New York. Performance of Cable Guiderail in WEI-SHIH YANG, LUIS JULIAN BENDANA, NICHOLAS J. BRUNO, AND WAYNE D. KENYON

Transcription

1 TRANSPORTATON RESEARCH RECORD Performane of Cable Guiderail in New York WE-SHH YANG, LUS JULAN BENDANA, NCHOLAS J. BRUNO, AND WAYNE D. KENYON Cale guiderail ith insffiient tension may deflet -exessively on n_npat, allowmg vehiles to ontat fixed objets behind the bamer. n 198 a two-phase study was initiated by the New York S_tate Department of Transportation to investigate auses of tension loss in able guiderail and formulate orretive measures. The study's first phase doumented performane of new able barriers in the field and the results of laboratory testing. Anhor movement and_ permanent le streth were identified as major auses of tension loss, suffiient to affet barrier performane adversely. Several hanges have already been made based on these results: onstrution speifiations and standard sheets were hanged to ensure proper soil ompation and better initial and long-term able tension. n the seond phase, field performane of seleted improved installations was doumented from 1984 to!987. n ad?ition.' presressed ble was used on some projets m 1985 to mvest1gate its effetiveness in reduing tension loss due to able streth. Laboratory streth tests were onduted uing normal and prestrethed able to determine any signifiant differenes appearing in able strain due to long-term loading. Rs1:11ts fro field and laboratory tests indiated that able guiderail mstallat1ons ontinually lose tension and need to be retensioned periodially and that substituting prestrethed for normal able does not redue the tension-loss problem. Lightweight able guiderail now in use in New York State was developed in the late 196s (J,2). The able is designed to separate from S3 x 5.7 steel posts on vehile impat, with tension in the able developing the fore neessary to retain and rediret vehiles. The tension in the barrier before impat affets the total able defletion that must our to develop this fore. The rail elements onsist of three %-in. galvanized steel ables mounted on the posts by hook bolts. The ables are seured to onrete anhor bloks at the ends of eah installation.to develop tension. Details of the able-guiderail system are shown in Figure 1. Spring-ompensator devies are inluded to allow for able length hange due to temperature hange. When properly adjusted, these spring ompensators should maintain a working range of able tension between 45 and 1,8 lb throughout the annual temperature yle without any need for periodi adjustment. The standard sheets require that in ases where the able run is 1, ft or moe, springs are required at eah end; otherwise they require spnngs at only one end (3,4).. During the 1979 New York State Department of Transportation (NYSDOT) Highway Safety Review, a problem was deteted related to the safety of able guiderails, onerning their inability to rediret traffi beause of insuffiient tension Engineering Researh and Development Bureau, New York State Department of Transportation, State Campus, Albany, N.Y in the ables. During inspetions almost every able installation observed was found to have insuffiient tension. nitially, there was onern that proper installation proedures had not been followed. After further investigation, it beame apparent that able guiderail ould beome slak even if installed and tensioned orretly. Besides being unattrative, slak able guiderail is a potential safety hazard beause it limits the ability of the barrier to rediret vehiles within the allowable defletion range. A vehile that impats an installation with insuffiient tension an be guided into an objet while being redireted. To address these onerns, a researh study was initiated in the spring of 198. ts objetives were fourfold: 1. Determine the extent of slak able guiderail, 2. dentify the auses, 3. Propose orretive ation, and 4. Verify that proposed solutions are effetive by onduting long-term follow-up surveys. The investigation began monitoring several able guiderail installations. t was determined that these installations quikly lost able tension, thus reduing their potential effetiveness. n the first phase, several possible auses were investigated, inluding anhor movement, post settlement, able reep, post movement, spring ompensator failure, aident impats, inadequate maintenane, inorret initial installation proedures, and turnbukle baking off. Major auses were identified as anhor movement, able reep, and nonuniform tension distribution throughout the barrier aused by fritional drag at the posts. Even after these installations were retensioned, they experiened unaeptable degrees of able tension loss. Based on the findings of Phase 1 (5), several hanges were made. First, onstrution speifiations were hanged in 1982 to ensure proper soil ompation during the plaement of onrete able anhors and to redue anhor movement to aeptable levels. Seond, the standard sheet for able guiderail was revised to ensure better initial and long-term able tension through improved installation proedures. Third, revised speifiations requiring prestressed able were used on some projets in With suh able, it was assumed that tension loss due to streth ould be redued. To determine the long-term performane of these orretive measures, a seond phase was initiated in The Phase 2 objetives were to monitor the effetiveness of the new speifiations and orretive measures and to doument the results of field and laboratory tests on prestrethed able.

