Skip Ribbon Commands
Skip to main content
Safety Recommendation Details

Safety Recommendation I-91-002
Details
Synopsis: ABOUT 0915 MOUNTAIN STANDARD TIME, ON SUNDAY, NOVEMBER 25, 1990, A FIRE ERUPTED AT A FUEL STORAGE AND DISPENSING FACILITY ABOUT 1.8 MILES FROM THE MAIN TERMINAL OF STAPLETON INTERNATIONAL AIRPORT AT DENVER, COLORADO. THE FACILITY, REFERRED TO AS A FUEL FARM, WAS OWNED BY UNITED AIRLINES AND CONTINENTAL AIRLINES. UNITED AIRLINES' PORTION OF THE FUEL FARM WAS OPERATED AND MAINTAINED BY AN INDEPENDENT CONTRACTOR, AMR COMBS. FROM THE TIME FIREFIGHTING EFFORTS WERE INITIATED IMMEDIATLEY AFTER THE FIRE ERUPTED UNTIL THE FIRE WAS EXTINGUISHED, A TOTAL OF 634 FIREFIGHTERS, 47 FIRE UNITS, AND 4 CONTRACT PERSONNEL EXPENDED 56 MILLION GALLONS OF WATER AND 28,000 GALLONS OF FOAM CONCENTRATE. THE FIRE BURNED FOR ABOUT 48 HOURS. OF THE 5,185,000 GALLONS OF FUEL STORES IN TANKS AT THE FARM BEFORE THE FIRE, ABOUT 3 MILLION GALLONS WERE EITHER CONSUMED BY THE FIRE OR LOST AS A RESULT OF LEAKAGE FROM THE TANKS. TOTAL DAMAGE WAS ESTIMATED BY UNITED AIRLINES TO HAVE BEEN $15 AND $20 MILLION. NO INJURIES OF FATALITIES OCCURRED AS A RESULT OF THE FIRE.
Recommendation: THE NTSB RECOMMENDS THAT THE NATIONAL FIRE PROTECTION ASSOCIATION: REVISE NATIONAL FIRE PROTECTION ASSOCIATION STANDARD 30 REQUIRE INTERNAL FIRE VALVES AND FAIL-SAFE CONTROL VALVES ON ALL ABOVE GROUND FUEL STORAGE TANKS.
Original recommendation transmittal letter: PDF
Overall Status: Closed - Unacceptable Action
Mode: Intermodal
Location: Denver, CO, United States
Is Reiterated: No
Is Hazmat: No
Is NPRM: No
Accident #: DCA91FZ001
Accident Reports: Fuel Farm at Stapleton International Airport
Report #: AAR-91-07
Accident Date: 11/25/1990
Issue Date: 10/9/1991
Date Closed: 6/2/2004
Addressee(s) and Addressee Status: National Fire Protection Association (Closed - Unacceptable Action)
Keyword(s):

Safety Recommendation History
From: NTSB
To: National Fire Protection Association
Date: 6/2/2004
Response: Mail Control #2040001 and 2040110 combined: NFPA indicated that the Safety Board was persuaded in previous correspondence by NFPA’s arguments against the use of internal fire control valves and agreed with NFPA that the problems with the piping system joints in the Stapleton fire contributed to its severity. NFPA also reported that it believed additional clarification would be helpful regarding what was meant by “fail-safe.” Safety Board staff contacted Mr. Robert Benedetti, the project officer for NFPA 30, to clarify this issue. Our staff explained that the intent of the recommendation was to ensure that valves on above-ground fuel storage tanks would close automatically in the event of a fire if all power were lost to them. In the Stapleton fire, the valves on three tanks had no automatic shutoff and could not be reached to be closed manually. This contributed to a continued supply of fuel for the fire. Mr. Benedetti indicated that this factor had been considered but the Flammable and Combustible Liquids Code Committee had decided against a revision. The Safety Board also received your letter of March 1, 2004, in which you state that “the technical question of requiring fail-safe control valves has been previously addressed in the full NFPA process and has not been adopted.” While your letter encourages a resubmission of the recommendation for the next revision cycle of NFPA 30, there is no new information that was not submitted to the technical committee in 1991. In your letter of April 18, 2001, NFPA stated, in part, “…every piping system is somewhat unique and [the committee] has, therefore, determined that it should not dictate piping system design.” The Board is disappointed that the committee does not agree that fail-safe valves are a part of managing flammable liquid flow between multiple tanks exposed to a fire situation. We continue to believe that fail-safe valves would provide an increased level of safety by preventing the inadvertent release of flammable or combustible materials in the event of a failure of the system that controls these materials. Accordingly, Safety Recommendation I-91-2 is classified CLOSED—UNACCEPTABLE ACTION.

