L'Ambiance Plazza
Staff posted on October 24, 2006 |
L'Ambiance Plazza
Cause of Collapse


There are several design and construction deficiencies that could have been the triggering cause of the collapse of the partially erected lift-slab structure at L'Ambiance Plaza. The most likely cause is slab failure resulting from incorrect drape of post-tensioning tendons and inadequate bottom reinforcement adjacent to the elevator opening in the east bay of the west tower. Tensile stresses at the bottom of the slab produced by slab weight and prestress only, as calculated by (NBS 1987) using finite element analyses, reach a maximum of about 600 psi (4.14 MPA) when the slab is lifted uniformly at all support points. This is in the range of the expected slab cracking strength, as discussed later.

The results are confirmed by an independent analysis described by (Poston, Feldmann and Suarez 1991). Even higher peak stresses result when support points have differential elevation, as can occur to a limited extent during lifting and from tolerances in the setting of support wedges. A maximum differential elevation between support points of 1/2 in. (12.5 mm) is permitted by ANSI A10.9.

Flexural tensile stresses in the range of about 8 to a maximum of will result in cracking. Using the specified concrete strength, of 5000 psi (35 MPA) the expected cracking stress range for the slab at L'Ambiance Plaza is 560 to 850 psi (3.8 to 5.7 MPA). Thus, the maximum stresses at about mid-length along the west edge of the elevator opening are sufficient to crack the slab in the tension zone. They are also near enough to the lower end of the expected cracking stress range to be consistent with the absence of failure during the slab lifting operations that took place before the time of collapse. Without sufficient reinforcing steel to restrain the opening of cracks or to resist the applied moments, it is likely that the slab finally broke in the region of highest tensile stress and deflected severely.

The arrangement of the north-south tendons with convex curvature in the top of the slab above the cracks maintained the downward load produced by tension in the unbonded tendons during deflection after cracking. The absence of bonded reinforcement and unbonded tendons in the bottom of the slab precluded the development of post cracking ultimate strength of the type expected after cracking with normal bonded or unbonded reinforcement systems. Furthermore, the offsets and sweep in the east-west tendon bands, the presence of the large elevator opening adjacent to Columns E.5 and C.5, and the trash chute opening in front E.5, and particularly, the precarious frame stability provided by the temporary connection between slab shear collars and steel columns, as described previously, rules out the ability of the tendon strand system to resist total collapse by "netting" action. The structural system in the elevator bay had little or no ductility or ultimate resilience to resist progressive collapse. The flexural cracking resistance of the unreinforced slab bottom represented essentially the ultimate strength of the slab. Once the slab developed flexural tensile cracks in the region of high tension near the edge of the elevator opening, the onset of rapid progressive collapse took place throughout the entire structure.

Based on its tests of lifting collars, NBS concludes that the most probable cause of the collapse is the inadequate strength and stiffness of lifting collars, resulting in sufficient deformation of the support angles to cause the lifting rods to disengage from the slotted angles in the lifting collar at one of two lifting points (Columns E-4.8 or E-3.8) (NBS 1987, Scribner and Culver 1988). This conclusion is vigorously denied by the lifting subcontractor, based on its own tests of the lifting collars and its extensive previous experience with the use of this system (Wonder 1988).

Certainly the practice of using open slots in rod support angles to facilitate positioning of rods in the lifting collar weakened the angles and increased their flexibility when loaded by the weight on the rods. It also provided opportunity for positioning rods with excessive eccentricity which could reduce the strength of the lifting collars to a dangerous level. The NBS tests show that if the rods were improperly positioned in the angle slots with significant eccentricity toward the column, the angles could have deformed sufficiently to permit a rod to slip out, or to bend and break, thereby triggering the collapse. On the other hand, there is no hard evidence that the rods actually were improperly positioned in the slotted angles. The collar manufacturer's tests and experience shows that the angles have sufficient strength and stiffness to support the load being lifted at the time of the collapse if the rods are properly positioned in the slots. In view of the lack of evidence that the rods were not properly positioned in the slots, failure of a lifting collar is not the most likely cause of the collapse.

