A visual condition assessment of a reinforced concrete railway bridge subject to alkali silica reaction (ASR) deterioration in Johannesburg

In the 1940’s, a reinforced concrete (RC) railway bridge located in central Johannesburg was constructed using Witwatersrand Quartzite, at the time, unknown to be an alkali-silica reactive aggregate. The bridge currently displays signs of distress in the form of severe map-cracking. This study presents preliminary results of a series of tests that have been conducted to characterise the severity of alkali silica reaction (ASR) deterioration in the bridge. Twelve concrete cores were extracted for testing from various locations of the railway bridge, which displayed varying degrees of distress. A visual assessment was conducted globally on the entire structure as well as locally (using modified damage rating index (DRI) approach) on the extracted cores. The findings showed that there was extensive map-cracking present on all thirteen railway bridge elements examined, revealing signs of distress in the bridge. Even elements that were sheltered from exposure to direct rain, such as the underside of three arch surfaces, also exhibited extensive map-cracking. Furthermore, the modified DRI method was a useful technique to assess and compare the relative extent of damage of concrete taken from various locations of a RC structure, using local images of the cylindrical surface of the concrete cores.


Introduction
Alkali-silica reaction (ASR) is one of the leading causes of concrete distress, resulting in loss of serviceability in affected structures [1]. ASR occurs when certain types of disordered or poorly crystalline silica in aggregates reacts with alkali hydroxides present in the concrete pore solution, resulting in the production of a secondary reaction product as "ASR gel" or "silica gel" [2,3]. Upon exposure to moisture, the above-mentioned reaction product swells and results in concrete cracking and expansion. Studies by Diamond [4] has shown that tensile stresses generated can reach 6 MPa to 7 MPa. Such cracking can significantly impact the mechanical properties of effected elements where a considerable reduction of tensile strength (i.e., up to 65%) and stiffness loss (up to 50%) at low/ moderate expansion (0.05-0.12%) while compressive strength losses can be up to 35% at very high expansion (around 0.30%) [5]. Despite many structures having been affected by ASR damage, case studies have shown that these structures can still meet and exceed their design requirements and design life. This has been achieved through correct diagnosis, successful repairs and preventive maintenance throughout and beyond the structure's design live [6].
A reinforced concrete (RC) railway bridge located in central Johannesburg which displayed signs of distress in the form of severe map-cracking, was selected as the field structure that was subject to ASR deterioration. The railway bridge is currently managed by the Passenger Rail Agency of South Africa (PRASA). Two conditions were required for the selection of the field structure in this study: the first being that the region contains aggregates that are reactive and secondly, that the in-service structure exhibits a degree of visual distress owning to ASR deterioration.
There is a known history [7] of aggregates prone to ASR in the Johannesburg region. The underlying geology is known to contain potentially alkali reactive aggregates such as quartzite and shales associated with the gold-bearing reefs. Furthermore, prominent RC structures based in the Johannesburg region such as the "double-deck" stretch of the De Villiers Graaff motorway (designed in 1963) [6] is known to be affected by ASR. Furthermore, ASR is an expansive reaction between * Janina P. Kanjee janina.kanjee@wits.ac.za 1 School of Civil and Environmental Engineering, University of the Witwatersrand, Johannesburg, South Africa alkalis and aggregates in the concrete and this reaction process take a long time to manifest [8]. Newly built structures will typically not show signs of visual distress owing to ASR in the first five years of the structure's lifetime (unless the structure contained highly reactive aggregates). If ASR has been initiated within the concrete, the first symptoms are usually noted from visual inspections. The most common observations are variable expansion, development of extensive cracking, surface discolouration, gel exudations and occasional pops-outs [9]. The selected RC railway bridge displays signs of distress in the form of severe map-cracking. Accurate assessments of the condition of deteriorating structures subject to ASR are essential to ensure safety, schedule the repair/replacement of structures and manage the costs associated with public infrastructure [10]. In structures such as the RC railway bridge (selected in this study), various non-destructive and destructive techniques were used to determine the structural health of the infrastructure. This paper is focused on using visual inspection techniques to determine the condition of a RC railway bridge subject to ASR deterioration and to characterise the extent of damage.
This work forms part of a broader research effort to establish reliable and effective ways to assess the condition of structures, owing to ASR deterioration through the use of various diagnostic techniques. The findings presented here, will later be compared to microscopic, mechanical and residual expansion tests results which will be performed on the extracted concrete cores. The aim is to consolidate all the results to develop a more effective management approach for concrete infrastructure subject to ASR deterioration.