From: National Fire Protection Association
To: NTSB
Date: 3/1/2004
Response: Letter Mail Controlled 3/9/2004 9:36:05 AM MC# 2040110: -From Casey C. Grant, P.E., Secretary, NFPA Standards Council: This is in response to your email transmittal dated 20 January 2004 with questions on the status of NTSB Safety Recommendation I-91-02. This safety recommendation focuses on the use of fail-safe valves for the storage of flammable liquids, and is the result of an NTSB investigation of an incident that occurred at Denver's Stapleton Airport on 25 November 1990, A relatively extensive trail of previous correspondence has been transmitted between the NTSB and NFP A on this item, as summarized in my letter to you dated 31 December 2003, Specifically, NTSB Safety Recommendation 1-91-02 is suggesting a revision to a particular section of NFP A 30, Flammable and Combustible Liquids Code, Your email transmittal dated 20 January 2004 includes the following statement (referencing a discussion between Ms. Nancy McAtee of NTSB staff and Mr. Bob Benedetti of NFP A staff): "Our understanding from ML Benedetti is that this requirement had been previously considered but was not adopted, nor were there plans in the A2006 revision cycle to do so," Your email requests clarification on the accuracy of this statement, and it is generally correct, although it implies that NTSB may not be intending to pursue a proposed change on this matter It is accurate that the particular technical question of requiring fail-safe control valves has been previously addressed in the full NFP A process and has not been successfully adopted (e.g. A93 TCR page 22), However, the second portion of your statement is postured in the past tense (i.e. "were"), and is not quite accurate since this topic may be addressed by public proposals submitted by NTSB or others, I would like to expand on this second point, and provide some further clarification on the NFP A Codes and Standards process, and the role of NTSB and all other process participants. The repeated correspondence between NTSB staff and NFP A staff on this specific technical issue clearly indicates the desire of NTSB to have this change adopted in NFPA 30. Although you've been communicating with NFPA staff on this issue, the primary way to effectively address this issue is through the full codes and standards process, which starts by submitting a public proposal. I urge NTSB to submit a public proposal to NFP A 30 and to pursue this matter through each stage of the full NFP A process. I note that such action is consistent with OMB Circular A-119 and Public Law 104-113, which encourages Federal government involvement and participation in private voluntary codes and standards making processes like NFP A's. The NFP A process is widely recognized for its open and accessible revision process, and anyone in any capacity can directly engage in this process except for NFP A technical staff Neither I, nor Mr. Benedetti or other technical staff, act as participants in the NFP A codes and standards development process and we have no adjudicatory or decision-making role within that process. NFPA technical staff act as facilitators for this process, and we are always available to discuss NFPA's rules and to make suggestions about how best to proceed. But an obligation resides with all participants in the process to advocate their own interests, and as such I urge the NTSB to consider Submitting a public proposal and to take other actions as you deem appropriate. As additional background, the NFP A codes and standards development process offers many opportunities to raise issues and it offers procedures to have concerns and problems addressed. It is through fully participating in and taking advantage of all avenues available to you within the NFP A codes and standards development process, as set forth in the NFPA Regulations Governing Committee Projects, that you must, if you desire to do so, raise any concerns you have with regard to revising NFPA 30. These avenues include, in addition to participation as committee members or as guests at NFP A committee meetings, the filing of public proposals and public comments setting forth your positions. In addition, a variety of motions can be made at the NFPA Membership Meeting Technical Report Session at which the next edition of NFPA 30 is presented for a recommendation to the NFP A membership (scheduled for May 2006). Further, various opportunities exist for appeals with the NFPA Standards Council. The public proposal closing date for the next revision cycle of NFP A 30 is 31 December 2004, and I urge NTSB to consider submitting a public proposal if they consider this issue unresolved and they wish that it be revisited. This will be the next appropriate step to assure that this matter will be fully addressed in the upcoming revision cycle of NFPA 30. We are ready to provide you with any guidance needed to submit a public proposal. It is NFPA's intent to be responsive to all NTSB recommendations, and thus we appreciate whatever steps are necessary to properly conclude this matter. Should you have any questions or concerns about your participation or the avenues available to you, I am available, as are Mr. Benedetti and other staff, to speak with you at any time now or in the future.

