Subject matter

Introduction

High amount of work has been performed during the last decade by 2 previous RILEM Technical Committees in the field of high temperature concrete behaviour.

RILEM TC 227-HPB dealt with the Physical Properties and Behaviour of High-Performance Concrete (HPC) at High Temperature. Two STARS have been published: one STAR on the properties (mechanical, thermal, transfer properties of HPC) at high temperature and one STAR on modelling of concrete behaviour at high temperature

RILEM TC 256-SPF dealt with the very peculiar behaviour of concrete exposed to fire: Spalling. One STAR on this behaviour and 2 recommendations on test methods will be published. In addition, several papers have been published based on the work performed in the frame of the TC.

The present proposal deal with the reassessment of the Framework. The concerned reference documents are principally RILEM recommendations on mechanical test methods published between 1995 and 2007 and the EUROCODE 2 Part 1-2 (Eurocode material models at high temperature).

These reference documents have never been assessed in the frame of RILEM TCs i.e. How accurately can fire resistance be predicted from this framework?

The topics covered by the different TCs during the last decade and the one presently submitted, even if they all deal with concrete at high temperature, are very different. The new proposal is complementary, but it is not a continuation of work of previous committees.

Context

Much scientific work on physical properties and behaviour of concrete at high temperature has been performed during the last half century. These research studies formed the basis of the work carried out by the following RILEM Technical Committees that have been active in the field of high temperature concrete behaviour over the past 40 years:

• RILEM COMMITTEE 44-PHT “Behaviour of concrete at high temperature”
• RILEM TC 129-MHT “Test methods for mechanical properties of concrete at high temperatures”
• RILEM TC 200-HTC “Mechanical concrete properties at high temperature - Modelling and applications”
• RILEM TC 227-HPB “Physical Properties and Behaviour of High-Performance Concrete at High Temperature”
• RILEM TC 256-SPF “Spalling of concrete due to fire: testing and modelling”

Most of these technical committees have produced State of the Arts and recommendations on test methods. The latest published State of the Arts have focused on the physical properties and behaviour at high temperature of Ordinary Concrete and High-Performance concrete [Pimienta, McNamee and Mindeguia, 2019] and on the Modelling of Concrete Behaviour at high temperature [Millard and Pimienta, 2019]. An upcoming State of the Art will be published on concrete spalling under fire.

The published recommendations on test methods for determining the mechanical properties at high temperature deal with stress-strain relationship, compressive strength, tensile strength, modulus of elasticity, thermal strain, transient state strain, steady-state creep and creep recovery, shrinkage, restraint stress and relaxation (see references).

In Europe, guidance on the fire design of concrete structures is given in the Eurocode 2 – Part 1-2 [EN 1992-1-2:2004]. This document contains various tables prescribing the minimum dimensions of the sections and the necessary reinforcement cover for different types of concrete elements according to the required fire resistance: columns, walls, tensioned elements, beams and slabs. The principles of simplified calculation methods are detailed for fire design. The general principle is based on the temperature profile of the cross-section of the considered element.

Section 3 of EC2 Part 1-2 describes the properties of the materials (concrete and steel reinforcement) as a function of temperature (from 0°C to 1200°C). The properties of concrete given in the Eurocode include strength and stress-strain relationship in compression, tensile strength, thermal elongation, specific weight, specific heat and thermal conductivity.

Thermal properties were determined from furnace fire tests by using inverse methods.

Part of the background experimental results related to compressive and tensile strength and thermal elongation were determined according to RILEM recommendations.

The temperature dependant stress-strain relationship given in the Eurocode was based on a model developed in the 1970s by Popovics for describing the behaviour of concrete at room temperature. In 1985, this model was suggested by the RILEM Committee PHT 44 as a way of describing the stress-strain relationship for concrete under compression at high temperature. In this recommendation, to keep it simple and generic, the effect of transient state strain was included in an implicit way. Transient state strain is an experimentally determined non-reversible strain component that develops in the opposite direction of thermal expansion during heating of concrete under compressive load. Moreover, this is the most complex and least (as compared to other strain components) understood strain component, affecting the behaviour of concrete at high temperature.

