Optimization check of certain recent practical formulations for concrete creep

Optimization algorithms that have recently become available in computer libraries revolutionize checking and identification of theoretical creep laws from test data. Much more detailed comparisons with creep test data are now feasible, and with little effort. Formulating the optimality condition in terms of a sum-of-squares objective function and expressing various positiveness constraints by quadratic substitutions, one can apply the Marquardt algorithm. In this manner, two recently proposed formulations are examined: (a) the viscoelastic model with reduced time, and (b) the rate-of-flow method. It is shown that none of these formulations is capable of giving a satisfactory description of creep data which cover the full range of interest in creep durations and ages at loading, even though an acceptable agreement has previously been demonstrated for creep data of narrow time range. Previously it has been found by the same method that the recently proposed creep formulation for C.E.B. Recommendations suffers by the same limitations. That formulation and the two formulations examined herein share the underlying concept of separating the total creep strain in reversible creep and irreversible creep. The present demonstrate that this (theoretically unfounded) concept is contradicted by creep data of not too limited time range.