Four aggregate classes were sourced from the batching plant stockpiles: fine sand 0/2, coarse sand 0/5, gravel 5/12.5, and gravel 12.5/25. The maximum aggregate size is Dₘₐₓ = 25 mm, as required by the structural specifications of the project.

The particle density of sands 0/2 and 0/5 is 2.65 and 2.68 kg/L, respectively, with fineness moduli of 2.0 and 2.8. The sand equivalent of the 0/2 sand is 90%, well above the minimum threshold of 65% required by the standard. The Los Angeles abrasion values of gravels 5/12.5 and 12.5/25 are 25.1% and 27.7%, respectively, both conforming to the maximum limit of 30% prescribed for structural concrete [10]. Aggregate properties are summarised in Table 1.

Table 1. Physical properties of aggregates.

Two Portland-based cements manufactured by LafargeHolcim were used, selected according to the structural function and environmental exposure class of each concrete element. The CPA-CEM I/52.5N is a pure Portland cement with a characteristic 28-day compressive strength of 58.7 MPa. The CHF-CEM III/B 42.5 L-LH/SR is a blast-furnace slag cement with low heat of hydration and high sulfate resistance, achieving 49.4 MPa at 28 days. The latter is particularly appropriate for elements exposed to aggressive environments (classes XS and XA), where alkali-silica reaction and sulfate attack are identified risks [11]. Chemical compositions and physical properties are presented in Table 2 and Table 3, respectively.

Table 2. Chemical composition of cements (Source: LafargeHolcim technical data sheets).

Table 3. Physical properties of cements (Source: LafargeHolcim technical data sheets).

Two admixtures were incorporated in all five mixes. PLASTIRETARD, a set-retarding admixture, was used at a dosage of 0.2% by mass of cement. Its role is to maintain workability during transit from the batching plant to the pour locations, given the severe road-congestion conditions in urban Abidjan. TEMPO-12, a polycarboxylate ether (PCE) superplasticiser, was dosed between 1.3% and 2.2% by mass of cement, depending on the target strength class. It enables a substantial reduction in mixing-water content while preserving the S4 consistency class (slump 160 - 210 mm) required by the project specifications. The positive effect of PCE superplasticisers on both workability and compressive strength through water reduction has been confirmed by multiple recent studies [8][12].

The Dreux-Gorisse method is a semi-empirical grapho-analytical approach that determines optimum constituent proportions from the reference grading curve and the target compressive strength [13]. Key design inputs include the target strength f_c28, the cement-to-water ratio (C/W), the NF EN 206-1 exposure classes, and the desired consistency class. The method systematically accounts for local aggregate grading characteristics and incorporates iterative corrections based on reception testing, making it particularly well suited to African construction sites where aggregate properties can vary significantly from batch to batch [9].

Five formulations were developed in accordance with the exposure and strength classes defined under NF EN 206-1, ranging from blinding concrete (C20/25, class X0) to prestressed box-girder concrete (C40/50, classes XC4-XF1-XS1). All mixes target the S4 consistency class (slump 160 - 210 mm). The resulting mix proportions are presented in Table 4.

2.5.1. Sampling and Specimen Preparation

Fresh concrete samples were collected directly from ready-mix trucks at the pour locations in accordance with NF EN 12350-1. Cylindrical specimens (∅160 mm × H320 mm) were cast in steel moulds and compacted with a vibrating needle in three equal layers (NF EN 12390-2). After demoulding at a minimum age of 16 hours, specimens were stored in water at 20˚C ± 2˚C until the test date.

Table 4. Mix proportions of the five concrete formulations (Dreux-Gorisse method).

2.5.2. Slump Test

Fresh concrete workability was assessed using the Abrams cone slump test (NF EN 12350-2). The standard cone (base ∅200 mm, top ∅100 mm, height 300 mm) was filled in three layers, each rodded 25 times. The slump was measured immediately after cone removal. The target range of 160 - 210 mm corresponds to consistency class S4.

2.5.3. Compressive Strength Test

Compressive strength was determined at 7 and 28 days of age in accordance with NF EN 12390-3, using a hydraulic press conforming to NF EN 12390-4. Loading faces were capped with molten sulfur to ensure planarity and parallelism. A constant loading rate of 0.5 MPa/s was applied throughout each test (within the permissible range of 0.2 - 1.0 MPa/s).

Figure 1 summarises the average compressive strengths recorded at 7 and 28 days for Mixes 1 (275 kg/m³), 2 (360 kg/m³), and 5 (400 kg/m³), together with the measured slump and w/c ratio.

Mixes 1, 2 and 5 are compared here as three complete mix formulations containing 275, 360 and 400 kg/m³ of cement, respectively; these formulations also differ in w/c ratio (0.62, 0.46 and 0.46), and, more generally, in aggregate proportions and admixture dosage, so the comparison reflects associations between complete formulations rather than the isolated effect of cement content. A strong positive association is observed between the mix formulations and 28-day compressive strength: the change from 275 to 360 kg/m³ of cement (accompanied by a reduction in w/c from 0.62 to 0.46) is associated with a strength gain of approximately 33%

Figure 1. Compressive strength as a function of cement content (mean values).