A third potential cause of the collapse is frame instability. Possible triggers of instability failure are wind, frame plumbing operations or excessive out-of-plumb, and other construction-related lateral loads. Assuming known or likely sources and magnitudes of lateral loads acting at the time of failure, approximate calculations indicate sufficient stiffness and the resisting moment capacity provided by the weight of the slabs to rule out frame instability as the primary cause of the collapse. Nevertheless, an appropriate design wind force could cause sufficient frame moment to seriously endanger the stability of the structure, because once the restoring moments from slab weight are exceeded by moments produced by lateral forces, frame joints lose stiffness and the frame becomes unstable. Regardless of the actual triggering mechanism, the use of temporary slab-column connections that relied on eccentric slab weight for stiffness and strength certainly contributed to the extent and suddenness of the total collapse of the structural frame.

A fourth potential cause of the collapse is the excessive maximum compressive stress and high non-uniform maximum shear stress that occurred in the vicinity of Column 2H as a result of the proximity of the shear wall slot to the column and the absence of concrete within the shear collar. The stress conditions at this location reported by (Poston et al. 1991) are well in excess of safe design limits. Also, several eyewitnesses reported that the collapse began at the southwest corner of the west tower (NBS 1987). However, the presence of tendons within the column shear zone and the compressive and shear strength imparted by the shear collars and mild reinforcing adjacent to the column probably provided enough strength and ductility in this area so that excessive stress caused by inadequate concrete area at this shear wall slot did not actually trigger the collapse of the structure.

The five other construction deficiencies described previously do not appear to be probable causes of the collapse. This is consistent with the conclusions given in (NBS 1987).

Public Safety Issues

The collapse of L'Ambiance Plaza can be blamed in part on a practice for design and construction of buildings in the United States that fails to recognize that the complexity of today's design and construction process requires significant improvements in minimum design standards, minimum professional qualifications, and typical building regulations for protection of public safety. The following four deficiencies in American building design and construction practice probably had a significant impact on the L'Ambiance disaster: (1) excessive reliance on performance specifications, allowing structural design responsibilities to be split among the engineer-of-record and multiple contractors, (2) building regulations that do not assure competent structural design for complex structures, or for projects where design responsibilities are split between an engineer-of-record and contractors, (3) inadequate safety factors and other technical provisions in building codes, and (4) construction phase quality assurance inspection by testing agency technicians in lieu of the engineer-of-record or a substitute engineer with knowledge of the structural system.

Excessive Reliance on Performance Specifications for Structural Systems


There are several reasons for an increase in the use of performance requirements in contract documents, giving the contractor significant responsibilities for design that impacts public safety. One of these is an increase in the complexity of modern building systems, resulting in specialty contractors for operations like post tensioning. Another reason is the increasing the use of competitive bidding for professional services, leading to fee cutting and a decline in engineering competency. In some cases there is justification for the use of performance specifications. However, safeguards to assure that design is performed by competent professionals are not always included in the performance requirements, and all too often uncertainly exists as to the scope of each party's responsibility.

With respect to the L'Ambiance Plaza project, several subcontractors were responsible for different components of the slab and structural frame design and installation. The result was inadequate designs for the slab in the bay with the elevator openings, for the slab at shear wall slots in close proximity to exterior columns, for the lifting collars, and for frame stability during the lifting operation. Also, if the engineer-of-record only reviewed the slab design shown on the shop drawings for "general conformance with the design concept of the project" (as stated on the review stamp), there is uncertainty as to the extent of his review. With this split of design responsibility, it is possible that no single engineer took responsibility for the entire slab design.

Another question is whether it even was possible to prepare a safe slab design within the requirements and constraints shown on the contract documents? The large openings and offset location of exterior columns in the elevator bays of the west tower imposed very special design requirements for this bay. Could the discontinuities associated with the large elevator opening adjacent to Column lines C.5 and E.5 and the trash chute opening directly in front of Column E.5-6 be accommodated safely in a 7 in. (178 mm) flat slab? Did the sub-subcontractor who provided tendon layouts and presumably designed the slab in this bay have the necessary expertise to design such a complex structure? Or did its staff lack engineering competence and just assume that this bay required the same prestressing as the typical bays? Are the ACI Code provisions permitting almost no minimum bonded reinforcement in flat slabs with unbonded tendons intended to cover this type of slab configuration? Could shear wall slots in such close proximity to columns 2H, 8A and 11A be accommodated safely during lifting. These complexities should have been foreseen by the engineer-of-record and, as a minimum, addressed in the performance specification, along with proper definitions of specific design responsibilities and requirements for professional qualifications by the specialist subcontractors.

