Prestressed concrete is a method to take advantage of the favourable properties of concrete in compression. Prestressing makes the structure compressed before loading. Some advantages of prestressed concrete are less deflections and eliminated or smaller cracks. Some of the prestressing is lost, both instantaneously and in the long term. Different design codes estimate these losses in alternative ways, and estimated residual prestressing forces thus turn out different. The code-based estimations differ often from measured prestresses. The estimations of residual prestressing forces with design codes are difficult according to assumptions of the prestressing systems properties, and long-term effects. The residual prestressing forces have a large influence on the response and capacity in both the Serviceability Limit State and the Ultimate Limit State, and thus estimated residual prestressing forces are important.

In this thesis, a literature study was first carried out. It covers the concepts of prestressing, and estimation of prestressing losses with focus on post-tensioned concrete. Design codes treated in this thesis were the one used in the beginning of 1960 (mostly based on the handbook from Lorentsen (1963)) and one from present time, Eurocode 2 (SS-EN 1992-1-1. 2005) (referred to as EC2 in this thesis). It was thought to be interesting to compare how residual prestressing forces once were estimated, with how they are estimated nowadays.

The study focussed on two simply supported beams (called Benchmark 1 and 2) found in the literature. For them, tests and instrumentation aimed to estimate the residual prestress force in laboratory were available. These experimental values served as a main point of comparison with the calculated prestressing losses according to EC2 and the handbook from Lorentsen (1963) for the respective case study. The beam in Benchmark 1 had a span of 6.62 m, and the beam in Benchmark 2 had a span of 4.2 m. The analysis was first done with the load case according to the original situations in the papers (linear analysis), and then extended until failure (non-linear analysis).

The background for this thesis was that a FE-model should be calibrated according to the measurement equipment at the Autio bridge. Estimations of residual prestressing forces should later be compared with code based residual prestressing forces. This however turned out to be too complicated and optimistic. The thesis does, however, provide a summary of the Autio Bridge and its tensioning system.

For Benchmark 1, a sensitivity analysis to estimate the influence of concrete material properties on the residual prestressing force and structural response of the beam was carried out. Updated material parameters were used instead of the initial ones in the codes for estimation of residual prestressing forces and structural response. The intention was to evaluate how much the age of the material affected the estimation of residual prestressing according to the codes, and the structural response. In the paper of Benchmark 1, instantaneous prestressing losses were estimated with an analytical method, based on displacements of the beam. The analytical method for the residual force was compared to the code estimated residual prestressing forces, and then modelled for comparison of structural response. Also, a new reference testing set with time-dependent (427 days) prestressing losses were evaluated for the codes.

In the paper for Benchmark 2, time dependent (356 days) losses were estimated both experimentally, with a novel formulation proposed by the authors, as well as with other design codes. These loses were compared with the time dependent losses calculated according to the 1963 code and EC2.  For benchmark 2 the total losses were calculated with both codes, to later be compared and modelled numerically.

The time-dependent losses of the prestressing force were 25.5% for the 1963 code, 10.3% for EC2 and 6.7% experimental for Benchmark 2. The measured losses for Benchmark 1 were also lower compared to the code estimated losses. The codes thus overestimated the losses compared to the measured ones. The losses where generally larger with the 1963 code compared to EC2 for both Benchmarks.

For the displacement, Benchmark 1 with instantaneous losses and time dependent losses (concrete parameters at day 28) loaded according to the paper had the following displacement in the middle: 1.4 mm for EC2 and 1.9 mm for the 1963 code (residual prestress of 75.6% respectively 65.1%). Less residual forces gave larger displacement, less compressed concrete, less tendon forces, and less load capacity. For Benchmark 1, with instantaneous and time-dependent losses loaded according to the paper, the following difference appeared in displacement in the middle, when concretes parameters were updated from 28 days to 427 days: 1.4 mm to 0.95 mm for EC2 (residual prestress 75.6% to 81.5%). Updating concrete parameters also gave larger concrete compressive stresses, larger tendon stresses and larger load capacity. Updating concrete parameters thus affects both structural estimated residual prestressing forces and structural responses.