Cement & Concrete Composites 47 (2014) 64–68

Contents lists available at SciVerse ScienceDirect

Cement & Concrete Composites

journal homepage: www.elsevier .com/locate /cemconcomp

Service life prediction of RC structures based on correlation betweenelectrochemical and gravimetric reinforcement corrosion rates

0958-9465/$ - see front matter � 2014 Published by Elsevier Ltd.http://dx.doi.org/10.1016/j.cemconcomp.2013.06.003

⇑ Corresponding author. Address: P.O. Box 1896, King Fahd University ofPetroleum & Minerals, Dhahran 31261, Saudi Arabia. Tel.: +966 3 860 2570; fax:+966 3 860 2879.

E-mail address: saghamdi@kfupm.edu.sa (S.A. Alghamdi).

Saeid A. Alghamdi ⇑, Shamsad AhmadDepartment of Civil and Environmental Engineering, King Fahd University of Petroleum & Minerals, Dhahran 31261, Saudi Arabia

a r t i c l e i n f o a b s t r a c t

Article history:Received 3 September 2012Received in revised form 30 May 2013Accepted 8 June 2013Available online 22 June 2013

Keywords:Reinforced concreteCorrosion current densityGravimetric weight loss methodLinear polarization resistance methodService life prediction

This paper outlines a comprehensive experimental investigation to establish relationships between elec-trochemically and gravimetrically measured reinforcement corrosion rates generated through a compre-hensive experimental program. The investigation is based on testing a total number of 486 reinforcedconcrete (RC) specimens prepared using coarse aggregates obtained from two distinctly different sourcesof aggregate and considering three levels of five design variables (namely: cementitious material contents,water to cementitious materials ratios, fine to total aggregate ratios, concrete-cover thicknesses, andexposure to chloride-solutions). The correlation between corrosion rates measured using electrochemicalmethod to that using gravimetric method is established and used to convert electrochemically measuredcorrosion rate into equivalent but more accurate gravimetric corrosion rate that would be valuable forservice life prediction of RC members RC structures in corrosive environments. The experimental pro-gram and sample numerical results obtained are outlined and summarized. Then a methodology for pre-dicting remaining service life of corroding RC members is illustrated through a numerical example.

� 2014 Published by Elsevier Ltd.

1. Introduction

The presence of chloride ions in a concrete mixture, contributedby its constituents, or due to natural transport processes has beenobserved to play a critical role in initiating and maintaining theprogress of reinforcement corrosion in RC structures. Corrosion ofreinforcing steel bars has been known worldwide to be the maincause of accelerated deterioration of many RC structures and somepremature structural failures even well-before their designed ser-vice life [1]. Numerous research reports and field studies haveidentified chloride products and chloride ions in particular as beingresponsible for the corrosion of steel bars in concrete [2]. Theresulting corrosion products have larger volume and induce strainsand stresses, which can cause cracking and spalling of the concretecover along with loss of bond between the steel and concrete. Areduction in the rebar diameter and loss of the bond due to rein-forcement corrosion can cause a significant loss of load-bearingcapacity of the corroded RC members [3] and may thereforeshorten the service life of the structure.

After corrosion initiates, corrosion rate is a key predictiveparameter to determine the remaining service life of corroded RCstructures exposed to a corrosive environment [4] and the method

of electrochemical linear polarization resistance (LPR) is often usedas a non-destructive method to monitor quantitatively generalcorrosion and galvanic corrosion. It is sometimes used to qualita-tively monitor localized pitting corrosion. The main advantagesof electrochemical techniques include sensitivity to low corrosionrates, short experimental duration, and well-established theoreti-cal basis. However, the electrochemical LPR technique is found tohave some experimental errors such as error due to Ohmic dropduring polarization. On the other hand, the method of gravimetricweight loss (GWL) measurement is an alternative destructive tech-nique to determine corrosion rate with more accuracy [5]. But thedestructive nature of the gravimetric weight loss (GWL) methodmakes the non-destructive LPR method, despite of its intricatepractical limitations, more commonly utilized for assessing therate of reinforcement corrosion.

