micrograph of the Crack 1 tip in Fig. 16 (~ 200X, etched 2% Nital)

Full size image
Fig. 18
figure 18

Backscattered electron (BE) micrograph of the oxide layer on the forward side of Crack 1 cross section

Full size image

A cross section of a thicker section of the Crack 1 lower edge is illustrated in Fig. 19. This section had been covered by oil, which was cleaned prior to mounting. In this cross section, the crack tip had not thinned to an edge, but exhibited a cup shape. Examination of the grains adjacent to the crack surface shows elongated grain flow, consistent with local plastic deformation (Fig. 20). Unlike the crack tip in Fig. 17, this crack edge did not exhibit a uniform surface oxide layer. Rather, pits containing oxide were observed, some of which were located along grain boundaries (Fig. 20). These features were consistent with this portion of the bulkhead having fractured from overstress. The pit-like corrosion was consistent with corrosion of the surface after fracture. Surface oxidation away from the fracture surface exhibited a uniform surface scale inconsistent with features observed in Fig. 20.

Fig. 19
figure 19

BF optical micrograph of a cross section through Crack 1 (lower region) at a thicker section than Fig. 16 (~ 25X, etched 2% Nital)

Full size image
Fig. 20
figure 20

BF micrograph of the right side of Fig. 19, exhibiting grain distortion and flow near the surface. Corrosion pitting was present on the surface (~ 200X, etched 2% Nital)

Full size image

The cross-sectioned welds were also mounted, polished, and etched. The weld shows a dendritic structure, consistent with solidified metal alloys. Figure 21 shows a cross section of the vertical weld between the inboard insert and the bulkhead. The weld exhibited ridges consistent with multiple passes during the weld operation. There were no features such as voids, cracks, or slag deposits that would be consistent with a lack of penetration or lack of fusion in the weld. Figure 21 demonstrates the thickness difference between the inboard insert and the bulkhead. Examination of the bulkhead portions adjacent to the vertical and horizontal welds showed that the bulkhead exhibited a smaller thickness at the time of welding than the adjacent inserts.

Fig. 21
figure 21

BF optical micrograph of the vertical weld cross section (~ 10X, etched 2% Nital)

Full size image

Figure 21 also shows typical microstructures of the original bulkhead and inserts. The bulkhead microstructure was consistent with that expected from a carbon steel, exhibiting ferrite and pearlite. This microstructure also exhibited elongated stringers, consistent with manganese sulfide phases typical of this steel composition. Micrographs of the upper insert and inboard insert were consistent with those typical of low carbon steels.

All the specimens examined exhibited chemical compositions consistent with carbon steel. The welded inserts exhibited lower carbon content and higher silicon, copper, and nickel content than the original bulkhead. The carbon content for the bulkhead was 0.17 wt.% versus 0.018–0.032 wt.% for the inserts. The bulkhead exhibited higher sulfur and arsenic contents than the inserts (although still typical for many carbon steels). The lower amount of pearlite seen in the inserts was consistent with their lower carbon than the bulkhead. There were no appreciable differences in chemical composition between bulkhead specimens taken near the cracks and away from the cracks, closer to the horizontal weld. As determined from XRD diffraction data, the specimens all contained distinct phases of iron oxide, including Fe2O3, FeO, Fe3O4, and FeO(OH). Also, FeS2 was identified on the aft side of Frame 2.

Analysis and Discussion

The bulkhead between the port side forepeak and #1 port side cargo hold exhibited multiple through-cracks. The presence of these cracks prior to the explosion would have allowed constituents from the cargo hold to enter the forepeak. In this condition, crude oil and/or constituents of crude oil (including flammable gasses) would have been able to enter the forward compartment [14]. These cracks and holes were consistent with having formed from corrosion prior to the explosion. Disproportionate one-sided corrosion and long, rusted cracks have been shown to be consistent with preexisting corrosion cracking [15, 16].

Examination of the through-cracks revealed crack tips with two broad categories of features: those with an enveloping oxide/corrosion layer and those without. Examined cross sections of the first crack tip category exhibited relatively thick (75 μm [0.003 inch]) oxidation layers. The thickness of this oxide layer was consistent not just over and around the crack tip, but also over the areas away from the crack tip.

There were cross sections of through-cracks that did not exhibit this uniform corrosion thickness at the crack tip. The microstructure of these crack tips exhibited features such as localized grain elongation and necking, consistent with plastic deformation. Corrosion product on these crack tips was no thicker than 25 μm (0.001 inch). The combined features of these crack tips areas were consistent with portions of the bulkhead that fractured from overstress. Therefore, these two types of crack tip profiles were consistent with areas that had corroded through prior to the explosion (thicker surface oxide), and portions of the cracks that had fractured due to the explosion (thinner or absent surface oxide).

The corrosion found on the forward side of the bulkhead exhibited corrosion severe enough to have areas of metal completely wasted away, manifest as uniform and pitting corrosion. The overall thickness of the bulkhead had corroded since the repair in 2008. The areas of the bulkhead inside the repair welds were assumed to be protected from the outside environment. This amount of material, at least 8.43 mm (0.332 inch), had corroded completely away at the location of the corrosion cracks above the two frame welds (an approximate corrosion rate of 0.84 mm/yr or 0.033 inches/yr). For comparison, typical corrosion rates for carbon steel in marine atmospheres tested at Cape Canaveral, FL are 0.13 to 1 mm/yr [17].

