Preliminary evaluation of Pier cap from an ASR affected bridge in Central Canada

,


Introduction
The selected bridge for this study, is a highway bridge, which was constructed in the early 1960's and in part traverses a waterbody.The structure includes the use of pre-stressed reinforced concrete spans.
Signs of deterioration in concrete members were observed in the Bridge with the cause of damage thought to be a combination of various damage mechanisms.Among those, the observation of cracks with a random nature following a polygonal pattern, known as "mapcracking" [1] on various elements of the Bridge attracted the attention of engineers conducting inspections as indicating the presence of alkali-silica reactions (ASR) ASR is a chemical reaction between un or poorly crystallized silica, including certain types of quartz, in the aggregate with alkali hydroxide from the cement paste.These produce a secondary product (the so-called ASR gel) which in the presence of water, can expand causing cracking in concrete [2,3].Such cracking could significantly impact the mechanical properties of effected elements where a considerable reduction of tensile (i.e., up to 65%) and stiffness loss (i.e., up to 50%) at low/moderate expansion (i.e., 0.05 to 0.12%) while compressive strength losses can be up to 35% at very high expansion (i.e., 0.30%) [4].Reactive aggregates have been found in many regions of eastern Canada [5], with ASR observed in many Quebec structures [2].Additionally, cement used in eastern Canada has historically had relatively high alkalinity [6] which makes concrete more susceptible to ASR.The presence of ASR in elements of the bridge was confirmed through visual inspections and the extraction of cores from several piers during the mid-1990's.
ASR-induced damage is frequently coupled with other damage mechanism (e.g., Freeze-thawing and/or steel corrosion) [5,7].The climate in the region includes freezing winter temperatures and warm, humid summers which make additional mechanisms likely.The use of air entrainment to mitigate freeze-thaw damage was a new practice which was still under development in the concrete industry during the period when the bridge was constructed.As a result due to the high level of damage in outdoor concrete internal expansion reactions such as ASR were frequently missed since damage was thought to be due to freeze-thaw [5].Additionally, the use of deicing salts for winter traction were introduced in the 1960's in Quebec and Ontario.Since then reinforcement corrosion due to chlorides has resulted in substantial changes in the design code in the neighbouring province of Ontario to prevent chloride ingress into concrete structures [8].Therefore, it is expected that ASR is occurring along with freeze-thaw damage and/or reinforcement corrosion in most elements.
It is important to note that the behaviour of coupled damage mechanisms is challenging to predict.Multiple mechanisms working together may significantly increase the rate of deterioration compared to a single mechanism working alone [9].As such, coupled ASR and freezethaw damage deterioration is generally linked to crack propagation from ASR and whether the cracks facilitate greater movement of liquid water which expands when frozen [9].On the other hand, in laboratory research, the coupled effects of chloride and freeze-thaw have been found to increase scaling on concrete surfaces as well as increasing the rates of crack propagation and chloride penetration [10][11][12].It is important to investigate field structures which are experiencing coupled deterioration mechanisms due to the different scale of the samples.
Accurate assessment of the current and future service life of a deteriorated structure is essential to ensure safety, schedule the replacement of structures and manage the costs associated with public infrastructure.In structures like the Bridge, visual and non-destructive techniques are the primary method to determine the structural health.This paper is aimed to use distinct visual inspection techniques and non-destructive tests (NDT's) to initiate the condition assessment of aged concrete and estimate the amount of damage on the selected Pier cap.

Scope of work
This work is part of a broader research effort to have more reliable and effective ways to assess the condition of the Bridge and similar structures through the use of distinct diagnostic and prognostic techniques.The in situ diagnostic techniques presented here, will later be compared to the various microscopic and mechanical testing which will be performed on the extracted cores.Thus, such results could be used to develop a more effective management protocol for deteriorated concrete infrastructures.

