Ordinary Portland cement (42.5R grade) was used as the binder. Natural river sand with a fineness modulus of 2.6 served as the fine aggregate. Expanded shale lightweight aggregate with a bulk density of 950 kg/m³ was employed as the coarse aggregate. A polycarboxylate-based superplasticizer was added to improve workability. Tap water was used for mixing and curing.

The mixture proportions of the lightweight aggregate concrete (LWAC) are summarized in Table 1.

Concrete was mixed in a pan mixer and cast into 100 mm × 100 mm × 100 mm cubes and 100 mm × 200 mm cylinders. After 24 h, the specimens were demolded and cured in water at 20 ± 2°C until testing at 28 days.

(1)

Compressive strength was tested on cubic specimens according to ASTM C39.

To ensure reliability of the literature data, the above experimental results were compared with reference values reported in previous studies on LWAC. The compressive strength of LWAC-1 was measured at 32.4 MPa, which falls within the reported range of 30–35 MPa for similar mix designs. The splitting tensile strength was 2.6 MPa, closely matching the literature-reported ratio of 7%–10% of compressive strength. The measured density was 1860 kg/m³, which also aligns with published LWAC values (1800–1900 kg/m³).

These initial experiments confirm that the mechanical performance and density of the prepared LWAC are consistent with literature data, thereby validating the reliability of the reported ranges used in this study.

The mechanical behavior of lightweight aggregate concrete (LWAC) was simulated using a two-dimensional mesoscopic finite element approach based on the Base Force Element Method (BFEM). The LWAC specimens were modeled as a three-phase material comprising lightweight aggregates (LWA), cement mortar, and the interfacial transition zones (ITZs). The random circular aggregate model [16] was extended to explicitly incorporate microscale variables, including aggregate density, particle size, Poisson’s ratio, water–cement ratio, and ITZ mechanical properties.

The random circular aggregate model was generated using a random sequential addition (RSA) algorithm in MATLAB R2023a. Aggregate diameters followed Fuller’s particle size distribution curve, ensuring realistic gradation within a 2D projection of the experimental particle range (5–20 mm). The probability density function P(D)=(D/Dmax)n with n = 0.5 was used to compute representative particle sizes. Each aggregate was randomly placed within the specimen domain, and overlap was prevented by enforcing a non-intersection condition: (1)dij≥Di+Dj2where dij is the center-to-center distance between aggregates i and j. When an overlap occurred, the particle was repositioned until the condition was satisfied. The final aggregate volume ratio was verified numerically to match the prescribed value (25%–40%).

Fig. 1 illustrates the overall process, including input parameter definition, random placement, overlap checking, ITZ generation, and meshing.

The specimens were discretized using a Delaunay triangulation mesh, providing high-quality triangular elements for accurate stress distribution and crack propagation analysis. The mesh size was chosen to ensure at least 8–10 elements along the diameter of each aggregate particle. A mesh sensitivity study was conducted to ensure that the predicted mechanical properties converged; no significant changes were observed when the mesh was further refined.

(1)

Compression Simulations: The bottom boundary of each specimen was fully constrained in both horizontal and vertical directions. A uniform vertical displacement was applied to the top boundary to simulate uniaxial compression. Lateral boundaries were free to move.

Each phase in the LWAC was assigned isotropic elastic–plastic damage properties: (1)

Cement Mortar: Elastic modulus, Poisson’s ratio, and tensile/compressive strengths were taken from experimental data [8].

These ratios were selected according to [8] and confirmed through parameter sensitivity analysis in Section 3.11.

The BFEM simulations employed an incremental displacement control approach. Nonlinear equilibrium iterations were solved using a Newton–Raphson scheme. Convergence was achieved when the norm of residual forces fell below 1 × 10⁻⁶ of the total applied load. Adaptive time stepping was used near peak stress and post-peak softening regions to ensure stable convergence and accurate capture of crack initiation and propagation.

The numerical model was validated against experimental data from [8], including uniaxial compression and tension tests for specimens with varying aggregate porosities (23.5%–48.5%) and volume fractions (25%–40%). The predicted stress–strain curves, peak strengths, and failure modes showed excellent agreement with experimental observations, confirming the reliability of the simulation methodology.

To evaluate the robustness and reliability of the proposed mesoscopic model, a parameter sensitivity analysis was conducted to quantify the influence of key modeling parameters on the predicted mechanical properties of lightweight aggregate concrete (LWAC). Three primary parameters were examined: (1) the thickness of the interfacial transition zone (ITZ), (2) the fluctuation of the elastic modulus of lightweight aggregates, and (3) the mesh size.

