The Real Service Life and Repair Costs for Bridge Edge Beams

The Real Service Life and Repair Costs for Bridge Edge Beams Hans-Åke MATTSSON*, Håkan SUNDQUIST ** *) MSc, PhD Student, Department of Civil and Architectural Engineering, Division of Structural Design and Bridges (KTH), SE-1 44, Stockholm, Sweden. Email: Hans-Ake.Mattsson@byv.kth.se **) Professor, Department of Civil and Architectural Engineering, Division of Structural Design and Bridges, The Royal Institute of Technology (KTH), SE-1 44, Stockholm, Sweden. Email: Hakan.Sundquist@byv.kth.se ABSTRACT: Bridges rarely just break down. It is the different structural members of the bridge that slowly degrade the whole structure. One of the most damaged structural members of a bridge is the edge beam. This paper is based on a survey of 26 repaired or replaced edge beams located in a region in Sweden during a 15-year period, 199-25. It describes the real life span for edge beams and associated costs to repair them. Survival analysis of replaced edge beams is presented. Keywords: edge beams, real life span, repair costs 1 INTRODUCTION 1.1 General Bridges are an important part of a nation s road system. It is necessary that the bridges are in good condition to ensure a safe passage of the traffic and therefore bridges are inspected regularly. The inspection intervals used in Sweden, see Table 1, do not differ much from similar systems in other countries, even if the denominations of the different steps are different. Routine inspections (regular) are performed with short intervals and more in-depth inspections (general and major) are performed with longer intervals. Table 1. Inspection intervals for bridges in some Western countries. After Silfwerbrand [1] and Sommer et al [2]. Country Regular inspection General inspection Major inspection Belgium 1 year 3 years When needed Denmark 1-6 years France 1 year 5 year Germany 3 months 3 years 6 years Italy 3 months 1 year Sweden 1 year 3 years 6 years Switzerland 15 months 5 years When needed USA 2 years The results of the bridge inspections in Sweden are reported in two ways, a functional classification (condition classes) and an economical classification. The condition classes, see Table 2, can be an integer number, 1, 2 or 3, dependant on how serious the damage at the time of inspection is and its impact on traffic safety.

CC CC 1 CC 2 CC 3 Table 2. Condition Classes (CC) for Swedish bridges Expected loss of designed function beyond 1 years Loss of designed function expected within 1 years Loss of designed function expected within 3 years Total loss of designed function at the time of inspection A bridge owner who has many thousands of bridges to manage knows that it is a complex task to handle all the different information about the bridges. Therefore a bridge management system (BMS) is a must for the effective planning of the maintenance of the existing bridge stock and for procurement of new bridges. The Swedish Road Administration (SRA) has since the mid 197s used a computerized BMS. The latest update of SRA s BMS is called Bridge and Tunnel Management System (BaTMan) and it supports the management of a bridge structure during its whole lifecycle, from the design phase to the demolishing stage and after. BaTMan [3] was introduced in 24 as an Internet based system, which means that users all the time have updated information online about the actual bridges. The methods used in Sweden with both a physical and economical classification makes it possible to consider many of the important factors such as the initial cost and the residual life cost including planned maintenance, preventive maintenance and replacement. 1.1 Previous research and investigations Racutanu [4] collected and evaluated inspection reports and other significant information from 353 bridges built between 193 and 1994. These bridges were systematically inspected and are located in different parts of Sweden. A total of 3 747 bridge inspection remarks were gathered from the performed inspections where the damaged bridge structural parts, types and causes of damage were stated. In that study, four bridge structural members were found to contain almost half of the deterioration reported: the edge beams, supports, slope and embankment end and deck slabs, see Fig. 1. Waterprofing 5% Primary loadbearing elements 6% Wing & retaining walls 6% Surfacing 7% Deck slab 8% Slope & embankment end 1% Other 9% a) Edge beam 21% Parapet 13% Support 15% Fig. 1. Percentage distribution of the 3747 damage remarks on structural members from 353 bridges in different parts of Sweden (Racutanu [4]). The edge beam is the most damaged part of a bridge. Eriksson and Swanlund [5] collected and evaluated data and information about bridges in three counties in Sweden. The study showed that the edge beam was the most damaged part of a bridge.