In LPR method, the steel bar is polarized by the application of asmall perturbation to the equilibrium potential through a counterelectrode. The polarized surface area of the steel bar is assumed tobe that area which lies directly below the counter electrode, butthere is considerable evidence that current flowing from the coun-ter electrode is not confined to the polarized area and may spreadlaterally over an unknown large area of the reinforcing steel, withthe natural consequence of inaccurate estimation of corrosion rate[4]. On the other hand, gravimetric (weight loss) measurement as adestructive test, when conducted in controlled laboratory condi-tions serves as the most reliable reference method. It is simple,but is also a time-consuming technique for the determination of

http://crossmark.crossref.org/dialog/?doi=10.1016/j.cemconcomp.2013.06.003&domain=pdfhttp://dx.doi.org/10.1016/j.cemconcomp.2013.06.003mailto:saghamdi@kfupm.edu.sahttp://dx.doi.org/10.1016/j.cemconcomp.2013.06.003http://www.sciencedirect.com/science/journal/09589465http://www.elsevier.com/locate/cemconcomp

Nomenclature

A exposed surface area of rebar (cm2)B stern–Geary constant (mV)ba anodic Tafel constant (mV)bc cathodic Tafel constant (mV)CL chloride concentration (%)Cs surface chloride concentration (%)Cth threshold chloride concentration (%)Cv concrete cover (mm)D diameter of rebars (mm)Dapp chloride diffusion coefficient (m2/s)F Faraday’s constant (96487 Coulombs)Icorr corrosion current density (lA/cm2)Icorr,e electrochemically measured corrosion current density

(lA/cm2)Icorr,g gravimetrically measured corrosion current density

(lA/cm2)

Jr corrosion rate (g/cm2/yr)Pr corrosion penetration rate (lm/yr)Qcr amount of corrosion products needed for causing

cracking (g/cm2)Rp linear polarization resistance (kX cm2)t age of structure (yr)tcr time to corrosion cracking (yr)tp corrosion initiation time (yr)tRL residual life (yr)T exposure time (h)W equivalent weight of steel (27.925 g)WL weight loss (g)Wi initial weight of the bars before corrosion (g)Wf weight of the bars after cleaning all rust products (g)

Fig. 1. Details of a typical test specimen and test specimens exposed to chloridesolution.

S.A. Alghamdi, S. Ahmad / Cement & Concrete Composites 47 (2014) 64–68 65

corrosion rate. The weight loss measured is converted to a uniformcorrosion rate over the exposure period. It has been proposed thatthe combination of the weight loss method and the polarizationresistance method offers means of quantitative corrosion analysis.However, studies on the relationship between the two methods arelimited and most studies were conducted on different sets of spec-imens [6].

Accurate value of in situ measured corrosion rate is needed to carryout the service life prediction of RC structures exposed to corrosiveconditions. As mentioned earlier, the accuracy of reinforcementcorrosion rate measured electrochemically is invariably unreliablebecause of the difficulties and errors frequently involved in experi-mental measurements. Therefore, it is highly desirable to correlatethe electrochemically measured corrosion rate with gravimetricallymeasured corrosion rate so that the electrochemically measured val-ues of corrosion rate can be converted into the more reliable equiva-lent gravimetric corrosion rate to be used for service life prediction.This paper presents: (i) an overview of the experimental program thatwas used to compare corrosion rate values using LPR and GWLmethods; (ii) statistical analysis and initial results on the correlationbetween electrochemically and gravimetrically measured reinforce-ment corrosion rates obtained using compiled experimental data;and (iii) a computational approach for service life prediction of RCstructures in a specified corrosive environments.

