3.2.1.1. Rheological behavior under steady-state conditions

To ensure the pumpability of mixes, the rheological behavior under steady-state conditions needs to be described by means of a rheometric test. As explained in Chapter 2, flowing cementitious materials exhibit viscoplastic behavior, which can be described by the Bingham, modified Bingham or Herschel–Bulkley models.

Several flow geometries can be used to analyze the rheological behavior of cementitious materials: parallel-plate geometry or coaxial geometry (Figure 3.1).

Parallel-plate geometry is often dedicated to cement paste, even though some planar rheometers allow concrete testing (necessary representative volume imposed by the size of the largest aggregate), whereas coaxial geometries are available at all scales (Feys et al. 2018; Perrot and Amziane 2024; Yahia and Perrot 2024).

The test protocol must allow the material to be destructured to erase its history of stress and material structuring. Tests generally start with shearing at high speed, which destructures the material. Then, stepwise ramps of rotational speeds are imposed, and resistive torques are measured to scan a sufficient range of shear speeds. A number of data analysis methods can then be used to transform the rotational speed–torque data into flow curves, that is, the evolution of shear flow as a function of shear rate. For example, the Reiner–Riwlin method (Feys et al. 2013, 2023, p. 266; Wallevik et al. 2015) or the method developed by Estellé et al. (2008a, 2008b) appear in the literature as effective solutions for obtaining rheograms.

However, it is important to note that rheological flow behavior is not directly related to pumping behavior due to the phenomenon of particle migration under shear (Secrieru et al. 2020). In this case, it is often the behavior of the lubricating layer that dictates the pumping pressure (Secrieru et al. 2018; Tavangar et al. 2022). It is then possible to set up tests directly on the lubricating layer (cement paste and fine sand) or to encourage the creation of a lubricating layer during the test, as with the linear shear test generated by a slip test in an instrumented tube (Mechtcherine et al. 2014; Nerella and Mechtcherine 2018).

3.2.1.2. Structural build-up of cementitious materials

The structural build-up of cementitious materials is often studied by measuring “static” shear yield stress after specific rest times. In this case, the age of the material and the aging conditions must be well controlled. Two time references can be used: the age of the material (after water–cement contact) and the resting time (after the sample has been placed in position for the test). For the study of structural build-up, the second time reference is used, which simulates the aging of the material after it has been deposited in layers (Perrot et al. 2016). Knowledge of the evolution of the static shear yield stress as a function of time enables stability conditions during printing to be verified.

The shear yield stress is often measured using a rheometer equipped with a vane geometry (scissometer) that limits the effects of wall slip. The test is traditionally carried out at a constant shear rate, slow enough to render viscous effects negligible. However, the speed must not be too slow to prevent the material from structuring during the test. In this case, the shear yield stress is calculated from the maximum torque recorded, representative of shear failure of the material (Roussel 2005). This test induces progressive destructuring of the material (transient regime) until it begins to flow. An increasing stress ramp can also be applied to the material, and the shear yield stress in this case corresponds to the sudden acceleration in material deformation recorded during the test. Typical curves for these tests are shown in Figure 3.2.

The cone penetration test is also often used to calculate the shear yield stress from the penetration of the material cone in the material at a given point in time (Lootens et al. 2009; Mazhoud et al. 2019). In this case, the shear yield stress is calculated by balancing the force of gravity acting on the cone and the shear forces generated on its surface.

It is interesting to note that versions of these tests which can be used continuously have been developed. A team at TU Dresden has set up a so-called “single batch” test, in which the same material sample is subjected to successive vane shear tests stopped at peak torque to limit material destructuring (Ivanova and Mechtcherine 2020). This method underestimates the rate of structural build-up by around 10%, but saves considerable time and reduces the dispersion of results caused by the variability of material samples. Similarly, slow-speed penetration tests have been developed to continuously monitor the evolution of the material’s shear yield stress (Pott and Stephan 2021; Reiter et al. 2022).

