U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590

Skip to content U.S. Department of Transportation/Federal Highway AdministrationU.S. Department of Transportation/Federal Highway Administration

Bridges & Structures

Technical Advisory

Cable Stays of Cable-Stayed Bridges

June 17, 1994

Technical Advisory 5140.25


To provide guidance on the following issues concerning cable stays for cable-stayed bridges:

  1. Main tension elements (cables);
  2. Sheathing;
  3. Corrosion protection; and
  4. Saddles.


Several issues have emerged in the United States relative to the primary member of a cable-stayed bridge, namely, the cable stay itself. The intent of this Technical Advisory is to document these issues and present recommendations for all future cable-stayed bridges.

  1. Main Tension Elements. Several types of "cables" are available for use as stays in cable-stayed bridges. The form or configuration of the cable depends on its make-up; it can be composed of parallel solid bars (ASTM A722), parallel wires (ASTM A421), parallel strands (ASTM A416), and single or multiple arrangements of structural strands (ASTM A586) or locked-coil strands. Structural strands or locked-coil strands are no longer used in the United States. Parallel wire (ASTM A421) is no longer commercially available in this country. Solid bars (ASTM A722) have a lower allowable stress and alower fatigue resistance than strand (ASTM A416). Therefore, parallel bar stays will be of a larger diameter than comparable parallel strand stays. The larger diameter will affect the area exposed to wind, the size of the anchorages, and attachment details.
  2. Sheathing
    • (1) The Schillerstrasse pedestrian bridge in Stuttgart, Germany, completed in 1961, is the first cable-stayed bridge to use a sheathed and injected stay system. The stays consist of a bundle of parallel bare prestressing wire encapulated in a polyethylene (PE) pipe and injected with cement grout. The stays have been inspected on numerous occasions and have shown no signs of corrosion. The first use of this system in the United States was for the Pasco-Kennewick Bridge completed in 1978. Polyethylene pipe is available in either a low, medium, or high density. It is current practice, for improveddurability, to specify a high density PE pipe for cable-stay applications.
    • (2) The purpose of the stay sheathing is twofold: to provide a form for the injected cement grout and to serve as an anti-corrosion barrier for the stay cables. A third purpose is sometimes mentioned as participation in live load carrying capacity. However, the Federal Highway Administration (FHWA) does not consider the sheathing as a load carrying member for the purpose of sizing the cable. Two types of sheathing have been used: a high-density PE pipe and a steel pipe.
    • (3) A PE sheath is an air-tight material. A 6 mm thickness provides the same vapor barrier as a 10.7 m thick concrete wall. However, the thermal coefficient of expansion of black PE is approximately six times that of the grout and steel cable and is, therefore, incompatible with other materials within the cable stay. Therefore, the PE pipe is wrapped with a white or light colored Polyvinyl Fluoride (PVF) tape to control temperature variations and reduce the magnitude of differential stresses.
    • (4) Steel sheathing for stay cables is typically black steel pipe. Generally, lengths of steel pipe arebutt welded together on the project site. The steel pipe in itself must be painted for corrosion protection and a light color should be used for thermal control. The paint system selected should consider susceptibility to ultra-violet ray exposure. Further, adequate surface preparation of the steel and weld surfaces is required to preclude lack of adherence and/or deterioration of the paint system.
    • (5) It is frequently stated that the steel pipe is more durable than the PE pipe and is less susceptible to cracking. With regard to durability, there were at least 53 cable-stayed bridges constructed worldwide in the period from 1961 to 1988, which have PE cable stay sheathing as corrosion protection. No stays have had to be replaced. Of the 48 completed structures (during that time period) only 2 (or 4 percent) of the projects have had distress associated with the PE pipes. Since not all of the stays in each of the two projects cracked, the actual percentage, based upon the total number of stays, would be a much lesser value.

      The cracks in these two projects were attributed primarilyto over stressing during grouting operations. The causes for this distress are known and are accounted for in current criteria. In both structures, the damaged PE pipes were successfully repaired or replaced. There were (during the same time period) six bridges, worldwide, that were constructed utilizing steel pipe cable stay sheathing. One of these, afterless than 10 years service, developed corrosion at the joints of the sheathing. Thus, 16 percent of the completed steel pipe sheathing projects have experienced distress. The field butt welds are subject to cyclic loading resulting from wind action on the individual stays and live load variations which eventually lead to fatigue cracking in the weld (the susceptibility to fatigue cracking is exacerbated by the presence of corrosion) and subsequent crack propagation. Field welding of cracks would be impractical from the standpoint of the close proximity of the high tensile steel strands. Welding heat input would adversely affect the metallurgy of the wires in the strands. Should cracks develop in the steel sheathing, there is no known practical retrofit procedure short of dismantling and replacing the stay(s). From the standpoint of durability of the sheathing, there is an insufficient track record of steel pipe used in this application to indicate its relative durability in comparison to PE pipe.

