Tank Wall Failure: Great Boston Molasses Flood of 1919

UNIVERSITY OF TEXAS AT ARLINGTON
MAE 3242: MECHANICAL DESIGN

Tank Wall Failure
Great Boston Molasses Flood of 1919

Talyna Morrison
12/03/2014




This paper is a brief overview and simple analysis of the mechanical failure of a commercially used cylindrical tank. The courts and subsequent scholastic researchers have proven the fatalities sustained could have been prevented with proper engineering ethics and review. This paper builds upon this by analyzing the basic component of the storage tank: a thin-walled cylindrical structure.


Introduction    One of the major debates of modern times is waged by dietitians and nutritionists over which sweetener is the best for human consumption; should it be an all natural, straight from the plant sap or a semi-refined product thereof? Should it be sugar in its purest form or a chemically perfect, lab created substitute? Throughout history, people of various regions have taken one side over another, most often out of convenience. At one point in time, the United States and Europe favored molasses as the sweetener of choice and its demand for its sweet properties, as well as its use in rum production, helped fuel the slave trade known as the infamous “triangular trade”. In the United States, even after the abolishment of slavery, the importation of molasses from the Caribbean continued. In Boston, Massachusetts, the Purity Distilling Company set up shop as a subsidiary of the United States Industrial Alcohol Company, and commissioned a steel tank to hold the imported molasses until it could be fermented and processed for alcohol to be used in overseas in the war. Less than four years after the tank was completed, it failed and flooded Boston with millions of gallons of molasses, killing or injuring over one hundred people and causing extensive damage to the surrounding community. The cause of the failure was disputed at first, with the Purity Distilling Company and the United Sates Industrial Alcohol Company claiming radical anarchist activists planted a bomb to protest the production of alcohol and push their prohibition agenda, however after a lengthy trial and in depth reviews, the true cause of the explosion was discovered to be poor design and negligence that resulted in a class action payout to all victims.

Cause    In 1915, the Purity Distilling Company contracted Hammond Iron Works to construct a massive tank to hold molasses that was already en route from the Caribbean. Arthur Jell, the man overseeing the project, reportedly did not file a permit for the tank, just the concrete slab upon which the tank would sit. Without a permit, there would be no engineering review by the city (The Engineering News-Record [1], however, claims a permit was filed, but the city approved it blindly because a civil engineer had stamped the submitted plans). Hammond Iron Works set about designing a tank fifty-foot high with a radius of forty-five feet that was, in theory, designed to withstand an explosion possibility, a requirement due to the in-progress World War I. The design also claimed to have a factor of safety of 3 [1]. The tank was to be finished by the time the ship arrived on December 31, 2015, so when severe winter weather caused delays, construction was rushed with spot-checks being missed. Even the final check, filling the tank with water to verify it would hold a liquid volume, was ignored; Mr. Jell only pumped enough water into the tank to a depth of six inches [2].
    In addition to the lack of engineering reviews, negligence was a key factor. There were plenty of warning signs that the design was insufficient, including molasses leaks at the rivets resulting in a pool of molasses at the base of the tank, noticeable undulating waves in the tank walls, and groaning noises echoing around the tank [2].
    Though these problems compile, the final cause of the failure was thought to be the sudden, unseasonable temperature spike that created a large pressure differential between the atmospheric air and a build-up of fermented molasses, which resulted in a force too great for the rivets to hold the sheets of metal. Whether this theory is true or not, the fact remains that the tank was designed poorly and the rivets sheared from exposure to a force far greater than the strength of the material.