3.2.1.3. Elongation behavior of materials

During the printing process, cementitious materials are subjected to compressive stresses related to the support of the layers placed above them. For certain geometries of printed structures, the material may also be in tension (Jacquet et al. 2021a, 2021b). It is therefore interesting to study the material’s behavior in compression or tension to more accurately predict the material’s stability with regard to the loading of the base layer and the occurrence of cracks in the material’s zones, which are under tension.

Regarding compressive behavior, uniaxial compression tests can be used to determine the material’s compressive yield point. This yield stress can be measured on materials of different ages in order to monitor the effect of structuring on compressive behavior. When carrying out this type of test, care must be taken to limit wall shear to ensure pure compression loading. It is also important to note that the specimen’s slenderness (ratio of height to diameter in the case of a cylindrical specimen) will play a role in the type of loading, and a slenderness of two will be more suitable than a slenderness of one for compressive loading. Finally, optimal specimen dimensions must also be chosen to limit the effects of gravity on the sample and to have a volume adapted to the dimensions of the largest particles.

For tensile behavior, it is possible to test the material under its own weight at the exit of a print head (Jacquet et al. 2020; Ducoulombier et al. 2021), using the so-called “slug test”, or to use dumbbell geometries to generate tensile stress in the shrinkage zone. As with compression tests, it is possible to carry out these tests after different resting times to study the evolution of the cementitious material’s tensile strength (Jacquet et al. 2021b).

Examples of the geometries used to study the material’s behavior in compression and tension are shown in Figure 3.3.

Finally, it is interesting to note that capillary extrusion tests can be used to study both compressive behavior (elongation failure yield stress in the shaping zone) (Perrot et al. 2012; Varela et al. 2023) and tensile behavior via self-weight failure tests by measuring the mass of filaments (called “slugs” in the literature) at the die exit.

3.2.1.4. Apparent elastic behavior

The apparent elastic behavior of cementitious materials in the fresh state is a key parameter for the 3D printing process. Indeed, one of the two modes of failure of structures during printing is related to the elastic buckling of the structure. Printable cementitious materials exhibit significant deformations before material flow, which can be considered at first glance as elastic deformations.

It is first possible to obtain information on the apparent elastic behavior of structures by analyzing simple compression tests. However, care must be taken to use a Hencky strain calculation to account for the large strains obtained in such a test (Jacquet et al. 2022). This methodology provides an apparent material stiffness that can be used pragmatically to ensure that the structure will not buckle during printing. This procedure can also be carried out sequentially to obtain the evolution of the material’s apparent elastic modulus over time.

Some authors have suggested using the deformation of a cantilevered beam to estimate the material’s elastic modulus in its fresh state (Jacquet et al. 2020). However, this type of measurement is limited to materials that are sufficiently firm for the test to be possible.

Finally, oscillatory rheometry enables a closer examination of the behavior of cementitious materials below the yield stress. Shear flow is often used in this context, and the relationship between the elastic modulus and the shear modulus is applied, assuming a value for the Poisson’s ratio. A sinusoidal strain amplitude is imposed on the material at a frequency often close to 1 Hz, and the resulting induced stress amplitude is measured by the rheometer. This method provides continuous access to the loss modulus G’’ (viscous part) and the storage modulus G’ (elastic part) (Yahia et al. 2023). These moduli are used to calculate the apparent shear modulus G* of the material, which is the ratio of shear stress to strain.

3.2.1.5. Non-destructive testing

During the various phases of concrete structure construction, a range of internal pathologies can be encountered. As these are generally undetectable when they first appear, they can adversely affect the quality of the structure. Hence, it has become essential to use non-destructive methods to monitor structures regularly, in depth, and in real time. These techniques also have the advantage of reducing costs and intervention times. In this context, several non-destructive investigation methods have been implemented since the 1960s and are certified by COFREND (Comité Français pour les Essais Non Destructifs).

