Practical cases in the application of the pullout method (LOK-TEST and CAPO-TEST) for in-place compressive strength

. The pullout methods LOK-TEST and CAPO-TEST for in-place compressive strength are presented with their theoretical analysis’ background and correlations from 30 major studies, made worldwide, showing robust general correlations between pullout force and strength by cylinders or cubes/ cores. The coefficient of variation of the systems are shown, reported in 1984. Practical cases using the systems are described: Case 1. In-Situ compressive strength testing of quarantined precast concrete tunnel lining segments using CAPO-TEST, UK; Case 2. Strength testing with CAPO-TEST on old bridges for further loading, Poland; Case 3. Safe and early form stripping with LOK-TEST, Canada; Case 4. Curing of the cover layer evaluated by pullout and bulk resistivity, Denmark.


Introduction
Reliable and quick testing of existing structures for strength may be important for purposes such as documentation of unknown strength, for upgrading, for further loading or for documentation of doubtful structures in cases where questions are raised in relation to compliance with code specifications. Pullout offers such advantages. Testing of cores from the structure is doubtful, depends on many factors such as moisture, planeness, aggregate size, L/D ratio and presence of reinforcement. Coring is also time consuming, expensive and causes large holes in the structure. The use of indirect methods such as the rebound hammer and/or ultrasound (UPV) requires for every structure many cores for establishing the correlation, and the relationships obtained are not sensitive.
On new structures, production control of the in place actual strength of the structure is essential, not only trusting laboratory strength, but also considering the effects on in-situ strength of the actual mix delivered, the transportation, the pumping, the casting, the compaction, and the curing of the cover layer, especially in aggressive environments. Again, rebound hammer and ultrasound pulse velocity require correlation testing for each mixture, involving lab testing. Estimation of strength by maturity requires a pre-established maturity-strength relationship for the mixture used and does not consider the effects of transportation, pumping, casting, consolidation and curing.
Pullout produces reliable estimates of strength in place, based on one correlation, the test systems are rapid and economical, only minor damage cause to the structure, and can be used without testing of cores, [10].
A special feature is testing of the cover layer, the "peel" of the structure protecting the reinforcement. The curing of this "Peel" is essential on new structures in terms of durability, not at least if chlorides are present from de-icing salts, the sea or airborne. Pullout can be used for this purpose for production control.

The pullout systems
Invented at the Danish Technical University (DTU) in the late 1960's and 1970's, [1,2], the LOK-TEST (the Danish name for "Punch-Test") uses a disc cast into the fresh concrete, and the CAPO-TEST (Cut And Pull Out-Test) uses a ring expanded in an undercut recess in existing concrete, [3,4]. Pullout is made through a counterpressure with dimensions as shown in Figure 1 and Figure 2, producing compression forces between the expanded ring and the counterpressure, hence the pullout force is a direct measure of the compressive strength.  Plasticity analysis of the failure was published in 1976, [5], as well as a comprehensive finite element analysis in 1981, [6]. The first one concluded that the pull-out force is proportional to the concrete compressive strength, and the second that the failure is caused by crushing of the concrete in the "strut" between the disc and the counterpressure.
In the following years, major correlation studies on specimens were performed in Denmark, Sweden, Norway, Holland. Canada, USA, Poland, England and KSA, investigating parameters such as types of cementitious materials, water-cementitious materials ratio (w/cm), age, air entrainment, use of admixtures, curing conditions, stresses in the structure, stiffness of the member, carbonation, as well as shape, type, and maximum size of aggregate up to 40 mm. Only for lightweight aggregate concrete, was another different correlation found.  In Figures 3 and 4, 30 such studies are summarized, 18 between pullout force and standard cylinder strength (Figure 3), and 12 between pullout force and standard cube or core strength (assuming that a 100 mm diameter by 100 mm long core gives the same strength as a 150 mm cube, Figure 4).
Furthermore, the pullout forces produced by the LOK test and CAPO test were found to be identical for the same concrete quality ( Figure 5). The robust general correlations are highlighted for cylinders (red) and for cubes/cores (blue) in Figure 6.
The findings in the two mentioned theoretical analyses, the plasticity theory modelling and the finite element "strut" theory, match well the cylinder correlation (red) for uniaxial compressive strength. Coefficients of variation were first reported on a large scale in 1984, [7]. Table 1 shows the results in laboratory conditions and Table 2 are from in-situ tests.