Available Information & Reports    There are plenty of records of the tank failure itself, with every newspaper in the region reporting the disaster and running pictures of the damage. Even recent years have seen an increase in coverage of the Great Boston Molasses Flood, with a retelling of the events that unfolded. However, the records used in the application for the building permit of the tank and the records used in the court proceedings are unattainable; none of these records are accessible online and must be reviewed in person. This is not to say that information and analysis on the tank explosion are unavailable.
    Less than a month after the disaster, Wilfred Bolster, Chief Justice of the Municipal Court, filed a report to the Grand Jury to investigate whether or not the United States Industrial Alcohol Company, Purity Distilling, or any person or persons could be found liable and accountable for the deaths and property damage resulting from the explosion. Bolster detailed his initial findings in the Boston Daily Globe, listing among the damages and deaths a strong case for negligence and lack of engineering oversight [3]. The theory of explosion due to anarchist was immediately thrown out and replaced by the possibility that fermentation may have been a contributing factor, though, if true, the design was ultimately at fault.
    One of the leading experts at the time of the trial was Charles Spofford, an MIT graduate and professor, who found the most likely root cause of failure to be the rivets [4]. Spofford calculated the molasses was exerting such a great force, the rivets were under double the stress their material would allow. He also made known the discovery that the tank was not built to the plans submitted; the sheets of metal making up the walls of the tank were thinner than those called for in the blueprints. A CEE technical Writer, Debbie Levey, chronicled Spofford’s testimony in an article for the website, Slice of MIT [4].
    Since the time of the trial, additional research studies have been undertaken by historians and professors alike. Many can be found online, but most repeat the same information found in the Engineering News-Record. Ronald Mayville, Ph.D, P.E., however, views the incident from a fatigue failure perspective, noting the fracture surface marks as indication that the failure, which appeared to be of the brittle variety, resulted from a rivet hole on a manhole near the bottom of the tank [5]. Mayville also writes in detail about the type of steel used, noting the manufacturing process (open hearth), the percentage of elements (low carbon, high manganese), and the effects thereof on the final resulting steel (lower ductile-to-brittle transition temperature).
    A public access television show called The Folklorist ran a segment on the molasses tank, which can be viewed online via YouTube, which supports the theory of a fatigue fracture. The segment discusses the groaning and rippling of the tank walls [2], which leads to speculation that the tank was a ticking time bomb, waiting for the right conditions to invite failure.
In the short version documentary “The Sweetest Thing: In the Wake of the Amber Tide”, director Chris Sonne interviews Perry Thoorsell, a member of Stanford University’s Mechanical Engineering department [6]. Thoorsell offers a short lesson on cylindrical tanks and the propensity for a “hoop stress” formation near the base. Though he does not go into any mathematical detail in his conceptual overview, Thoorsell does mention imperfections in the material, manufacturing, and construction, all factoring in to fatigue cracking.
    By all accounts, there was a major temperature swing from January 14 to January 15, 1919. None of the reports referenced herein credit this drastic change as a contributing factor to the failure. It should be noted that this disaster resulted in more stringent permit approval and construction management processes [7].

Analysis    Using Spofford’s investigational finding that the tank was built with thinner material than the original plans detailed [4], this analysis seeks to determine if the steel thickness used may have been a severe factor in the tank failure: was the material of the tank sufficient to hold the pressure of the molasses within?
    The molasses tank was built to the overall dimensions of 90-foot diameter by 50-foot tall, giving a total volume of 318,086.26 cubic feet, or 2.38 million gallons. It was noted that the tank was only filled 48-feet 10-inches deep [1], giving a liquid volume of 310,089.00 cubic feet, or 2.32 million gallons. The walls of the tank were made with half-inch thick steel sheets, which creates a thin-wall cylinder situation, per the criteria that the wall thickness be less than one-tenth the radius, as suggested in Norton’s Machine Design textbook [8]. This criteria still holds true at the rivet locations, where the sheets overlap to create a wall thickness of 1 inch, which is still far less than .1 at a ratio of 1:540.
    By taking the tank as a thin-walled cylinder, the stress equations simplify to straight forward ratios of pressure, radius, and thickness. The radial stress will be negligible, the axial stress (the “hoop stress” [6]) will be half of the tangential stress, and the tangential stress will be the pressure multiplied by the radius over the wall thickness, as shown in Equations 1 through 3 in Appendix A. This assumption is used with the understanding that the equations are not valid at locations of local stress concentrations, such as at the rivets of the steel sheet overlaps; however, it is important to check for failure potential in the initial design before adding more complex aspects. Per Norton,

Pressure vessels can be extremely dangerous even at relatively low pressures if the stored volume is large . . . Large amounts of energy can be released suddenly at failure, possibly causing serious injury.