For 3D concrete printing, Acoustic Emission AE, a technique originating in the 1950s (Drouillard 1994), is undoubtedly the most widely used. The instrumentation and components of the AE apparatus must meet the conditions required by the standards (Fournier 2018). As a passive method, AE can be sensitive to any active process (e.g. stressed materials) and depends in particular on several material-specific factors: when a material is stressed, internal defects (micro-displacements) develop, releasing energy in the form of transient elastic waves. These waves, originating from the material itself, are detected by piezoelectric sensors active in a frequency range extending from 20 kHz to 1 MHz (ultrasonic domain). These sensors transform mechanical waves into electrical signals, which are amplified and read by an acquisition system (Mix 2005; Ohtsu and Grosse 2008), enabling a comprehensive assessment of the internal behavior of materials. Evaluation can be carried out by tracking events in real time, locating a source in 2D or 3D, based on propagation speed, and characterizing the type of source (mechanism produced) by distinguishing the shape of the emitted elastic waves. Standards have been developed to provide certain guidelines specific to the use of AE in concrete (RILEM Technical Committee (Masayasu Ohtsu) 2010; Shen et al. 2020) and to classify damage based on parameters extracted from waveforms, such as the classification of cracks into cracks resulting from tension or other mechanisms (Ohtsu 2006, p. 200; Ohno and Ohtsu 2010; Aggelis 2011). Broadband sensors are recommended for hardened concrete (resonant frequency sensors can also be used with frequencies above 60 kHz) (RILEM Technical Committee (Masayasu Ohtsu) 2010). The choice of frequency is crucial, as it is directly related to the wavelength in the material – the main characteristic of ultrasonic waves.

One of the major advantages of AE is its sensitivity to very low-energy phenomena. The use of AE has therefore been extended to the study of several intrinsic phenomena occurring during concrete setting, which play a crucial role in concrete performance. However, it is important to note the high attenuation of fresh concrete, which is of the order of 2 dB/mm (Iliopoulos et al. 2016), varying according to the frequency used. This implies a signal attenuation of 20 dB for a propagation of 1 cm, which is significant. Drying shrinkage, a phenomenon occurring in relatively fresh concrete, has been studied by several authors (Shiotani et al. 2003; Bacharz et al. 2015). Shiotani et al. (2003) monitored the formation of microcracks due to drying shrinkage in concrete containing glass beads (replacing aggregates) using broadband sensors at 12–960 kHz. They noted the effect of aggregates on AE activities, showing their beneficial effect on shrinkage limitations. Certain AE parameters were used to determine when microcracks develop into macrocracks. McLaskey et al (2007) monitored the formation of microcracks due to drying shrinkage using 200–800 kHz frequency transducers at the surface and depth of the concrete. They established a correlation between the frequency of acoustic emissions and setting time, identifying and classifying shrinkage microcracks by AE. With the same objective of classifying and identifying types of shrinkage microcracks, Bacharz et al. (2015) studied the evolution of microcracks in cementitious paste, on the concrete surface, and at the paste-aggregate interface using 30–80 kHz frequency bandwidth transducers. They noted higher acoustic energy at the concrete surface, compared to other locations.

Some authors have focused on the microstructural activity of concrete. Skal’s’kyi et al. (2004) linked acoustic emissions to the development of the crystalline structure, which was almost complete after around six hours, noting a parallel drop in acoustic activity. Van Den Abeele et al. (2009) tested two mortars with different W/C ratios and showed an increase in acoustic emissions for a lower W/C, explained by faster hydration. In particular, they asserted that AE activity begins after temperature peaks, a finding that was also highlighted in Chotard et al. (2001) and Assi et al. (2018). However, other studies (Iliopoulos et al. 2016; Dzaye et al. 2018, 2020) observed high AE activity just after mixing. Further investigation is needed to understand the effect of fresh mixing and the resonant frequency of the sensors used on AE activity to explain the origin of acoustic emissions in the fresh state more deeply. Chotard et al. (2001) linked acoustic parameters (amplitude and frequency) to hydrate formation, using 50–300 kHz wide-bandwidth sensors connected to waveguides. The effect of the W/C ratio observed by Lura et al. (2009) demonstrated the presence of air bubbles by AE, highlighting the phase when mechanical properties begin to develop. Dzaye et al. (2018) studied segregation mechanisms, showing the effect of the density and viscosity of the cementitious paste on the decrease in acoustic energy and propagation velocity. They also demonstrated the sensitivity of AE to the movement of aggregates simulated by falling glass beads.

