Technical Paper



The Full Scale Test Of The American Underpressure System (AUPS)
On Board The USNS Shoshone

The Authors:
Mo Husain, President MH Systems, Inc., San Diego, CA
Robert E. Apple, Sr. Vice-President, MH Systems, Inc., San Diego, CA
Randy Sharpe, Sharpe Surveying, Alameda - Consultant to MH Systems, Inc.
Gary Thompson, Vice-President, M. Rosenblatt & Sons,


The American Underpressure System (AUPS), an active inert gas controlled system utilizing vacuum technique, drastically reduces or totally prevents spillage from an accidental rupture of the hull. Existing inert gas systems are modified to provide underpressure in tank ullage space during the voyage so that if tank rupture occurs, oil spill above the rupture is prevented. The actively and dynamically controlled underpressure is set to insure that pressure forces inside and outside the tank are equalized at the line of rupture and then there is no inducement for flow to occur.

The underpressure requirement is generally moderate (-2 to -4psi), and is well within the structural limit of the tanker as has already proven by the recent Full Scale test of the AUPS on USNS Shoshone in the Bayship & Yacht Shipyard in Richmond California on June 11, 2001. Also, the test has shown that the AUPS can maintain an inerted atmosphere in the ullage space at all times to prevent explosion hazards. Additionally, the laboratory tests with three different crudes shows that AUPS closed loop system in the operational mode can prevent emission and loss of cargo due to evaporation even when subjected to moderate negative pressures.

The AUPS in conjunction with the existing inert gas system can be used as a primary containment system for existing single hull tankers until all single hull tankers are retired or as a backup to other costly approaches. It is a simple, practical approach that can be easily retrofitted in a timely manner into existing tankers at minimal costs and logistical impacts.


The desire for a system to provide spill protection for single hull tankers was first expressed in the U.S. Congress Oil Pollution Act of 1990. This act mandated the exclusive use of double hull tanker construction after 2015. The Act also specified studies on spill prevention for existing tankers and specifically mentioned research and development on the use of vacuum in tanks.

MH Systems (MHS) initiated the development of the American Underpressure System, (AUPS), in 1989. The system reduces or virtually eliminates oil outflow from an oil tanker damaged in grounding or collision by applying a partial vacuum to the empty space above the oil cargo, the ullage space. This negative pressure eliminates the pressure difference at the rupture point to prevent outflow. Analysis predicts that this underpressure system will reduce average outflow per incident by approximately 65%, a reduction approaching that predicted for the double hull. The system is also expected to improve further the effectiveness of double hulls.


In the operation of the underpressure system, negative pressure is applied to the ullage space, to control the pressure balance at the point of rupture to reduce or eliminate outflow. Figure 1 shows the midship cross section of a tanker. The ullage space above the cargo is supplied with an inert gas mixture, derived generally from flue gases, in order to prevent the formation of an explosive mixture. In the event of damage at the point "X", cargo would flow out of the conventional vessel until the pressure at "X" inside the tank is equal to that outside the tank. The underpressure system provides a reduced, sub-atmospheric, pressure above the cargo. The inside and outside pressures at "X" are equalized, preventing outflow of cargo.

Figure 1


Funding for a "Full Scale Test of the American Underpressure System" (AUPS) was released by the Advanced Research Project Agency in May 1995 as a part of the MARITECH competition. A three-phase test of the AUPS, administered by the Maritime Administration, was implemented.

Phase I, Concept Definition, was completed in April 1996. It consisted of extensive documentation of the previous six years of design studies, experiments and technical papers. Also prepared were several plans including the preliminary master plan for the analyses, test and full validation of the AUPS.

Phase II, Development of a Detailed Test Plan, including a Validation Plan, was completed in March 1997. The submittals included system schematics, preliminary analyses, test memos for testing the tanker with, first fresh salt water, and then crude oil, and descriptions of the extensive analyses required for testing and for the subsequent generic system design.