Methodology
The Pier cap discussed in this paper, (Fig. 1 and Fig. 2) was part of an over-land approach portion of the Bridge near a roadway on the western side and a large waterbody on its eastern side.Due to the proximity of the roadway, the Pier cap may have experienced increased exposure to de-icing salts and vehicle exhaust.
The preliminary evaluations results presented in this work have been gathered from the five Pier cap segments (each with dimensions 1 m x 2 m x ~2.0 to 3.3 m) removed from the Pier cap, as shown in Fig. 1, and made available to the University of Ottawa team at an outdoor storage site.A number of visual inspection techniques (conventional, semi-quantitative and quantitative using the cracking index) as well as non-destructive techniques (Schmidt hammer, ultrasonic pulse velocity and surface resistivity) were performed a preliminary assessment of the condition of the above-mentioned element.
The results of the preliminary assessment will be used to decide the appropriate locations (i.e., demonstrating the lowest/ highest deterioration) to retrieve a number of cores for further evaluation.

Visual Inspections
Detailed visual inspections were performed on all Pier cap segments.This included qualitative, semi-qualitative (i.e., measuring crack widths) and quantitative (i.e., conducting the cracking index) assessments of observable damage.
An alternative to traditional qualitative visual inspection, the Cracking Index (CI), a crack mapping process that averages crack widths and frequencies along the vertical and horizontal directions displayed by a grid drawn on the surface of the evaluated concrete member having minimum dimensions of 0.5 x 0.5 m [13] can be used.The cracking index is determined by finding the summation of crack widths and dividing it by the base length (usually 0.5 m) to find a final value in mm/m (Eqn.Therefore, for the sake of this work, the CI was performed in the most cracked area of each of the west, east and bottom surface of each segment since these would have been accessible while the element was in service.To control for inconsistencies between practitioners all cracking indexes were performed by the same person using a crack comparator card.

Non-destructive Testing
Several non-destructive techniques were chosen to investigate concrete quality (i.e., Schmidt hammer, ultrasonic pulse velocity and surface electrical resistivity).The selected tests require a smooth surface so the top surfaces (the load bearing surface in contact with the girders), which had a rough finish, were omitted.The west and east facing sides of Pier cap segments were prioritized for testing since they would be most easily accessed.
Schmidt hammer is a portable rebound testing apparatus which records the energy absorbed by the concrete on impact [14].The value produced in this test represents the elastic properties the concrete and can be correlated to the compressive strength of concrete using equations [15].This test was performed on 10 points grouped in a single location on each of the west and east facing along with the bottom (diagonal) surfaces (as per Fig. 1) following the procedure in ASTM C 805 [16].A Proceq Schmidt rebound hammer device was used with tests performed with the device held perpendicular to the test surface.
Ultrasonic Pulse Velocity (UPV) is used to assess concrete quality and detect defects in concrete such as voids and cracks [17].This is not an exhaustive method for cataloguing defects but can give a general impression of the material's condition and whether further tests are required [17,18].The technique is described in ASTM C 597-16 [17].Tests were performed using a Screening Eagle Pundit UPV device on a 0.6 m strip along the top of the west and east faces with spacing of 0.2 m between the transmitter and receiver; contact to the surface was conducted using petroleum jelly.The indirect method was used due to the large size of the segments and the presence of reinforcing bars which can impact the results.In 1992 the Indian Standards association [23] proposed a concrete quality grading system in which pulse speeds below 3.0 Km/s were described as indicating 'doubtful' quality requiring further investigation.
Schmidt hammer and UPV results were used to estimate concrete strength using Surface resistivity measurements indicate the ease of ion movement through concrete with a low resistivity value (in KΩꞏcm) indicating easier movement of ions in concrete pore solution [19].Measurements were performed on the west, east and bottom surfaces as well as the cut surface (i.e.perpendicular to the white dashed lines in Fig. 1) using a four-pronged Resipod concrete electrical resistivity device.Surface resistivity is strongly influenced by concrete moisture content which is difficult to control when testing outdoors.The AASHTO-T-358 standard [20] outlines a procedure to achieve SSD conditions in laboratory cylinders while RILEM TC 154-EMC [23] suggests the use of wetted sponges for measuring surface resistivity in the field emphasizing that a water film on the surface should be avoided.Readings were performed on all elements on the same day following a week of periodic rain with wetting of surfaces occurring 10-20 minutes before testing to approximate SSD condition.Following AASHTO-T-358 [20] and ACI 228.2R-13 [14], surface resistivity values <12 kΩꞏcm would indicate high probability of chloride ion penetration (low durability concrete) while surface resistivity values >254 kΩꞏcm indicate low probability of chloride ion penetration (high durability concrete).

