Effect of Various Fly Ash and Ground Granulated Blast Furnace Slag Content on Concrete Properties: Experiments and Modelling

Concrete is known as the most globally used construction material, but it releases a huge amount of greenhouse gases due to cement production. Recently, Supplementary Cementitious Materials (SCMs) such as fly ash and Ground Granulated Blast Furnace Slag (GGBFS) have been widely used in concrete to reduce the cement content. However, SCMs can alter the mechanical properties and time-dependent behaviors of concrete and the early age mechanical properties of concrete significantly affect the concrete cracking in the engineering field. Therefore, evaluation of the development of the mechanical properties of SCMs-based concrete is vital. In this paper, the time development of mechanical properties of concrete mixes with various fly ash and GGBFS was experimentally investigated. Four different cement replacement levels including 0%, 20%, 30%, and 40% by fly ash and GGBFS as well as ternary binders were considered. Compressive strength, splitting tensile strength, flexural strength, and elastic modulus of concrete were measured until 28 days. Three additional concrete mixes with ternary binders were also cast to investigate the early-age autogenous shrinkage development until 28 days. In addition, prediction models in existing standards were used and compared to experimental results. The comparison results showed that the prediction models overestimated the compressive strength but underestimated the splitting tensile strength development and autogenous shrinkage. As a result, a model capturing the effect of fly ash and GGBFS on the development of compressive and splitting tensile strength is proposed to improve the prediction accuracy for current standards and empirical models.


Introduction
Concrete is known as the most globally used construction material due to its excellent performance of compressive strength and workability. The tensile strength of concrete can also be improved using rebar and fibres [1]. Due to the rapid development of urbanisation, the consumption of concrete has increased significantly, since it is used as the primary construction material. In addition, Ordinary Portland Cement (OPC) is utilised as the primary cementitious material in concrete. From 1880 to 1990, it was reported that the global consumption of OPC has increased to 1.3 billion tons. Annual global cement consumption is expected to increase up to 5.2 billion tons by 2050 [2]. Each tone of clinker produced releases an average of 850 kg of carbon dioxide into the environment [3]. To reduce the negative impact of production cement on the environment, it is urgent to find alternative materials which can fully or partially replace cement. This can be achieved by using geopolymer concrete, partial replacement of cement with Supplementary Cementitious Materials (SCMs), and incorporation of nanomaterials for sustainability [4][5][6][7]. The mechanical properties of concrete at early age significantly affect the concrete cracking in the engineering field [30]. In existing standards, Eurocode 2 and FIB 2010 are commonly used prediction models to evaluate the early age mechanical properties of concrete [31,32]. However, the prediction model does not distinguish between conventional concrete and SCMs-based concrete. As a result, numerous researchers have developed various prediction models to investigate the time development of concrete mechanical properties. Singh et al. [33] created an artificial neural network (ANN) to predict the concrete compressive strength based on nonlinear regression analysis. Bhaskara et al. [34] adopted a reaction-kinetics-based strength development model to predict the strength development for OPC and fly ash concrete. Liu and Wang [35] proposed a model capturing the strength factor and maturity method to better reflect the strength development of fly ash concrete. However, owing to the complex procedures and more experimental parameters such as heat of hydration, these models are not suitable for implementation in existing codes. Therefore, it is important to examine the adaptability of existing models in predicting the time development of mechanical properties of concrete.
In this study, 12 concrete mixes including various fly ash and GGBFS content were cast to investigate the mechanical properties and autogenous shrinkage of concrete. Four different cement-replacement levels (0%, 20%, 30%, 40%) by fly ash and GGBFS were used to examine the influence of cement-replacement levels on the mechanical properties of concrete until 28 days. Three additional concrete mixes including ternary binders were also utilised to evaluate the early age concrete autogenous shrinkage. Prediction models in existing standards were utilised and compared to the time development of mechanical properties and autogenous shrinkage of concrete.