Inadequate Building Regulations


The safe design of many structural systems used in modern buildings requires specialized knowledge and experience far beyond that needed for registration as an engineer or an architect. Thus, it takes more than a requirement that drawings have the seal of a registered engineer or architect to assure that the designer is qualified to produce a safe design. Furthermore, when contract documents delegate a significant portion of the design to the contractor, building authorities must recognize the split responsibilities and the need to examine, either directly or through a review engineer, the design documents produced by the contractor, along with his qualifications to perform those design responsibilities. These contractor design documents should be included in the building permit process, in addition to the record drawings submitted by the engineer. Also, there needs to be better definition of what the engineer-of-record is responsible for and what the contractor's engineer is responsible for. This definition should place significant burden on the engineer-of-record to assure that the contractor's engineer has interpreted correctly the contract requirements and that there is no communications error.

Another undefined area of responsibility in building regulations is design required solely for construction methods and means. Traditionally, construction methods and means are the sole responsibility of the contractor. In a lift-slab project, however, structural design requirements for the completed structure and requirements for strength and stability during lifting are related. This muddies the question of who is responsible for what and requires careful coordination of design responsibilities between the structural engineer-of-record and the contractors. Building regulations do not include requirements that qualified professional engineers must design significant structural systems or components related solely to the construction process. At L'Ambiance Plaza this meant that there were no specific design submittals showing that the lifting system and frame had adequate strength and stability during all phases of the lifting operations.

Before the collapse of L'Ambiance Plaza there were no provisions in Connecticut building regulations for an independent of building structural designs. Generally, building authorities do not have an adequate staff to review structural designs and there were no requirements for peer (second engineer) review. One of the commendable improvements made in Connecticut as a result of the collapse is to require review of structural designs by an independent engineer who is acceptable to the building authority.

Code Provisions


ACI 318-83, "Building Code Requirements for Reinforced Concrete" gives the code requirements for design of the reinforced and prestressed concrete components at L'Ambiance Plaza. Section 18.12 contains special requirements for prestressed slabs and Section 18.9 contains requirements for minimum bonded reinforcement. In two-way slabs with unbonded tendons, at least two tendons in each direction must pass through the shear zone of every column, but no bottom bonded reinforcement is required when tensile stress is less than , (although a minimum quantity of bonded reinforcement is required in the midspan slab bottom for one way slabs). This irrational difference is justified on the basis that tests (ACI-ASCE Committee 423 1989), (Burns and Hemakon 1985), and (Kosut, Burns and Winter 1985) have been conducted on half scale models to substantiate the omission of bottom bonded reinforcement in two way slabs, but not for one way slabs. Some engineers also believe that two way slabs have greater redundancy than one-way slabs, but this reasoning obviously is incorrect for two-way slabs without perimeter beams. No provisions are included that limit the omission of bonded bottom reinforcement to slabs with regular rectangular bays without large openings (ie: the test conditions).

Because tests were conducted on model structures with regular bays and a limited number of variables, the reasoning used to justify the elimination of bonded reinforcement in two way slabs at least should include provisions that limit such design to the conditions and variables of the test program. Furthermore, codes often require much higher safety factors for designs that are justified by test results than for designs based on accepted design practice with a satisfactory experience record. For example, the BOCA code requires a minimum safety factor of 2 for components whose safety is demonstrated by proof-of-design tests. This compares with the ACI Safety factor of 1.4/.9 = 1.55 for flexural effects of dead load. The above considerations lead to a conclusion that the ACI Code does not provide a reliable or complete basis for assuring structural safety of post-tensioned flat slab designs for some building structures, such as the post-tensioned flat slabs with large openings and offset columns at L'Ambiance Plaza.

