A Simple Method of Gripping Prestressing Strands for Tension Tests

Transcription

1 A Simple Method of Gripping Prestressing Strands for Tension Tests Presents a simple technique for gripping a prestressing strand to measure its breaking strength, modulus of elasticity or stress-strain behavior. A combination of prestressing chucks and strand splicing sleeves are used to anchor the strand in an ordinary universal testing machine with V-grips. Test results from using this method are compared with those employing a special strand testing machine. Chandu V. Shenoy Structural Engineer Blakeslee Prestress, Inc. Branford, Connecticut Gregory C. Frantz Associate Professor Department of Civil Engineering University of Connecticut Storrs, Connecticut 58 A tension test is, conceptually at least, perhaps the simplest of mechanical tests. However, performing a tension test on a length of prestressing strand is very difficult. Due to the nature of both the strand and the common universal testing machine, substantial effort and/or expense is needed to produce reliable test results. This paper presents a simple way of gripping a prestressing strand to perform a tension test of the strand. The method can be used with a relatively short length of strand and can be utilized with any universal testing machine. CURRENTLY USED TEST METHODS ASTM A3701 states, "the true mechanical properties of the strand are determined by a test in which fracture of the specimen occurs in the free span between the jaws of the testing machine." However, the serrated teeth of the gripping devices of ordinary testing machines, such as those used to test solid steel tensile coupons, will bite deeply into the outer surface of the wires and will usually result in a premature failure of one of the wires in the grips. This same ASTM Standard lists several procedures that can be used for testing strand. These methods consist of using a cushioning material such as aluminum foil between the grips and the strand, encasing the gripped portion of strand in tin, using epoxy to bond the strand inside metal tubing, using smooth grips with an abrasive slurry, using sockets with zinc to anchor the strands, or using dead end eye splices. They caution that prestressing chucking devices should not be used because the sharp teeth of the chuck will also nick the wires and lead to premature failure just as with standard PCI JOURNAL