Phase III The American Underpressure System Test and Validation Project
In late September 1998, in the Defense Appropriations Bill, Congress directed the Secretary of the Navy to provide funding to the Maritime Administration. In 1999 Congress redirected funding to the Office of Naval Research. The contract was initiated July 11, 2000. The project involved four elements of investigative effort. The ultimate objective of the project was to develop the information necessary to permit rulemaking changes so that the system can be installed on tankers.

There are four elements of investigative effort as shown below:.
This paper, however, primarily discusses the full-scale tests of AUPS on USNS Shoshone.

1. Conduct full-scale tests on a tanker.
2. Conduct supporting analyses.
3. Conduct Laboratory tests.
4. Develop the Generic Design.

The Full Scale Test

Test Plan

The full-scale test was planned to be more than a simple demonstration and validation. It afforded a means of exploring critical design issues; particularly those in which uncertainty about scaling effects required validation of conclusions arrived at by model-scale or laboratory testing. The key issues investigated were the following.

  • Retention of cargo-- effectiveness of ullage underpressure in reducing cargo outflow (spillage).
  • Underpressure control-- the dynamic stability of ullage space pressure control using pressure measurements, fan characteristics, valving, and digital control algorithms.
  • Gas/Vapor evolution-- the rate of evolution of hydrocarbon gas/vapor under reduced ambient pressure. Correlation with laboratory data.
  • Ullage inerting by inert gas mixing-- the ability to maintain an inert atmosphere while simultaneously maintaining a reduced ullage pressure. Verification of transient and steady state mixing analyses.
  • Structural loads-- validation of finite-element structural analyses of pressure-induced loads.

Test Design

The test design was driven by the need to use a reasonably contemporary tanker configuration, at full scale and using minimum modifications. A suitable vessel, the USNS Shoshone, of about 35,000 dwt, was made available from the reserve fleet. One tank on the port side was designated as the nominal cargo tank for the purposes of the test. An adjacent centerline tank was used as a simulated ocean, thus avoiding contact with the environment. The simulated ocean tank was filled with fresh water, rather than salt water, so as not to contaminate the vessel.

The intervening bulkhead between the two tanks was pierced about one and one-half feet above the tank bottom, and fitted with a pneumatically operated 12" valve. This arrangement permitted the valve to be used to simulate a damage incident. Nitrogen gas was drawn from a liquid supply to serve as a simulated inerting medium. M. Rosenblatt & Sons, Inc prepared the design of the required tanker modification and AUPS system installation. The test setup is shown in Figure 2.

Figure 2

Control System

The control system utilized in the full-scale demonstration test on USNS Shoshone was required in order to control the ullage space underpressure and inertness during routine and spill containment operations. A computer simulation was developed for the test to verify that stable and conservative operation would prevail during the test. These test series confirmed the control system's capabilities.

Control System Description

A schematic of the control system arrangement for the test is shown above in Figure 2. There are three flow streams interfacing with the test tank. The test tank contains water as cargo and an adjoining tank simulates the draft of the tanker. A rupture valve 12 inches in diameter and 18 inches above the tanker bottom is located in a common bulkhead so that spillage is contained when the valve is opened, thereby preventing possible impact on the environment. To preclude the formation of localized high concentrations of oxygen in the ullage space due to potential air leakage into an underpressured tank, a continuous supply of inert gas is circulated through the ullage space and then vented to the atmosphere.

Nitrogen Supply Stream

Nitrogen is supplied dockside from a cryogenic liquid. It is stored, vaporized, and regulated to approximately 17.0 psia and 70 F. The nitrogen enters the AUPS distribution piping through a flow meter and a single-loop control valve. The control valve is based on electro-pneumatic technology and operates in conjunction with a remotely- located Remote Electro-Pneumatic Controller. The controller is a single-loop, fluid control that compares the ullage pressure to the desired set point. It uses air to operate the valve opening to maintain the controlled variable. A response speed adjustment can be used to vary the rate at which the valve will open and close. There are three adjustments for tuning the controller - the desired set point, the speed of response, and the width of the dead band. Conventional pressure and temperature transducers are used.

Air Supply Stream

Air is introduced into the ullage space at rates in excess of the expected air leakage into a tank assembly. The test will validate the oxygen dilution capability of the system. The piping arrangement provides for a remotely operated manual valve followed by a flow meter, and by pressure and temperature sensors.