Construction Phase Quality Assurance


Another question related to structural safety arises concerning inspection of field construction. In a complex post-tensioned reinforced concrete structure such as L'Ambiance Plaza, can any organization other than the designer (or a structural engineer with detailed knowledge of the design requirements) have enough understanding of design requirements to give assurance that the structure is being built in accordance with the design? In this project, the inspection function was given to a testing organization; the engineer-of-record was not retained for this purpose and resided in a distant state. Only an inspector who is intimately familiar with the design concept of post-tensioned concrete flat plates could have been expected to discover the improper drape of the post-tension tendons in the elevator bays, and only then if he observed the slabs after tendon installation and prior to placement of concrete.

Under the current system, there are no legal requirements that the designer observe construction of the work, and often an owner perceives that there is cost saving by having the testing laboratory that he has engaged to sample and test materials also inspect structural components. The result is a disturbing trend toward engaging testing laboratories, rather than the structural engineer-of-record, for inspecting structural components for conformance to design requirements. This is particularly dangerous for reinforced and prestressed concrete construction where deficiencies are concealed after the concrete is cast.

Conclusions: Structural Deficiencies


At the time of the collapse, L'Ambiance Plaza had the following serious deficiencies which rendered the building structurally unsafe.

    1. The post-tensioned slabs in the elevator bay at the east end of the west tower were designed improperly because the north-south tendons had improper drape and the east-west tendon bands had excessive sweep. The improper drape of the north-south tendons caused excessive flexural tensile stresses in the bottom of the slab that were within the range of the expected cracking strength of the concrete. The excessive sweeps of the east-west tendon bands placed the tendons far outside the required shear zone around certain columns. Because of the incorrect tendon drape and absence of tendon bands over adjacent columns, the slab had no reserve ultimate strength and the structure had no resistance to progressive collapse, once the slab experienced flexural cracking.
    2. Portions of the slab at three columns that are in close proximity to shear wall slots were designed improperly for the temporary conditions during lifting. The prestress force at these locations produced an excessive and unsafe level of compression in the top of the slab. Shear strength was also questionable at these locations because of the sharp discontinuity caused by the slot in close proximity to the column and the slab edge.
    3. The steel channel lifting collars were functionally unreliable. The slotted lifting rod support angles in the collars were excessively flexible and subject to serious overstress if lifting rods were not accurately positioned within the support angle slots.
    4. Frame stiffness for lateral or torsional-lateral stability of the structure above completed sections of the shear walls was unreliable and inadequate under the temporary construction conditions that occurred during lifting of the slabs. The design of the temporary connections of slabs to columns relied on the stabilizing effect of the slab weight distributed between supports on column flanges for stiffness and strength. Under sufficient lateral force, the entire joint stiffness and strength could be lost.
Conclusions Related to Public Safety Issues


Regardless of which of the four major deficiencies that were described above actually may have caused the collapse of L'Ambiance Plaza, the existence of so many life threatening deficiencies demonstrates a serious failure of American practice for designing and constructing major buildings to protect public safety. Specifically, the L'Ambiance tragedy demonstrates the following inadequacies in our system for assuring public safety:

    1. The slab design was severely flawed and the improper design probably was not detected because responsibility for design was fragmental by multiple subcontractors, because the engineer-of-record did not take overall responsibility for the design, and because of the absence of requirements for a review by a second engineer. The con- tract documents did not provide a complete slab design. Significant design respon- sibilities were given to the contractor, who in turn passed them on to three or more subcontractors. Some of the resulting designs by subcontractors were grossly improper. These errors were not discovered by the engineer-of-record who reviewed the subcontractor designs only for general conformance to the design concept.
    2. Building authorities accepted an incomplete design on record drawings. Only the contract drawings by the engineer-of-record contained the seal of a registered professional engineer, and these did not show the slab and slab-column connection design requirements. The authorities did not require design documents showing designs for the temporary lifting conditions.
    3. The ACI code is deficient because sections relating to two-way slabs with unbonded post-tensioning systems do not limit the use of such systems to slabs having approximately the regular rectangular bays used in tests to determine design requirements. Also the code should recognize that such slab systems should have either a larger safety factor, or requirements for minimum amounts of bonded reinforcement in both the top and the bottom of the slab. This is because slabs without adequate bonded reinforcement have less reserve capacity and ductility under overload or dimensional deviation and less resistance to progressive collapse than slabs with bonded reinforcement, and because certain key design requirements are based solely on tests.
    4. Although the new (1989) ACI 318 code requirements for structural integrity (Section 7.13) are intended to increase the toughness and resistance of structures to progressive collapse, the more explicit provisions in Section 18.9.3.1 exempt two-way slabs from any requirement for minimum bonded reinforcement in the slab bottom, in effect overriding the intent of the new integrity provisions.
    5. Building regulations do not require adequate consideration of structural safety related to construction methods and means.
    6. The Connecticut State Building Code which incorporates the BOCA Code, did not include any reference to ANSI Standard A10.9, which covers safety requirements in construction including specific requirements for the lift-slab construction method. Thus, compliance with this standard was not subject to enforcement by city building authorities.
    7. Inspection of structural parts during the construction process by a testing agency that is not intimately familiar with the structural design does not provide adequate assurance that the structure is being constructed substantially in accordance with the design and safe structural practice.
Recommendations


Based on lessons learned from the L'Ambiance Plaza tragedy and other similar incidents of structural collapse, the following improvements to our system for designing and constructing building structures would result in better assurance that public safety is being protected:

    1. Building regulations should require that all structures (beyond those of a routine nature such as simple wood-framed residences and small one and two story commercial and farm buildings) must be designed by a qualified professional structural engineer. The practice of accepting structural drawings sealed by an architect or by an engineer who is not a qualified professional structural engineer should be discontinued.
    2. The written examination, and the education and experience necessary to qualify for registration as a professional structural engineer should insure that the applicant has a fundamental knowledge and understanding of modern structural engineering practice. These qualifications should be upgraded substantially beyond the knowledge obtained in undergraduate civil engineering programs and should be comparable to the educational requirements of other professions, such as law, medicine and dentistry. (Suitable grandfathering provisions for practicing structural engineers should accompany such a change).
    3. Building regulations should require review of the structural design of significant structures by a second engineer, independent of the engineer of record, and acceptable to the building authority.
    4. When certain structural components in a project are to be designed and constructed by a specialty contractor, that contractor should be required to obtain a separate permit for its part of the construction, to be given only after it has been determined that a qualified professional structural engineer, employed or retained by the specialty contractor, has designed the specialty components and assumed responsibility for their design adequacy. Also this engineer should be required to submit an affidavit that he or she has performed the design and that it complies with building regulations. As a part of this special permitting process, the building official also should require that the professional structural engineer for the overall project has reviewed and accepted the design of the specialty contractor.
    5. When construction methods and means require specific design for structural safety, such as the lift-slab erection process, building regulations should require a separate permit for that particular construction process. This permit should be issued to the contractor based on a determination by the building official that the structural safety aspects of the special construction process have been designed by a qualified professional structural engineer who is responsible for the adequacy of structural safety during the construction process. Since this involves construction methods and means and is not related to the final structural conditions, the professional structural engineer who is responsible for the overall final structure should not be required to review the structural safety of construction methods and means. When the construction methods and means affect the final structure, the engineer of record should review their effects on the final structure, but his or her responsibility and corresponding legal liability should be limited to this scope.
    6. Building regulations should include a requirement that the construction of structural components shall be observed either by the structural engineer responsible for their design, or by another qualified structural engineer who is familiar with the design requirements of the project.
    7. The structural engineering profession should initiate a comprehensive review of the basic safety factors that are incorporated in national structural engineering codes (design specifications) for steel, reinforced concrete, wood, and bridges, and in national standards for design loads, such as the new ASCE Standard 7-88 Minimum Design Loads for Buildings and Other Structures (formerly ANSI A58.1 1982). This review should take into account as much data about actual structural failures as is possible. It should be performed by practicing structural engineers with extensive knowledge and experience with existing design practice, investigation of structural failures, and probability theory. Because such a study is urgently needed to improve protection of public safety, the profession should seek financial assistance from the National Science Foundation with the objective of implementing the results of this review within the next 3 to 4 years. The professional liability insurance industry also should contribute to such a review.
    8. Practicing structural engineers on existing code committees should encourage the committees to review all provisions related to public safety. Consideration should be given to increasing safety factors for systems of limited ductility or subject to progressive collapse and to limiting the application of systems that are designed based on test results to the conditions tested. Furthermore, the ACI Code should be changed to extend and increase requirements for minimum bonded reinforcement in zones subject to tension under factored load for all components which rely on unbonded tendons for strength.