Moreover, in recent years, the increased emphasis on considering fires with cooling phase (also referred to as natural fire or real fire scenarios or physically based fire exposure) and performance-based design, has called attention to the limitation of the Eurocodes. Although Eurocodes allow the use of advanced modelling approaches to evaluate the performance/behaviour of RC structures, it leaves many open questions on how to implement such advanced design approaches. The limitations of the current constitutive (stress-strain) law for concrete at elevated temperature has long been criticised (and under review) for not being suitable for concrete members with high compressive loads and for fires with a cooling phase [Bratina et al., 2005; Gernay 2012; Bamonte and Lo Monte 2015; Kodur and Alogla 2017]. These limitations are mainly attributed to the implicit consideration of transient state strains in the stress-strain law of concrete.

Fire spalling is an important parameter when designing concrete structures. However, and unfortunately, there are no validated theoretical models that can be used to replace fire testing assessment. Very important work has been performed in this area in the frame of RILEM TC SPF that includes the production of a State of the Art, 2 test recommendations and several publications in journals and conferences. These publications are in preparation. In the frame of the new TC, we propose to extend that work by the creation of a database summarizing results from fire spalling tests around the world. Although, there have been a few publications aimed at collecting the details of large numbers of experiments of this type, these works have barely scratched the surface and need to be strongly pursued.

Aim of the TC

The aim of this proposed Technical Committee is to conduct a re-assessment of the existing framework for assessing concrete under fire, from material behaviour to real structural behaviour. This also includes the creation of a RILEM database for fire spalling test results. To fulfil this aim, the TC will work on three main tasks, A-C illustrated in the Figure 1.

Figure 1: Summary of the tasks A, B and C of the TC

Task A aims to assess the existing framework for structural fire design of concrete structures. It consists of the 3 following sub-tasks:

• Sub-task A1: Compare fire resistance test results with calculations based on models and material properties given in the Eurocode
• Sub-task A2: Compare fire resistance test results with calculations using the calculation framework (e.g. stress-strain curves) from the Eurocode but using mechanical properties obtained from experiments that follow the RILEM recommendations
• Sub-task A3: Define the limitations of existing Eurocode constitutive law and suggest refinements to account for different scenarios and concretes

Task B aims also to look at the whole framework from material characterization to modelling to real behaviour during fire exposure of elements. But, in this task, we are studying in more detail concrete structural behaviour under fire. The field of investigation is narrowed down to the behaviour of structures made of one or two selected mixes, where the characterization and modelling approaches are extended to the state of the art methods and not limited to the RILEM recommendations and Eurocode framework. The main activities will be to coordinate this type of research study among the members of the TC.

Task C focuses on the spalling behaviour of concrete structures under fire. It aims to generate an open access database containing the conditions and detailed outcomes for past experimental test data available in literature.

The library will focus on studies where the main objective was to investigate spalling of concrete due to fire and on studies where spalling was observed despite the study being focussed on other behaviours of concrete during/after fire. Additionally, this task will intend to include results from fire tests performed by industry.

Due to the scale and scope of this task, a RILEM TC presents the ideal conditions. Information on key publication and sources of information will be made available by the TC’s members.

Terms of reference

The TC organization will be based on parallel working groups (WG) corresponding to the 3 tasks A to C and sub-tasks. Members of the new TC can take part and contribute to more than one WG. Two annual meetings (2 days) will be organized.

Detailed working programme

Task A Review of existing framework from material characterization to real behaviour

The main activity will be to review the available literature such as scientific literature, RILEM publications and “Background documents” of Eurocodes on the following:

• studies comparing fire resistance experiments on different structural members with calculations based on the material models given by the Eurocode (A1)
• studies comparing fire resistance experiments with calculations based on the framework in the Eurocode but substituting mechanical properties experimentally determined by following the RILEM recommendations (A2)
• Assessment of the use of the test protocols from the RILEM recommendations (A2). What are the deviations? Why are there deviations? Do we need to update the recommendations?

The aim of task A3 is to provide a simplified, generic and practical solution to improve the existing material constitutive law, in contrast to the other complex possibility of implementing the transient state strain explicitly as suggested by some studies in literature.