The location of the most severe through-corrosion was located 2.5 to 5 cm (1 to 2 inches) above the frame welds. There were no noted metallurgical or microstructural issues noted at those locations in the bulkhead or the welds, such as porosity, severe heat-affected zones, or “hard spots.” Studies performed on corrosion in crude oil carrying vessels have investigated corrosion from exposure to saline environments [18,19,20,21], thickness variations of plating [22, 23], and repeated buckling enhancing scale spallation. It was possible that any of these factors may have exacerbated the local corrosion rate.

The bulkhead underwent repeated cyclic stresses, such as localized buckling from repetitive filling and emptying of crude oil from the compartment aft, as well as typical stresses from water batter when underway [24]. During the corrosion process, both metal dissolution and formation of protective corrosion products occur. Repeated cyclic stresses can crack or spall the protective corrosion-product layer, particularly iron oxide, which is not fully compact or adherent. Spalling and cracking of the iron oxide would expose the underlying steel to the environment, allowing this corrosion mechanism to repeat itself [25].

From the EDS and XRD results, most of the corrosion products were iron oxides. While the Fe2O3 (hematite) phase of iron oxide is common in room temperature corrosion, goethite and lepidocrocite phases of FeO(OH), ferrihydrite (Fe2O3∙H2O), and Fe3O4 (magnetite) have all been reported on corrosion of mild and weathering steels in marine applications [26,27,28,29]. The compounds identified in this investigation were comparable to those found in low-alloy rebar steels subjected to a marine environment [30].

Previous studies have shown the presence of wüstite and magnetite is typically associated with higher temperature oxidation (> 573 °C) of iron and steels [31]. Decarburization is common with high-temperature steel exposure above 700 °C, but below that temperature carbon can accumulate CO/CO2 gasses causing scale blistering [32], or as solid-state enrichment near surface of the alloy [33]. Blister-free iron oxide scale is rare above 800 °C at long exposure times [31, 34, 35]. Figure 17 appears to show some qualitative increase in carbon near the scale/alloy interface. The scale thickness, generally consistent with 60–80 mm where adherent in this investigation, was more than oxide thicknesses in other studies (33 mm oxidized in air at 600 °C for 24 h [36]). The presence of these phases and microstructural feature indicates high-temperature oxidation may have occurred during the 6.5-hour fire after the explosion.

There was a small amount of iron sulfide found, and its location was underneath the main oxide layer. This sulfide corrosion could be due to constituents from sour crude and could increase the overall corrosion rate of the steel. The hydrogen sulfide (H2S) in sour crude can slightly solubilize in water to a weak acid, promoting hydrogen absorption in steel, and allow elemental sulfur to become mobile in steel [37]. The formation of iron sulfide can accelerate or retard corrosion rates in H2S-containing environments, depending on oxygen partial pressures and the presence of various microbes (no microbial testing was performed in this investigation) [38].

In addition to the bulkhead cracks discussed above, investigators found several areas of material wastage and corrosion on the main deck of the barge. The wasted areas discovered during the investigation were determined by the Coast Guard to predate the accident. Investigators also discovered epoxy compound that was used to fill in wasted areas in various locations on the cargo deck plating as a temporary repair. These epoxy repairs were not approved by the Coast Guard or the classification society.

After the discovery by investigators of the poor material condition of the B. No. 255 after the accident, the Coast Guard completely evaluated the tank barges run by the operator. Twenty-five barges were inspected, of which one required three-quarters of every cargo tank bulkhead to be replaced, and another found areas of a cargo tank completely wasted and areas down to “paper-thin steel.”

As a result of these expanded inspections, the Coast Guard issued 251 deficiencies to 25 barges, placing operational controls on 10, which limits or prevents their use until corrected. One barge was issued 66 deficiencies, and another was issued 33 deficiencies, despite both being previously inspected and classified. In addition, five of the 10 barges that were issued operational controls had been transporting oil in bulk, exhibiting extensive corrosion discovered within cargo tanks and along the hull.

Conclusions

The National Transportation Safety Board determined that the probable cause of the explosion aboard the barge B. No. 255 was the lack of effective maintenance and safety management of the barge by Bouchard Transportation, which resulted in crude oil cargo leaking through the corroded bulkhead into the forepeak void space, forming vapor, and igniting. Contributing to the accident were the ineffective inspections and surveys by the Coast Guard and the American Bureau of Shipping [39].

The NTSB made several recommendations because of the investigation. The agency recommended that Bouchard Transportation Co., Inc. has their safety management system (SMS) be evaluated by an independent third party to determine how the B. No. 255 could deteriorate into such poor mechanical and structural condition. The agency recommended that the shipper revise their SMS to remedy any identified deficiencies. The NTSB recommended that the US Coast Guard and American Bureau of Shipping (the classification society) establish joint policies and procedures. This would allow them to share information from audits, surveys, examinations, inspections, and other applicable activities related to vessel safety.