1.2 Typical edge beam The main purpose of the edge beam is to hold the railing in place and stiffening the bridge deck. Edge beams are normally composed of concrete. In contrast to the bridge deck slab, which normally is protected with waterproofing and surfacing, the edge beam is directly exposed to environmentally induced degradation, see Fig. 2. 1 2 Fig. 2. A typical edge beam that is exposed to airborne pollutions (1) and water with chloride content (2). There are three major kinds of environmental actions, where salt is present that are considered to be of importance for the determination of the technical service life of an edge beam: frost attack, corrosion of reinforcement and carbonation. De-icing salts is used to keep the roads skid-free during the winter season and thus save lives. The salt is spread on relatively large areas around the road with the help of precipitation, wind and traffic. The salt will remain on the road during the winter until the spring when the edge beams will be washed free from salt. The edge beams are being affected negatively by the spread salt, especially elder bridges with high water-cement ratio concrete. More detailed descriptions about degradation processes can be found in e g the LIFECON-project [6]. 1.3 Purpose Since edged beam is one of the most damaged structural members of a bridge it is of interest for a bridge manager to know how long it normally will last. The purpose of this study is to try to find an answer to the following three questions; What is the real service life of an edge beam? Which factors are decisive for the real service life of an edge beam? What costs are associated to repairing a damaged edge beam?

2 METHOD 2.1 Choice of area in the investigation The total bridge stock owned by the Swedish Road Administration (SRA) consisted, in the autumn 26, of 15 3 bridges of which about 1 85 were located in SRA s Mälardalen Region (VMN). SRA has divided Sweden in seven regions. The reason to choose VMN for the study was that the first author is involved in an ongoing Ph.D. project on bridge maintenance in Uppsala County, Sundquist et al [7] and Mattsson [8]. Uppsala County which is situated some 2 km to 2 km north of Stockholm is a part of VMN. The region (VMN) contains some large and smaller cities but not the largest ones (Stockholm, Göteborg) and not vast wilderness, why it could be considered as a typical representative both for Sweden and Europe. 2.2 The information gathering process Information has been gathered from BaTMan during the spring 27. For every repaired or replaced edge beam it has been recorded year of construction and year of corrective action. The material of the primary load bearing structure has also been recorded. SRA has four responsible bridge engineers in VMN, one for each county (Uppsala, Västmanland, Södermanland and Örebro). Interviews of these SRA s bridge engineers have also been performed when there was some need for clarification. The gathered data about the edge beams have been used as input in a database. 3 SURVEY RESULTS 3.1 Material of the primary load bearing structure The material of the primary load bearing structure for bridges, where edge beams have been repaired or been replaced, have been concrete (95 %) and steel (5 %). 3.2 Replacement and minor repairs of edge beams In this survey 135 replacements and 125 minor repairs of edge beams have been included, see Fig. 3. Replacement is to demolish the old edge beam and construct a new edge beam. The average age for 135 replacements of the edge beam was 45 years with a standard deviation of 11 years. When an edge beam is replaced normally the waterproofing, the surfacing and the railing are also replaced. Minor repair is, for example, to patch repair the old edge beam with concrete and repair the parapet post fixing. The average age for 125 minor repairs of the edge beam was 28 years with a standard deviation of 15 years.

Number of bridges 3 25 2 15 1 5 Minor repair Replacement -4 5-9 1-14 15-19 2-24 25-29 3-34 35-39 4-44 45-49 5-54 55-59 6-64 65-69 7-74 Age of edge beam (years) Fig. 3. 135 replaced and 125 minor repaired edge beams in VMN during 199-25 sorted by age of edge beam. 3.3 Location of replaced edge beams Of 135 replaced edge beams 3 have been located on major European roads in Sweden (E roads): E4, E18 and E2, and 15 have been located on other roads, see Fig. 4. The average age for 3 replacements of the edge beams on E roads was 37 years with a standard deviation of 11 years. The average age for 13 replacements of the edge beams on other roads was 48 years with a standard deviation of 1 years. The major reasons for the shorter life span of edge beams located on E roads could be due to more wear caused by heavy traffic and use of more de-icing salt. Number of bridges 25 2 15 1 5 E roads Other roads -4 5-9 1-14 15-19 2-24 25-29 3-34 35-39 4-44 45-49 5-54 55-59 6-64 65-69 7-74 Age of edge-beam (years) Fig. 4. 135 replaced edge beams in VMN during 199-25 sorted by road type and age of edge beam. Data about the average annual daily traffic (AADT) for the bridges with replaced edge beams have been gathered from BaTMan, see Fig. 5. Regression analysis indicates that the life span for the edge beam gets longer when the AADT-value decreases.