Many other aspects of current structural design and building construction practice are relevant to the full assurance of public safety, and a number of other significant improvements are needed. Discussion of these, however, is beyond the scope of this paper.

Acknowledgements

Information about the status of construction at the time of the collapse, observations of eyewitnesses, the results of finite element analyses of stresses in a typical floor slab during lifting, and the results of load tests on lifting collars were obtained from the NBS 1987 report. The results of additional finite element analyses of a typical floor slab during lifting were obtained from (Poston, Feldmann and Suarez 1991). Results of other load tests performed by the lifting subcontractor on lifting collars were obtained from (Wonder 1988). The author gratefully acknowledges the review and valuable suggestions of his associates, Glenn R. Bell and Donald O. Dusenberry.

References


American Concrete Institute (1983 edition). "Building Code Requirements for Reinforced Concrete, ACI 318-83," with Commentary. Detroit, 111 pp.

ACI-ASCE Committee 423 (1989). "Recommendations for Concrete Members Prestressed with Unbonded Tendons." ACI Journal, V. 86, No. 3, May-June, pp. 301-318.

ANSI A10.9 (1982). "American National Standard for Construction and Demolition Operations Concrete and Masonry Work Safety Requirements" "Ch. 10. Lift-Slab Operations," ANSI A10.9 1982, American National Standard Institute, New York, N.Y.

ASCE 7-88 (1990). "American Society of Civil Engineers Standard Minimum Design Loads for Buildings and Other Structures." ASCE, New York NY (Revision of ANSI A58.1-1982).

Building Officials and Code Administrators International, Inc. (1986 Edition). The BOCA National Building Code/1990, Country Club Hills, IL, 552 pp.

Burns, N.H., and Hemakon, R. (1985). "Test of Post-Tensioned Flat Plate with Banded Tendons." ASCE Journal of Structural Engineering, V. 111, No. 9, September, pp. 1899-1915.

Hawkins, N.M. (1981). "Lateral Load Resistance of Unbonded Post-Tensioned Flat Plate Construction." PCI Journal Jan.-Feb., pp. 94-116.

Kosut, G.M., Burns, N.M., Winter, C.V. (1985). "Test of Four-Panel Post-Tensioned Flat Plate." ASCE Journal of Structural Engineering, V. 111, No. 9, September, pp. 1916-1929.

National Bureau of Standards Report NBSIR 87-3640 (1987). "Investigation of L'Ambiance Plaza Building Collapse in Bridgeport, Connecticut." Center for Building Technology, National Engineering Laboratory, U.S. Department of Commerce, 309 pp.

Poston, R.W., Feldmann, G.C. and Suarez, M.G. (1991). "Evaluation of the L'Ambiance Plaza Post-Tensioned Floor Slabs." ASCE Journal of Performance of Constructed Facilities, Vol. 5, No. 2, May.

Post-tensioning Institute (1977). Design of Post-Tensioned Slabs, Phoenix, 55 pp.

Scribner, C.F., and Culver, C.G. (1988). "Investigation of the Collapse of L'Ambiance Plaza," ASCE Journal of Performance of Constructed Facilities, Vol. 2, No. 2, May, pp. 58-79.

Wonder, D.R. (1988). "What Happened at L'Ambiance Plaza?" Civil Engineering, ASCE, October, pp. 68-71.

Page: 2 Of 2First  Previous  1  2  Next  Last  

Recommended For You