Materials and Mix Proportion
The raw materials in this study included Ordinary Portland cement (strength class 42.5) with a Blaine fineness of 375 m 2 /kg in accordance with Chinese Standard GB 175-2020 [36]. Two types of SCMs, namely fly ash and GGBFS, were used in experiments in accordance with Chinese Standard GB/T 1596-2017 [37] and GB/T 18046-2017 [38], respectively. Figure 1 presented the microstructure of fly ash and GGBFS using scanning electron microscopy (SEM) technology. Table 1 provided the chemical composition of cement, fly ash, and GG-BFS using X-ray fluorescence (XRF). Two kinds of coarse aggregate were used including the crushed stone with a maximum nominal size of 20 mm and 10 mm, respectively. The ratio of 20 mm coarse aggregate to 10 mm coarse aggregate is 1.5:1 by weight. Fine aggregate was manufactured sand with a maximum aggregate size of 2.6 mm. Polycarboxylate high-range water reducer was utilised to improve the workability of the mixtures in this study.
Therefore, it is crucial to evaluate the mechanical properties as well as autogenous shrink-age of concrete containing fly ash and GGBFS.
The mechanical properties of concrete at early age significantly affect the concrete cracking in the engineering field [30]. In existing standards, Eurocode 2 and FIB 2010 are commonly used prediction models to evaluate the early age mechanical properties of concrete [31,32]. However, the prediction model does not distinguish between conventional concrete and SCMs-based concrete. As a result, numerous researchers have developed various prediction models to investigate the time development of concrete mechanical properties. Singh et al. [33] created an artificial neural network (ANN) to predict the concrete compressive strength based on nonlinear regression analysis. Bhaskara et al. [34] adopted a reaction-kinetics-based strength development model to predict the strength development for OPC and fly ash concrete. Liu and Wang [35] proposed a model capturing the strength factor and maturity method to better reflect the strength development of fly ash concrete. However, owing to the complex procedures and more experimental parameters such as heat of hydration, these models are not suitable for implementation in existing codes. Therefore, it is important to examine the adaptability of existing models in predicting the time development of mechanical properties of concrete.
In this study, 12 concrete mixes including various fly ash and GGBFS content were cast to investigate the mechanical properties and autogenous shrinkage of concrete. Four different cement-replacement levels (0%, 20%, 30%, 40%) by fly ash and GGBFS were used to examine the influence of cement-replacement levels on the mechanical properties of concrete until 28 days. Three additional concrete mixes including ternary binders were also utilised to evaluate the early age concrete autogenous shrinkage. Prediction models in existing standards were utilised and compared to the time development of mechanical properties and autogenous shrinkage of concrete.

Materials and Mix Proportion
The raw materials in this study included Ordinary Portland cement (strength class 42.5) with a Blaine fineness of 375 m 2 /kg in accordance with Chinese Standard GB 175-2020 [36]. Two types of SCMs, namely fly ash and GGBFS, were used in experiments in accordance with Chinese Standard GB/T 1596-2017 [37] and GB/T 18046-2017 [38], respectively. Figure 1 presented the microstructure of fly ash and GGBFS using scanning electron microscopy (SEM) technology. Table 1 provided the chemical composition of cement, fly ash, and GGBFS using X-ray fluorescence (XRF). Two kinds of coarse aggregate were used including the crushed stone with a maximum nominal size of 20 mm and 10 mm, respectively. The ratio of 20 mm coarse aggregate to 10 mm coarse aggregate is 1.5:1 by weight. Fine aggregate was manufactured sand with a maximum aggregate size of 2.6 mm. Polycarboxylate high-range water reducer was utilised to improve the workability of the mixtures in this study.  The mix proportion of concrete in this paper is shown in Table 2. A total of 12 groups of concrete mixtures were designed. For the first nine groups, a fixed w/b ratio of 0.38 was used, and they were used to examine the influences of fly ash and GGBFS on the development of the mechanical properties of the concrete. For the last three groups, different w/b ratios and SCMs replacement levels were used to investigate the effect of ternary binders on the autogenous shrinkage of the concrete. Four different cement replacement levels, 0%, 20%, 30%, and 40%, by fly ash and GGBFS and ternary binders were adapted. According to different proportions of fly ash and GGBFS, the concrete mixtures were designated as 'C100', 'C80F20', 'C80S20', 'C60F20S20', etc. Mixture 'C100' was the reference mixture mixes without SCMs, while 'C80F20' and 'C80S20' were the concrete mixes with 20% fly ash and 20% GGBFS as a cement replacement by weight, respectively. 'C60F20S20' represented concrete mixes with ternary binders (20% fly ash and 20% GGBFS) were used to replace cement.