Lightweight aggregates (LWA) are characterized by complex internal pore structures that critically influence the mechanical behavior of lightweight aggregate concrete (LWAC). In this study, the lightweight aggregate is idealized as a dense honeycomb closed-pore structure, consistent with microstructural observations reported in prior experimental investigations [16]. To verify the reasonableness of this assumption, representative data from mercury intrusion porosimetry (MIP) and X-ray computed tomography (CT) scans in the literature were analyzed to characterize the pore morphology and distribution.

Low-density lightweight aggregate particles are a type of light porous material. Firstly, the internal pores of lightweight aggregate particles are dense, honeycomb-like micropores, and they are all closed, which contributes to the lower unit weight of lightweight aggregate than normal concrete. Secondly, for brittle materials such as concrete, internal pores are one of the primary causes of failure. Therefore, this study will investigate the effect of the porosity of lightweight aggregate particles on the mechanical performance of lightweight aggregate concrete.

This study considers only the case when the lightweight aggregate particles have pores themselves. Fig. 6 shows multiphase lightweight aggregate concrete with pores. It is assumed that the lightweight aggregate concrete consists of lightweight aggregate, cement mortar, interface layer, and pores in the lightweight aggregate.

Because light aggregate particles have very tight pores inside them, which are very heterogeneous in volume and mostly fall at the microscale, the investigation of pores as a discrete phase in concrete is difficult to do. Studies [20–23] with the three-phase sphere model and the hollow cylindrical model have been employed to estimate the effective bulk modulus and the effective shear modulus of porous composite materials assuming the matrix material and pores are isotropic and estimated the effective elastic modulus and Poisson’s ratio of the materials. This study applies it to investigate the porosity of lightweight aggregates, exploring how different internal porosity degrees of lightweight aggregate particles affect lightweight aggregate concrete mechanical performance. Fig. 7 is a model of pore meso-equivalent analysis for lightweight aggregates. However, Fig. 8 shows the three-phase spherical model of lightweight aggregate composite material containing pores.

In modeling the mechanical behavior of lightweight aggregate (LWA), it is essential to account for the effect of porosity on the elastic properties of the material. Since LWA contains numerous pores and micro-defects, its stiffness and deformation characteristics differ significantly from those of natural dense aggregates. To describe these effects, both the bulk modulus and shear modulus must be modified to incorporate the influence of porosity. Eqs. (3) and (4) are therefore introduced to derive the effective bulk modulus K∗ and shear modulus G∗, which form the foundation of the elastic response of porous aggregates. Once these effective parameters are obtained, the elastic modulus E∗ and Poisson’s ratio υ∗ can be determined using classical isotropic elasticity relations, as expressed in Eq. (5) [22].

The inclusion of Poisson’s ratio is particularly important because it governs the lateral deformation behavior of the aggregate under loading. For lightweight aggregate, which typically exhibits higher deformability compared with dense aggregates, Poisson’s ratio directly influences the stress distribution, volumetric changes, and cracking resistance of the composite material. Eqs. (6) and (7) provide explicit relationships between porosity and the effective elastic modulus/Poisson’s ratio, enabling a quantitative assessment of how pore volume alters the stiffness and transverse deformation of LWA. These expressions are critical for accurately predicting the meso-scale response of lightweight aggregate concrete in subsequent simulations. (3)K∗=4K0G0(1−V)4G0+3K0V

The effective shear modulus G∗ derived from the hollow cylindrical rod model is: (4)G∗=G0(1−V2)where V is the porosity of the lightweight aggregate, K0 and G0 are the bulk modulus and shear modulus when the porosity of the lightweight aggregate V=0. For isotropic materials, from the effective bulk modulus and effective shear modulus, the effective elastic modulus and Poisson’s ratio can be obtained as: (5)E∗=9K∗G∗3K∗+G∗,υ∗=3K∗−2G∗6K∗+2G∗

Substituting the shear modulus G0=E02(1+υ0), the bulk modulus K0=E03(1−2υ0) into Eq. (5), the relationship between the effective elastic modulus and effective Poisson’s ratio with porosity can be obtained as follows: (6)E∗=6(1−V2)6+3V−3υ0V+V2+υ0V2⋅E0 (7)υ∗=6υ0−3V+3υ0V−V2−υ0V26+3V−3υ0V+V2+υ0V2where υ0 is the Poisson’s ratio when the porosity of the lightweight aggregate V=0; E0 is the elastic modulus when the porosity of the lightweight aggregate V=0.