2 1 TRANSPORTATON RESEARCH RECORD 1419 NORMAL END SECTON.-----Conrete Anhor Blok Fae of Posts 18'" Plan View 6'" 6'" 6'" 6'" 16'" TO" Anhor Blok Elevation DRVEWAY END SECTON 4"x6" Wood Post 8 1 " 16, 'O" Normal Spaing S3x5.7 Steel Posts Front Fae Of Post 4"x6" Wood Post Plan View 4-3/4" + 1/2" 16'" Normal Spaing Conrete Anhor Blok---- SPRNG TURNBUCKLE DEVCE 2 'O" 14-1/8" 1'6" Standard Turnbukle (12" Takeup) FGURE 1 Cable-guiderail detail (1 mm =.4 in.). Periodi ondition surveys were arried out for 3 years, monitoring both normal and prestrethed ables. Laboratory tests were onduted fo determine the differene in streth between the two able types. The effetiveness of the orretive measures and use of prestressed able are doumented in a final report (6). NVESTGATON AND RESULTS Phase 1 n the first phase, possible auses of tension loss were investigated and identified. A total of 53 new installations on 12 onstrution projets were monitored during 198 to observe and doument installation proedures, and referene systems were installed to monitor hanges in the barrier. n addition, individual site parameters were reorded for eah sample to determine what effet, if any, they had on tension loss. These parameters inluded the presene and degree of horizontal and vertial urves, length of run, ontrator who installed the barrier, temperature at the time of tensioning, and whether inspetions were by state or onsultant personnel. Minor alterations in the tensioning proedure were also tried on 15 of these runs in an attempt to remedy the problem of tension loss. After promising orretive measures were formulated, these modifiations were inluded on 21 new installations in 1981 to evaluate their effetiveness. Some of the ables plaed

3 Yang et al. in 198 were also retensioned in 1981 to see if setting proper tension a seond time would aid in maintaining tension. Cable tension, ourrene of aident damage, and any hanges ourring with respet to the referene systems were monitored throughout the year by onduting spring, summer, and fall surveys. Snow and ie made it impossible to obtain winter measurements, but reonnaissane surveys were made throughout winter months to keep trak of aident and snowplow damage. f any sample installations were readjusted or experiened damage that would affet tension, they were no longer surveyed. All data obtained from the surveys were organized in tabular form to failitate statistial analysis. As disussed later, the tensioning proedure, anhor movement, and permanent able streth were identified as major auses of tension loss and were found suffiient to harm barrier performane (5). Evaluation of Tensioning Proedure Four tensioning proedures were ompared to verify whether any benefits were obtained by hanging normal tensioning pratie: (a) a normal-tension group with 154 samples, (b) a 198 revised-tension group with 78 samples, () 1981 revisedtension group with 63 samples, and (d) a retensioned group with 73 samples. The sequene for installing able guiderail using these four proedures was as follows: 1. Normal-tension proedure: After posts were driven and anhors plaed, the able was unrolled and ut to the approximate length. The able was strung through the J-bolts, unloaded spring lengths marked on the ompensator rod, and the anhor hardware attahed to the able at one end and seured to the anhor. With the able now fixed at one end, it was pulled straight by applying tension at the opposite end with a hand winh or pulling with a truk to remove the slak. With this tension held by the truk, or winh, or loking pliers lamped on the able against posts, the able was ut to the final length. The anhor hardware was then installed at this end and seured to the anhor, and the slak between the anhor and the point at whih the tension was being held was taken up with the turnbukle. The temporary lamps were then released, leaving the able seured to both anhors with some intial tension present revised-tension proedure: This proedure involved plaing 1,6 lb of initial tension on eah able. This value equates to 3.5 in. of spring ompression, slightly below the upper limit of 4 in. After 2 to 3 weeks, tension was set to the standard-sheet value if the ables had not already relaxed to that level revised-tension proedure: This proedure inluded two additional modifiations of the 198 revised-tension proedure. First, to overome the problem of fritional drag, the springs were ompressed by applying tension at the opposite end of the barrier. By pulling