8 Age of edge beam (years) 7 6 5 4 3 2 1 E roads Other roads R 2 =,28 5 1 15 2 25 Average Annual Daily Traffic (AADT) Fig. 5. 135 replaced edge beams in VMN during 199-25 sorted by AADT and age of edge beam. 3.4 Year of construction and replacement of edge beams Year of construction for replaced edge beams located on E roads have mainly been 1955-1974 and for other roads 193-1969, see Fig. 6. Number of bridges 6 5 4 3 2 1 193-34 1935-39 194-44 1945-49 195-54 1955-59 196-64 1965-69 197-74 R Other roads R E roads C Other roads C E roads Year of construction / replacement Fig. 6. 135 replaced edge beams in VMN during 199-25 sorted by road type and age of edge beam. C denotes year of construction and R denotes year of replacement. 1975-79 198-84 1985-89 199-94 1995-99 2-5 The life span of replaced edge beams have been plotted against year of construction, see Fig. 7. Regression analysis indicates that edge beams constructed during the 193-194 s have a longer life span compared to edge beams constructed during the 196-197 s.

8 Age of edge beam (years) 7 6 5 4 3 2 E roads Other roads R 2 =,88 1 193 1935 194 1945 195 1955 196 1965 197 1975 198 Year of construction Fig. 7. 135 replaced edge beams in VMN during 199-25 sorted by age of edge beam and year of construction 3.5 Bridges located on E roads and other roads in VMN Since the survey indicates that edge beams located on E roads have a shorter service life compared to other roads a more detailed study of bridges located on E roads and other roads was performed. The four most common types of bridges located on E roads and other roads in VMN are concrete beam and slab bridge, concrete slab bridge, concrete slab frame bridge and steel beam and slab bridge, see Table 3. Culverts have not been recorded since they don t have any edge beams. Bridges that have been constructed before 193 have also not been included in this study. Table 3. Type of bridges located on E roads and other roads in VMN. E roads Other roads TOTAL Type of bridge Material (No) (%) (No) (%) (No) (%) Beam and slab bridge concrete 67 16,8 81 9,1 148 11,5 Slab bridge concrete 42 1,6 145 16,3 187 14,6 Slab frame bridge concrete 267 67,1 545 61,4 812 63,2 Other concrete 12 3, 72 8,1 84 6,5 Beam and slab bridge steel 8 2, 41 4,6 49 3,8 Other steel 2,5 3,3 5,4 398 1, 887 1, 1 285 1, Fig. 8 shows an example of a slab frame bridge constructed of concrete. This type of bridge is the most common of all small sized bridges in Sweden and has been built with an increase in numbers and areas since the 193s.

Fig. 8. Bridge C293-1 and C293-2 located on E4 in Uppsala County. These slab frame bridges were built in 1972 and have a bridge area of 181 m 2 and 179 m 2, respectively. For 398 bridges located on E roads there have been noted 3 performed replacements (8 %), 32 minor repairs (8 %) and 336 without repairs (84 %) of edge beams, see Fig. 9. Most replacements and minor repairs have been noted for bridges constructed 1955 1974. The figure shows also actual Condition Classes for the edge beams. About 34 % of the edge beams have a Condition Class of 1-3. Number of bridges 1 8 6 4 2 Replaced Minor repaired CC 1-3 CC 193-34 1935-39 194-44 1945-49 195-54 1955-59 196-64 Year of construction Fig. 9. 3 performed replacements and 32 minor repairs of edge beams in VMN during 199-25 sorted by year of construction, with a total number of 398 bridges located on E roads. 1965-69 197-74 1975-79 198-84 1985-89 199-94 1995-99 2-5 For 887 bridges located on other roads there have been noted 15 performed replacements (12 %), 93 minor repairs (1 %) and 689 without repairs (78 %) of edge beams, see Fig. 1. Most replacements and minor repairs have been noted for bridges constructed 193 1974. The figure shows also actual Condition Classes for the edge beams. About 31 % of the edge beams have a Condition Class of 1-3.

1 8 6 4 2 193-34 1935-39 194-44 1945-49 195-54 1955-59 196-64 1965-69 197-74 1975-79 198-84 1985-89 199-94 1995-99 2-5 Number of bridges Replaced Minor repaired CC 1-3 CC Year of construction Fig. 1. 15 performed replacements and 93 minor repairs of edge beams in VMN during 199-25 sorted by year of construction, with a total number of 887 bridges located on other roads. 4 SURVIVAL ANALYSIS 4.1 General Survival analysis has been used for a long time in areas like medical research, Altman [9], and health sciences, Daniel [1]. In clinical studies, an investigator may wish to monitor the health change of patients from some fixed starting point in time, such as surgery, until the occurrence of some well defined event, ultimately death. The time elapsing between enrolment in the study and the event (death) is referred to as the patient s survival time. The statistical treatment of survival times is known as survival analysis. Although the patients will be followed up for several years there will be many who are still alive at the end of the study. Their survival time from surgery will be longer than their time in the study. Such survival times are called censored. From a set of observed survival times and censored times from a sample of individuals one can estimate the proportion of the population of such people who would survive a given length in time. One common method to use is the Kaplan-Meier procedure. The procedure involves the successive multiplication of individual estimated probabilities and it is sometimes referred to as the product-limit method of estimating survival properties. If p k is the probability of surviving k years, r k is the number of objects still at risk immediately before the kth year, and f k is the number of observed failures on year k, then rk f k pk = pk 1 (1) rk The equation above can be used for a population of edge beams. Edge beams can be monitored from a fixed starting point (construction year) until some well defined event (replacement). For example a population of 5 edge beams of which one edge beam was replaced after 3 years. We have p = 29 1, and r 5 because all subjects are still at risk at 3 years. There was one failure at 3 years, so 3 = f = 1 3 and we can calculate the proportion surviving 3 years as (5 1) p 3 = p29 =,98 5

