The most important thing to remember is that concrete strength gain is very temperature-dependent. If you cure concrete at 20°C instead of 10°C, the rate of strength gain doubles. If you go from 20°C to 30°C, it doubles again. And if you go from 30°C to 40°C, it doubles again, and so on. Therefore, concrete curing at 40°C will be gaining strength eight times faster than concrete curing at 10°C.
This is because the hydration reactions between cement and water, which are what causes the strength gain, are exothermic: a little bit of energy (heat) is needed to make each reaction happen, but then those reactions produce even more heat. The higher the temperature, the faster the rate of reaction. For a given concrete mix, this is the single most important variable. The amount of hydration reactions that have taken place, as a proportion of the total possible reactions, can be quantified by a parameter called the “degree of hydration”, which goes from 0 to 1. If we cure concrete at different constant temperatures, and then plot the rate of hydration against degree of hydration, we get a family of curves as shown in Figure 1, page 40.
Figure 1 shows that the rate of hydration approximately doubles every time the curing temperature increases by 10°C. It is virtually impossible for degree of hydration to reach 1, as even in very old concretes, there is still some unhydrated cement. Degree of hydration is a very useful parameter, and is a far better measure of maturity than age. All concrete properties, including compressive and tensile strength, stiffness, shrinkage and creep, correlate much better with degree of hydration than they do with age, particularly during the first few hours and days.
Compressive strength has been found to have a linear relationship with degree of hydration. An example of this relationship from tests at Bond Street Station Upgrade is shown in Figure 2. Each concrete mix will have a unique linear relationship. Generally, stronger concretes have a steeper gradient. This linear relationship only applies during early age, perhaps up to a few days or approximately 0.6-0.7 degree of hydration.
Figure 1 also illustrates the kinetics of concrete hydration. The slow rate of hydration at the start is known as the ‘dormant period’. After this there is a rapid increase in the rate of hydration, which slows towards a peak and then decreases. A maturity relationship that takes account of temperature sensitivity can be used to ‘affine’ these curves, in other words, to make them all line up over each other. Therefore the shape of these curves is a characteristic of the concrete mix and depends on the properties of the cement, the admixtures used, and also the influence of cement replacement materials such as micro silica or limestone powder.
Figure 2 also shows that a critical degree of hydration is required before strength gain begins, usually between 0.05 and 0.08. This means that between 5 and 8 per cent of cement hydration reactions are needed for an initial skeleton to form and further reactions contribute to strength gain.
The rate of hydration can be found using the Arrhenius equation: dt is the rate of hydration, Ã(??) is the normalised affinity, Ea is the activation energy, R is the ideal gas constant and T is the temperature.
The normalised affinity is the shape of the affined curves from Figure 1, which can be thought of as the signature of the concrete mix. It is a function of degree of hydration. It describes how easy it is for reactions to take place at different stages of hydration. When normalised affinity is high, it is easier for reactions to occur, and when it is low, it is more difficult. The normalised affinity is not temperature dependent.
An example of a normalised affinity curve for an accelerated shotcrete mix is shown in Figure 3. Note the difference from the shape of the curves in Figure 1, which is due to modelling the accelerator reaction. The accelerator causes a very high initial value of normalised affinity, making it very easy for reactions to occur. Then there is still a dormant period, followed by the long hump of the main silicate clinker reactions which is where most of the strength gain happens. The activation energy can be thought of as the small amount of energy required to make each reaction happen. This is a constant and has units of Joules per Mole. This is the temperature sensitive part, and so it can be calculated from isothermal (constant temperature) tests at different temperatures. Figure 4 shows the activation energy calculation from Bond Street Station Upgrade. At each temperature, there are four tests, and the closeness of the points demonstrates the excellent repeatability of isothermal calorimetry.
Using the Arrhenius equation means that at any time, the rate of hydration can be calculated if we know the current degree of hydration and the current temperature. Therefore, we can use a time-stepping algorithm to calculate rates and by integration over time, degrees of hydration, if we have the temperature history of the concrete.
STRENGTH TESTING AND MONITORING
Strength testing of concrete can be thought of as having two distinct purposes, which are often conflated:
¦ Testing for conformity – i.e., does the concrete delivered to site have the properties we expect it to have, does it meet the specification? Often compressive strength is used as an index test (or ‘identity test’) to verify conformity of a whole range of properties on a day-to-day basis, such as flexural tensile strength, durability, etc., during construction.