Visual Inspection
The conducted conventional visual inspection of the cut surfaces of each segment revealed that all sections had an exterior layer of concrete containing a different concrete mix design which is thought to be a repair concrete.The repair concrete depth varied between 130 -230 mm.
Several layers of rebar, arranged in the longitudinal directions, were used; rebar diameters observed ranged between 60 to 15 mm ɸ bars.Larger diameter rebar was used longitudinally along the top (closest to the concrete girders of the Bridge) while medium diameter rebar is used in other areas of the Pier cap.The smallest diameter rebar (15 mm ɸ) was used in the repair concrete which also contains smaller aggregate.These can be easily observed on the cut surfaces of the elements (Fig. 3) but rebar condition would be challenging to assess through non-destructive testing when the element was in service.Stirrups were also observed to enclose the longitudinal rebars.

Fig. 3 View of rebar arrangement inside Pier cap section
The interior concrete had visible signs of ASR reactions including reaction product within coarse aggregates (Fig. 4) while some cracking appeared to originate from corroded reinforcing bars especially along the interface between internal and repair concrete.The qualitative visual inspection has revealed that the average aggregate size of the repair concrete is around 8-12 mm and between 35-55 mm in the main concrete.The top, bottom, west and east facing surfaces of each segment had distinct appearances.The surfaces which were facing west and east had a smooth finish with patterns of fine map-cracking as shown in Fig. 5, which could be the sign of ASR in the Pier cap.The surfaces corresponding to the top, where the Pier caps were connected to the girders, had a rough finish as shown in Fig. 6.The rough finish on these surfaces limited the types of non-destructive tests that could be performed.
Fig. 6 The top surface of section 46.3.
The diagonal portions of the Pier cap (the bottom surfaces) had a distinct appearance and were darker in colour and were finished with a corduroy type ridge texture (Fig. 7).White efflorescence, which could be dried salt or reaction product, was presented on various bottom surfaces of the Pier cap.The discolouration may be due to the different finishing method used on this surface and the geometry of the Pier cap.

Cracking Index (CI)
Cracking index measurements are used to quantify the damage on different areas of an element or structure or to track cracking at the same location at different times.More detailed investigations are recommended if CI is greater than 0.5 mm/m and/or cracks have widths greater than 0.15mm [21].
In the segments investigated, crack widths as large as 1 mm were observed during the cracking index measurements but the majority of crack opening were between 0.10 to 0.35 mm.The results of CI measurements on different surfaces (Table 1) indicate that more detailed investigation is required for this structure.UPV measurements, summarized in Table 3, indicate poor concrete quality warranting further testing [18].During measurements it was sometimes necessary to move the transmitter and receiver to get measurements over a 0.2 m segment such that it did not travel over a substantial crack because these could sufficiently block signal transmissions.During testing, weather at the outdoor site included frequent showers so concrete is thought to have been at approximately SSD condition.
UPV measurements had a large variability which was to be expected given the deterioration of the concrete.UPV readings varied between 4.89 km/s and 1.43 km/s with an average reading of 2.74 km/s.

Surface Resistivity
While water saturation can cause large variability in results, surface resistivity data indicated a clear difference between the repair concretes on the exterior (average readings around 200 KΩꞏcm) and the internal concrete (average readings around 9 KΩꞏcm).This indicates that the repair concrete has a better quality than the main concrete.Surface electrical resistivity readings on the repair concrete were found to have high variance depending on the location on the concrete surface where measurements were taken.Due to the size of the element variability between concrete quality and the on-going deterioration mechanisms may impact these readings.To show this the results of 24 resistivity measurements from the west face of segment 46.2 (Fig. 8) are shown in the contour plot illustrated in Fig. 9. Resistivity values were higher in the region in vicinity of load bearing surface (i.e., left side of the segment) while it has the lowest value in the center of the segment possibly indicating small differences in concrete permeability in the structure.

External Concrete Quality
The Pier cap segments show signs of cracking and discolouration, but concrete has not experienced substantial spalling or pop-out of aggregates.Due to the location of the Pier cap, below and adjacent to roadways which had heavy winter salt usage, evidence of salt on concrete surfaces was expected.A small amount of white efflorescence was observed on the diagonal (bottom) surface of segments (e.g., Fig. 7) but these were not observed on the west or east facing surfaces.Although it is worth noting that the elements were not studied in situ, chloride exposure may have been primarily due to runoff from the bridge surface rather than spray from the roadway adjacent to the Pier cap.