Materials and methods
The following section outlines the experimental approach used to characterise the distress in the railway bridge using a visual inspection approach both globally on the entire structure as well as locally on the extracted concrete cores.

Visual inspection of railway bridge
Once the RC railway bridge was identified, an on-site visual assessment was undertaken in accordance with the Guide for Conducting a Visual inspection of Concrete in Service [11]. The results of the visual inspection were then used as input, to evaluate the probability of presence of ASR of the field structure as either a low, medium or high [12]. Depending on the outcome of the probability assessment, a decision was taken whether to continue with the experimental programme by extracting cores for further laboratory testing or select another field structure that is already subject to ASR deterioration.

On-site sample collection: number of specimens, specimen type and size
Owing to the heterogeneity of concrete within a large structure, various areas were identified on the bridge and marked as "damaged" or "undamaged" areas. Factors such as element exposure conditions, access to moisture, and visual evidence of likely ASR deterioration (map-cracking, light, gel-like exudations) were considered when establishing areas that were "damaged" and "undamaged". A minimum of three concrete core specimens were selected from each of the identified "damaged" and "undamaged" areas. The minimum core diameter of 100 mm as well as maximum core barrel length of 300 mm was specified to meet the minimum specimen requirements to conduct the bulk expansion tests [12]. Coring was carried out under "wet" conditions (using water), with acknowledgement of the possibility that any ASR gel present on the exterior core surface may have been washed away. Concrete cores were labelled and transported to the laboratory for further testing.

Assessment of as-received cores
Concrete cores were first examined in an "as-received condition". A visual, macroscopic inspection of the cores was performed and a core log was completed. Observations such as size and distribution of aggregates, compaction and voids, presence and condition of reinforcement were recorded.
A 2D image of the circumferential surface of each concrete core was created. The Damage Rating Index (DRI) (a semiquantitative) method [13] was used to evaluate the extent of damage in the concrete. In this study, the approach was modified to characterise the 2D images of the circumferential surface of each concrete core. A grid of 2 cm squares (as opposed to the recommended 1 cm squares) were overlaid on the 2D images. The 2 cm size of the grid was used as the examination of the image was done with the naked-eye. The examination involved counting the number of the features noted in Table 1 for each square of the grid (examples of certain petrographic features are shown in Fig. 1). After these features were recorded for the sample, the total is multiplied by the appropriate weighting factor (representing the relative severity of the feature in the overall ASR deterioration process, Table 1), to produce a gross DRI value. This is value was then normalised to a standard area of 100 cm 2 , to produce a final DRI for the sample. The higher the DRI number, the higher the damage due to ASR [15].