2. Experimental program

2.1. Materials

ASTM C 150 Type I Normal Portland cement was used with 8%replacement by silica fume for all the mixtures. Coarse aggregatesfrom distinctly different sources located in the eastern region andin the western region of the Kingdom (namely: Abu Hadriyah quar-ries and Taif quarries) were used [7]. The specific gravity and waterabsorption, needed for mix design, were determined using ASTMC128 [8] and abrasion test results of coarse aggregates were ob-tained in accordance with ASTM C131 [9]. Dune sand was usedas fine aggregate.

2.2. Test specimens and concrete mix design

Steel bars of diameter 16 mm were placed centrally in each oneof the 486 cylindrical concrete test specimens [10] of height150 mm and diameters 66 mm, 91 mm and 116 mm, with threedifferent cover thicknesses 25 mm, 37.5 mm and 50 mm. The testspecimens were prepared to evaluate corrosion rate and epoxy

coating was applied to the steel bar at the bottom and at the inter-face between the concrete and air as shown in Fig. 1a.

The absolute volume method was used for the concrete mixdesign and the quantity of each constituent was calculated onthe basis of weight. All the concrete specimens were prepared withcementitious materials (92% Normal Portland cement and 8% silica

Fig. 2. Samples of corroded steel bars before cleaning.

66 S.A. Alghamdi, S. Ahmad / Cement & Concrete Composites 47 (2014) 64–68

fume). The groups of water–cementitious ratio, cementitiousmaterial content and fine to total aggregate ratio used to preparethe specimens were respectively (0.4, 0.45 and 0.5), (350, 375and 400 kg/m3), and (0.35, 0.4 and 0.45). Three sodium chloridesolutions with concentrations of 3%, 7% and 12% were used to sim-ulate corrosive conditions. A homogenous concrete was obtainedwith all constituents mixed together with the addition of potablewater and with a superplasticizer mixed uniformly with the con-stituents to enhance workability, and to achieve uniform consis-tency and cohesiveness without segregation of concrete mixturea revolving drum-type mixer was used for one minute at the endof each mix design.

Test specimens were de-molded after 24 h of casting and testspecimens were then cured for a period of 28 days in water tanksunder laboratory conditions, and were subsequently partlysubmerged in chloride solutions to allow corrosion to take place.Samples of test specimens exposed to a chloride solution areshown in Fig. 1b.

2.3. Experimental techniques

2.3.1. Electrochemical linear polarization resistance (LPR) methodThe LPR method was utilized to determine the corrosion current

density (Icorr) utilizing PARSTAT 2273 potentiostat equipment [7].As a method of evaluating the instantaneous corrosion rate of rein-forcing steel in concrete, the method is used to measure corrosioncurrent density using a three-electrode system that includes: (i)reference electrode; (ii) counter electrode (steel plate) was con-nected to the respective terminals of the potentiostat; and (iii)steel reinforcement in the concrete specimen (often known asthe working electrode) was polarized to ±20 mV from its equilib-rium potential at a perturbation scan rate of 0.166 mV per second.

After a suitable initial delay, typically 60 s, the steel was polar-ized. The product of surface area of rebar under polarization andthe slope of the applied potential versus measured current plotwas taken as the linear polarization resistance Rp (kX cm2)[5,11]. Corrosion current density was calculated using the follow-ing relationship:

Icorr ¼BRp

ð1Þ

in which: Icorr = corrosion current density (lA/cm2), and

B ¼ babcðba þ bcÞð2Þ

expressed is expressed in terms of: anodic Tafel constant ba andcathodic Tafel constant bc that are generally obtained from the Tafelplot [12], or in the absence of the plot, a value of B equal to 52 mVfor steel in passive state and a value equal to 26 mV for steel inactive state can be used. The value of B used in this test was 26 mV.