2 t I"± 24" PLP GRIP _j MIN CLEAR STRAND-- 28 " I 24" PLP GRIP 5".t ''v.. GRIPS IN TESTING MACHINE 5"! -- I. 24" -, 15" 15" EXTENSOMETER ~ M_I_N_IM U_M L_E_N_G_T_H _J.I I" t ~-- -- O F_T_E_ S_T S_A_M P_L_E_=_S_'-_4_" ~ ~1 USE TESTING MACHINE "v" GRIPS FOR 3 /4" DIA. TO 1" DIA. BARS. THESE DETAILS APPLY TO 3 /e" AND 7 1t6" DIA. STRANDS. Fig. 1. Details of the PLP splice sleeve (1 in. = 25.4 mm) (from Ref. 2). test machine grips. A paper by Preston 2 evaluated several of these methods. The following techniques produced the best results: 1. Sand grip - The strand is gripped by specially fabricated long, smooth U-grips with a sand slurry coating applied to the strand to help prevent slipping. This method yields only about 50 percent clear breaks and is not preferred over the following three methods, which yield about 90 percent clear breaks. 2. Aluminum insert - Aluminum angles grip and cushion the strand in machines that use tapered mechanical or wedge grips. 3. Tinius-Olsen grip - These are long, toothless U-grips that have been developed and fabricated by the Tinius-Olsen Testing Machine Company. The grips fit only special kinds of test machines. This method is usually used only by strand producers or large testing companies. 4. PLP grip - These grips are made from devices for splicing seven-wire strand that are manufactured by the Preformed Line Products Company. The splice is composed of two sets of three wires and one set of four wires that are helically preformed so that they will fit tightly around the strand. The inside surfaces of the splice wires are coated with a grit which also helps to prevent slippage. The PLP splice is then gripped in a testing machine. About 23 in. (580 mm) of splice at July-August 1991 each grip will develop the full ultimate strength and elongation of a ~ in. (12.7 mm) Grade 270 strand. Fig. 1 shows details for the PLP grip for % and Yl6 in. (9.5 and 11.1 mm) strands. METHOD DEVELOPED BY THIS RESEARCH As part of a research projectl.4 studying the strength of old prestressed concrete bridge beams, it was necessary to measure the stress-strain curve of some 'X6 in. ( 11.1 mm) prestressing strands that were removed from the beams. Some preliminary testing was done using standard 4 in. (102 mm) V grips or using standard chucking devices to anchor the strand in a universal testing machine. Premature failures resulted with both methods due to fracture of wires in the gripping devices. The PLP splices recommended by Preston are no longer being manufactured. However, Florida Wire and Cable Company manufactures the FLO-LOC strand grip system, which is very similar to the PLP splice and commonly used in cross member bracing systems of buildings. This FLO LOC strand grip system, which is sized for various strand diameters, uses a helical winding of two groups of five wires interwoven around the strand to encase it. The wires are also coated with a grit to help prevent strand slippage. Tests were conducted on Yl6 in. ( 11.1 mm) Grade 250 and ~ in. (12.7 mm) Grade 270 strands using the appropriate sizes of the FLO-LOC gripping devices. Testing was done in a Young universal testing machine equipped with standard 4 in. (102 mm) long V grips typically used for testing % to 1 in. (19.0 to 25.4 mm) diameter coupons. The strands with the sleeves had outside diameters of about 0.69 and 0.80 in. (18 and 20 mm) for the Yl6 and ~ in. (11.1 and 12.7 mm) strands, respectively. Fig. 2 shows a sketch of the test setup. Preliminary testing showed that the strand slipped in the sleeve at a load of about 7 kips (31 kn) for the ~ in. (12.7 mm) strand in 14 in. (356 mm) long splice sleeves, and at about 23 kips (102 kn) for the 'X6 in. (11.1 mm) strand in 17 in. ( 430 mm) long splice sleeves. It appeared that the V -grips could bear on a larger outside portion of the smaller diameter sleeves of the 'X6 in. (11.1 mm) strand than they could on the larger diameter sleeves of the ~ in. (12.7 mm) strand. This probably accounted for the much higher load at slipping for the 'X6 in. (11.1 mm) strand. Tests were not conducted to determine how much length of sleeve, if any, would prevent slipping and yield clear center breaks of the strand as reported using the PLP splices. Various methods to increase the frictional resistance between the 59

3 V - GRIPS ( 4") MACHINE HEAD CHUCK V- GRIPS (4") MACHINE HEAD SLEEVE SLEEVE... STRAND STRAND Fig. 2. Original gripping method using sleeves and V-grips (1 in. = 25.4 mm). Fig. 3. Modified gripping method using sleeves, V-grips and chucks (1 in. = 25.4 mm). strand and the sleeves were tried, including inserting carborundum cloth between the sleeves and the strand. However, none of the methods prevented slipping of the strands at relatively low loads. MODIFIED TEST METHOD It was clear that using these lengths of sleeves with our particular test machine would not produce satisfactory results. Nor would using prestressing chucks as gripping devices be successful. Therefore, the test method using the FLO-LOC sleeves was modified (Fig. 3). The sleeves were attached to the strand so that about 4 in. (102 mm) of strand extended beyond the ends of the sleeves. As before, the sleeves were gripped using the 4 in. (102 mm) long V -grips. The specimen was carefully aligned so that the ends of the sleeves did not extend beyond the outside ends of the V -grips. A load of about 5 kips (22 kn), which was below the slipping load, was applied to seat the sleeves in the grips and to seat the wedge-shaped grips in the machine head. The machine was then held at constant load while standard prestressing chucks were attached to each end of the strand with the chucks bearing on the V -grips. The test was continued by slowly increasing the load. Figs. 4 and 5 show details of the modified test setup. The steel plate below the chuck was attached to the test machine for safety in case of a specimen rebounding out of the machine after failure (none did). The sleeves prevent the grips from nicking the strand and resist part of the Fig. 4. Modified test setup (unloaded). 60 Fig. 5. Detail at end of specimen. PCI JOURNAL