    • (6) Other items of concern regarding steel pipe sheathing are as follows:
      • (a) Internal backing bars are required to obtain a full penetration weld at the joints of the pipe sheathing. The internal diameter of the sheathing pipe will have to be increased so that the backing bar thickness does not obstruct the free passage of the strands in the cable stay. This increased diameter will increase the volume of grout and, thus, the dead weight load on the stays which might have an affect on the equivalent modulus of the stay.
      • (b) It is difficult, if not impossible, to isolate the steel pipe sheathing from participation in live load carrying capacity.
    • (7) The National Bridge Inspection Standards (NBIS) require inspection of bridges every 2 years. The current state-of-the-art of stay construction/fabrication does not produce a stay that is accessible for inspection. An FHWA funded research program has successfully developed amethod for non-destructive investigation for stay cables by use of the magnetic field perturbation method (MPC). However, the presence of steel pipe cable stay sheathing seriously inhibits or prevents thesuccessful use of MPC equipment.
  3. Corrosion Protection Systems
    • (1) Corrosion protection systems may be either two-phase or single-phase. In the two-phase method the permanent corrosion protection is applied as the last operation of construction of the structure. This means that a temporary corrosion protection is required during a construction period that may be 2 to 4 years in duration. In cases of unforseen delays in construction, this time period may be extended to 6 years or longer. The effectiveness of most temporary corrosion protection methods is short lived. Therefore, replenishment is required at specified time intervals. If replenishment is overlooked or not accomplished, for whatever reason, there is a distinct risk of the onset of corrosion occurring before the permanent corrosion protection can be applied, which could lead to having to replace the cable. There is currently a trend towards single-phase corrosion protection systems that provide both the temporary and permanent system simultaneously, i.e., a system that provides protection from manufacture of the cable throughout its service life.
    • (2) Cable stays of most, if not all, early cable-stayed bridges consisted of zinc-coated or galvanized locked-coil strand, e.g., the Lake Maracaibo Bridge in Venezuela (constructed in 1962). In many cases, these strands also had apaint coating. Zinc coating or galvanizing has the advantage that it is not easily damaged in handling and installation and is relatively inexpensive. However, it is a sacrifi-cial coating, i.e., it is consumable with time in an aggressive environment. In the harsh environment of Lake Maracaibo, the galvanized locked-coil strands had to be replaced in 1980 after 18 years of service and currently, the replacement strands are being threatened once again by corrosion. The Kurt Schumacher Bridge in Mannheim, Germany, began to show evidence of corrosion after approximately 17 years of service. The deterioration of numerous other structures with this type of corrosion protection system for the stays has been reported.
    • (3) The galvanizing process affects the mechanical properties of steel. It has been reported that ultimate tensile strength can be reduced by as much as 5 percent and fatigue strengthby 20 percent. With some steel formulations there may also be changes in ductility. The possibility of hydrogen embrittlement due to the galvanizing process has been suspected. However, there is no hard evidence to support this suspicion. In the case of galvanized wire or strand encapsulated in grout, the zinc coating may react with some cements releasing hydrogen gas. This reaction is apparently dependent upon the cement alkalies, type of steel, and the composition of the zinc coating. When galvanized strand is embedded in cementitious grout, the corrosion rate of zinc itself is accelerated. As a result, zinc as an anode or sacrificial metal coating does not perform the same as in atmospheric conditions.
    • (4) Grouting of the stays is not accomplished until the structure is completely erected and final adjustment of all stay forces is completed. The period of time between prestressing steel installation and grouting may be several years, especially if there are unexpected project delays, and temporary corrosion protection of the prestressing steel is required. The life expectancy of the temporary corrosion protection system and means for its replenishment should be considered in design and construction. Water soluble corrosion inhibitors may require repeatedapplication every 3 months in a moderately corrosive environment.
    • (5) Both types of sheathing normally have a telescopic type sleeve detail at one end of the cable stay (either the top or the bottom) to preclude inducing dead load and erection stresses into the sheathing. The prestressing steel at this location is exposed during construction, and a temporary protection measure is required until such time as the sleeve is permanently attached to the sheathing and the stay is grouted. If an inadequate enclosure is provided or the temporary corrosion protection system is exhausted and not replenished, corrosion of the prestressing steel may occur. Very costly replacement of the affected strands in terms of material cost, labor cost, and project delays can be the result of inattention to this detail.
    • (6) Cementitious grout with its alkaline properties provides an active corrosion protection to the prestressing steel. However, recent autopsies of grouted cable stay fatigue test specimens have confirmed that under cyclic loading the grout fractures and cracks. The hairline cracks occur in the grout every 25 to 50 mm. Thus, should the sheathing be compromised by a propagating crack emanating from a defective butt-weld in a steel sheathing or a crack resulting from circumferential overstrain in a PE sheathing, a direct path to the prestressing steel is available for aggressive corrosive agents.
    • (7) In recent years, to overcome the potential problems of a sheathed and grouted system, as described above, multiple barrier systems have been developed. The concept simply provides increased redundancy in the corrosion protection system. Therefore, in connection with the outer sheathing and grouting material, systems of three, four or as many as five barriers are created. Generally, these additional barriers are provided by one of the following two methods: (a) individual greased and sheathed strands (the monostrand method). It should be noted that the word grease as used in this context is generic; the material may be grease, wax, epoxy-tar or some other appropriate material, and (b) a coating applied directly to the strand such as galvanizing, epoxy, or possibly a ceramic material. Both of these systems are installed or applied prior to shipment. Thus, they are not only incorporated into the final total corrosion protection system but also provide the temporary corrosion protection during shipping, storage, after installation until the final grouting operation and during service life.
    • (8) The monostrand system as used for cable stays is a adaptation or transfer of technology of the monostrands that are used for parking garage or flat slab construction. The stay consists of a parallel bundle of 15 mm diameter unbonded prestressing strands that are individually greased and sheathed, enclosed in a PE pipe and grouted.