In the majority of the studies cited above, sensors are placed on waveguides or formwork molds. However, Thirumalaiselvi and Sasmal (2019) conducted their study by placing the sensors directly in contact with the material. This prevented wave propagation for the first two hours after mixing, due to the high attenuation of fresh concrete (Iliopoulos et al. 2016). Nevertheless, and based on the evolution of the acoustic signals, they classified four mechanisms developing in concrete: the formation of CSH, the evaporation of water from the pores, chemical shrinkage and the appearance of microcracks.

Recently, starting in 2022, three studies (Liu et al. 2022; Skibicki et al. n.d.; Yue et al. 2022) have monitored and evaluated the behavior of concretes obtained by 3D printing compared with those cast by the conventional method (using formwork), in the hardened state under mechanical loading. Yue et al. (2022) studied the behavior of printed concretes subjected to splitting and three-point bending, using 35–100 kHz bandpass sensors. They recorded the peak frequencies of the AE, roughly indicating the onset of damage. Peak frequencies were lower for printed concrete than for cast concrete, and higher for tensile cracking modes than for shear cracking modes for both types of concrete. In particular, they integrated the acoustic parameters AF (average frequency) and RA (the ratio between wave propagation time and amplitude), collected during fracture, into an automatic learning algorithm. This allowed them to classify and distinguish between tensile and shear microcracking modes for each type of concrete. Following this calculation based on AE parameters (Yue et al. 2022), it was shown that the percentage of micro-shear cracks in printed concrete is higher than in formwork concrete. Liu et al. (2022) studied concrete reinforced with mesh placed as an internal core in the concrete. They printed five different lattice structures to evaluate the mechanical properties of concrete under three-point bending using AE and DIC (digital image correlation). A positive correlation was found between acoustic parameters, notably number of hits and energy, and the deformation process during the first three phases of concrete failure. The dominant frequency domain also provided information on internal damage. Low-frequency values (around 100 kHz) were caused by cracking of the hardened cementitious paste, and high-frequency values (around 480 kHz) were caused by fracture of the fine aggregate particles. It has been observed that the number of hits and the dominant frequency of lattice-reinforced concrete were lower than those of control concrete (Liu et al. 2022).

In addition to the energy dissipated by a material during an active process, AE can be used in transmission–reception mode, where energy is generated by one sensor and received by another (resonance frequencies can exceed 1 MHz). Skibicki et al. (n.d.) examined the effect of different distributions (linear and pyramidal) and the number of layers in printed concrete, on the material’s mechanical and acoustic properties. Acoustic analysis showed that ultrasonic (US) velocities and resonant frequencies are similar for printed concrete and reference concrete (similar result for US velocities in Liu et al. (2021)). No effect of the number of layers on US velocity was observed. Nevertheless, the US velocity indicated that failure of multi-layer printed samples occurs in different directions, in contrast to reference samples where cracks propagate mainly in the middle of the material. The effect of the presence of fibers (Skibicki et al. n.d.) in printed pyramidal samples was also illustrated by AE, with the highest US velocities, thus inducing high mechanical strengths (a similar result was obtained for formwork concrete by Kouta et al. (2021)).

Liu et al. (2021) conducted a study on the analysis of the anisotropy characteristics of printed concrete loaded in different directions using US propagation velocities. The results revealed a decrease in US velocities of 5.38% and 8.22% in the Y and Z directions, respectively, with loading applied in the longitudinal direction only. Results obtained by Sun et al. (2021) in the same context demonstrated that the US velocity was higher for printed samples compared to reference samples. According to the authors, this is due to a more compact microstructure and lower porosity; printed paste achieves a higher degree of final hydration, explaining the US speed result which is directly associated with high compressive strength. Ma et al. (2020) evaluated the mechanical properties of printed mortar during the initial curing phases. They obtained a result similar to the previous study, showing that printed elements reach an age of 3–7 days with a stiffness 35% higher than that of reference samples.

