Characterisation of normal and high-strength plain and fibre-reinforced concretes by means of strength, fracture and combined fracture/relaxation tests.
The main objective of this study was the application of strength, fracture and
creep/relaxation tests to plain and fibre-reinforced high-strength concretes.
Initially, five grades of concrete were developed and evaluated. Target 28 day
compressive strengths were 40,60,80 100 and 120 N/mm2, the latter three being
high-strength concretes (HSCs) containing a süperplasticiser and 10% silica fume.
The others were normal-strength mixes used for comparison purposes. Each grade
was made with 10 mm maximum-sized crushed limestone and gravel coarse
aggregates making ten mixes in total. All were required to have sufficiently high
workability and stability to accept reasonable amounts of fibre reinforcement. The
data reported allows estimates of mix proportions for a range of HSC mixes to be
Various amounts of steel and polypropylene fibres were then added to the ten mixes
to determine their optimum and maximum practical concentrations. The traditional
type of toughness test based on un-notched beams in four-point loading was not
employed. Instead, notched beams in three-point loading (equivalent to the RILEM
work-of-fracture arrangement) and compact compression specimens were used.
Both were tested under closed-loop conditions using crack-mouth opening
displacement (CMOD) control. Post-cracking toughness was determined by means
of the 15a nd 110to ughness indices given in ASTM C 1018. It was found that though
fracture-based tests under CMOD control were an improvement on more traditional
techniques, 15 and 110 were too insensitive to allow fibre type and volume to be
Next, work-of-fracture tests to measure the fracture energy, GF, were carried out on
the plain concretes, initially under quasi-static loading. Both load/deflection and
load/CMOD curves were recorded. GF showed little change with strength for a given
aggregate type. Even though similar grades of crushed limestone and gravel HSCs
had different GF values, the measure was still considered unsuitable for
characterising the fracture properties of concrete. Similar experiments were then
carried out on all ten mixes at five orders of magnitude of test duration (30 seconds
to 2 days). GF appeared to be independent of strain rate. Both types of test
highlighted the greater suitability of load/CMOD rather than load/deflection curves
when evaluating GF.
Finally, combined fracture and relaxation tests were undertaken in an attempt to
obtain medium term fracture parameters. Though the CMOD was locked at 90,70
and 50% of the peak load in the strain-softening region, the deflection, when
measured, showed a noticeable reduction over the seven days of each experiment,
suggesting that significant cracking and stress redistribution within the fracture
process zone was taking place. This finding has opened up a major area of
important future research by confirming that the use of quasi-static fracture
parameters in finite element studies is suspect.