    • (9) A recent innovation from conventionally sheathed strands is the application of a corrosion inhibiting material directly to each of the seven individual wires of each 15 mm diameter strand and extruding a polyethylene jacket over each strand. During the application of the corrosion inhibiting material the seven-wire strand is put through a destranding operation ("bird-caging" over a finite length), a coatingoperation which covers the entire surface of each wire, restranding to the original configuration and jacketing the strand with PE. The corrosion inhibiting material is a soft petroleum base wax that can be applied at ambient temperature, displaces any moisture on the surface of the steel, has a melting point over 260oC, and offers superior corrosion protection.
    • (10) In the search for corrosion protection methods and materials, consideration has been given to coatings applied directly to the prestressing steel. Coatings for prestressing steel are much more demanding than that for a passive conventional reinforcing steel. For prestressing steel the coating must be capable of withstanding the elongation that occurs during stressing without cracking, i.e., strain compatibility. Galvanized prestressing strand (or wire) has been used in some multibarrier systems. As previously discussed, galvanized prestressing steel should never be used where it is in direct contact with cementitious grout and the designer must be cognizant of the effects of galvanizing on the material properties of the steel. For a short time, epoxy coated strands were considered the answer to the corrosion protection challenge. During fatigue tests of stay assembles for several bridges, questions were raised about the long-term corrosion protection provided to the interior interstices of the individual stands. A relatively recent development is an epoxy filled strand whereby the interstices between the wires are filled with epoxy. This eliminates the concern for corrosive agents gaining access to the interior of the strand and also provides some improvement to fatigue resistance. It should be noted that the epoxy formulation and thickness used for prestressing steel is not the same as that used for conventional reinforcement. Epoxy-coating of the individual strands of the cable stay provides both the temporary and permanent corrosion protection to the strand.
  4. Cable Saddles
    • (1) Cable saddles are devices used to allow the cable to be continuous over a support (pylon), i.e., there isno termination of individual stays (on each side of the pylon) in an anchorage at the pylon. This means that stressing of the cable requires simultaneous and coordinated jacking of the cable at both ends (at the deck stay anchorage points on each side of the pylon), not only during initial construction and installation but also at any future time should cable force adjustment be necessary for any reason. It also means, that in the event cable stay replacement is required, twice the cable length must be replaced as compared with cable stays individually anchored at the pylon.
    • (2) Because of corrosion and other considerations, there are basically only two systems available for cable stays: the epoxy-coated and filled strand or the monostrand (greasedand sheathed) strand systems. The epoxy coated and filled strand is a proprietary system, therefore; an alternative system is required--the monostrand system. The epoxy coated and filled strand should contain sufficient embedded grit in the surface of the epoxy coating to provide enough friction to preclude any slip of the strands in the saddles under any loading condition. However, this may be detrimental to the ease of stay removal. The monostrand (greased and sheathed) strand, by its composition, is unable to provide a non-slip condition in the saddle and requires individual termination anchorages at the pylon for each stay. Therefore, alternate design details based upon the cable stay system and saddle or individual anchorages, for each cable-stayed bridge design, would be required when saddles are used. Therefore, it appears appropriate to design both cable stay systems so that they terminate at the pylons with individual stay anchorages, i.e., stay saddles should not be used.
    • (3) For a large single saddle (where all stays converge to the top of the pylon), no load transfer occurs along the pipe to the pylon. All load must be transferred at a local detail. Therefore, one cannot use the steel pipe along its length for any appreciable load transfer, for to do so will load the pipe up in tension. If bond is established too early, very high stresses in the pipe will occur.
    • (4) When all stays converge to the top of the pylon to a common saddle, another problem arises in that as the stays converge and the space between stays decreases, there is insufficient space to accommodate the Tedlar tape wrapping machine or the (MPC) inspection equipment.
    • (5) Wind buffeting loads during construction are a severe threat to the open bridge (center span closure not consummated). During construction in balanced cantilever, one must "overpull" the stay to prepare for unbalanced dead loads for one half the balanced cantilever cycle. For a light steel/concrete composite bridge, the cables are small enough that the saddle friction resistance is just about reached during application of the dead load. Wind load on top of the dead load imbalance can cause slip, which can destroy all control on the bridge, if not the bridge itself. In this regard, it is important to note that a factor of safety on this slip load is hard to attain. Slip is a function of the friction coefficient µ, and the angle of inclination (_), which is a constantfor a givencase. It takes a very large factor of safety on µ to get a normal factor of safety on the actual cable force imbalance across the pylon. For this reason, it is not practical for lighter bridge decks to achieve a conventional factor of safety (say above 1.5), even for dead load conditions.
    • (6) The advantage extolled for the use of cable saddles is cost. While material itself may be cheap, the system is not. Construction must be tightly balanced, which requires full operations at both ends of the bridge, and a "split" of cantilevering (i.e., steel up on one side, then on the other; then back to the first side for the panels, then back to the other, etc.). In addition, and most significantly, deck work must be done from the water rather than from the deck. In most jurisdictions, labor and equipment costs on water are much more expensive than over land. In some cases the difference can be 30 percent or more, when all fringes and insurance are considered. These costs can easily balance any material savings in the pylon.
    • (7) The most effective way to erect a cable-stayed bridge is by close geometric control. Unless the deck is stiff enough to affect cable force, erection control can not be effectively accomplished by force alone (The process is analogous to having a weight on a string--how do you control elevation by measuring force in the string?) Cable length is themost foolproof way to control bridge construction. With saddles, cable length cannot readily be controlled. With prefabricated cables, length control is relatively simple. With in-situ strand cables, it is much more difficult but not impossible when an upper anchor at the pylon exists. With saddles, it is impossible. Geometry control must be forced through deck elevation, which is better than no control, but varies with deck loads and temperature. Both of these effects are avoided when unstressed cable length can be controlled.