the able the entire length of the barrier, the amount of tension at any point had to be at least the value indiated by the springs at the far end. Seond, these runs were tensioned aording to l F temperature intervals orresponding to Y4-in. spring ompression inrements as presented in Table 1. TABLE 1 Spring Compression Settings for Cable Tensioning Previous Settings (2 deg F, -in. inrements) Temperature Range, F -2 to to to to to to to Spring Compression, in. Note: t = (tp - 32)/1.8-1 N = lbf 11 Revised Settings (1 deg F, -in. inrements) Temperature Range, F -2 to to to to to to to to to to to to to to Spring Compression, in. 4. Retensioned proedure: Samples installed with the normal tensioning proedure were retensioned aording to the updated temperature-spring ompression settings to see whether setting proper tension a seond time would help maintain it better. The ables in the normal-tension group experiened an average 26 perent loss (298 t) between the time they were tensioned during late summer and fall 198 and the fall 198 survey. This loss is based on the differene between atual measured values and theoretial tension values at the measurement temperature. After the first winter, an average 46 perent loss (465 lb) had ourred, and by spring 1983 average loss was 62 lb, or 56 perent. Seventy-seven perent of the total loss measured in spring 1983-after three winters in servie-had ourred by spring 1981, and thereafter the gap between theoretial and measured tension widened at a muh slower rate. Also, it was found that tension was not distributed uniformly through the able. Figure 2 shows measured tension values throughout a 1,946-ft barrier just after tensioning by the normal proedure. Tension in the middle portion of the run was about 5 perent less than that indiated by spring ompression at the ends. Fritional drag at the able-post onnetion aused nonuniform tension throughout the barrier. The ables in the.198 revised-tension group experiened an average loss of 19 perent, or 315 lb, during the 2- to 3- week period when tension was left at the high initial level. By spring 1981, after one winter in servie, average tension loss was 196 lb, or 19 perent of the theoretial value, referened to the tension value at the end of the pretensioning period. By spring 1983, atual losses averaged 32 perent of the theoretial value, or 325 lb. Losses ourring by spring 1981 averaged 78 perent of the 1983 loss. Forty-nine perent of the spring 1983 total loss-that was ourring during the 2- to 3-week pretensioning period, plus the loss taking plae thereafter-ourred before final adjustments were made at the end of the pretensioning period. The high loss during the high initial tension period was not ritial sine it ourred after final adjustments were made. After one winter, tension loss in the normal-tension group was 137 perent greater than

4 12 TRANSPORTATON RESEARCH RECORD 1419 \ \ \ \ Top Cable i Middle Cable - _J Bottom Cable llo f t Barrier Setions FGURE 2 Distribution of tension through barrier after tensioning by former normal proedure (1 N =.225 lbf, 1 m = 3.28 ft). in the 198 revised-tension group, but by spring 1983 tension loss in the normal-tension group was 85 perent greater. Figure 3 shows that by spring 1983, all samples from both the normal-tension group and the 198 revised-tension group experiened tension losses greater than 2 lb, onsidered the maximum aeptable loss. However, only 1 perent of the 198 revised-tension group had experiened losses greater than 4 lb by 1983, ompared with 95 perent of the normaltension group. The 2-lb maximum tension loss was established so that the minimum tension in the able would not drop below 45 lb. The rationale for this riterion was developed as follows (5): First, initial tensions and field temperatures were reorded for the maximum length of the able used, 2 ft. Seond, these tensions were adjusted to a maximum design temperature of 95 F. Third, maximum tension loss was alulated so that barriers may not drop below the minimum desired tension of 45 lb at 95 F. The 1981 revised-tension group, installed in fall 1981, experiened a 483-lb (28 perent) loss during the high-initialtension period. After final adjustments, additional losses of 43 and 527 lb (37 and 46 perent) ourred by spring 1982 and spring 1983, respetively. By spring 1983, 94 perent of the sample experiened tension loss greater than 4 lb, ranging up to 681 lb. This group lost onsiderably more tension than the 198 revised-tension