The estimated proportion surviving stays the same until the next failure time. The probability of surviving to time t, S (t), is estimated by multiplying the survival probabilities across the time periods r 1 f1 r2 f 2 rt f t S( t) = (2) r1 r2 rt It is usual to present survival properties as a graph. The median survival time can be read from the plotted curve, being the time corresponding to a cumulative survival proportion of,5. The tail of the survival curve is often unstable due to the small numbers at risk. 4.2 Survival analysis of replaced edge beams located on E roads and other roads in VMN The survival analysis of the replaced edge beams located on E roads in VMN is based on 368 bridges with no recorded replacements and 3 bridges with recorded replacements during the period 199-25. In the survival analysis the edge beam is defined as dead when it is replaced. Minor repair is not calculated since the old edge beam is still alive. The survival analysis of the replaced edge beams located on other roads is based on 782 bridges with no recorded replacements and 15 bridges with recorded replacements during the period 199-25, see Fig. 11. 1 Cumulative Survival Proportion,75,5,25 E roads Other roads 1 2 3 4 5 6 7 8 Age of edge beams located on E roads and other roads (years) Fig. 11. Survival curve for edge beams located on E roads and other roads in VMN. The survival curves have been curtailed when there were only five objects still at risk. From the above figure one can read median survival time, the time when the curve crosses the probability of,5. The median survival time for edge beams located on E roads is 58 years. The median survival time for edge beams located on other roads cannot be calculated since the survival curve doesn t drop below,5, see Table 4. The table also shows the different ages of bridges at the cumulative survival proportion of,95 and at the end of the curve.

Table 4. Real service life of edge beams located on E roads and other roads. Road Min (,95) Median (,5) At the end of the curve European 3 years 58 years 6 years (,48) Other 4 years N/A 75 years (,64) For young edge beams (about 25 years) no difference could be noted for edge beams located on European roads compared to edge beams located on other roads. As the edge beams gets older the difference increases between edge beams located on European roads (shorter service life) compared to other roads (longer service life). 5 COSTS FOR REPAIR AND REPLACEMENT OF EDGE BEAMS In BaTMAn there is a list of costs for different kind of bridge work, excluding the contractors establishing costs. In this list the cost to replace 1 m of edge beam is about 5 ksek (1 ksek is about 11 ). Costs for minor repairs like patch repair of the old edge beam and repair of the parapet post fixing can also be found in this list. If one converts the costs for the minor repairs to replacement costs, the minor repair will be about 1 % of the replace cost of the edge beam. If we use the findings, in section 3.2, that the average age for replacement of the edge beam was 45 years with a standard deviation of 11 years and that average age for minor repair of the edge beam was 28 years with a standard deviation of 15 years, and assuming that minor repair cost is 1 % of replacement cost we can construct a relative cost diagram for the edge beam, see Fig. 12. From the figure one can see that the relative cost for the edge beam is low at early age and increases as the edge beam gets older. One can also see that there is some 2-3 years difference between an early need and a later need for repair. This means that the knowledge and judgement of the responsible bridge engineer is important. It also indicates that it is cost effective to keep the edge beam in good condition by means of preventive bridge maintenance before the deterioration process starts to accelerate. To compare the results from VMN general data for edge beams in the Stockholm area have been gathered from Stockholm Konsult. Edge beams with an age of about 5 years need to be replaced if they have not been repaired during the last 2 years. The replacement cost of 1 m edge beam is about 5 ksek (Replacement). For an edge beam with an age of about 3-4 years it could be enough to dismantle the railings, remove only the damage part with water jet and cast replacing concrete. If the edge beam is about 4 years one can assume that the damaged part is bigger and the cost can be estimated to be about 5 % of the replacement cost (Repair). If the edge beam is about 3 years one can assume that the damaged part is smaller and the cost can be estimated to be about 3 % of the replacement cost (Minor Repair). For young edge beams it could be enough with normal maintenance (Maintenance), see Fig. 12. Both the data from VMN and Stockholm area show similar relative cost development for the edge beams. The relative cost for the edge beam is low at early age and increases at an accelerating rate as the edge beam gets older.