Internal Concrete Quality
As per Fournier et al. [21], concrete elements with CI > 0.5 mm/m (and/or cracks of width > 0.15 mm) require more detailed investigations to identify the cause and extent of damage.Analyzing the visual inspection and non-destructive tests results, it was clear that cores should be taken to further investigate concrete quality.As an example, the results gathered on UPV, attest the 'poor quality' of various segments of the Pier cap.Analysis of this concrete element is challenging due to the thick layer of repair concrete which significantly restricts access to the interior concrete.The layer of external concrete also makes identification of the source of damage more challenging since cracking in the interior concrete does not necessarily match the cracking observable on the exterior.Concrete quality is seen to differ depending on the location in the structure.East facing surfaces on segments 46.2 and 46.5 have lower quality confirming by higher CI value.This could be due to increased moisture due to its proximity to the river which can accelerate both the transport of chlorides to initiate corrosion in reinforcement as well as contributing to ASR.Moisture transport may also be facilitated by predominant wind directions.Otherwise, comparing the results of various NDTs, one sees that due to the more localized nature of Schmidt hammer and UPV tests, data trends were less clear by location.For these tests Schmidt hammer is lower on the West and East faces compared to the diagonal (bottom) surfaces indicating that the diagonal surface is less damaged.It is important to note that the different surface finishes may impact Schmidt hammer results which may be responsible for the difference between horizonal surfaces and the diagonal surface.For UPV (Table 3) the results were consistently poor quality indicating further investigation should occur for the West facing surfaces.Other surfaces were also of poor quality except one surface on the bottom surface of 46.1 which was of medium quality (value of 3.32).
Surface resistivity on the exterior and interior concretes, possible along the cut surfaces of the Pier cap sections, indicate substantial differences in the potential for the ingress of harmful ions like chlorides.Investigation of the surface in situ would indicate a low probability for the penetration of chloride ions (high quality concrete) which differs substantially from the poorer quality of interior concrete.A higher permeability concrete allows more variable moisture at a larger depth from the surface which could increase the development of AAR gel.However, larger pores which are typically present in higher permeability concrete may reduce damage to the concrete matrix due to expansive gel produced in ASR since gel can fill pores without creating further cracking.The use of a low permeability outer concrete layer would provide increased protection and increase durability of the element but may not halt deterioration which has already started.As such, further information is needed on repairs and initial concrete design of the structure.
In order to analyze the results obtained from the Schmidt hammer, the combined Sonic Rebound (SonReb) method was used to further evaluate the internal concrete quality.This method calculates an approximate compressive strength using UPV and Schmidt hammer results.According to information provided to the authors, the Pier caps used in the bridge had an approximate 28 day strength of 24 MPa and therefore, repair concrete should be within a similar range but is expected to be somewhat higher due to the lower permeability of the external concrete which suggests a low w/c ratio.Many authors have developed equations to utilize the SonReb method which are discussed in a review paper by Cristofaro et al. [22] who looked at 17 different published equations.Most authors utilized cubes or cylinders made in a laboratory environment with samples utilized differing based on the specific goal of that research.For example, differences in the investigated compressive strengths, aggregate sizes and types etc resulted in different equations being developed to fit the data.The majority of equations either strongly underpredicted or overpredicted strength based on inputs obtained from the Pier cap.The following Four equations were found to result in strength predictions in the expected range and were chosen to evaluate the data: Meynink, Samarin [15]; Ramyar and Kol [23]; Kheder [24] and Menditto et al. [25].The equations (Table 4) uses the variables 'SH' for the Schmidt hammer compressive strength and 'v' for the speed in obtained from UPV [Km/s].  = 0.0158ꞏ 1.1171 ꞏ 0.4254 cubes Menditto et al. [25]   = 0.00004ꞏ These techniques give an estimate of concrete strength which will be compared to compressive tests at a later date.The values found indicate lower quality of the concrete elements in the West and East facing surfaces (Table 6) compared to the bottom surfaces (Table 5).Ramyar and Kol routinely estimate higher values which is reasonable given that their equation is based on cores extracted from larger elements which likely experienced higher curing temperatures.