¦ Monitoring of strength development for safety of temporary works – i.e., does the concrete have sufficient strength to allow a shutter to be struck and falsework to be removed? Or for shotcrete, does it have sufficient strength to support its own weight and the ground loading?
In Europe, testing of early age strength of shotcrete is mainly done with needle penetrometer and stud driving tests to EN 14488-2: 2006. In the UK, due to concerns about the risk to operatives of falling shotcrete or ground near to the face, these tests are often performed on a test panel sprayed just after a section of lining has been sprayed.
In effect, when we are testing the strength of shotcrete panels, we are testing conformity. But because the panels may experience a very different temperature history to the lining, we are not actually monitoring the strength of the lining. This is the same problem that is encountered for normal concrete – when we test cubes or cylinders cured in a water bath at 20°C, is it representative of the concrete in the pour? In cold weather, this could lead to overestimating the strength of the concrete. For mass pours or in warmer weather, perhaps construction progress could be accelerated if we knew the actual strength of the concrete in situ.
A real example of measured shotcrete temperatures at different locations in the same tunnel is shown in Figure 5. Figure 6 shows the strength development calculated by SMUTI based on these temperature histories. In this case, the lower temperature was caused by groundwater inflow. A panel sprayed at the same time would not necessarily identify this disparity, which at 10 hours is approximately 10 MPa. The difference at very early age is not large, and this is because the fresh concrete temperatures were similar. On sites where the fresh concrete temperature may vary, significant retardation at very early age may be experienced, and this may not be evident in a panel sprayed after the lining, when the equipment and concrete have had a chance to warm up.
The converse of this is that SMUTI is not as direct a measure of conformity as strength testing. Although significant deviations from expected temperature profiles would indicate a quality issue and intelligent triggers can be programmed into the software, the calculation assumes that the shotcrete delivered is the same shotcrete that has been tested in the laboratory, and that the only variable is temperature. Therefore, conformity has to be separately assured according to the requirements of EN 206-1: 2000.
USING SMUTI
SMUTI uses a thermal imaging camera to collect the temperature history of the shotcrete lining. The temperatures go into the software, and the timestepping algorithm calculates the degree of hydration, based on the normalised affinity and activation energy obtained from laboratory tests on the shotcrete mix.
In order to know the strength, the linear relationship between degree of hydration and strength is required. Since shotcrete needs to be sprayed under site conditions to obtain realistic strength results, panels are sprayed and simultaneously monitored for strength and temperature (Figure 7). The degree of hydration found by SMUTI is then calibrated against the strength testing results. An example is shown in Figure 2. This linear relationship and the temperature history is used by the software to provide strength results in real time.
SMUTI is currently used on representative areas of the shotcrete lining for each advance. For instance, the left axis, left shoulder, crown, right shoulder, right axis is usually the default. Other areas of concern, perhaps identified by having a significantly different temperature to the rest of the lining, can be easily added into the software.
Currently, research and development is focussed on integrating SMUTI with photogrammetry or laser scanning, so that the temperature history of every point on the lining is known, every scan being positioned in 3D space. This would give the engineer a near-real-time 3D view of shotcrete thickness, age and compressive strength.
The software can work offline, and synchronises with a server database when connected, enabling the data to be viewed from anywhere in the world. It tracks which user has entered which data, and does not permanently delete anything, making it easy to undo any edit or delete errors.
The experiences gained on site so far indicate that using SMUTI is very quick and intuitive with engineers reporting no problems using it. The training, not just on how the software works, but also on the theoretical background, upskills the engineers and miners, increases motivation and ownership and improves safety, efficiency and quality.
CONCLUSIONS
SMUTI allows us to monitor the strength gain of the whole shotcrete lining (as against local tests on a panel that may not be representative) from a safe and remote position. There is no other method or technology that can achieve this; it represents a step-change in safety and quality control of shotcrete tunnelling. Most in situ strength testing is conformity testing, and is not actually monitoring the safety of the temporary works. SMUTI monitors the early age strength of the shotcrete lining to allow better decision-making. If the actual strength is below the specification, then a disaster may be averted. If the strength is better than the specification, faster progress may be made without compromising safety.
SMUTI doesn’t need to replace traditional strength testing to be beneficial. When it is used in parallel with existing methods, it provides increased confidence in the extrapolation of test panel strengths to the lining and helps engineers and operatives understand the strength development better.
The software is very easy to use and taking readings is very quick and easy. It provides rich, accessible and traceable data.