2.3.2. Gravimetric weight loss methodUpon completion of four rounds of corrosion rate measurement

using the LPR method, each one of the 486 concrete test specimenswas broken in preparation for the determination of corrosion rateby gravimetric weight loss method. Preparation, cleaning and esti-mation of the weight loss were done according to the ASTM G1-03standard practice [13]. The cleaning solution used was 1000 ml ofhydrochloric acid with 20 g of antimony trioxide and 50 g of stan-nous chloride. The weight loss WL was calculated as:

WL ¼Wi �Wf ð3Þ

where Wi = initial weight of the bars before corrosion (g), and Wf =weight of the bars after cleaning all rust products (g).

Then the corrosion rate (lm/yr) was determined using the fol-lowing equation:

Pr ¼1:116� 107 �WL

A� T ð4Þ

where WL = weight loss (g); A = exposed surface area of rebar (cm2);and T = exposure time (h).

The corrosion penetration rate Pr was then converted to equiv-alent corrosion current density Icorr (expressed in lA/cm2) usingthe following formula [7,11]:

Icorr ¼Pr

11:7ð5Þ

Samples of the corroded steel bars extracted from test speci-mens before and after cleaning are shown in Figs. 2 and 3.

3. Results and discussion

Values of reinforcement corrosion rates determined electro-chemically and gravimetrically (designated herein, respectively,as Icorr,e and Icorr,g) were utilized to obtain correlation between Icorr,eand Icorr,g. In order to see the effect of mix variables, cover thick-ness, and different concentrations of chloride on the correlationbetween Icorr,e and Icorr,g, the Icorr,e versus Icorr,g data were groupedand plotted separately. However, it was found that there is nosignificant effect of mix variables, cover thickness and chlorideconcentration on the correlation between Icorr,e and Icorr,g. Hence,an overall plot of Icorr,e and Icorr,g values obtained for all the 486specimens was obtained as shown in Fig. 4. It is observed fromthe Fig. 4 that a good correlation exists between the Icorr,e and Icorr,g,as evident from a higher value of R2 (0.92). From the correlationequation shown in Fig. 4, the value of gravimetric corrosion currentdensity can be taken, on an average, as 86% of the value of electro-chemical corrosion current density. The correlation between Icorr,eand Icorr,g may be used to convert Icorr,e into more accurate Icorr,girrespective of cover thickness, chloride concentration, andconcrete quality.

4. Service life prediction for RC structures

The total service life is assumed to be the sum of the time to cor-rosion initiation and the time to corrosion-cracking. It is assumedthat corrosion is initiated when chloride content at the level of rein-forcing steel reaches the corrosion threshold. The estimated time ofcorrosion initiation was determined using Fick’s second law of dif-fusion, in which chloride ingress into concrete was assumed to bemainly through diffusion [3,4,10–12]. The estimated time to corro-sion cracking was determined using the empirical model developedby Morinaga [14], utilizing the equation correlating Icorr,e and Icorr,gdeveloped in this study. Experimental data representing an existingstructure and recommended prediction models can be used to

Fig. 4. Correlation between Icorr,g (gravimetric) and Icorr,e (LPR).

Fig. 3. Samples of corroded steel bars after cleaning.

S.A. Alghamdi, S. Ahmad / Cement & Concrete Composites 47 (2014) 64–68 67

perform the predictive calculations. The following methodology isutilized to estimate the service life for a RC structure:

(i) Determine the diameter of the bar (D), concrete cover (Cv),corrosion current density (Icorr,e), chloride concentration(CL), threshold chloride concentration (Cth), surface chlorideconcentration (Cs), chloride diffusion coefficient (Dapp) andthe age of the structure (t).