Cable stay design and construction practices for cable-stayed bridges have been evolving in the United States since the early 1970's. The recommendations listed below for design and construction practices are basedboth upon current state-of-the-art, United States experience with this evolving technology, and the experience gained from knowledge of international projects.

  1. Main tension elements. The "cables" of cable-stayed bridges should consist of parallel strands, approximately 15 mm in diameter, conforming to ASTM A416-90a (or later edition) "Standard Specification for Steel Strand, Uncoated Seven-Wire for Prestressed Concrete," and should be weldless, low-relaxation grade. Epoxy coated strand used in cable stays should conform to ASTM A882, "Standard Specification for Epoxy Coated Seven-Wire Prestressing Strand." For use in stay cables, epoxy coated strand should be of the type in which the interstices of the strand are filled with epoxy, and the strand should be approximately 15 mm in diameter, weldless, low-relaxation grade. Parallel solid bars should not be used for cable stays.
  2. Sheathing. Sheathing should be a high density polyethylene pipe (HDPE) conforming to requirements of ASTM D3035 or ASTM F714 depending upon nominal diameter. Subject to the approval of the engineer, the contractor may propose a white or light colored HDPE pipe, provided it has a demonstrated U.V. resistance equivalent or better than that of a black HDPE pipe. Steel pipe should not be used for the sheathing of cable stays.
  3. Corrosion protection. The corrosion protection system should bea multi-barrier, single-phase system that provides both the temporary and permanent system simultaneously, i.e., a system that provides protectionfrom manufacture of the cable throughout its service life. This can be provided by either an epoxy coated and filled strand or the monostrand system (greased and sheathed).
  4. Saddles. All stays should terminate at the pylon in appropriate anchorages. Saddles should not be used for cable-stayed bridges.


The following existing guidance serves in part as the basis for the recommendations set forth in paragraph 3. "Post-Tensioning Institute Recommendations for Stay Cable Design, Testing and Installation," August 1993. The FHWA has accepted these recommendations for the design of cable-stayed bridges, with the following exceptions:

  1. Solid bars should not be used for the main tension element of the cable stay,
  2. Steel pipe sheathing should not be used for cable stays,
  3. Saddles should not be used for cable stays, and
  4. Fatigue testing should be required for all cable-stayed bridges; results of previous fatigue testing will not be accepted.

William A. Weseman, Director
Office of Engineering

Updated: 07/02/2013
Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000