Exhaust Stream

The exhaust steam vents the ullage gases to the atmosphere. The piping arrangement consists of a flow meter; blower and blower trim valve, and associated pressure and temperature sensors. The gas vents to the atmosphere through a vent riser containing a Pressure/Vacuum (P/V) relief valve and a flame arrestor. The blower is a 2-stage, regenerative, radial compressor driven by a constant-speed electric motor. The blower pressure rise matches the overall resistance to the flow venting to the atmosphere. The trim valve secured the required flow and underpressure and was not used as a control variable when final perturbations were introduced.

The Control System Features

  1. The data acquisition console is capable of real time monitoring displayed and data capture of the test parameters.
  2. A simulated tank rupture using a 12-inch butterfly valve with spring return and manual override.
  3. Remotely located test control centers on the ranker housing the test conductors and data acquisition and control consoles, and a center for the display of spill containment.

Results (Control System)

The computer simulations modeled the main flow components, duct loses, and interactions to yield the transient and steady state test parameters versus time.

The output provided assurance to proceed with the test. The test results showed conformance with the simulation model

  1. Controls performed conservatively over a range of perturbing conditions such as air leakage and inert gas circulating rates.
  2. Inertness and underpressure were maintained during routine and spill containment.
  3. Flow stability was present on all these operations.

Planning of AUPS Test on USNS Shoshone

Use of Crude Oil

The planned use of crude oil resulted in the requirement for many permits and technical reviews. Due to the anticipated use of crude oil for the test on board SHOSHONE it was necessary to involve the Coast Guard in plan approval for the project. Meetings were held with the local Coast Guard Marine Safety Office for San Francisco Bay to outline the scope of the project and determine the extent of Coast Guard involvement. It was agreed that the SHOSHONE would not require full inspection for issue of a Certificate Of Inspection. The Coast Guard agreed to limit their inspections to the installation of the test equipment. Plans were developed for the test installation by M. Rosenblatt and Sons and sent to the Coast Guard Marine Safety Center in Washington DC. There was some initial hesitation on the part of the Marine Safety Center and the staff of Coast Guard headquarters technical branch concerning the project. After meeting with these individuals it was agreed that they would review the plans for compliance with Coast Guard regulations for installation of the system on board a tank vessel.

Structural Analysis

During the review process additional information was provided to the Marine Safety Center of the USCG on structural calculations for the SHOSHONE. This was required due to initial concerns over the vacuum on the ullage causing structural damage to the cargo tanks to be involved in the test. Elaborate finite element analyses were carried out by Dr. Alaa Mansour of University of California, Berkeley. These calculations, approved by the American Bureau of Shipping (ABS), found that the cargo tank could be subjected to 12 PSI of negative pressure without sustaining any damage. After some modifications to the plans to meet all of the Coast Guard regulations for installation of the system on a tank vessel, the plans were approved by the Coast Guard and the Navy approved the delivery of the SHOSHONE to MH Systems to install the system and conduct the test.

Approvals; U.S. Coast Guard (USCG), Office of Naval Research (ONR), Office of Naval Operations (OPNAV)

When work commenced on board SHOSHONE the Coast Guard Marine Safety Office inspections department was involved in inspection of the installation of the system in accordance with the approved plans. This involved several call outs to have an inspector witness welding and testing of the installation. Prior to conducting the tests the Coast Guard was satisfied that the system was installed in accordance with the approve plans. When the test was completed, it was required to return the SHOSHONE to her original condition. The Coast Guard Inspector was again called to witness fit-up and welding of the inserts of the cargo tanks and deck where piping penetrations were removed. Upon completion of these repairs the Coast Guard issued a final inspection report noting that the ship was returned to satisfactory condition for MARAD to accept it back into the reserve fleet.