(32.8 → 43.5 MPa), while the further change from 360 to 400 kg/m³ of cement (w/c held at approximately 0.46) is associated with an additional 16% gain (43.5 → 50.7 MPa). This behaviour is consistent with the greater quantity of calcium silicate hydrates (C-S-H) and calcium aluminate hydrates (C-A-H) formed during hydration, which constitute the binding matrix responsible for mechanical resistance [14], although part of the gain between Mixes 1 and 2 may also reflect the concurrent reduction in w/c ratio.

The diminishing marginal return observed at higher dosages (275→360 versus 360→400 kg/m³) is consistent with findings in the literature [4], indicating that excess unreacted clinker in highly dosed concretes contributes little to ultimate strength. Across all three grades, the measured mean compressive strengths comfortably exceed the target strength values prescribed by NF EN 206-1 for each class, a favourable indication of the Dreux-Gorisse formulations’ performance; this observation, however, is based on mean values only and does not by itself constitute the formal conformity assessment defined by the standard, which additionally requires data on the number of batches tested and the associated variability.

To examine the association between the w/c ratio and compressive strength, Mixes 1 (w/c = 0.62), 3 (w/c = 0.44), and 4 (w/c = 0.47) were compared. These mixes also differ in cement content (275, 380 and 380 kg/m³), cement type (CEM III/B for Mixes 1 and 3, CEM I for Mix 4), aggregate proportions, and admixture dosage, so the comparison should be read as an association between mix formulations rather than an isolated w/c effect. Figure 2 summarises the results.

A clear inverse association is found between the w/c ratio and compressive strength at both test ages. Reducing the w/c ratio from 0.62 (Mix 1) to 0.44 (Mix 3) is accompanied by an increase in 28-day compressive strength from 32.8 to 49.3 MPa (a 50% gain); however, Mixes 1 and 3 also differ in cement content (275 vs 380 kg/m³), so this gain cannot be attributed to the w/c ratio alone. Comparing

Figure 2. Effect of w/c ratio on compressive strength.

Mix 3 (w/c = 0.44) to Mix 4 (w/c = 0.47) shows only a marginal variation in 28-day strength (49.3 vs 50.5 MPa); differences in cement type, aggregate proportions, and superplasticiser dosage between Mixes 3 and 4 are additional confounding factors, so this comparison likewise does not isolate the effect of the w/c ratio and must be interpreted with care.

These results are consistent with the Féret-Abrams law relating strength to the C/W ratio [5]. Further evidence from ultrasonic pulse velocity measurements [6] shows that concretes with lower w/c ratios exhibit lower initial porosity and a less distinct pore-compaction stage during loading, resulting in superior mechanical performance. Reducing the w/c ratio also diminishes the volume of capillary pores that remain unfilled after hydration, thereby improving durability against the chloride and sulfate ingress characteristic of the marine coastal environment of the Abidjan site [7].

The association between cement type and compressive strength was examined by comparing Mixes 3 (CHF-CEM III/B 42.5 L-LH/SR) and 4 (CPA-CEM I/52.5N), which share an identical cement dosage of 380 kg/m³ but differ in w/c ratio (0.44 vs 0.47), aggregate proportions, and superplasticiser dosage (6.84 vs 8.36 L); consequently, the strength differences reported below cannot be attributed solely to cement type. Figure 3 presents the results.

At 7 days, CEM I concrete (Mix 4) develops a compressive strength of 38.8 MPa compared with 32.0 MPa for CEM III/B concrete (Mix 3), a differential of approximately 21%; as noted above, this differential reflects an association between two mix formulations that also differ in w/c ratio and superplasticiser dosage, not a controlled cement-type effect. Part of this early-strength difference is nonetheless consistent with the faster hydration kinetics of high-C₃S Portland cement clinker relative to the pozzolanic reaction of ground granulated blast-furnace slag,

Figure 3. Effect of cement type on compressive strength at equal dosage (380 kg/m³).

which requires the presence of portlandite (Ca(OH)₂) released by clinker hydration and proceeds more slowly, particularly at low curing temperatures [11].

By 28 days, however, the strengths converge: 50.5 MPa for CEM I versus 49.3 MPa for CEM III/B, a difference of less than 3%. This pattern aligns closely with findings reported in the literature [7], showing that, at equal dosage and with the same aggregate type, CEM III concrete achieves 28-day compressive strengths comparable to CEM I while providing substantially lower gas permeability and initial water absorption—properties of particular importance for durability in chloride-rich coastal environments. In the context of the Abidjan Bridge, the use of CEM III/B for submerged and buried elements (foundations, piles) where rapid early-strength gain is not required is therefore well justified from a durability standpoint, while CEM I is appropriately deployed for the box-girder elements requiring fast formwork removal.

All five concrete mixes achieved slump values in the range 160 - 210 mm, conforming to the target S4 consistency class. This confirms the effectiveness of the combined use of TEMPO-12 superplasticiser and PLASTIRETARD set retarder in maintaining workability over the transit duration from plant to site without increasing the w/c ratio. Optimal PCE superplasticiser dosing enables water reductions of 15% - 30% while preserving target slump, with a net positive effect on 28-day compressive strength [12]. The consistency results thus validate the admixture selection strategy adopted for the project.