In conclusion, acoustic emission has been further developed for monitoring damage in hardened concrete. The application of this technique to the assessment of concrete in the fresh state has seen significant growth, highlighting the complex process of hydration. Nevertheless, the explanation of the sources of acoustic emissions is not yet fully understood. It depends on the conditions and assumptions made by the researchers during the experiments. It is therefore suggested that future research vary and record several input data (resonance frequency and sensor positions, concrete formulation and mixing process, environmental conditions). Note that in fresh concrete, in its liquid state, only longitudinal modes (where the particles in the medium vibrate parallel to the direction of propagation) propagate. Later, as the shear modulus increases (where the particles in the medium move perpendicular to the direction of propagation), shear modes and surface (Rayleigh) waves also develop. It would therefore be interesting to place AE sensors in contact with fresh concrete, so as to easily separate the different waves propagating in it. It should be noted, however, that direct contact results in strong attenuation of the waves (Iliopoulos et al. 2016), which suggests that the resonant frequency of the sensors should be chosen carefully. As far as the application of AE for 3D printing of cementitious materials is concerned, there are very few studies on this subject, spanning the period from 2020 to 2023. This research is a preliminary step toward future work to develop the basic acoustic parameters of printed structures and to deepen the analysis of crack location during loading, taking into account the anisotropy of printed concretes. It is also recommended to extend the analysis to the evaluation of mechanisms evolving during the setting of the printed material, particularly the development of microcracks at the interlayer of the material.

3.2.2.1. Controlling ingredients on receipt

Controlling ingredients on receipt is a vital step. In the case of printable mortars, the ingredients received by the manufacturer depend on the type of process used (see Table 3.2). Controls focus primarily on solid components: cement, secondary binders and the granular skeleton in general, since these are natural components. Additives, on the other hand, are synthetic products, and their consistency is rarely called into question.

3.2.2.2. Initial rheology and pumping control

Again, depending on the type of process, the number and possibilities of rheology control differ. Initial rheology can be measured at the mixing outlet in the case of continuous mixing, or even during mixing in the case of batch mixing. Control can be carried out using the same tests as those used for suitability, or with the aid of sensors included in the printing system.

As far as pumping is concerned, there is currently no reliable universal indicator of the pumpability of a cementitious material. However, pumping is monitored using pressure transducers to track potential anomalies. Accurate in-line measurement of flow using sensors is often relatively complex to achieve. As a result, most large-scale 3D printing technologies use progressive cavity pump technology, enabling smooth dosimetric pumping. However, the efficiency of these pumps depends on the rheology of the material. It is therefore customary to calibrate the flow rate at the start of printing.

When using a variable print flow rate, it may also be important to check the linearity of the flow rate with pump rotation, which may not be achieved due to the rheo-thickening nature of mortars at high shear rates (potentially achieved at high flow rates). This calibration procedure can be carried out by simple sampling, or automatically if the system is fitted with a flow sensor.

3.2.2.3. Control of material stiffening and printing stability

Once pumping has started, it is necessary to adjust the consistency of the extruded material before printing begins, and then to check the quality of the extrusion and ensure that it can print a given design. To do this, it is necessary to measure the consistency of the material leaving the extruder and its structuring. Once again, this involves measuring a shear yield stress and its evolution immediately after extrusion. However, extruder exit measurements are challenging since printable materials inevitably undergo rapid changes in shear yield stress over time, making this a very active field of research.

Most often, manufacturers, especially those using single-component technologies, aim to check that the consistency of the mortar and its evolution is repeatable and greater than the minimum value expected for a given material. In the case of multi-component technologies, it is possible to adjust these values by adding the accelerator to target the structural build-up rate required for a particular design.

3.2.2.4. Shrinkage control of printed samples

The 3D concrete printing process essentially requires no formwork to guarantee geometric conformity. On the other hand, the printed material is directly exposed to variations in humidity, temperature and air flow. When the material is young, it is likely to be subject to heavy drying, particularly in dry weather. This evaporation of water from the surface of the printed material layer can have repercussions not only on cement hydration, but also on water distribution in the core of the printed material. A number of technical solutions exist to reduce the influence of these physical phenomena on the final durability of printed cementitious materials.