Mechanical Properties Investigation
The mechanical properties including compressive strength, splitting tensile strength, flexural strength, and elastic modulus of concrete for all mixes were conducted according to Chinese Standard GB/T 50081-2019 [39]. The compression and splitting tensile strength were tested using a hydraulic universal testing machine as shown in Figure 2. Cubic specimens were cast for determining the compressive strength and splitting tensile strength. The typical size of cubic specimen is 150 mm × 150 mm × 150 mm. Prismatic specimens with a dimension of 150 mm × 150 mm × 600 mm were cast for flexural strength, and 150 mm × 150 mm × 300 mm concrete prisms were cast for determining the elastic modulus, respectively. The specimens were cured in standard conditions at a temperature of 20 • C ± 1 • C and relative humidity of 95 ± 5%. The mechanical properties tests were carried out at the age of 2, 3, 7, and 28 days. In addition, the workability of concrete mm × 150 mm × 300 mm concrete prisms were cast for determining the elastic modulus, respectively. The specimens were cured in standard conditions at a temperature of 20 °C ± 1 °C and relative humidity of 95 ± 5%. The mechanical properties tests were carried out at the age of 2, 3, 7, and 28 days. In addition, the workability of concrete was evaluated by slump test and the slump of all concrete mixtures was controlled as 180 ± 20 mm.

Figure 2.
Concrete specimens and test apparatus.

Autogenous Shrinkage Measurement
The size of autogenous shrinkage specimens is 100 mm × 100 mm × 400 mm. The specimens were demoulded at 24 h after casting and stored in an environmentally controlled room at a temperature of 20 °C ± 2 °C and RH of 60 ± 5%. As shown in Figure 3, the specimens were immediately wrapped with the plastic film after demoulding to avoid moisture loss. Thermocouples were embedded in the specimens to monitor the change of the internal temperature. A shrinkage dial gauge was utilised to measure the autogenous shrinkage of concrete. The readings were recorded at 1, 3, 7, 14, and 28 days after the benchmark reading.

Autogenous Shrinkage Measurement
The size of autogenous shrinkage specimens is 100 mm × 100 mm × 400 mm. The specimens were demoulded at 24 h after casting and stored in an environmentally controlled room at a temperature of 20 • C ± 2 • C and RH of 60 ± 5%. As shown in Figure 3, the specimens were immediately wrapped with the plastic film after demoulding to avoid moisture loss. Thermocouples were embedded in the specimens to monitor the change of the internal temperature. A shrinkage dial gauge was utilised to measure the autogenous shrinkage of concrete. The readings were recorded at 1, 3, 7, 14, and 28 days after the benchmark reading. mm × 150 mm × 300 mm concrete prisms were cast for determining the elastic modulus, respectively. The specimens were cured in standard conditions at a temperature of 20 °C ± 1 °C and relative humidity of 95 ± 5%. The mechanical properties tests were carried out at the age of 2, 3, 7, and 28 days. In addition, the workability of concrete was evaluated by slump test and the slump of all concrete mixtures was controlled as 180 ± 20 mm.

Figure 2.
Concrete specimens and test apparatus.

Autogenous Shrinkage Measurement
The size of autogenous shrinkage specimens is 100 mm × 100 mm × 400 mm. The specimens were demoulded at 24 h after casting and stored in an environmentally controlled room at a temperature of 20 °C ± 2 °C and RH of 60 ± 5%. As shown in Figure 3, the specimens were immediately wrapped with the plastic film after demoulding to avoid moisture loss. Thermocouples were embedded in the specimens to monitor the change of the internal temperature. A shrinkage dial gauge was utilised to measure the autogenous shrinkage of concrete. The readings were recorded at 1, 3, 7, 14, and 28 days after the benchmark reading.     Figure 4 presents the development of compressive strength for all concrete mixes. An average of 3 specimens was used for each displayed point. As shown in Figure 4a, it was observed that for concrete mixes with fly ash, the compressive strength was about 30% lower than that of the control mix at an early age. It also observed that the higher fly ash content in concrete, the lower the compressive strength of concrete at an early age.