group. The 1981 revised-tension samples were set at higher average tension on final adjustment than the 198 revised-tension sample for two reasons. First, the 1981 proedure pulled out all slak from the total length of barrier at one end, so average tension throughout those runs was higher 'than in the 198 group. Seond, the 1981 group was set to l F intervals rather than the 2 F intervals used in 198. Beause the 1981 runs were tensioned to a higher level, the greater loss was less ritial. Average measured tension values are in the same range as the 198 revisedtension group, even though the theoretial loss for the 1981 group was greater. The retensioned group experiened signifiant tension loss after being readjusted in the fall of Relative to the final adjustment, this group of 15 runs experiened average tension losses of 253 lb (26 perent) and 38 lb (28 perent) by spring 1982 and spring 1983, respetively. This group ontained six runs-two 198 normal-tension and four 198 revisedtension-that were retensioned by the ontrator in late spring These six runs thus were retensioned twie. Considering

5 Yang et al. 13..: (/) ti) ::l..-i u > u..: E--< µ NORMAL TENSON 1 9:: ' " \ 'e-.. \ ', " \. \ ' 2.,-,.,-.::\ , 'r-11= ,- 198 REVSED TENSON - Spring 1983 Spring Spring 1981 ti) ti)...l r-i ti) E--<..:.µ r-i (a) 1981 REVSED TENSON () \ RE TENSONED..-i P.. (/) 4-l 'r\-.. -1< ;, ' (b) (d) / Tension Loss, lb FGURE 3 Distribution of tension loss after final adjustments 198 revised tension (1 N =.225 lbf). the six separately reveals a relatively small tension loss by spring 1983 of only 119 lb, or 11 perent. Tension throughout two barriers tensioned by the revised proedure is shown in Figure 4. The expeted tension distribution did not always our, and Barrier A shows a slightly lower tension at mid-run than at the ends. However, variation throughout this run is only about 1 perent, ompared with as muh as 5 perent for barriers tensioned by the normal proedure. Tension in Barrier Bis more uniform, with the top and middle ables having tensions higher than the spring-indiated value, and only the bottom able has some measured tension values slightly lower than the springs indiate. Even though this proedure did not preisely dupliate the expeted results, it produed a more uniform distribution of tension throughout the barrier than the normal proedure. Anhor Movement The four groups just desribed were also monitored for anhor movement. Some movement ourred during tensioning of most of the sample runs, and resulted in a visible gap between the anhor and the earth behind it as springs were ompressed. By the time the last able was tensioned, this movement was sometimes large enough to result in dereased spring ompression for the other two ables. Researh personnel told the ontrator about this movement so that orretions ould be made. Anhor-movement measurements presented here are the ombined movements of both anhors relative to their position immediately after the barrier was tensioned. Anhor movement for eah sample group is summarized in Figure 5. The urves represent average measured anhor movements for the samples in eah survey. Most movement ourred immediately after tensioning and between the first fall and spring surveys. After the first spring, a slight derease in average anhor movement was often noted, probably aused as the anhors settled bak. During and after tensioning, the anhors tipped forward, pressing against the fresh fill and ompating the soil beneath the front portions of the anhors. Most guiderail is installed during late summer and fall, and falling temperatures maintain a load on the anhors. Movement probably eases during winter beause soil around the anhor freezes, but high moisture levels and thawing in spring result in low soil support. This poor support is normally oupled with relatively high able tension from the ool temperatures, ompared with installation, and this situation probably produes the signifiant movement measured in the first spring