California Bay Area Air Quality Management District Approval

In addition to Coast Guard approval, approvals had to obtain from the California Bay Area Air Quality Management District (BAAQMD). All hydrocarbon emissions from the SHOSHONE had to be determined and a plan approved by the BAAQMD. During the test of the system while the AUPS system was maintaining a vacuum on the ullage space of the #5 port cargo tank there would be a small amount of hydrocarbon emissions from the exhaust of the vacuum blower. Calculations were made as to the amount of the emissions and found to be under the 15-pound threshold in the California regulations. A permit was applied to conduct the test and it was granted. Steve Hill of BAAQMD was very helpful in working with us on the permitting process for this test.

Vapor Recovery System Approval

SHOSHONE was built prior to the regulation of vapor emissions for crude oil and did not have a vapor recovery system installed. The original plans for the installation included piping for a vapor recovery system. However during the search for a barge to deliver crude oil to SHOSHONE it was determined that there were no barges capable of conducting vapor recovery while loading the SHOSHONE. There were barges in the industry which could reclaim their own vapors while being loaded but they could not vapor balance with SHOSHONE. MHS approached the BAAQMD and requested a one-time permit to load SHOSHONE from a barge without vapor recovery. They granted MHS the one time permit to load the crude oil. However the loading of the barge both from the terminal and then when offloading SHOSHONE still had to comply with the vapor recovery regulations. This somewhat simplified the test installation on SHOSHONE, as the vapor recovery piping was no longer required. However it limited the choices of barges to load SHOSHONE with crude oil. The Barge JOVALAN operated by Public Service Marine a Harley Marine Services company was identified as being certificated by the Coast Guard and the BAAQMD to conduct this operation. The Coast Guard agreed to the use of this vessel for litering the crude oil to SHOSHONE. Unfortunately scheduling problems and cost uncertainties combined at the last minute and prevented us from loading crude oil for this test.

Final Approval

When all reviews and approvals were finally received, then the Office of Naval Research presented the results to the cognizant offices in naval operations for final approval.

Approval of the Dead Ship Tow

The Coast Guard was also involved in the approval of the dead ship tow of SHOSHONE from the Suisun Bay Reserve Fleet to the Richmond Point facility where the work was to be carried out. A towing plan was provided to the Coast Guard by Foss and the tow completed using three tugs. There was a San Francisco Bay Harbor Pilot onboard to direct the tow as well as a riding crew from Bay Ship & Yacht. Both dead ship towing operations were uneventful.

Survey, Tow, Test Set-up Installation & Testing

After a survey of the vessel, it was towed to the shipyard and modified in accordance with the test plan. Ship modifications included provision of the simulated rupture, the addition of gas and fluid inlet/outlet piping, connection and test of the underpressure fan, strain gage instrumentation, and pressure, flow and temperature instrumentation. A data acquisition and computer display and storage facility was set up on deck. Figure 3 shows approximate spatial relationship of measurement parts.

There were approximately two weeks of setup and testing. These operations were open to witnessing by USN and others. The final test, involving a formal demonstration of the simulated rupture, was observed by a large party or representatives from a number of organizations.

Approximate spatial relationship of measurement points


Figure 3


  1. Tank access and inert gas inlet
  2. "Center" oxygen gauging point, Hand level indication and ullage pressure measurement
  3. Inboard Oxygen measurement point, sometimes referred to as "steam" connection
  4. Blower suction point, blower oxygen concentration measured from here
  5. Air leak entrance point
  6. Tank level indicator and ullage pressure measurement point

Test Sequence

The testing sequence consisted of the following phases:

  • Structural loads testing-- installation of strain gages and monitoring of stress changes under varying fluid levels and underpressure levels. Measured strains were converted to stress using parameters obtained from a test coupon. The loads' testing was required before further testing was permitted.
  • Leak testing-- the sea tank remained at atmospheric pressure during all tests and was not leak tested. The cargo tank, however, had to be leak test and, as necessary, rehabilitated to achieve a leak-tight enclosure.
  • Gas mixing tests-- a series of tests was conducted with preset inert gas inflow rates and with differing levels of simulated air leakage. These tests were designed to demonstrate transient mixing behavior, to provide a basis for validating theoretical analyses of mixing effects.
  • Spill tests-- two separate, but essentially identical spill tests were carried out. Pre-calculated levels of underpressure, chosen for outflow prevention at the observed cargo and outflow depths, were used. In the second test, the underpressure was discontinued after the 30 minute test period in order to observe cargo outflow.