(ii) Convert the corrosion current density, Icorr,e measured usingelectrochemical technique to equivalent actual corrosioncurrent density, Icorr,g using the relationship shown in Eq. (6)

Table 1Sample input and output values of a spreadsheet program for automated service life pred

Input specified values

Measured parameters

Diameter of tension rebars, D (mm)Concrete cover, Cv (mm)Corrosion current density measured using electrochemical method, Icorr,e (lA/cm2)

Output values

Corrosion current density, Icorr,g (lA/cm2)Corrosion initiation time, tp (year)Corrosion rate, Jr (g/cm2/yr)Amount of corrosion products, Qcr (g/cm2)Time to corrosion cracking, tcr (yr)Residual life, tRL (yr)

iction of R

16500.

Icorr;g ¼ 0:86 Icorr;e ð6Þ

(iii) Determine the time to corrosion initiation, tp (using Eq. (7))based on the values of Cv, Cth, Cs and Dapp.0 1

tp ¼1

12DappCv

1� CthCs� �0:5B@ CA

2

ð7Þ

(iv) Determine the time to corrosion-based cracking, tcr using thevalues of Cv, D and Icorr,g in terms of Qcr and Jr where� �0:85

C

5

Q cr ¼ 0:602 1þ2CvD

D ð8Þ

� �

Jr ¼WF

Icorr;g ð9Þ

Then the time to corrosion cracking tcr is

tcr ¼Q crJr

ð10Þ

where Qcr is the critical mass of corrosion products (10�4 g/

cm2); Cv is the thickness of concrete cover (mm); D is thediameter of steel bar (mm); while W = 27.925 g; andF = 96487 Coulombs (Faraday’s constant).

(v) Determine the residual (expected remaining) service life tRLof the RC structure, by subtracting the given age of the struc-ture t from the total service life using Eq. (13).

tRL ¼ tp þ tcr � t ð11Þ

This procedure was implemented in a spreadsheet program de-signed for automated computation of service life evaluation of a RCmember in a specified corrosive environment. Sample values oftypical input and output results for prediction of service life forRC members are presented in Table 1.

5. Conclusions

From this study, following conclusions may be drawn:

� A linear correlation exists between values of reinforcementcorrosion rates measured electrochemically using LPRmethod and gravimetrically using weight loss method, irre-spective cover thickness, chloride concentration, and concretequality.� The correlation developed in this study can be utilized in con-

verting the electrochemically measured value of reinforcementcorrosion rate to the more accurate equivalent gravimetric cor-rosion rate to be used for remaining service-life prediction.

structure.

Other variables

Chloride concentration CL (%) 3Threshold chloride concentration Cth (%) 0.1Surface chloride concentration Cs (%) 0.3Chloride diffusion coefficient Dapp (m2/s) 10�12

Age of structure, t (year) 10

0.4336.98

0.00390.00521.32

28.30

68 S.A. Alghamdi, S. Ahmad / Cement & Concrete Composites 47 (2014) 64–68

� A methodology for estimating the remaining service-life of a RCstructure exposed to corrosive environments is proposed and itsutilization is illustrated through a typical example using aspreadsheet program.

Acknowledgments

This research work was conducted with financial funding fromKing Abdulaziz City for Science and Technology (Riyadh) underKACST Project AT-23-21. The researchers would also like toacknowledge with appreciation the logistical support and utiliza-tion of research facilities at King Fahd University of Petroleumand Minerals (KFUPM), Dhahran, Kingdom of Saudi Arabia.

References

[1] Bentur A, Diamond S, Berke NS. Steel corrosion in concrete: fundamentals andcivil engineering practice. London (UK): E&FN Spon; 1997. 201.

[2] ACI 222R-01. Protection of metals in concrete against corrosion. AmericanConcrete Institute; 2001: p. 2–4.

[3] Page CL. Corrosion of reinforcement in concrete construction. Cambridge: TheRoyal Society of Chemistry; 1996. p. 55.

[4] Andrade C, Martinez I. Calibration by gravimetric losses of electrochemical ratemeasurement using modulated confinement of the current. Mater StructNovember 2005;38:833–41.