6 14 TRANSPORTATON RESEARCH RECORD B_A_RR_l_E_R_A_C_2_8_M_(4_2o_F_n_ "O. ;::l C 9--P- -<.-1 :::::: (/) (/) E-<. u :6 6 u u. u u.-1 s H µ., µ.) (/)..-1 H p.. Cf.l --- Top CAble Middle Cable --- Bottom Cable BARRER B (98 M (32 Fnl #-. -::: ,... -:..:....,,...,.,,,..... "' (/). "O.-1 Cl H p.. Cf.l ;::l p... H :::::: (/). (\j u. u.-1 :i s H µ.. "O µ.) o Anhor Anhor 16-ft Barrier Setions Anhor ft Barrier Setions FGURE 4 Distribution of tension in two barriers after final adjustments for four sample groups (1 N =.225 lbf). survey after installation. As temperatures moderate, able loads derease, allowing the anhors to press against the soil behind and beneath them. As this soil ompats, the anhors settle bak, thus negating some of the forward movement. After the first spring, anhor movement generally flutuated slightly, but signifiant movement rarely ourred unless the barrier was retensioned, at whih point movement showed another major inrease. For the 198 revised-tension group, 56 perent of the total spring 1981 anhor movement ourred during the high-initialtension period, and for the 1981.revised-tension group, 54 perent of the total spring 1982 movement ourred during this period. Total anhor movement for the revised-tension groups-whih is the sum of the values above and below the datum line-is greater than the total amount experiened by the normal-tension group, but movement affeting working tension in the barrier is less. Anhor movement for the retensioned samples is also shown in Figure 5. This group of 15 samples ontains 6 installations that were retensioned twieone by the ontrator, and then one by researhers. Before retensioning, the anhor movement trend for the whole group was similar to the other sample groups, with substantial movement totaling 2.82 in. by the first spring. Movement then leveled off until the samples were retensioned by researhers in fall 1981, at whih point signifiant movement again ourred, totaling an additional 1.9 in. by spring The movement that ourred before retensioning in fall 1981 appears below the datum line in Figure 5. Thirty-seven perent of the normal-tension sample experiened anhor movement exeeding 2 in., ranging up to 6Y4 in. By omparison, none of the 198 revised-tension group, the 1981 revised-tension group, or the retensioned-group experiened anhor movement greater than 2 in. by the first spring after final adjustment. The revised proedures thus resulted in very signifiant redution in ritial anhor movement ompared with the normal-tension group. Although the revised tensioning proedures redued the effet of anhor movement on tension loss, additional measurements were desirable to stabilize the anhors. Anhor plaement was observed on 12 projets to determine typial proedures used. Generally, the hole was dug with a bakhoe, the anhor plaed, and soil then bakfilled and ompated around the anhor; however, time and effort spent plaing and ompating the fill varied. Some rews plaed the bakfill in five or six lifts, tamping eah, while others dumped the bakfill around the anhor and simply dropped the bakhoe buket around the top of the fill to ompat it. Roks and large hunks of asphalt pavement were sometimes inluded in the bakfill, making ompation diffiult. Careful bakfill proedures generally resulted in less movement, but did not guarantee stable anhors. Plaing the anhor in the ground a few weeks before tensioning also seemed to have the favorable

7 Yang et al Normal-Tension Group Revised-Tension Group Revised-Tension Group Re tensioned Group r-1.j..j E > ;::;::,.... (J 1 4 : ;:::---:.-:::;. :: : ;' Datum Line --:-;_i_i_i t-/-/-,-,-/-1 =--::...:.==--,... -:=--./.../--./ o O'\ o o O'\ Cl) r-1 h. Cl) Note: Surveys are not shown on a relative time sale. o o N o ("") o O'\ O'\ O'\ O'\,... ) OJ) o r-1 r-1,...,... :l.. Cl) Cl) UJ Survey FGURE 5 Summary of anhor movement for four sample groups (1 mm =.4 in.). effet of reduing movement. Constrution speifiations permitted onsiderable latitude in plaement proedure, and firm guidelines for bakfilling and ompation were not inluded. This problem was brought to the attention of the NYSDOT Soil Mehanis Bureau, and standard sheets were revised in the hope of ensuring that adequate ompation is ahieved around all anhors. Speifially, limits of exvation have been provided, and the revised speifiation requires suitable fill material to be paed in 15-mm (6-in.) lifts and ompated to 95 perent of standard Protor maximum density. Permanent Cable Streth Construtional streth is an inherent property of all wire rope produts. This deformation is permanent and remains after load is released. Construtional streth an be removed by prestrething, whih involves subjeting the able to repetitive loadings of up to 5 to 6 perent of its ultimate strength (7). Manufaturers of a wire rope ontated in this study laimed that guiderail able may experiene a permanent streth of.25 to.5 perent of its unloaded length (personal orrespondene, S. E. Chehi, Bethlehem Steel