Figure 4 shows the simulated test set up with real time sensor inputs. This was presented to the observers that were in the ship's mess hall so that they could watch the activities in real time.

Figure 4

Fresh Water Substitution of Crude Oil

The design of the test, the design of the installation and all of the regulatory approvals in the Bay Area and in Washington D.C. were to allow crude oil to be used in the test. 6,300 bbls were required. At the last moment the schedule for delivery of the crude oil to the USNS Shoshone and the cost of delivery and return became unacceptable uncertainties. With the approval of the Office of Naval Research, fresh water was substituted for crude oil. As a result some credibility was lost but fortunately very little information was sacrificed. This was because laboratory tests with crude oil were being conducted simultaneously. Three crude oil samples covering most of the natural range were (laboratory) tested at a number of temperatures and negative pressures, and extensive analysis was performed

Summary of Full Scale Test Results

  • The spill test validated the effectiveness of ullage underpressure in restraining cargo loss. In two separate tests, a 30-minute exposure to potential spillage evidenced no cargo loss.

  • The simple underpressure control system was stable and was able to set and maintain a specified ullage pressure even with controlled leakage. No real time adjustment was required.

  • The laboratory test, covering a wide range of crude oil density, obtained consistent measurements of evaporant composition. As expected, the substances evaporated were mostly C7 and below.

  • The Equation of State (EOS) model developed by the laboratory, typical of many similar computer programs used in the petroleum industry, correlated well with measurements. The EOS model can be used as a design tool in the detail development of the underpressure system.

  • The transient and steady state gas mixing predictions were reasonably confirmed in testing. The calculated time constants were well supported. These results suggest that the analytical approach is reasonably good, but that some further analysis or mockup testing may be required for some geometrical arrangements.

  • The test validated that safe inert gas level in the ullage space can be maintained in spite of air leakage

  • Strain measurements indicated low stress increments due to pressure loading, and agreed reasonably well with predictions.


This project, the "American Underpressure System (AUPS) Test and Validation Project", does not complete the program - it initiates the program. The single focus of the program is to reduce the potential for catastrophic events from casualties to single hull tankers, as well as from double hull tankers. The project, which has just completed, has verified that these catastrophic events can be prevented or, at least, significantly minimized. The technical information obtained from the full-scale test, from the laboratory experiments, and from the accompanying analyses was used to proceed to initiate the design installation of AUPS in a 70,000 dwt operating crude carrier.

The design and technical data must satisfy many questions and requirements that have been set forth by the U.S. Coast Guard. This procedure is required to allow the rulemaking change that will permit crude carriers to maintain an inerted vacuum in all of the ullage spaces. There are four principal areas of effort that still must be accomplished. The design must proceed to a level approaching in detail and definition that of a contract design. The control system must be defined to the same level. A sophisticated simulation program must be developed to permit the investigation of all aspects of all catastrophes, and the reliability, economic and other analyses required by the Coast Guard must be made.

There is a strong sense of urgency to try to protect the single hull tankers with AUPS before any additional catastrophes might occur. One important additional benefit to the U.S. Maritime Industry is the additional work that these installations will provide to the repair shipyards,

MH Systems is scheduled to have all of the designs and technical data available for Coast Guard review by the Spring of 2002.


Eight (8) photographs Test Installation on board USNS Shoshone

Description (click to view image)
USNS Shoshone
Nitrogen Regulator Valve (with "Leak" Valve below) and associated instrumentation
American UnderPressure System Full Scale Test: Typical Oxygen Sensor Unit
Top View of Nitrogen Supply Valve Area
American UnderPressure System Full Scale Test: Exhaust Blower with Associated Instruments and Trim Valve
USNS Shoshone, Main Deck Port Side Midship: Test Control Room Located Aft of Cargo Boom
American UnderPressure System Full Scale Test: Test Tank P/V Valve Assembly
USNS Shoshone Under Tow

Your further inquiries are invited. Write to:

Copyright 2009 MH Systems, Inc. All rights reserved. | Terms of Use | Last Modified 11/11/02