1,2 1, Replacement MR&R relative costs,8,6,4,2, Early cost Maintenance Minor repair Avg cost 1 2 3 4 5 6 Age of edge beam (years) Repair Late cost Fig. 12. Relative costs for maintenance, repair and replacement of edge beams in VMN (exponential curves) and Stockholm area (step curve), respectively. 6 CONCLUSIONS This study has found following answers to the three questions in the purpose section; 1) What is the real service life of an edge beam? The average age for replacement of the edge beam was 45 years with a standard deviation of 11 years. For E roads the average age was 37 years with a standard deviation of 11 years and for other roads the average age was 48 years with a standard deviation of 1 years. Survival analysis shows that the real minimum life span for edge beams located on E roads was 3 years and for other roads 4 years. Survival analysis shows also that the real median life span for edge beams located on E roads was 58 years and for other roads at least 75 years. There is two ways to describe the real service life for an edge beam population. The first way is to analyse only the replaced edge beams and estimate a service life. This estimated service life will probably be too low for the whole edge beam population since only bad edge beams counts. The second way is to analyse both replaced edge beams and edge beams that have not yet been replaced, survival analysis. This seems to be a better approach since all available data about the edge beam population is used. 2) Which factors are decisive for the real service life of an edge beam? For young edge beams (about 25 years) no difference could be noted for edge beams located on E roads compared to edge beams located on other roads. As the edge beams gets older the difference increases between edge beams located on E roads (shorter service life) compared to other roads (longer service life). The major reasons for the shorter life span of edge beams

located on E roads is probably due to more wear caused by heavy traffic and use of more deicing salt. 3) What costs are associated to repair a damaged edge beam? The cost for minor repairs is about 1 % of the cost for replacement of the edge beam. It indicates that it is cost effective to keep the edge beam in good condition by means of preventive bridge maintenance before the deterioration process starts to accelerate. 7 FURTHER RESEARCH An area for further research could be to predict long term behaviour (e g the next 2-5 years) of the studied population using survival analysis. This could be done by estimating the remaining service life of the existing edge beam population and estimate the number of new bridges that will be constructed and added to the edge beam population in the future. This will result in some different scenarios depending on the estimates. Maybe the end result will be a worst case scenario, a very optimistic scenario and a realistic scenario. 8 REFERENCES 1. J. Silfwerbrand. Aktivt brounderhåll en förstudie. (Active bridge maintenance a pilot study). TRITA-BKN Rapport 65, Dept. of Structural Engineering, Royal Institute of Technology (KTH), Stockholm (22). 2. A. M. Sommer, A.S Novak and P. Thoft-Christenssen. Probability-Based Inspection Strategy. ASCE Journal of Structural Engineering, Vol. 119, No. 12, pp. 352-3536 (1993). 3. https://batman.vv.se/ 4. G. Racutanu. The Real Service Life of Swedish Road Bridges A Case Study. TRITA-BKN Bulletin 59, Dept. of Structural Engineering, Royal Institute of Technology (KTH), Stockholm, PhD thesis (2). 5. D. Eriksson and S. Swanlund. Förebyggande underhåll. (Preventive bridge maintenance). Examensarbete 1993:9, Dept. of Geotechnical Engineering, Chalmers University of Technology, Gothenburg, Sweden (1993). 6. http://lifecon.vtt.fi/ 7. H. Sundquist., H-Å. Mattsson and G. James. Procurement of bridge management based on functional requirements. Second International Conference on Bridge Maintenance, Safety and Management, IABMAS, Kyoto, Japan (24). 8. H-Å. Mattsson. Funktionsentreprenad brounderhåll. En pilotstudie i Uppsala län. (Bridge Maintenance based on Functional Requirements. A Case Study in Uppsala County). TRITA- BKN Bulletin 82, Dept. of Structural Engineering, Royal Institute of Technology (KTH), Stockholm, Lic thesis (26). 9. D. G. Altman. Practical Statistics for Medical Research. Chapman & Hall, London, ch. 13, 365-395 (1991)

1. W. W. Daniel. Biostatics. A foundation for analysis in the health sciences. John Wiley & Sons, Hoboken NJ, 8 th edn., ch. 12, 647-658 (25).