Corporation, Marh 1981). This muh streth would produe large tension loss in able guiderail. Aording to industry representatives, guiderail working loads of up to 2, lb would never remove all the potential streth, whih would ontinue as long as the able was loaded (personal orrespondene, S. E. Chehi, Bethlehem Steel Corporation, Marh 1981). They further indiated that even setting initial tension in the upper range of the working load will not remove all potential streth, but probably would help redue tension loss from subsequent able streth. Prest rething the able at loads of 1,2 to 15, lb would be the only way to remove all the streth, but this was not onsidered feasible by one major manufaturer beause of lak of failities to perform the work. The manufaturer did prestreth one reel ontaining 1,78 ft of able. This material was then made available for field instal-

8 16 TRANSPORTATON RESEARCH RECORD 1419 r-l t-----::\ =---::.,,,,,; """---t r-l.µ (ij o r-l i:i.:i r-l (ij.µ E--< (ij ::r: ].. u E--< (J :6 '.'.5!-< (ij ::r:...µ "Cl...g ul Days FGURE 6 Permanent able streth over 12-week period (1 mm =.4 in.). lation to monitor its performane. However, no data were reorded during prestrething to determine able length hanges, so testing to determine the behavior of guiderail able under working loads was initiated as part of this study. Figure 6 shows the streth that ourred over a 12-week period under a load of about 1,9 lb. Muh of the total streth ourred during the first 2 days that the able was plaed under load. Additional signifiant streth ourred when the able was tapped with a hammer, reating vibrations that apparently helped seat the wires and lays. By the beginning of the twelfth week, a total of.52 in. of permanent streth had ourred over the instrumented length of 8 ft 1 in., equating to strain of.5 in.jin. This able streth testing was initiated late in the projet shedule, after learning that the manufaturer did not doument the prestreth testing. Thus, time allowed for only one test, whih is insuffiient for firm onlusions. These preliminary test results, based on only one sample of able, plus information supplied by manufaturer of wire rope indiate that permanent streth is a major ontributing fator in tension loss. This problem is not easily remedied in the field beause normal working loads are not large enough to remove all potential onstrutional streth (the streth during installation). More testing is neessary to determine the total range of onstrutional streth to be expeted for a large sample of able. The reel of prestrethed able donated for this projet was installed in the fall of This 25-ft run was tensioned aording to the 1982 revised proedure, exept for exluding the period of high initial tension, beause all pending onstrutional streth was supposed to have been removed. The barrier thus was tensioned aording to the revised tensioning table and left at this value. The primary interest in this test was the reation of other barrier omponents if the able did not undergo onstrutional streth. By spring 1983, an average tension loss of 681 lb had ourred for the three ables, aused mostly by anhor movement. The ombined anhor movement of 2V2 in. is equivalent to a theoretial loss of 723 lb for this run. These results support the hypothesis that if able streth does not our, the anhors or some other portion of the barrier yield instead to relieve the load. n 1984 seleted projets were onstruted with normal able guiderail using improved installation proedures. Conditions were surveyed periodially for 3 years from 1984 to 1987 to determine their effetiveness. Methods similar to those used previously and desribed for Phase 1 again doumented installation and long-term performane. Phase 2 To determine long-term performane of prestrethed able, the NYSDOT Materials Bureau responded to a request by researh personnel by issuing a speial speifiation for pre-