From: National Fire Protection Association
To: NTSB
Date: 12/31/2003
Response: Letter Mailed Controlled 1/6/2004 2:12:33 PM MC# 2040001 A noteworthy amount of review has occurred with this item and this activity is summarized in transmittals between NTSB and NFPA as follows: · Letter dated 9 October 1991, from James L. Kolstad (NTSB). to Anthony R. O'Neill (NFPA) · Letter dated 29 October 1991 from Anthony R. O'Neill (NFPA) to James L. Kolstad (NTSB) · Letter dated 6 December 1991. from James L. Kolstad (NTSB) to Anthony R. O'Neill (NFPA) · Letter dated 6 March 2001 from Elaine M. Weinstein (NTSB) to Georze D. Miller (NFPA) · Letter dated 19 July 2001 from Carol Carmody (NTSB) to George D. Miller (NFPA) · Letter dated 19 April 2001, from Robert J . Vondrasek (NFPA) to Carol Carmody (NTSB) This correspondence exchange reflects the on-going activity on this subject. In particular, the applicable NFPA TC has addressed this matter in various revision cycles of NFPA 30, Flammable and combustible liquids code (e.g.. reference: Comment 30-1 7. A93 TCR, page 21). However, all of this dialogue can best be summarized in the final paragraph of the letter dated 19 July 2001 from Carol Carmody of NTSB to George Miller of NFPA which states: The Safety Board is persuaded by NFPA's argument against the use of internal fire control valves and agrees with NFPA that the problems with the piping system joints in the Stapleton fire contributed to the fire's severity. The Board is pleased to learn that NFPA has revised Subsection 3.4.3 of NFPA 30 to address this issue. However, the Board disagrees with NFPA on the issue of fail-safe control valves. It is a basic lire safety principle that in order to extinguish a fire fuel for the fire needs to be isolated from the fire. In the absence of electrical or pneumatic power needed to operate control valves, the valves should prevent the flow of a flammable liquid. The Board believes that NFPA 30 should clearly require fail-safe control valves on all above-ground fuel storage tanks to prevent the inadvertent release of flammable or combustible materials in the event of a failure of the system that controls these materials. Pending such a revision being made to NFPA 30, Safety Recommendation I-91-2 is classified "Open--Acceptable Response." Upon further review, it appears that additional clarification would be helpful on what is meant by "fail-safe." This becomes important when considering the philosophical question of managing flammable or combustible liquid flow between multiple tanks exposed to a fire situation. To fully resolve this matter, by copy of this letter I am notifying the NFPA Staff Liaison for the project, Mr. Robert Benedetti, so that a more direct dialogue can be established with NTSB staff. NFPA 30 is currently in the A2006 revision cycle, and public proposals on this document will be accepted for this next revision cycle until 31 December 2004. I urge the NTSB staff to contact Mr. Benedetti to clarify this matter and he can then assist in preparing the necessary submittal so that this matter will be properly and fully addressed. Mr. Benedetti will also be able to clarify when the next meeting will be held of the applicable NFPA committee so that NTSB staff can attend if available and if they deem appropriate.