Results & discussion
The results and analyses of the data obtained from the visual assessment of the RC railway bridge subjected to ASR deterioration are presented in this section. All data generated or analysed during this study is available from the authors upon reasonable request. Table 1 Weighting factors assigned to petrographic features to determine the damage rating index [14] No Petrographic feature Weighting factor Comment i Cracks in coarse aggregate 0.25 A low factor is given as such cracks are likely produced by aggregate processing operations (quarried aggregate) or weathering (gravel) ii Opened cracks in coarse aggregates 2 A "network" of cracks is likely caused by expansive reactions within the aggregate particles iii Crack with reaction product in coarse aggregate 2 Cracks containing secondary reaction products (whitish, glassy or chalky in texture) Sometimes, the secondary products do not fill all the cracks (material lost during the preparation of the section) iv Disaggregate/corroded aggregate particle 2 Aggregate particle that shows signs of disintegration, "corrosion" or disaggregation v Coarse aggregate debonded 3 Crack showing a significant gap in the interfacial zone between the aggregate particle and the cement paste Would likely cause debonding of the particle when fracturing the concrete vi Cracks in cement paste 3 Crack visible but with no evidence of reaction products vii Cracks with reaction product in cement paste 3 Cracks containing secondary reaction products (whitish, glassy or chalky in texture) Sometimes, the secondary products do not fill all the cracks (material lost during the preparation of the section)

Structural review of railway bridge
Existing construction drawings of the RC railway bridge were signed in the year 1946, this provided an indication of the period when the Railway Bridge was possibly constructed, making the structure approximately 76 years old. The drawings indicated that the RC Railway Bridge consists of three 32 feet (9.7 m) RC arches, furthermore details of the volumetric concrete mixture used were also noted ( Table 2).

Visual inspection of railway bridge
Three structural elements of the railway bridge were selected to be visually examined: the abutments, the piers and the arches. This resulted in a total of thirteen different elements that were inspected across the entire railway bridge. In summary the results of the visual inspection indicated that globally, there was extensive map-cracking present on all thirteen railway bridge elements examined, revealing signs of distress in the bridge. Elements that were sheltered from exposure to direct rain, such as the underside of three arch surfaces, also exhibited extensive map-cracking. Based on the observations noted in the visual assessment report, each of the thirteen different elements were then evaluated for the probability of ASR deterioration. Table 3 is a summary of all thirteen elements inspected. It incorporates six features that are typically associated with the ASR deterioration accompanied by descriptions of the severity of the feature as the deterioration mechanism progresses over time. The description of the features that was most applicable to the RC Railway bridge structure based on the visual assessment observations were noted in the appropriate column in italics.
It can be seen that four of the six features in the table fall under the medium and high probability category, indicating that the rail bridge structure is likely prone to ASR deterioration. The outcome of the visual assessment provided sufficient motivation for concrete cores to be extracted from the structure for further laboratory testing.

Assessment of as-received cores
Twelve concrete cores were extracted from various locations of the railway bridge, which displayed varying degrees of visual distress. Table 4 summarises the extent of visual distress observed on the RC railway bridge and the respective concrete cores that were extracted from the various locations.
Examination of the concrete cores showed that the concrete mixture used consisted mainly of quartzite and shale aggregates, with the largest aggregates size being approximately 50 mm. With regard to macro-features observed, cracking was found to be more pronounced in the coarse aggregates as compared to in the cement paste. There was also evidence of debonded aggregates in various concrete cores. There were no visible signs of any gel exuding from the aggregate cracks, cement paste or surrounding  Outdoor exposure but sheltered from wetting Parts of components frequently exposed to moisture, eg., rain ground water, water due to natural function of the structure the aggregate particles. This result may be attributed to the gel being washed-out when the cores were extracted from the structure during the coring process (water was used as a coolant).  Figure 2 is an example of a 2D local image of concrete core C6 cylindrical surface that was generated and overlaid with 2 cm grid.