Based on the literature reviews done under task A1 and A2, complimented by the numerical investigations, the limitations of the mechanical model in the Eurocode will be defined (A3). For the numerical investigation, well documented experimental studies will be identified and selected during the review process, which will then be used to serve as a benchmark to validate the possible proposals to improve the constitutive stress-strain law of concrete at elevated temperature.

Finally, based on the findings from the literature review and numerical investigations, task A aims at providing guidelines to improve the existing constitutive model of concrete at elevated temperature to address the existing limitations.

Task B: Detailed study of mixtures from material characterization to real behaviour

Task B.1 Choice of one or two reference concrete mixtures based on available experimental data

Within this subtask the available experimental data for specific concrete mixtures will be collected by the members. One or two mixtures will be selected to carry out further experimental investigations.

In the best case, well-documented experimental data is already available for the selected formulations. In this case, the members of the RILEM TC can supply the experimental data or confirm it through complementary experiments, depending on their capabilities.

The experimental data should include both extensive material characterisation and fire tests on components. Due to the high effort and costs involved, it is recommended to use concrete formulations with existing results of component tests and, if necessary, to carry out additional material tests within the framework of the RILEM TC. Due to the large and diverse expertise of the numerous members of the RILEM TC, a wide range of different material parameters for the selected concrete formulations may be created and compiled.

For some selected material properties, by determining the same characteristics under comparable conditions, round robin test can be carried out between the individual test laboratories.

Task B.2 Additional experiments based on B.1

The following preliminary list of additional possible experiments to be performed in the frame of task B.2 has been established. This list will be updated by the TC on the basis of the analyses performed in B.1.

• Materials characterisation
  • Mechanical properties at high temperatures (stress-strain relation, compressive strength, tensile strength, modulus of elasticity, thermal strain, transient state strain, steady-state creep, shrinkage, restraint stress)
  • Thermal properties (conductivity, heat capacity, density)
  • Microstructure characterisation (Permeability, Hg-Porosity, Microscopic Investigations, Sorption-Isotherm)
  • Dehydration and decomposition behaviour (TGA and DSC, Microscopy (SEM)

• Intermediate scale
  • Screening tests to assess the spalling behaviour

• Large scale
  • Column-tests
  • Wall tests
  • Slab tests

The realization of these works will also depend in particular on the means of the laboratories in relation to the large-scale tests, which are very costly.

Task B.3 Advanced numerical investigations

Based on the compiled and extended material knowledge, the material behaviour will be modelled on different scales in this task.

• Microscale --> dehydration behaviour, development of porosity and cement phase decomposition
• Mesoscale --> crack evolution, development of permeability, moisture transport
• Macroscale --> this scale of modelling is used for evaluating the structural behaviour. A numerical round robin exercise is proposed to recalculation/simulate the behaviour of structural members tested under Task B.2 and other well documented experimental studies from literature

Important points that will be focused on in the round robin exercise will include:

• Effect of transient state strains on structural response:
  • Comparing the effect of different transient state strain models available in literature on the fire resistance (only considering heating phase)
  • Comparison of implicit and explicit transient state strain models on the structural behaviour (including cooling phase)
• Effectiveness of the improved simplified concrete constitutive model proposed in Task A, considering technical feedback from different researchers.

Task C Fire spalling library

As written above, the RILEM TC presents the ideal conditions to perform this work as key publications and sources of information may be made available by the TC’s members.

The following information will be collected (this preliminary list will be updated by the TC):

• Type of concrete (strength, density, avg. moisture content, type of mix, admixtures, age of concrete, fibres, curing conditions)
• Scale/dimension of the samples
• Heating apparatus used (standard fire resistance furnace, non-standard gas furnace, electrical oven, radiant panel, heating blanket)
• Heating condition (temperature measured inside the furnace/oven with plate thermometer, shielded thermocouple, bare thermocouples or incident radiant heat flux)
• Sustained stress-strain conditions (e.g. compressive stress state [MPa])
• Any other specific detail of the experiment
• Outcomes of the experiment
• Method of spalling observation (video and/or audio recordings and level of explosiveness)
• Time-to-first spalling and to last-spalling [min]
• Type of spalling observed (surface spalling, explosive spalling, progressive spalling)
• In-depth temperature inside sample (if measured)
• In-depth pore pressure inside sample (if measured)
• Mean and max depth of spalling [mm]