Case 1: In-Situ compressive strength testing of quarantined precast concrete tunnel lining segments using CAPO-TEST, UK
Tunnel elements were produced at the Translink Joint Venture, in the Isle of Grain, UK, and hardened in a heating tunnel on a moving conveyer belt. For strength estimation, cubes were placed alongside. The production took place in large numbers, automatically. The cube strength, after heating, was specified to be 60 MPa. During a period, the cube strength dropped, but production continued until the drop was realized.
All the elements produced in that period were quarantined. Scrutinizing, it was later established that the reason for the drop was a change in the cement used in the mix; the gypsum component in the cement had been changed.
To test the final strength of the quarantined elements, two systems were selected as candidates, coring or CAPO testing. Testing with cores was too time consuming, uncertain and costly, and considered to cause too much destruction. CAPO testing was selected due to the minimal damage.
A calibration program was conducted in relation to cube strength ranging from 35 MPa to 100 MPa, partly between production cured cubes and CAPO testing, and partly between standard cured cubes and CAPO tests ( Figure 7). Testing was made in relation to maturity at 4, 7, 28, 154 and 329 actual days. Subsequently, the quarantined elements were tested at random in a statistical valid manner with three CAPO tests in each element as shown in Figure 8.
All the quarantined elements older than 150 days old were accepted for erection in the tunnel, as the strength with CAPO tests related to cube strength showed strength over 60 MPa from 150 days and onwards. Variation in the CAPO test was on average 9.6%, ranging from 7.9% to 11.5% for all the elements. Testing on each element lasted about 1 hour for each 3 CAPO tests.

Case 2: Strength testing with CAPO-TEST on old bridges for further loading, Poland
Fifty bridges in Poland needed to be upgraded for increased loading (Figure 9). Fifteen of the bridges, ranging in age from 25 to 52 years old, were investigated initially, for establishing the correlation curve between cores and the CAPO test (ASTM C-900, [17], and EN 12504-3, [19]), with special focus on the effect on carbonation. The depth of carbonation varied from 2 mm to 35 mm. The strength of the bridges ranged from 20 MPa to 50 MPa. The testing in detail is reported in [9]. As can be seen from Figure 10, the best fit curve (purple) match almost identically the general correlation for cubes (red) showed in Figure 6. The average COV on the cores was 7.4% and 8.8% on the CAPO-TEST.
Most interestingly, the effect of carbonation was only minimal (2.8%) on the CAPO test and there was no correlation found between the depth of carbonation and the relative error of the estimate based on the CAPO test.
The Schmidt Hammer rebound technique was also performed. The estimated strength from this test showed about 80% higher strength than cores using the general correlation published by the manufacturer of the instrument. Subsequently testing of the remaining bridges was performed only by CAPO testing.

Case 3: Safe and early form stripping with LOK-TEST, Canada
Not only for accelerating construction schedules, but also for safety, the LOK test pullout system is used extensively for testing the strength of slabs during construction on high rise residential and office structures in Canada.
This test system is used in conjunction with optimized concrete mixes, by which a scheduled time of construction can be reduced, saving interest, costs of formworks, reshoring, winter heating and earlier rental, [14,15].
In a 100 m 3 slab pour, 10 to 15 LOK test inserts are installed equally distributed on the bottom of the slab through prepared port holes in the flying form systems. Inserts can also be installed as floating inserts on the top, but the bottom installation is preferred due to simplicity ( Figure 11). At the time of testing, a couple of inserts are tested, evaluated by maturity, and if meeting the expectations, the remaining inserts are tested. 10 inserts can be tested in about 1 hour. The LOK test pullout forces are converted to equivalent cylinder strength in MPa by means of a preestablished relationship. The standard deviation is calculated, followed by calculation of the "Minimum inplace strength" as: Average Strength less a K-factor times the Standard Deviation. The "K" factor relates to the 10% fractile of the T-distribution. If the "Minimum in-place strength" is higher than 75% of the f'c, stripping/reshoring takes place, otherwise, testing of remaining inserts is performed later, e.g., after another half-a-day, and the "Minimum in-place strength" is recalculated.
This procedure has been adhered to in many cases for safe and early loading of slabs in high rises as the Scotia Plaza in Toronto (Figure 12), Canada, where earnings due to speeding up construction schedule was reported to be 1.5 million dollars. Optimized concrete mixes were used, allowing forms to be removed as quickly as after 1.5 actual days, even in cold winter conditions. On the other hand, in the substructure, strength is not needed that quickly. Here e.g., fly-ash, slag cement, or other supplementary materials may be used in the mix, reducing the costs of the concrete mix. A full description including correlations, variability and reports from 18 projects is given in [11,12,13,14]. On projects as reported in Trinity Square, [13], the building officials allowed elimination of the usual mandatory standard cylinder tests, only relying on LOK testing for in-place strength.