4 Table 1. Description of tests. Strand gripping method Breaking Specimen Area With With With First slip Seat chuck strength No. sq in. chucks V-grips sleeves kips kips (kips) Comments on failure li in. diameter Grade 270 strands 1 NA X NS 39.2 At top at firs t notch in grip 2 NA X NS 35.0 At top at fi rst notch in chuck jaw 3 NA X X NS 34.5 At top in chuck jaw X X 12 in. 7* At bottom, inside sleeve, 2 in. outside grips 5 NA X X 14 in. 7* Two wires fai l inside top sleeve, li in. ins ide grips, remaining fi ve wires fail in center X X 20 in. NS Clean break at center y., in. diameter Grade 250 strands X X 20 in. NS Clean break at center 8 NA X X 17 in. 23.5* Clean break at center X X 14.5 in. NS Clean break at center Tested by Florida Wi re and Cable II Tested by R orida Wire and Cable Tested by R orida Wire and Cable Metric (SI) conversion factors: I in. = 25.4 mm; I kip= len. Notes: "NA" means not available. "NS" means no slip was observed; for Specimens 6, 7 and 9 the chucks were installed before slipping occurred. These specimens were first installed without chucks and were loaded to first slipping; then the specimen was unloaded, chucks were installed and seated, and then loading continued to failure. tension force. The chucks prevent the strand from slipping through the sleeves and provide the additional resistance needed to properly test the strand. Furthermore, the bearing of the chuck on the V -grips wedges the V grips even tighter against the sleeve and strand to help prevent further slippage. The length of the sleeve needs to be only long enough so that the sum of the chuck strength and the sleeve strength exceeds the strand capacity. TEST PROGRAM A series of twelve tests were done on both ;.{6 in. ( 11.1 mm) Grade 250 and ~ in. (12. 7 mm) Grade 270 strands (see Table 1). Various methods of gripping arrangements, as well as various lengths of sleeves, were tried. Only a small number of tests were performed due to the limited number of available ;.{6 in. (11. 1 mm) strands (which had been removed from an old bridge beam) and a limited supply of FLO-LOC grip systems. To check our results, Florida Wire and Cable Company tested three pieces of the ;.{6 in. (11.1 mm) Grade 250 strands using a Tinius-Olson test- July-August 1991 ing machine specially designed for testing seven-wire strands and using an extensometer. INSTRUMENTATION A suitable extensometer was not available for our tests. On several strands, strains were measured in the center region between the sleeves using either three or four SR-4 electrical resistance strain gages of Ys in. (3.18 mm) gage length bonded to different wires of the seven-wire strand. The clear length between sleeves varied from 5 to 24 in. (127 to 610 mm). Load was applied with a 60 kip (270 kn) hydraulic universal testing machine. The machine was held at constant load for about 15 seconds while the gages were read. Readings were taken at load increments of 2 kips (9.0 kn) except near the failure load where the increment was 1 kip (4.5 kn). TEST RESULTS Table 1 summarizes the results. Specimens 1 to 6 were Y, in. (12.7 mm) Grade 270 strands used to develop the test method. When these strands were gripped using only machine V -grips (Specimen 1) or only prestressing chucks (Specimen 2), no strand slippage was observed; however, the strands failed prematurely at the location of the first nick caused by the V -grip or the chuck jaw. Specimen 2, using only chucks, carried about 85 percent of the full strand strength. Using a combination of chucks and machine V-grips (Specimen 3) produced the same mode of failure as using only chucks, probably because the chuck jaw teeth were much sharper than the worn V -grip teeth. When Specimens 4 and 5 were encased with 12 and 14 in. (305 and 356 mm) lengths of sleeves, respectively, and then gripped with only the machine V -grips, the strands slipped in the sleeves at very low loads. When using these same specimens with the modified procedure with a chuck attached, no slippage was observed, but the break occurred inside the sleeve at a load slightly below the full strand strength. Specimen 6 with chucks and sleeve lengths of 20 in. (508 mm) produced a clear break in the center region of the specimen. 61