With over two million gallons of molasses, this case would qualify as potentially dangerous storage. The pressure inside the tank, neglecting the theory of pressure buildup from fermentation, would be the hydrostatic pressure from the stored molasses. The resulting tangential stress is 33.1ksi. A copy of The Making, Shaping and Treating of Steel from 1920 lists a range of ultimate strength of open hearth steel based on carbon percentage, with the lowest strength as 56.7ksi [9]. This table was chosen based on the steel specification listed in the analysis by Mayville [5]; all other resources stated the tank was simply steel. The resulting factor of safety, based solely on the tank wall design, is 1.7, far lower than the 3 that was claimed [1]. Though misrepresenting a factor of safety is not a criminal act, the design was not up to the Safety Factor of 4 that Mayville references as the engineering standard recommended for static loads in steel structures in 1914 [5].
    Taking into consideration the fact that the tank was intended to be filled, rested, drained, and filled again, it would not be a stretch to place it into a repeated, cyclical loading category. Under this condition, however, the failure strength would be greatly reduced. With the main concern still being the tangential stress, performing the cyclical analysis learned in chapter 6 of Norton’s Machine Design, using a conservative reliability factor and considering the sheets to be forged or cast, the new factor of safety drops below 1.
    From this simplistic analysis, the tank was an unsafe design from the very beginning. Other engineers and mathematicians and researchers have delved deeper, looking at the shear strength of the rivets [1] and the fracture patterns of the steel plates for clues about sudden brittle failure of ductile material [5]. Their conclusions are the same: the design was flawed and unsafe.


APPENDIX A: FIGURES & EQUATIONS


Figure 1. Cylinder Tangential Stress Calculation from Internal Hydrostatic Pressure


Figure 2. Cyclical Cylinder Tangential Stress Calculations for Factor of Safety


Figure 3. Cover Image: “View of Wreckage, Showing Roof and Base of Collapsed Tank” [1]



Works Cited

1	Brown, Burtis. 1919, “Details of the Failure of a 90-Foot Molasses Tank,” Engineering News-Record, vol. 20 (issue 20), pp. 974-976.
2	“The Folklorist: The Great boston Molasses Flood.” 30 Mar 2012. Online video clip. YouTube. Accessed on 30 Nov 2014. < http://youtu.be/UMgfHmY7o3k >
3	Bolster, Wilfred. 1919, “Public Blamed For Tank Disaster In Findings of Justice Bolster,” Boston Daily Globe, 8 Feb 1919.
4	Levey, Debbie. 2012, “Solving the Great Molasses Flood Mystery,” Slice of MIT, from < http://slice.mit.edu/2012/04/06/great-molasses-flood >
5	Mayville PhD PE, Ronald A. 2014 “The Great Boston Molasses Tank Failure of 1919,” from < http://cenews.com/article/9864/the-great-boston-molasses-tank-failure-of-1919 >
6	Sonne, Chris. “The Sweetest Thing: In the Wake of the Amber Tide.” 6 May 2009. Online video clip. YouTube. Accessed on 26 Nov 2014. < http://youtu.be/fti-bleren4 >
7	2011. “Boston Molasses Flood, 1919,” from < http://www.celebrateboston.com/disasters/molasses-flood.htm >
8	Norton, Robert L. 2014, Machine Design: An Integrated Approach, 5th ed., Pearson, Upper Saddle River, NJ, Chaps. 4 and 6.
9	Camp, J.M. and Francis, C.B., 1920. The Making, Shaping and Treating of Steel, Carnegie Steel Company, Pittsburgh, PA, pp. 290.