Three major types of shrinkage are commonly distinguished in 3D printing: endogenous shrinkage, plastic shrinkage and drying shrinkage:

• Chemical shrinkage is mainly due to the chemical activity induced by cement hydration. It occurs throughout the mass of the material, not just in contact with the surrounding environment.
  
• Plastic shrinkage develops in concretes that lose water while still in a plastic state, a state commonly recognized for printable cementitious materials.
  
• Drying shrinkage develops during hardening. The main cause is the evaporation of water contained in the porosity of the concrete from the free surfaces of the structure exposed to the ambient air.

In the case of 3D-printed concrete elements, only the final two points differ from a cast-in-place element. Drying shrinkage is conditioned by the porosity of the concrete, the humidity of the ambient air and the ratio between the exposed surface and the volume of the element. For 3D-printed structures, the multi-layer nature of the material means that the exposed surface area is considerable. What is more, the use of VMAs (viscosity modifying agents) often induces high porosity in materials during curing.

Despite the significant contribution of drying shrinkage, plastic shrinkage remains predominant, as it is mainly related to the rheological properties of the printed material. This loss of water in the fresh state is induced by surface evaporation, which can be accentuated by the difference in ambient humidity. This plastic shrinkage can also be followed by a delayed migration of water from the core of the material to areas with lower water content, those having previously undergone drying by surface evaporation. This induces a pressure gradient dominated by capillary pressure, causing water to migrate from the center of the printed material layer toward the surface exposed to air, with the aim of returning to a state of equilibrium (Keita et al. 2019). This water migration is also determined by the rheological properties of the printed materials, in particular, the permeability of the material and the viscosity of the interstitial fluid. Optimizing the compactness of the granular skeleton and adding viscosity-modifying agents helps to reduce this permeability and the associated plastic shrinkage. The addition of cellulose ethers has now become commonplace in the formulation of printable cementitious materials, as it not only achieves the rheological properties required for extrusion, but also significantly reduces surface drying (Brumaud et al. 2013).

In addition to the surface effects (cracking, crazing, dimensional variations, etc.) caused by evaporation, the lack of hydration can negatively affect the cement’s hydration, thereby impacting the final strength of printed structures.

There are four main ways of curing printed materials:

• optimizing the formulation of the printable material by increasing the compactness of the granular skeleton and the viscosity of the interstitial fluid. Surface drying is possible, but is significantly reduced;
  
• preventing water evaporation by placing covers over the printed structures, which can also take the form of a hermetically sealed enclosure positioned on-site;
  
• using a curing agent, or shrinkage-reducing admixtures to reduce the surface tension of water within capillaries (Lura et al. 2007; Bentz et al. 2015; Federowicz et al. 2020);
  
• implementing a spraying system for the previously printed layer via a module attached to the print head (Moelich et al. 2021).

Considering the direct environment (temperature, ambient hygrometry, air flow, etc.) may eventually involve co-simulation of the additive manufacturing process during component design, so as to ensure that the required mechanical performance can be achieved. This co-simulation will consider aspects related to rheological properties in the fresh state (water migration/evaporation kinetics), properties during hydration (shrinkage and plastic deformation) and the prediction of the structure’s service-life mechanical properties.

3.2.2.5. Homogeneity control of printed materials

Homogeneity during printing can be checked by analyzing data collected during printing. For example, droplet mass dispersions collected before printing can be used to judge material consistency during printing (Ducoulombier et al. 2021). Similarly, it is possible to use direct visual inspection or via an on-board camera to analyze defects during deposition and ensure that material extrusion remains constant during the process. The quality of the bead (geometry and surface appearance) is often a very good approximation of its homogeneity, as material heterogeneities generally result in inhomogeneous flows.

Finally, the many sensors integrated into the systems, such as pressure transducers and motor torque measurement, enable us to obtain stable curves over time during printing, highlighting the stability and homogeneity of the material’s behavior during the manufacturing process.