5 Specimens 4 and 5 were first tested without chucks to determine their slipping loads. This initial slippage may have worn off some of the grit coating on the sleeves and may have lowered the effectiveness of the sleeves when these same specimens were subsequently tested to ultimate capacity. It is possible that even these shorter sleeve lengths may have been effective for this modified method if slipping tests had not been done first on these specimens. Specimens 7 to 12 were Y.6 in. (11.1 mm) Grade 250 strands that had been removed from the same bridge beam. Specimens 7 to 9, tested by the modifled test method, and Specimens 10 to 12, tested by Florida Wire and Cable _... ~ 200 CIJ CIJ 150 (]) : (fj o Gage 1 6 Gage 2 o Gage 3 * Gage Fig. 8. Stress-strain curve for Specimen 7 (1 ksi = MPa). 300 Fig. 6. Typical break in center region. 250 t:,... ~ 200._... CIJ CIJ (]) 150 :... 0 Gage (fj Gage Gage 3 Fig. 7. Closeup view of strand showing pure tensile failure Fig. 9. Stress-strain curve for Specimen 9 (1 ksi = MPa). PCI JOURNAL

6 Company with a special test machine, were randomly selected from the six available samples. Specimens 7 to 9 had measured areas less than the nominal area, while Specimens 10 to 12 had measured areas greater than the nominal area (Table 1). To eliminate the effect of variations in the strand area measurements, it is probably best to interpret the test results using the measured load or the stress and modulus calculated using the nominal strand area (as used in the figures in this paper). Specimens 7, 8 and 9 used the modified procedure with sleeve lengths of 20, 17 and 14.5 in. (508, 432 and 368 mm), respectively. Overall strand lengths and the corresponding clear spans between sleeves were 66 and 18 in. (1680 and 457 mm), 48 and 6 in. (1220 and 279 mm), and 48 and 11 in. (1220 and 279 mm), for Specimens 7, 8 and 9, respectively. Each specimen failed with a clear break in the center region (Figs. 6 and 7) and showed the characteristic pure tensile failure surfaces. It is possible that even shorter sleeve lengths may be satisfactory; however, this minimum length was not determined. Specimens 7 and 9 were instrumented with strain gages. The test results, using nominal strand area, are shown in Figs. 8, 9 and 10. These curves are similar to the load-deformation curves typically recorded for strands. The curves are best fit curves through the average strain values. The two tests compare well with each other. The last data point in each figure corresponds to the last load stage for which a strain reading was available, not the maximum load or strain sustained by the specimen. The total strain at failure was not recorded. Data were recorded past the elastic and yield points well into the inelastic region of behavior. By using only one strain indicator to monitor all of the strain gages, it was not possible to get reliable strain readings past the points shown. Although the readings from the individual strain gages varied, using several gages and the best fit curve through all of the data seemed to give a reliable stress-strain curve. Of course, an extensometer rather than strain gages could be used with this July-August _... ~ 200 (/) (/) Q) 150 '- +-1 en 100 Specimen 7 50 Specimen Fig. 10. Comparison of stress-strain curves for Specimens 7 and 9 (1 ksi = MPa) ~ 250 en 200 en Q) 150 ' ~ (f) Specimen 10 Specimen 11 Specimen Fig. 11. Stress-strain curves for Florida Wire and Cable Company Specimens 10, 11 and 12 (1 ksi = MPa). 63