9 Yang et al. 17 TABLE 2 Field Survey Results Avg Tension Loss, lb (aeptable loss = 2 lb) Loation Fall Spring Fall Spring Fall Spring Fall Group Tested '87 '87 '86 '86 '85 '85 '84 NSTALLED Sag Spring Avg Sag Spring Avg Sag Spring Avg Sag 661 Spring 154 Avg 48 5 Sag Spring Avg Sag Spring Avg Sag Spring Avg NSTALLED (Normal) Sag Spring Avg (Prestrethed) Sag Spring Avg Note: 1 N.225 lbf Average = Average = Average = 412 strethed able guiderail in Suh able should experiene smaller degrees of permanent streth, plaing greater loads on guiderail omponents over longer periods. This may or may not have a long-term effet on guiderail omponents, and ultimately on tension. Using this speial speifiation, additional test setions were installed in 1985, and were also surveyed twie a year in 1986 and 1987 using the same proedures. During eah field ondition survey, able tensions were measured at two positions for sag at a low point on eah able's run between posts and at the spring-ompensator. The weight of the able and able defletion were used to ompute able tension. Tensions of all three ables (top, middle, and bottom) at these two positions were measured and averaged to represent tension at that partiular position. The differene be- tween design tension and measured tension is the tension "loss" given in Table 2. Average tension loss in Table 2 is the average of tension loss measured at sag and at the springompensator. Groups 1 through 7 were installed in 1984 and monitored for 3 years. Groups 8 and 9 were installed in 1985 using the revised speifiation to ompare the differene between normal and prestrethed able. These new installations were also monitored twie a year for 2 years. From the field results, overall average tension loss for the first set of installations (Groups 1 through 7) was about 33 lb. For the other installations (Groups 8 and 9) it was greater than 4 lb. Both exeed the aeptable level of 2 lb. Laboratory streth tests were onduted using normal and prestrethed able to find any signifiant differenes in able strain due to long-term loading. Bethlehem Steel Corp. pro-

10 18 TRANSPORTATON RESEARCH RECORD 1419 Plate Welded to Sleeve for Dial-Gage Plunger to Contat Sleeve Mounted on Cable with Set Srews 2" Round Aluminum Strutural Tubing Tubing Welded to Sleeve Mounted on Cable with Clamp 19-lb Conrete Blok FGURE 7 Cable streth tests: as able speimen elongates under load, strutural tubing fixed to able bottom moves down with it, and gauge mounted on top of tubing registers this movement with respet to plate fixed above able (1 m = 3.28 ft, 1 mm =.4 in., 1 kg = 2.21 lb). vided a reel of prestrethed able for this purpose. A series of tests was onduted in the laboratory to doument the amount and nature of permanent streth ourring during normal working loads to whih guiderail is subjeted. This involved suspending a length of new able and loading it with a onrete blok weighing about 1,9 lb-near the upper extreme during the annual temperature yle, if the spring onstant were at the upper aeptable limit of 5 lb/in. The spring limit is speified as 45 ± 5 lb/in. Thus, extreme able loads would vary from 4 to 2, lb with a spring ompression range of 1 to 4 in. over the antiipated temperature range. Amount of streth was measured with a dial gauge reading to thousandths of an inh. Figure 7 shows the apparatus and experimental setup. To ensure that the tests simulated field onditions, the following observations were made. During field tensioning, spring ompression is set at the ompletion of the tensioning proedure, so that streth ourring during tensioning does not affet barrier performane. Only streth ourring after tensioning is of onern in terms of effet on barrier performane. n the laboratory, dial-gauge readings thus began immediately after the 1,9-lb load was applied. Results of laboratory able streth tests are given in Table 3, whih gives results of tension loss if the same able streth experiened in the laboratory ourred in the field. The data show that able samples tested under a onstant 1,9-lb loading elongated to an unaeptable length. f the same onstant tension were plaed on guiderail in the field, an unaeptable level of able streth would our. The degree of elongation experiened by these laboratory samples would translate into unaeptable tension loss, if able used in guiderail installations elongated the same amount, as shown in Table 4. Sine the able used in suh installations does not experiene onstant tension, these laboratory results annot be used to predit how muh tension will be lost; however, they an predit that the able is apable of strething to unaeptable lengths if the tension is maintained. This experiment may show that another less plasti material might be used in plae of steel able, although this imaginary produt would still have to be plasti enough to allow the barrier to deflet if impated. Results of the laboratory streth tests show that prestrethed able elongates less than normal able, but that elongation is enough that guiderail installations would lose all their tension in a relatively short time.