Mechanical Properties of Concrete
Similar results can be found in [40,41]. The reason can be attributed to the dilution effect on cement hydration affected by the addition of fly ash [40]. Moreover, the pozzolanic reaction between fly ash particles and calcium hydroxide also retarded the hydration degree [42]. The rate of compressive strength development of fly ash concrete was higher than that of OPC mix at a later age. Excluding the C80F20 mix, the compressive strength of fly ash concrete was only marginally lower than that of the reference mix at 28 days. This can be attributed to the pozzolanic reaction leading to an ever-increasing reaction degree. Fly ash particles can also serve as nucleation sites to accelerate the hydration of cement, leading to an insignificant difference from that of the control mix [43,44]. As shown in Figure 4b, it was observed that the compressive strength of GGBFS concrete was lower than that of the OPC mix at 2 and 3 days (early age). The higher the GGBFS content in concrete, the lower the compressive strength. At 3 days, the concrete mixes with the 20%, 30%, and 40% GGBFS were observed to be 74%, 71%, and 60%, respectively, of that of plain concrete. Similar results were reported by Shariq et al. [45]. However, at the age of 28 days, the compressive strength of GGBFS concrete was 11%, 17%, and 8% higher, respectively, than that of control concrete. This was attributed to the late pozzolanic reactivity of GGBFS contributes to the rate of strength gain at later ages [46]. For the ternary binder group, the concrete mixes with fly ash and GGBFS were lower than reference concrete at all different ages. Such a behaviour was also noted by Wang and Chen [47]. In addition, the measured compressive strength for autogenous shrinkage specimens including C65F35, C60F25S15, and C65F15S20 were 30.1, 40.3, and 45.5 MPa at 28 days. observed that for concrete mixes with fly ash, the compressive strength was about 30% lower than that of the control mix at an early age. It also observed that the higher fly ash content in concrete, the lower the compressive strength of concrete at an early age. Similar results can be found in [40,41]. The reason can be attributed to the dilution effect on cement hydration affected by the addition of fly ash [40]. Moreover, the pozzolanic reaction between fly ash particles and calcium hydroxide also retarded the hydration degree [42]. The rate of compressive strength development of fly ash concrete was higher than that of OPC mix at a later age. Excluding the C80F20 mix, the compressive strength of fly ash concrete was only marginally lower than that of the reference mix at 28 days. This can be attributed to the pozzolanic reaction leading to an ever-increasing reaction degree. Fly ash particles can also serve as nucleation sites to accelerate the hydration of cement, leading to an insignificant difference from that of the control mix [43,44]. As shown in Figure 4b, it was observed that the compressive strength of GGBFS concrete was lower than that of the OPC mix at 2 and 3 days (early age). The higher the GGBFS content in concrete, the lower the compressive strength. At 3 days, the concrete mixes with the 20%, 30%, and 40% GGBFS were observed to be 74%, 71%, and 60%, respectively, of that of plain concrete. Similar results were reported by Shariq et al. [45]. However, at the age of 28 days, the compressive strength of GGBFS concrete was 11%, 17%, and 8% higher, respectively, than that of control concrete. This was attributed to the late pozzolanic reactivity of GGBFS contributes to the rate of strength gain at later ages [46]. For the ternary binder group, the concrete mixes with fly ash and GGBFS were lower than reference concrete at all different ages. Such a behaviour was also noted by Wang and Chen [47]. In addition, the measured compressive strength for autogenous shrinkage specimens including C65F35, C60F25S15, and C65F15S20 were 30.1, 40.3, and 45.5 MPa at 28 days.  28 days. Such behaviour was also noted by Zhang et al. [48]. For the GGBFS group, the average value of splitting tensile strength of GGBFS concrete at 28 days was decreased by 15%, 18%, and 21% compared to that of OPC mix when the GGBFS content increased from 0% to 20%, 30%, 40%, respectively. Similar results were obtained by Shen et al. [49]. For concrete mixes with ternary binders, it was observed that the higher ternary binders content results in a lower splitting tensile strength. The reason was attributed to the water required for hydration being insufficient, leading to the reduction of the hydration products [50].  Figure 5 showed the development of the splitting tensile strength of concrete. For concrete mixes with fly ash, the splitting tensile strength at the age of 28 days after casting for C80F20, C70F30, and C60F40 were 5.35, 4.93, and 4.02 MPa, respectively. It was seen that the higher fly ash content, the lower the average value of splitting tensile strength at 28 days. Such behaviour was also noted by Zhang et al. [48]. For the GGBFS group, the average value of splitting tensile strength of GGBFS concrete at 28 days was decreased by 15%, 18%, and 21% compared to that of OPC mix when the GGBFS content increased from 0% to 20%, 30%, 40%, respectively. Similar results were obtained by Shen et al. [49]. For concrete mixes with ternary binders, it was observed that the higher ternary binders content results in a lower splitting tensile strength. The reason was attributed to the water required for hydration being insufficient, leading to the reduction of the hydration products [50]. The development of flexural strength of concrete was provided in Figure 6. It can be seen from Figure 6 that the introduction of fly ash decreased the flexural strength at 28 days. As the replacement of fly ash increased, the difference between fly ash concrete and reference concrete in flexural strength was larger at 28 days. Results on flexural strength The development of flexural strength of concrete was provided in Figure 6. It can be seen from Figure 6 that the introduction of fly ash decreased the flexural strength at 28 days. As the replacement of fly ash increased, the difference between fly ash concrete and reference concrete in flexural strength was larger at 28 days. Results on flexural strength of fly ash concrete were in accordance with the results in [51]. A similar trend was observed for GGBFS and ternary binder group. It was reported that GGBFS improved the early age flexural strength, but GGBFS also reduced the calcium ion concentration between cement and aggregate, leading to a decrease in flexural resistance at a later age [52].
The 7-and 28-day elastic modulus of concrete was provided in Table 3. The elastic modulus was calculated from the compressive test. It was observed that fly ash and GGBFS can improve the 7-day elastic modulus by about 3% to that of the control mix. While the ternary binders lead to a 19% decrease in elastic modulus at 7 days. For 28 days results, it can be seen that the addition of fly ash, GGBFS, and ternary binders results in a marginally reduction of 5%, 10%, and 10% lower than that of the reference mix. Similar test results can be found in [53,54]. of fly ash concrete were in accordance with the results in [51]. A similar trend was observed for GGBFS and ternary binder group. It was reported that GGBFS improved the early age flexural strength, but GGBFS also reduced the calcium ion concentration between cement and aggregate, leading to a decrease in flexural resistance at a later age [52].
(a) (b) (c) The 7-and 28-day elastic modulus of concrete was provided in Table 3. The elastic modulus was calculated from the compressive test. It was observed that fly ash and GGBFS can improve the 7-day elastic modulus by about 3% to that of the control mix. While the ternary binders lead to a 19% decrease in elastic modulus at 7 days. For 28 days results, it can be seen that the addition of fly ash, GGBFS, and ternary binders results in a marginally reduction of 5%, 10%, and 10% lower than that of the reference mix. Similar test results can be found in [53,54].    Figure 7 showed the autogenous shrinkage of concrete mixes with ternary binders. As mentioned in the previous section, three different concrete mixes with different w/b ratios and cement replacement are cast to investigate the autogenous shrinkage. It should be noted that the autogenous shrinkage is related to the compressive strength of concrete. As a result, it is difficult to compare the autogenous shrinkage of concrete with different SCMs replacement levels and compressive strength. However, it was reported that the autogenous shrinkage of fly ash and GGBFS concrete was higher compared to that of OPC concrete due to pore structure refinement [55,56]. This point was also supported by this study. It was clear that despite the compressive strength of C60F25S15 being higher than C65F35, the autogenous shrinkage of C60F25S15 was lower than that of C65F35, indicating the incorporation of fly ash and GGBFS increased the autogenous shrinkage of concrete which compensated the effect of compressive strength.