7 300 - ~ 250 en 200 ~ 150 to... en Modified Specimen 7 FWC Specimen Fig. 12. Comparison of stress-strain curves for Specimens 7 and 12 (1 ksi = MPa). Table 2. Comparison of test results for V.s in. strands. Nominal * Actual t Actual Breaking Breaking Modulus of Breaking Modulus of Specimen a rea load strength elasticity strength elasticity No. sq in. (kip) (ksi) (ksi) (ksi) (ksi) Using modified test procedure Average Using special strand testing machine lo II Average Metric (S l) conversion factors: I sq in.= sq em; I kip= kn; I ksi = MPa. Using nominal area of0.108 sq in. t Using measured strand area. modified gripping method. Fig. 11 shows the stress-strain curves for Specimens 10 to 12, tested by the Florida Wire and Cable Company using their special testing machine. Fig. 12 compares the stressstrain curves of Specimen 12 and our Specimen 7. The stress values in Figs. 11 and 12 are based on nominal strand area. Table 2 compares the essential 64 I results from all of these tests using both the actual and nominal strand areas. There is good agreement between the two methods, especially when using the nominal strand area. The two test methods gave essentially identical breaking strengths. As expected, because of the tendency of the strand to straighten under loading, the modulus of elasticity determined by using strain gages applied directly to the individual wires was higher than the modulus determined by using an extensometer attached to the entire strand. Generally, it is very difficult to accurately measure the stress-strain curve for strands without special extensometers. If strain gages are used, experienced personnel are required to install the very small strain gages. It is best to verify the testing technique by comparing some test results with those from a strand fabricator, as was done in this study. ADVANTAGES OF THE MODIFIED TEST METHOD These tests have demonstrated a relatively simple procedure for gripping a prestressing strand for performing an accurate tension test of the strand. This gripping method is a significant improvement over the previous gripping methods. Some of these methods require fabrication of special machine grips or the use of special test machines. When gripping uses only sleeves, such as the PLP or FLO-LOC sleeves, the required sleeve length is affected by many variables, such as (1) the type of sleeve, (2) how well the sleeve fits around the strand, (3) the direction of the helical twist pattern of the strand and sleeve, (4) how well the V grips fit around the sleeve, and (5) the length of the V -grips. In this modified method, different types of commercially available sleeves and common universal testing machines can be used. The length of the sleeve is not nearly as critical since the chucks can carry most of the load. The strand gripping technique consists of a combination of encasing the ends of the strand with splicing sleeves (such as the FLO-LOC strand gripping system or a similar system) and using standard prestressing chucks attached to the strands and bearing on the ends of the V -grips. The sleeves prevent the grips from nicking the strand and resist part of the tension force. The chucks prevent the strand from slipping through the sleeves and provide the additional resistance needed to properly test the strand. The chucks \ I PCI JOURNAL

8 also wedge the V -grips even tighter against the sleeves. The length of the sleeve needs to be only long enough so that the sum of the chuck strength and the sleeve strength exceeds the strand capacity. Since the sleeve does not carry the full tension force, the sleeve length can be relatively short, permitting tests of shorter strand lengths in shorter testing machines. With this technique, laboratory personnel can easily solve the difficult problem of how to grip a strand by using a common universal testing machine and readily available sleeve gripping devices. RECOMMENDATIONS The performance of this gripping technique depends on many variables, such as the type of strand, splicing sleeve and test machine used. It is recommended that initial tests be done with a universal testing machine using standard V -grips, with sleeve lengths of about 15 in. (380 mm) long, and without chucks. If slipping does occur, chucks can be added to provide positive anchorage of the strand. ACKNOWLEDGMENTS This work was performed as part of the research project, "Tests of Prestressed Concrete Bridge Beams," sponsored jointly by the Connecticut Department of Transportation and the Precast/Prestressed Concrete Institute. The authors gratefully acknowledge the assistance provided by Blakeslee Prestress, Inc., in Branford, Connecticut, and, Florida Wire and Cable Company in Jacksonville, Florida, for donating the FLO-LOC grips and for performing stress-strain tests, H. Kent Preston, and graduate_student Valerie Murray. REFERENCES 1. ASTM, " A370, Methods and Definitions for Mechanical Testing of Steel Products," Annual Book of ASTM Standards, V , American Society for Testing and Materials, Philadelphia, P A, Preston, H. K., "Testing 7-Wire Strand for Prestressed Concrete - The State of the Art," PCI JOURNAL, V. 30, No. 3, May-June 1985, pp Shenoy, C. V., and Frantz, G. C., "Tests on Prestressed Concrete Bridge Beams, Part I - Structural Tests of the Bridge Beams," Research Report JHR 91-l98a, Project 87-3, Department of Civil Engineering, University of Connecticut, January 1991, 120 pp. 4. Murray, V. E., and Frantz, G. C., "Tests on Prestressed Concrete Bridge Beams, Part II - Chloride Penetration in the Bridge Beams," Research Report JHR b, Project 87-3, Department of Civil Engineering, University of Connecticut, January 1991, 125 pp. July-August