10. Conclusion
In this paper, an existing model allowing calculation of the average moment of inertia of RC beams during cycles of loading and unloading is modified to take into account the effect of historical cracking damage on the serviceability reliability of RC members. The suitability of the average moment of inertia model for reliability analysis is verified by considering experimental tests on a total of eleven reinforced concrete beams. The model errors associated with both the effective and the average moment of inertia are calibrated using the experimental data. Combining the sources of uncertainty of RC-members and MCS, an quantitative analysis approach is presented to evaluate the loss in serviceability reliability due to the historical cracking damage for the reinforced concrete beam.
By using the proposed approach, both short-term and long-term serviceability reliability of a cracked reinforced concrete beam was analyzed. The results confirm that the effect of historical cracking damage on short-term serviceability reliability should be taken into account, when the deflection induced by historical loading is larger than the deflection limitation. In such a scenario, neglecting the historical cracking damage leads to overestimation of the serviceability reliability of cracked RC-members. Light historical damage (e.g. Ppre = P1, in this case) has no influence on the short-term serviceability reliability, although it affects the probability density distribution of the deflection of the beam. However, even when the historical damage is light, the long-term serviceability reliability index is decreased as the cracking damage affects the time-dependent deflection. Additionally, the later the damaging load is applied to the reinforced concrete beam, the less influence the cracking damage has on the long-term serviceability reliability.