11 TABLE 3 Laboratory Cable Streth Tests Strain in Normal Cable, in. (orreted for temperature) Test Test Test Test Test Test Test Test Days Average * Strain in Prestrethed Cable, in. (orreted for temperature) Test Test Test Test Test Test Test Days Average *Value used in Table 4. Note: 1 mm =.4 in. TABLE 4 Guiderail Tension f Cable Streth in Laboratory Ourred in Field Normal Cable Barrier Length, ft Prest rethed Cable Barrier Length, ft Days Dimension o, in..79* P, lb 318* o, in P, lb o, in P, lb o, in LS P, lb o, in P, lb *Sample Calulation: 6 =.13 (from Table 3). A=.22 in. 2 E 11 x 1 6 lb/in. 2 e: = 6 in./99 in. o = PL/AE = L x 12 in./ft x e: in.fin. = 5 ft x 12 in./ft x.13/99 =.79 in. p = oae/l = (.79 in. x.22 in x 11 x 1 lb/in. )/5 ft x 12 in./ft 318 lb Note: 1 m = 3.28 ft

12 2 TRANSPORTATON RESEARCH RECORD 1419 CONCLUSONS From the results of this study, the following onlusions may be drawn: Barriers installed using either the normal or revised tensioning proedures experiened greatest tension loss soon after the barrier was first tensioned and over the first winter. Tension losses generally ontinued after this point, but at a muh slower rate. f the barrier is retensioned, a new yle of tension loss ours. Even with the proposed retensioning proedures, tension loss will ontinue to our in able guiderail, regardless of the installation proedures used. The revised proedure oupled with at least two retensionings will probably be neessary to onfine losses to 2 lb. Substituting prestrethed for normal able in these guiderail installations does not solve the tension-loss problem. nstallations using prestrethed able lose tension at almost the same rate as those using normal able. Cable guiderail installations ontinually lose tension, and thus must be retensioned periodially. The data, however, were insuffiient to estimate how often this must be done. REFERENCES 1. Burnett, W. C., J. L. Gibson, and R. H. Freer. New Highway Barriers: The Pratial Appliation of Theoretial Design. Researh Report Bureau of Physial Researh, New York State Department of Publi Works, Albany, May Whitmore, J. L., R. C. Piioa, and W. A. Snyder. Testing of Highway Barriers and Other Safety Aessories. Researh Report 38. Engineering Researh and Development Bureau, New York State Department of Transportation, Albany, De Standard Speifiation-Constrution and Materials. Offie of Engineering, New York State Department of Transportation, Albany, Jan Standard Sheet 66-JR2. New York State Department of Transportation, Albany, Ot Kenyon, W. D. Cable-Guiderail Tension. Researh Report 124. Engineering Researh and Development Bureau, New York State Department of Transportation, Albany, July Yang, W., N. J. Bruno, and W. D. Kenyon. Tension Loss in Cable Guiderail. Speial Report 14. Engineering Researh and Development Bureau, New York State Department of Transportation, Albany, Marh Bethlehem Wire Rope for Bridges, Towers, Aerial Tramways, and Strutures. Catalog Bethlehem Steel Corporation, Bethlehem, Pa. Publiation of this paper sponsored by Committee on Roadside Safety Features.