Selection of Coring locations
The CI and Schmidt hammer were decided to be the best parameters for selecting coring locations due to the clear trends exhibited.There are no limitations on the number of cores that can be extracted, however it was desired to replicate core extraction locations which might be used insitu so cores were extracted primarily from the West and East facing surfaces.The equivalent of a minimum of seven cores of 100 ɸ mm x 200 mm were extracted per segment for further analysis.
For further assessment, it is important to recognize which concrete segments experienced the lowest amount of damage.Thus, based on the cracking index results, the West facing surfaces of segments 46.2 and 46.3 were selected as the best candidates.The higher Schmidt hammer values for 46.3 resulted in choosing this segment as the lowest damage section of the Pier cap (Table 5.3the numbers were highlighted in blue).Conversely, a highest damage segment was also identified as being segment 46.4 East having the highest CI and lowest Schmidt Hammer value (The numbers highlighted in red in Table 7).

Preliminary findings from the extracted cores
Conventional visual inspection has been performed on the extracted cores which confirmed the presence of ASRreaction product throughout the coarse aggregate particles of the interior concrete of Pier cap (Fig. 10).Coring was performed using extenders with the coring depths as deep as 1.5 m.Additionally, the depth of repair concrete (Fig. 11) in the initial portion of cores varied between 130 mm and 230 mm.

Conclusions
Preliminary investigation of the five segments of the Pier cap clearly identified the need for further investigation of this concrete element.
For visual assessment: the Cracking Index (CI) was successful in comparing the amount of cracking at different locations in the structure.Rankings were reasonable considering the exposure conditions of each area.
For Non-destructive testing (NDT): The NDT techniques used, successfully identified broad trends within the collected data.
• The Schmidt hammer test identified a clear difference between the horizontal (West and East facing) surfaces and the diagonal bottom surface; the compressive strength of the latter was found to be approximately 10 MPa higher than the former.• Ultrasonic Pulse Velocity (UPV) identified concrete as being mainly of 'poor' quality which agrees with the extent of cracking observed in the visual inspection.• Electrical surface resistivity measurements clearly distinguished the inner quality of the internal and repair concrete, however there was high variation when used solely on the repair concrete.
In conclusion, the use of visual and NDT techniques was successful in identifying the damage experienced by the studied Pier cap.Without destructive testing such as microscopic and mechanical procedures, identifying the precise cause and extent of the deterioration might not be possible.Visual inspection of the cut surfaces of each segment were beneficial in providing more insight on the state of the damage on each segment where a number of cracks could be observed both originating from the aggregate particles and reinforcing bars.

Future work
Testing of extracted cores will take place to better correlate visual and non-destructive test methods to concrete condition.A multi-level assessment approach including the use of mechanical (compressive testing and Stiffness Damage Test-SDT) and microscopic procedures (the Damage Rating Index-DRI and SEM) will be used.

Fig. 2
Fig. 2 East facing side of the Pier cap (Pier cap photo from 2021).

Fig. 1
Fig. 1 Original locations of Pier cap segments shown on the West facing side of the Pier cap (photo from 2016).

Fig. 4
Fig.4 Presence of ASR reaction product within the coarse aggregate particle on the cut surface of a Pier cap segment.

Fig. 5
Fig. 5 East facing surface of section 46.3

Fig. 9
Fig. 9 Plot of surface resistivity values for the west facing side of segment 46.2.

Table 1
CI results for Pier cap segments (mm/m).
the material absorbs more energy.Analyzing the results gathered from the Schmidt hammer (Table2), one sees that they can be correlated to results from the Cracking Index (Table1) with higher Schmidt hammer results corresponding to lower Cracking Index values which indicates better quality concrete.Correlating the results from these two methods, one notices that the west facing side of section 46.3 demonstrated the lowest damage segments among the other.

Table 2 average
Schmidt hammer rebound numbers for Pier cap sections

Table 4 :
[22]tions used to calculate concrete strength using the SonReb method as presented in[22]

Table 5 :
1.88148 ꞏ 0.80840 cubes Calculated compressive strength (MPa) of bottom face of Cap Beam segments using SonReb method.

Table 6 :
Calculated compressive strength (MPa) of based on tests on West (W) and East (E) facing Cap Beam segments

Table 7 :
Correlation of visual inspection and NDT results; the highlighted numbers in red and blue are representative of the segments direction experiencing the most and least damage.