The effective strength f∗ of lightweight aggregate containing pores and the strength f when the porosity of the lightweight aggregate V=0 is related as follows: (8)f∗=f0(1−V23)

The equivalent critical strain derived from this is: (9)ε∗=f∗E∗=f0(1−V23)E∗where V is the porosity, and f0 is the strength of the fully dense aggregate. The equivalent critical strain ε∗ is obtained by combining the effective strength with the effective elastic modulus E∗:

To clearly relate this to the reference strain of the fully dense aggregate, we note that ε0 = f0/E0, yielding the proportional relationship: (10)ε∗=E0(1−V23)E∗ε0

It is known from the experiments in [16] that the porosity of lightweight aggregate is V = 43.5%. In this study, the method of controlling variables is adopted. At V = 43.5%, the specimen is completely the same as specimen 1 in [8]; other specimens only perform equivalent calculations related to porosity, and other parameters remain consistent with study of [16].

Based on the material parameters taken in [16], according to the formulas, it can be deduced that when V=0, υ0 = 0.427, E0 = 20,196 MPa, f₀ = 3.89 MPa. Compared with ordinary concrete, the porosity of lightweight aggregate concrete is generally larger. After consulting related literature [17,18], the general range of porosity of lightweight aggregate particles was determined. This study selected porosities V of 23.5%, 28.5%, 33.5%, 38.5%, 43.5%, 48.5% for research. The equivalent parameters for different porosity of lightweight aggregates are shown in Table 2.

It is worth noting that literature data inevitably involve variations in cement types, water sources, and aggregate origins, which may lead to differences in reported results. Nevertheless, the comparative analysis of these studies provides valuable insights into the overall performance trends of LWAC.

For ordinary concrete, aggregate is the material with the highest strength in the multiphase medium of concrete and occupies a high-volume proportion. The mechanical properties of concrete are not only related to the type, strength, particle size, and shape of the aggregate but also to the volume ratio of the aggregate. The volume ratio of the aggregate significantly affects the workability and durability of concrete [25,26].

Lightweight aggregate concrete material can be considered as a three-phase material composed of the matrix phase, the dispersed phase, and the bonding interface between the two. Its mechanical properties are mainly influenced by the matrix and dispersed phases. The dispersed phase, that is, the lightweight aggregate, is generally the weakest part of the three-phase material and has a very apparent effect on the strength and crack resistance of lightweight aggregate concrete [19]. Karimi et al. [27] found through research that when the mechanical properties of the matrix are weaker than the dispersed phase, the volume ratio of the dispersed phase has a significant impact on the fracture energy of concrete. When the mechanical properties of the matrix are stronger than the dispersed phase, the volume ratio of the dispersed phase has a greater impact on the strength and brittleness of the concrete.

This study establishes random aggregate mesoscopic models with lightweight (coarse) aggregate volume ratios Pk of 25%, 30%, 35%, and 40% and analyzes the impact of the volume ratio of lightweight aggregates on the macroscopic mechanical properties of lightweight aggregate concrete.

The simulations reveal a clear trend in which increasing LWA porosity from 23.5% to 48.5% leads to a shift in weak points from the interfacial transition zone (ITZ) to the aggregate itself. At low porosity levels, the aggregates are relatively strong, and cracks predominantly initiate and propagate along the ITZ. This observation aligns with prior studies indicating that the ITZ is the weakest phase in concrete’s heterogeneous microstructure [5–7].

As porosity increases, the internal microvoids within the aggregates reduce their intrinsic strength. Consequently, the aggregates become the primary weak zones, resulting in more severe damage concentrated within the LWA particles during both compression and tension. For instance, at 48.5% porosity, the damage at peak stress is primarily localized inside the aggregates, whereas the ITZ exhibits comparatively minor cracking. These findings confirm that aggregate porosity is a critical factor influencing LWAC failure, consistent with experimental observations in [16,17].

Stress-strain curves from the simulations further illustrate the impact of porosity. Peak stresses decrease monotonically with increasing porosity, while peak strains slightly reduce due to reduced aggregate stiffness. This behavior demonstrates the negative correlation between LWA porosity and LWAC strength, emphasizing the importance of controlling aggregate quality in practical applications.

The volume fraction of lightweight aggregates also plays a crucial role in LWAC behavior. Increasing the aggregate content enlarges the weak regions in the concrete, thereby reducing overall compressive and tensile strength. The numerical models show that higher volume ratios lead to more tortuous failure paths, as cracks navigate around or through weaker aggregates.