Determination of damage rating index
In general it can be seen in Fig. 3, that for all twelve concrete cores analysed, the majority of the petrographic features observed were cracks present in the coarse aggregate (CCA), followed by cracks in the cement paste (CCP) and debonded coarse aggregates (CAD). None of the other petrographic features noted in Table 1, could be observed using the images of the respective cores due to the lower magnification of the generated images.
Core C1 had the highest number of petrographic features with a value of 217, followed by core C6 and C12 with a value 205 and 154, respectively. The extent of damage apparent in core C1 appears to be uncharacteristic in the set of results, as the value was the highest and not comparable to Core C2 which was extracted from the same wall but adjacent to a construction joint. The result can possibility point to a variation in the composition of concrete used in the casting of that element of the railway bridge.
Concrete extracted from various locations on the Abutment 1 inner wall (cores C2, C3 and C4) exhibited a comparable number of petrographic features noted with values of 130, 157 and 113, respectively. Core C6 and C12 were expected to have a high number of petrographic features with a value of 134 and 205 respectively as the cores were extracted from Pier 1's outer walls, which are directly exposed to rain. The possible reason that the extent of damage observed in Core C12 was lower than that of core C6 may be attributed to the steel pipe gantry running alongside the South-facing side of the railway bridge, which limit's the concrete exposure to direct rain. The extent of damage of core C11 which is also located on the outer wall of Pier 1 was also noted as another uncharacteristic result as the core exhibited the lowest number of petrographic features with a value of 45 amongst all twelve cores analysed. Concrete extracted from various locations on Pier 1's inner wall (cores C7, C8, C9 and C10) had a comparable number of features noted, with values of 85, 97, 81 and 139, respectively.
The damage rating index (DRI) value for each core was then determined and plotted as seen in Fig. 4.
From Fig. 4 it can be seen that owing to the higher weighting factor of 3 assigned to cracks in the cement paste, the CCP portion is the largest aspect contributing to the DRI values of all twelve concrete cores. Debonded coarse aggregates (CAD) were also assigned a weighing factor of 3, however this feature was not as prominent in the cores examined. Despite the high number of cracks observed in the aggregates of the cores, the weighting factor is only 0.25, thus the lower contribution to the DRI value calculated.
There is agreement that higher DRI values correspond with higher damage and low DRI values reflect comparatively little damage. However the DRI value obtained in this study could not be compared to the established DRI scaled values [17] owing to the modified DRI approach adopted. The choice of conducting the visual examination using the naked eye as opposed to using a stereomicroscope (which provided a 16-times magnification), makes direct comparisons difficult as this may result in a lower number of petrographic features being observed and counted in the analysis. Nevertheless, the approach was found to be valuable in comparing the relative extent of damage observed between the twelve core samples.
The contribution of CCA to the DRI values were fairly similar across all twelve cores examined, while the contribution of CCP and CAD features varied. This variation may be linked to the locations from where the cores were extracted. Cores taken from areas with higher moisture exposure, such as the edge of Abutment 1-Inner wall, Pier 1-North-facing, and Pier 1-South-facing (C1, C6, and C12), showed more prominent CCP and CAD features, indicative of severe ASR degradation.  It is expected that the findings presented here will be compared to microscopic, mechanical and residual expansion tests results which will be performed on the extracted concrete cores. The goal is to consolidate all the results to develop a more effective management approach for concrete infrastructure subject to ASR deterioration.

Conclusion
A summary of the visual assessment of the railway bridge subject to ASR deterioration are presented and a number of trends and conclusions are generalised to draw useful information.
• A visual inspection of thirteen elements of the railway bridge showed signs of extensive map-cracking, white exudations around cracks and dampness in certain areas, indicating a medium to high possibility of deterioration due to ASR. • The concrete used in construction of the railway bridge consisted mainly of quartzite and shale coarse aggregate, both of which are known reactive aggregates in the region. • The modified DRI method used in this study has shown that it is possible to assess and compare the relative extent of damage of concrete taken from various locations of a RC structure, using local images of the cylindrical surface of the concrete core. • Concrete extracted from elements that were exposed to direct moisture exhibited higher degrees of internal damage (cracks in the aggregate, cement paste and debonded aggregates) than those from drier areas of the structure.
Acknowledgments The authors would like to acknowledge the support of this work by the National Research Fund (NRF) of South Africa, Grant no. 113843.
Funding Open access funding provided by University of the Witwatersrand.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.