PROBABILISTIC STUDY OF TIME DEPENDENT PROPERTIES OF PLAINT &, FLY ASH CONCRETE

Abstract

Considerable research activity is currently in progress to tackle the problem of evolution of structural design principles based on probability and reliability considerations. This viewpoint is gaining momentum because of the fact that the nature of loads and material resistances is probabilistic. Revisions of Building Codes are underway in various countries including India. In this context, there is an established need for extensive collection of data on material strengths and loads, its statistical analysis and probabilistic modelling so that optimum values of various safety margins and load factors may be arrived at. The present study is aimed to systematically generate data on time dependent properties of plain and fly ash concrete, namely, strength, shrinkage, creep and creep recovery, to subject the data so generated to statistical analysis, and, time-variant and timeinvariant probability modelling. The two other related problems have also been included in this study, namely, effect of the use of fly ash as a part replacement of cement and the autoclaving of concrete as a quick curing method for concrete blocks and prefabricated com ponents. The experimental data has been generated for two mixes, MIX land MIX U, with compressive strength of 15 and 40 N/mm2 respectively. Fly ash concrete having equal consistency, with 5and 15 per cent partial replacement of cement, by equal weight of cement sized fly ash particles, are also included. The tests for various properties were conducted on specimens subjected to controlled environment, having a temperature of (27 +2)°C and relative humidity of (90 +5) per cent. Strength and shrinkage were studied upto the age of 180 days and creep upto a period of 150 days. To study the batch effect variance, the number of batches taken were ten for strength and four for other properties. MIX JH, having 28-day normal cured strength of 35 N/mm , was autoclaved at pressures varying from 0. 6N/mm gauge to 2.2 N/mm2 in steps of 0.4 N/mm2, for curing periods of 2 to 24 hours. The periods of humid air curing after autoclaving were 28-day, 90-day and two years. The present investigation shows that the variance in compressive strength, creep, shrinkage and creep recovery due to batching, under laboratory conditions, is insignificant at 1.0 per cent probability level for all the plain and fly ash concrete mixes. Further, with increase in age, the population mean of all the above properties increases, while C.O.V. decreases at a diminishing rate. The decrease in C.O.V. with ageis greater in case of creep and shrinkage. The relative histograms and frequency curves for the properties studied indicate clustering of the data around the mean with increase in age. The age effect is more pronounced in case of creep and shrinkage. Both shrinkage and creep appear to be stabilizing within the 28-day period while creep recovery stabilizes within a 10-day period. The information that these properties show a relatively lesser spread at later ages is very useful from the designer's view xi point, whose interest lies in later age values. The time-invariant probability modelling indicates that most of the generated data, for all the properties, follow the Gaussian and lognormal distribution, with 1.0 per cent level of significance. The time-variant probabilistic modelling of all the four properties studied has been done using a special class of the maximum likelihood method, that is, the 'Chi-square minimization'method. The scatter in the observed properties at each age has also been considered for evaluation of the constants of the regression models, and the 'best estimate' parameters computed along with their associated errors. The mean function of each of these properties is found to be linearly related with the logarithm of time. The proposed probability model fits the observed data at 5 per cent level of significance. Further, computed 99 per cent confidence envelopes contain most of the observed data. The size of these envelopes is around +12.5 per cent of mean, and it varies with time. A semi-stochastic model on the lines of linear homogeneous Markov's model is found to be satisfactory for predicting long term creep and shrinkage. The effect of partially substituting cement with fly ash is to lower the 7-day and the 28-day strength, and increase 180-day strength, the latter two being only marginal. The replacement reduces creep and shrinkage. For both plain and fly ash concrete, the effect of sustained load is to increase the compressive strength. xii The effect of increasing pressure and time of autoclaving on compressive strength is to increase the strength at a diminishing rate. Power expression (sc • a. th) fits the data well with high coefficient of correlation. Further, significant increase in strength has been observed due to subsequent humid curing of autoclaved concrete. From probabilistic design point of view, the findings of the present broad-based investigation thus give useful information on the time-dependent properties of plain and fly ash concre