From: NTSB
To: National Fire Protection Association
Date: 11/11/2003
Response: "Sent via email: In 2001 the Safety Board asked that NFPA Standard 30 be revised to more clearly require fail-safe valves on above-ground tanks. Was NFPA Standard 30 revised?"

From: NTSB
To: National Fire Protection Association
Date: 7/19/2001
Response: The Safety Board is persuaded by NFPA's arguments against the use of internal fire control valves and agrees with NFPA that the problems with the piping system joints in the Stapleton fire contributed to the fire's severity. The Board is also pleased to learn that NFPA has revised Subsection 3.4.3 of NFPA Standard 30 to address this issue. However, the Board disagrees with NFPA on the issue of fail-safe control valves. It is a basic fire safety principle that in order to extinguish a fire, fuel for the fire needs to be isolated from the fire. In the absence of electrical or pneumatic power needed to operate control valves, the valves should prevent the flow of a flammable liquid. The Board believes that NFPA Standard 30 should clearly require fail-safe control valves on all above-ground fuel storage tanks to prevent the inadvertent release of flammable or combustible materials in the event of a failure of the system that controls these materials. Pending such a revision being made to NFPA Standard 30, Safety Recommendation I-91-2 is classified OPEN--ACCEPTABLE RESPONSE.

From: National Fire Protection Association
To: NTSB
Date: 4/19/2001
Response:

From: National Fire Protection Association
To: NTSB
Date: 4/19/2001
Response: Mail Control # 2010361: Robert J. Vondrasek, P.E., Vice President, Codes and Standards Operations: George Miller has asked me to respond to your inquiry regarding the subject NTSB recommendation, which resulted from your investigation of the November 25, 1990 fire at Denver's Stapleton International Airport. Our response is predicated on our understanding that an "internal fire valve" is a valve mounted in a pipe connection to a tank in such a way that the working parts of the valve are actually inside the tank and that the valve assembly includes a mechanism on the outside of the tank that responds to an exposure fire by closing the valve. We understand a "fail-safe control valve" to be any valve in a piping system that is remotely controlled (e.g., air-operated, motor-operated) that closes automatically upon loss of the normal motive power that operates the valve. Regarding so-called internal fire valves and whether they should be mandated for storage tanks holding flammable and combustible liquids, whether such valves should be required by NFPA 30, Flammable and combustible Liquids Code, has been discussed by NFPA's Flammable and Combustible Liquids Code Committee on several occasions. Indeed, the subject had been discussed prior to the incident that resulted in the NTSB recommendation and was discussed again as it related directly to the Stapleton Airport fire at the Spring, 1991 meeting of the Committee, some time before the release of the NTSB recommendation. NFPA provided its response to the NTSB Safety Recommendation based upon the Committee's review of this incident and the October 29, 1991 letter to then Chairman James L. Kolstad (see attached). While the Committee recognizes that these valves can be a component of a safe storage tank and piping installation, they were not convinced that NFPA 30 should mandate their use. The reasons for their conclusion follow: • These valves are difficult to maintain and repair. especially in the larger sizes, because of their more complex design and because the working parts of the valve are mounted inside the tank itself. This includes the mechanism that closes the valve automatically if the fusible link operates. The concern expressed during the Committee discussion focused on the inability to repair a minor leak without taking the tank out of service. Another concern related to any malfunction that left the valve "frozen" in a full- or partially-open condition, at a time when the tank contained liquid. In order to repair the valve, the tank would have to be completely emptied, perhaps from the top of the tank, an operation that was felt to be of some risk. • Use of such valves could interfere with the ability to limit the extent of a fire. It is sometimes possible to move product from a fire-exposed tank to another tank removed from the fire area, perhaps one in a different diked area. Doing so does not suppress the fire, but it does limit the amount of fuel involved. This tactic could not be employed on tanks fitted with internal fire valves. Once the valves reacted to an exposure fire, the contained fuel is "locked in" the fire area. Regarding fail-safe valves, here again the Flammable and Combustible Liquids Code Committee has previously and on several occasions discussed the issue. The Committee has always recognized that every piping system is somewhat unique and has, therefore, determined that it should not dictate piping system design. Note that Subsection 3.5.6 of the 2000 edition of NFPA 30 does require a "sufficient number of valves to operate the equipment properly and to protect the equipment". But, this is a performance requirement; it is the responsibility of the piping system design engineer to determine where control valves are required and of what type they should be. Finally, the decisions of the Flammable and Combustible Liquids Code Committee on these issues, particularly since they were raised yet again by the Stapleton Airport Fire, were influenced by the fact that the piping system for this installation included numerous pipe joints that did not meet the basic requirements of Chapter 3, Piping Systems, of NFPA 30. The pipe joints in question relied on the frictional characteristics of a combustible component, which burned out under fire exposure. This caused the joints to completely disengage, as shown in the accompanying photo. It was the Committee's conclusion that the failure of these joints contributed to the magnitude of the incident. The Committee has since developed more definitive requirements for use of similar joints that do not completely disengage, as given in Subsection 3.4.3 of NFPA 30, a copy of which is enclosed. We are certainly open to discussing these issues to greater length and would welcome having a representative of NTSB attend the next meeting of the Committee. NFPA technical committee meetings are open to anyone who wishes to attend and anyone is permitted to address the Committee on relevant issues.

