Application of Artificial Neural Network (ANN) Technique to Reduce Uncertainty on Corrosion Assessment of Rebars in Concrete by NDT Method

  • M. Bal
  • A. K. ChakrabortyEmail author
Conference paper


The basic objective of this study is to assess corrosion behavior of steel bars in concrete members by nondestructive method of testing. Corrosion possibility is assessed by half-cell potential method (using CANIN) while resistivity meter (RESI) is used to estimate risk of corrosion of rebars in concrete members. Interestingly, higher half-cell potential indicates more possibility of corrosion, but higher value of resistivity indicates lower probability of corrosion. An extensive research program was undertaken in order to assess risk of corrosion using type 1 (Fe415:TATA-TISCON), type 2 (Fe500:TISCON-CRS), type 3 (Grade Fe415:ELEGANT steel), type 4 (Fe415:VIZAG steel), and type 5 (Fe500:SRMB steel) TMT steel bars of 16 mm diameter with M20, M40, and M60 grade concrete samples, prepared with OPC exposed both in AIR and NaCl for a period of 900 days. Taking only average values of experimental data, initially huge uncertainties were found. Secondly, after applying standard statistical method, uncertainties were slightly reduced. Third analysis was then taken up by modifying standard statistical process; satisfactory results were still not obtained. Then, fourth analysis was carried out with optimum values to minimize the uncertainties. Three-dimensional graphs for each case were plotted using MATLAB, an ANN-based software. More appropriate ANN-based software is required for better correlation.


Corrosion of rebar in concrete Half-cell potential Concrete resistivity Artificial neural network 


  1. 1.
    Solomon T, EERI M, Murat S (2008) Vulnerability of seismically deficient older buildings, risk parameters, significant threat to life safety and its survivability. Earthq Spectra 24(3):795–821CrossRefGoogle Scholar
  2. 2.
    Glass GK, Buenfeld NR (2001) Chloride-induced corrosion of steel reinforcement bars embedded in reinforced cement concrete by applying neural network techniques. Prog Struct Eng (Struct Control Health Monit) 2(4):448–458CrossRefGoogle Scholar
  3. 3.
    Eswari S, Raghunath PN, Suguna K (2008) Ductility performance of hybrid fibre reinforced concrete by applying neural network techniques. Am J Appl Sci 5(9):1257–1262, ISSN 1546–9239CrossRefGoogle Scholar
  4. 4.
    Rasa E, Ketabchi H, Afshar MH (2009) Prediction on density and compressive strength of concrete cement paste containing silica fume using artificial neural networks. Trans A Civ Eng 16(1):33–42Google Scholar
  5. 5.
    Neven U, Ivana BP Velimir U (2004) Application of neural network in predicting damage of concrete structures caused by chlorides. Published in proceedings of international symposium ASFCACT, pp 187–194Google Scholar
  6. 6.
    Mansour NJ (1996) Prediction of stress-strain relationship for reinforced concrete sections by implementing neural network technique. J King Saud Univ Eng Sci 9(2):169–189Google Scholar
  7. 7.
    Rose AL, Suguna K, Ragunath PN (2009) Strengthening of corrosion-damaged reinforced concrete beams with glass fiber reinforced polymer laminates. J Comput Sci 5(6):1549–3636, ISSN 1549–3636Google Scholar
  8. 8.
    Cardenas H, Kupwade-Patil K, Eklund S (2011) Corrosion mitigation in mature reinforced concrete using nanoscale pozzolan deposition. J Mater Civil Eng 23(6):752–760Google Scholar
  9. 9.
    Hayfield PCS (1986) The cathodic protection of reinforcing steel bars using platinised-type materials. Platin Metals Rev 30(4):158–166Google Scholar
  10. 10.
    Behzad B, Lisa R (2002) Corrosion protection of steel rebar in concrete using migrating corrosion inhibitors MCI 2021 & 2022. A report published by College of Engineering and Computer Science, California State University, Northridge, CA, pp 1–10Google Scholar

Copyright information

© Springer India 2013

Authors and Affiliations

  1. 1.Department of Civil EngineeringBengal Engineering and Science UniversityHowrahIndia

Personalised recommendations