Case 4: Curing of the cover layer evaluated by pullout and bulk resistivity
For resistance to chlorides from, e.g., the sea or deicing salts, the cover layer is the "Peel of the Orange", Figure  13, protecting the reinforcement against corrosion. Similar with carbonation. This "peel" is the essential part of a new structure when it comes to durability, not the interior. To achieve a good, durable cover layer, the right mix has to be used, it has to be well compacted, have a sufficient thickness and be well cured. Optimal curing is providing water or keeping the formwork in place during hydration, alternatively, using internal curing with LW fine aggregates or water absorbent polymers, while less efficient curing is achieved if curing compounds, or plastic sheets are applied. Fig. 13. The cover layer is the "peel of the orange" for concrete.
No curing has significant detrimental effects, as does exposure to high temperature and wind (miscuring). Early Danish research in 1969 at DTU showed a 31% reduction in LOK test pullout strength for a w/c-ratio of 0.36, and 40% for a w/c-ratio of 0.50 when concrete is miscured compared to water curing at 20 o C. But how about the resistance to chlorides?
Recently, a comparison between LOK test pullout strength and bulk conductivity has been performed for estimating the chloride diffusivity and service life for wet cured concrete and air cured, for simplicity.
The two standards applied were: ASTM C900-19: "Standard Test Method for Pullout Strength of Hardened Concrete" (LOK-TEST, Figure 14) and ASTM C1876-19: "Standard Test Method For Bulk Electrical Resistivity or Bulk Conductivity of Concrete" (using the Merlin device, Figure 15). The last one was performed on slices of cores, estimating the chloride diffusion coefficient and the service life, defined here as the estimated time it takes to build up a critical chloride content at the depth of the steel reinforcement.  The concrete used was a C40/C50 class concrete (f'c 40 MPa on cylinders, 50 MPa on cubes) which was tested after 56 days.
The average results from testing of three sets of specimens (wet and air cured) are shown in Table 3. The LOK test showed a 23% compressive strength reduction. The bulk resistivity testing with the Merlin device on the 50 mm cover layer resulted in 166 Ω•m for wet curing and 111 Ω•m for air curing, which represents a 33% reduction. With simplified assumptions, these resistivity values can be transformed to a chloride diffusion coefficient, D a , using the Nernst-Einstein relation. This way, wet curing would correspond to a chloride diffusion coefficient of 27.2 mm 2 /y and air curing to 41.5 mm 2 /y. For a 50 mm cover layer and sea water splash exposure condition, the estimation shows a 40% reduction of the service life regardless of whether the critical limit for corrosion of the reinforcement is considered to be 0.050% Clor 0.100% Clby concrete mass (Table 4 and Figure  16). For miscured concrete (wind and higher temperature), the reduction would be much larger.
In this manner, a quick on-site strength test, the LOKtest or the CAPO test, will immediately indicate the cover layer quality. If lower than expected, cores may be drilled out from the cover layer, sliced and water saturated for further testing with the MERLIN for bulk resistivity (or its inverse, conductivity) and estimating the remaining service life in chloride environment.
More examples of pullout used for cover layer quality are shown in Figure 17.

Conclusions
Pullout test is a physical test, as are compression of standard cylinders, standard cubes or cores. The compression and the crushing of the concrete happens between the cast-in disc (Lok-Test) or the expanded ring in a recess (Capo-Test) and the counterpressure on the surface. Robust correlations between pullout force and standard cylinders, or standard cubes or cores exist, and can be used with great confidence without further correlations involving traditional laboratory testing. In this manner a portable compression machine -the hydraulic Lok-Test and Capo-Test pull machine -can be brought to the site for testing. In several cases the lab testing compression machine has been eliminated, e.g., as reported in [13] or on remote sites where the lab compression machine was too troublesome to bring along.
If needed, for potential strength in the lab, 200 mm cubes with inserts are installed centrally in the vertical faces, compacted, water cured and tested at specific time intervals and using the general correlations, eliminating the need of the traditional compression machine in the laboratory. This paper illustrates the successful use of pullout for testing of dubious structures, for testing of old structures before further loading, for safe and early loading of structures involving use of optimized mixes and for evaluation of the curing of the cover layer protecting the reinforcement against aggressive environment to maximise the service life.
Capo-Test has proven to be a useful inspection tool in condition evaluation of old structures, especially when the reinforcement is densely installed and cutting of reinforcement (for example in bridge columns) has to be avoided. Also, on slim columns where cores would be weakening the column, Capo-Test has been preferred.