From: National Fire Protection Association
To: NTSB
Date: 10/29/1991
Response: -From Anthony R. O’Neill, Vice President and Chief Operating Officer: In reply to your safety recommendation 1-91-02 concerning the November 25, 1990 fire at an aircraft fuel storage and dispensing facility at Stapleton International Airport in Denver, Colorado, we have reviewed the safety recommendation and provide this response. After this incident occurred, the NFPA Committee on Flammable and Combustible Liquids who are responsible for NFPA 30 the Flammable and Combustible Liquids Code, met in March, 1991 to review information regarding this incident. A presentation was made to the Committee by NFPA's fire investigator which included a preliminary report and video tapes of the incident. The Committee also heard from a Committee member who was on-site during the incident. The Committee discussion focused on the use of pipe joints that relied on resilient gaskets for tightness. In the Committee's opinion, the severity of the fire, especially the resurgence of fire at Tank #4 midway through the incident, was directly related to complete disengagement of at least one major pipe connection to Tank #4 which then allowed Tank #4 to completely empty its contents into the fire area. NFPA 30 subsection 3-4.2 does allow the use of such couplings in above ground liquid service with the provisions that either the piping be secured in some fashion so that the joint does not disengage, or that the piping be designed with a sufficient number of isolation valves so that the adverse impacts of a release resulting from disengagement can be controlled. In this case, the Committee felt that neither provision existed in the Denver incident. Another consideration raised by the Committee was that the resilient couplings in question are listed for water service. To our knowledge, no such couplings are listed for flammable liquids service. The Committee was concerned that the use of an internally mounted fusible link operated valve can cause additional complications. Operation of such a valve on a tank exposed to fire would prevent transfer of the contents of that tank to other remote storage safe from the effects of the fire. Though this subject has been considered by the Committee this spring, we will nevertheless bring forth this NTSB safety recommendation as a public proposal to amend NFPA 30 and will move it through the Committee process for the next revision cycle. NFPA 30 is scheduled for revision in our Annual '93 revision cycle. The proposal closing date is January 24, 1992. The next Committee meeting is scheduled for November 12 - 15, 1991 at the Holiday Inn, Beaumont Plaza, Beaumont, TX. If you or your staff wish to attend the November meeting or a subsequent meeting, please contact our staff liaison, Robert Benedetti, so that the Committee can budget time for a presentation. Thank you for your continued interest in NFPA codes and standards. We appreciate the feedback and concern displayed by NTSB in fire related safety issues.