[5] Sathiyanarayanan S, Natarajan P, Saravanan K, Srinivasan S, Venkatachari G.Corrosion monitoring of steel in concrete by galvano-static pulse technique. JCem Concr Compos 2006;28:630–7.

[6] Ganesan K, Rajagopal K, Thangavel K. Evaluation of bagasse ash as corrosion-resisting admixture for carbon steel in concrete. J Anti-corros Method Mater2007;54:230–6.

[7] Alghamdi SA, Ahmad S, Yusuf MO. Analysis of durability-based designperformance of coarse aggregates used for RC structures in corrosiveenvironments. Proceedings, the fourth int. conf. on structural engineeringmechanics and computations, University of Cape Town, Cape Town, South,Africa. vol. 6–8; September 2010.

[8] ASTM C128. Standard test method for density, relative density (specificgravity), and absorption of fine aggregate. West Conshohocken (Pa): ASTM;2007.

[9] ASTM C131. Standard test method for resistance to degradation of small-sizecoarse aggregate by abrasion and impact in the los angeles machine. WestConshohocken (Pa): ASTM; 2006.

[10] Alghamdi SA, Ahmad S. Multi-criteria optimal design methodology for durablerc members in corrosive environments – an experimental investigation.Proceedings, international conference on durability of concrete structures(ICDCS 2008). China: Hangzhou; November 2008. 26–27.

[11] Alghamdi SA, Ahmad S, Multi-criterion optimal designs of R/C beams andcolumns: experimental and analytical studies. KACST Research Project AT-23-21, Final Report. September 15; 2010.

[12] PowerCORR User’s Manual, Corrosion Measurement Software. PrincetonApplied Research. USA; 2001.

[13] ASTM G 1-03. Standard practice for preparing, cleaning, and evaluatingcorrosion test Specimens. West Conshohocken (PA); 2003.

[14] Morinaga, S. Prediction of service lives of reinforced concrete buildings basedon the corrosion rate of reinforcing steel. Proceedings of building materialsand components. Brighton (UK); 7–9 November 1990; p. 5–16.

http://refhub.elsevier.com/S0958-9465(13)00104-2/h0005http://refhub.elsevier.com/S0958-9465(13)00104-2/h0005http://refhub.elsevier.com/S0958-9465(13)00104-2/h0010http://refhub.elsevier.com/S0958-9465(13)00104-2/h0010http://refhub.elsevier.com/S0958-9465(13)00104-2/h0015http://refhub.elsevier.com/S0958-9465(13)00104-2/h0015http://refhub.elsevier.com/S0958-9465(13)00104-2/h0015http://refhub.elsevier.com/S0958-9465(13)00104-2/h0020http://refhub.elsevier.com/S0958-9465(13)00104-2/h0020http://refhub.elsevier.com/S0958-9465(13)00104-2/h0020http://refhub.elsevier.com/S0958-9465(13)00104-2/h0025http://refhub.elsevier.com/S0958-9465(13)00104-2/h0025http://refhub.elsevier.com/S0958-9465(13)00104-2/h0025http://refhub.elsevier.com/S0958-9465(13)00104-2/h0030http://refhub.elsevier.com/S0958-9465(13)00104-2/h0030http://refhub.elsevier.com/S0958-9465(13)00104-2/h0030http://refhub.elsevier.com/S0958-9465(13)00104-2/h0030

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具

http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/vip.htmlhttp://www.xuebalib.com/db.phphttp://www.xuebalib.com/zixun/2014-08-15/44.htmlhttp://www.xuebalib.com/

Service life prediction of RC structures based on correlation between electrochemical and gravimetric reinforcement corrosion rates1 Introduction2 Experimental program2.1 Materials2.2 Test specimens and concrete mix design2.3 Experimental techniques2.3.1 Electrochemical linear polarization resistance (LPR) method2.3.2 Gravimetric weight loss method

3 Results and discussion4 Service life prediction for RC structures5 ConclusionsAcknowledgmentsReferences