3.2.2.6. Printing conditions and ageing of the printed structure

As in the precast concrete industry, controlling the aging conditions of the material is crucial. Limiting drying and controlling curing speed can be achieved in a precast-type industrial printing context to ensure optimum curing of the cementitious material. Concrete curing is a major element often used in prefabrication plants, particularly those using ultra-high-performance cementitious materials. For example, it can enhance mechanical performance and limit shrinkage and creep. The latter is more complex to apply when printing on site.

3.2.2.7. Summary

Due to the need to control the evolution of consistency and setting, quality protocol is naturally more important for additive manufacturing processes and is often an argument for the higher cost of this type of material. However, from our point of view, this increased control is an opportunity, particularly in the case of two-component batch printing processes. This is because the knowledge acquired about the material during preliminary trials or during the manufacturing process not only reveals any potential anomalies, but also enables the formula to be adjusted for optimum use of the material. There is even reason to believe that this could enable the use of materials whose cement content could be reduced in comparison with cementitious materials used in traditional processes, where excess cement provides the robustness required for their use in a context of low control. In the case of printable materials, a significant part of the robustness could be provided by the industrial process and quality control. Controlling curing is also an essential element in guaranteeing the quality and optimum performance of the cementitious material used.

When 3D printing in an industrial context, many parameters need to be controlled: it may be worthwhile to set up a database and analyze it to build a predictive tool for print quality and the influence of process parameters on the properties of interest. These data can also be used to build digital twins of the parts, including all the manufacturing data for the elements at each point of the printed part.

3.3.3.1. Compressive strength

For the mortar scale, the specimens tested are 40 × 40 × 40 mm cubes, with a tolerance of ±1.0 mm. In this study, these cubes were not derived from previous bending tests (EN-196-1). The cubes are therefore extracted by cutting the printed part (Figures 3.5 and 3.6). The three possible bead orientations are tested, and each is called a series. These three series are broken down into three locations, depending on where the cube was extracted from: bottom, middle and top. As each test has to be repeated three times, there are a minimum of nine specimens per series, making a total of 27 cubic specimens.

For the concrete scale, the test specimens are 150 × 150 × 150 mm cubes, with a tolerance of 3.75 mm. All three spatial directions are also tested. As each test has to be repeated five times, this makes a total of 15 cubic test specimens for the concrete scale.

The tests in this study (Figure 3.7) are recommended to be carried out on stress-controlled presses. After each specimen has been weighed and measured, it is tested in compression to failure, at a slow, constant loading rate (2,400 N/s for mortar, 0.6 MPa/s for concrete).

These tests are used to determine the material’s characteristic modulus of compressive strength in all three directions.

3.3.3.2. Flexural and tensile strength

Three- and four-point bending tests are recommended. For the mortar scale, the specimens tested are rectangular prisms measuring 40 × 40 × 160 mm, with a tolerance of ±1.0 mm. All three possible bead orientations are tested. Each orientation test is repeated five times, for a total of 15 specimens.

For the concrete scale, the test specimens are rectangular prisms measuring 100 × 100 × 400 mm, with a tolerance of ±3.0 mm (Figure 3.8). All three possible bead orientations are tested. Each orientation test is repeated five times, for a total of 15 specimens.

The tests in this study are designed to be carried out on stress-controlled presses. After each specimen has been weighed and measured, it is tested in three-point (Figure 3.9) and four-point (Figure 3.10) bending to fracture, with a length between supports of 100 mm for the mortar scale and 300 mm for the concrete scale.

These tests are used to determine the characteristic flexural strength modulus in all three directions and to calculate a tensile strength modulus.

3.3.3.3. Tensile strength and interlayer qualification

For direct tensile tests, the specimens are of the same type and size as the prismatic specimens used for bending tests (see section 1.3.3.2).

The tests in this study are designed to be carried out on universal testing machines which can be controlled in terms of displacement. After each specimen has been weighed and measured, it is tested in tension. These tests are used to determine the characteristic tensile modulus.

To characterize the impact of the 3D printing process on a given material, the test protocol developed by RILEM has made it possible to compare results from numerous laboratories and companies around the world. The principle is to study the impact of superimposed beads in all three spatial directions, mainly by testing cubic specimens or rectangular prisms. However, this test protocol is more varied than summarized in this section.