In compression tests, cracks are primarily concentrated around large aggregates, forming preferential paths, while smaller aggregates influence crack distribution less significantly. In tension, higher LWA volume ratios accelerate crack coalescence, promoting earlier specimen failure. These results highlight the combined influence of aggregate size, distribution, and volume fraction on the macroscopic mechanical response of LWAC, which is consistent with experimental findings reported in [18,25].

Our study also examines the combined effects of porosity and volume ratio. At moderate porosity, increasing the volume ratio exacerbates damage concentration in weaker zones, accelerating failure. At higher porosity levels, even lower volume ratios result in significant aggregate damage, suggesting that porosity dominates failure behavior at extreme values. These observations emphasize the need to consider both microscale properties when optimizing LWAC for strength and durability.

The numerical results have been compared with literature-reported experimental data for LWAC under uniaxial compression and tension [16,17,24]. Peak strengths, stress-strain behavior, and failure modes show good agreement, with deviations generally within 5%. This comparison validates the reliability of the numerical framework and confirms that the model accurately captures the effects of aggregate porosity and volume ratio. Notably, the failure modes observed in simulations, including crack initiation in the ITZ and subsequent propagation through weaker aggregates, closely match those observed experimentally.

The findings provide practical guidance for LWAC mix design. Controlling aggregate porosity is critical to ensuring mechanical strength, particularly in applications requiring high durability and load-bearing capacity. Similarly, the aggregate volume fraction must be optimized to balance lightweight performance with structural integrity. These insights can help engineers design LWAC mixes with improved reliability and predictability, reducing the likelihood of premature failure.

While the present study provides detailed insights into LWAC microscale mechanics, it is limited to 2D numerical models and uniaxial loading conditions. Future work could extend the modeling approach to three-dimensional simulations and include environmental effects such as high temperature, freeze-thaw cycles, or long-term creep. Additionally, investigating the influence of aggregate shape, surface texture, and novel sustainable lightweight aggregates could further enhance understanding and application of LWAC in construction.

To highlight the novelty and academic contribution of this study, a comparative analysis was conducted between the present work and relevant previous studies, particularly [16,19,24,27]. Table 7 summarizes the methodological differences, modeling scopes, and main findings.

The present study improves upon prior BFEM-based LWAC models by extending the range of aggregate porosity, integrating sensitivity analysis of key mesoscale parameters, and refining the representation of the interfacial transition zone (ITZ). While study [16] provided the foundational modeling framework, its limited porosity range restricted generalization. Here, the porosity interval was extended to 23.5–48.5%, enabling a broader investigation of mechanical behavior transitions from dense to highly porous structures.

Furthermore, compared with experimental findings in [24], the current numerical simulations show strong consistency in both compressive strength degradation and stress–strain nonlinearity trends. The effects of aggregate volume ratio on brittleness, discussed experimentally in [19,27], were also quantitatively analyzed using the BFEM model, confirming that higher aggregate content increases peak strength but reduces post-peak ductility.

Overall, this study establishes a more comprehensive and experimentally validated framework for analyzing LWAC behavior, offering improved generality, interpretability, and parameter sensitivity compared to existing literature.

The present study focuses exclusively on the uniaxial mechanical behavior of lightweight aggregate concrete (LWAC) under standard laboratory curing and loading conditions. While these conditions provide a fundamental understanding of the intrinsic material response, they do not encompass the diverse environmental scenarios encountered in real engineering applications. In practice, LWAC structures are often subjected to complex environmental stresses such as elevated temperatures, freeze–thaw cycles, and moisture fluctuations, all of which significantly affect mechanical performance and durability.

Previous studies have demonstrated that temperature rise can alter both the physical and mechanical characteristics of lightweight aggregates and the surrounding cementitious matrix. Roufael et al. [3] reported that exposure to high temperatures leads to gradual degradation in compressive strength and stiffness due to microcrack propagation and expansion of internal pores. Similarly, Herki [4] observed that the residual strength and integrity of lightweight concretes strongly depend on the type and stability of the aggregates used under thermal exposure. Moreover, repeated freeze–thaw cycles can induce microstructural fatigue, weaken the interfacial transition zone (ITZ), and accelerate crack coalescence.

Consequently, the current conclusions regarding stress–strain behavior, damage evolution, and brittleness are primarily valid within the standard curing regime and should be cautiously extrapolated to extreme service environments. Future extensions of this research will include coupling mesoscale modeling with thermo-mechanical and hygro-thermal effects to quantitatively assess the degradation mechanisms of LWAC under such environmental influences.