Autogenous Shrinkage of Concrete
SCMs replacement levels and compressive strength. However, it was reported that the autogenous shrinkage of fly ash and GGBFS concrete was higher compared to that of OPC concrete due to pore structure refinement [55,56]. This point was also supported by this study. It was clear that despite the compressive strength of C60F25S15 being higher than C65F35, the autogenous shrinkage of C60F25S15 was lower than that of C65F35, indicating the incorporation of fly ash and GGBFS increased the autogenous shrinkage of concrete which compensated the effect of compressive strength.

Mechanical Properties of Concrete
The development of the compressive strength of concrete in Eurocode 2 and FIB 2010 models [31,32] is expressed as a function of time and compressive strength at 28 days, as shown in Equation (1): where depends on the cement type and is equal to 0.2, 0.25, and 0.38 for cement strength classes of 32.5N, 42.5N, and 52.5N, respectively. Figure 8 showed a comparison between predicted and measured development of compressive strength using = 0.25 for all concrete mixes with SCMs. It was obvious that the prediction model cannot well predict the development of compressive strength for all concrete mixes. The estimated development of compressive strength was lower than that of experimentally measured values. As a result, the value of should be reconsidered when the fly ash and GGBFS were presented in concrete.

Mechanical Properties of Concrete
The development of the compressive strength of concrete in Eurocode 2 and FIB 2010 models [31,32] is expressed as a function of time and compressive strength at 28 days, as shown in Equation (1): where s 1 depends on the cement type and is equal to 0.2, 0.25, and 0.38 for cement strength classes of 32.5N, 42.5N, and 52.5N, respectively. Figure 8 showed a comparison between predicted and measured development of compressive strength using s 1 = 0.25 for all concrete mixes with SCMs. It was obvious that the prediction model cannot well predict the development of compressive strength for all concrete mixes. The estimated development of compressive strength was lower than that of experimentally measured values. As a result, the value of s 1 should be reconsidered when the fly ash and GGBFS were presented in concrete.
The development of the splitting tensile strength of concrete in Eurocode 2 and FIB 2010 models [31,32] is expressed as a function of time and splitting tensile strength at 28 days, as shown in Equation (2): where s 2 depends on the cement type and is equal to 0.2, 0.25, and 0.38 for cement strength classes of 32.5N, 42.5N, and 52.5N, respectively. Figure 9 displayed the prediction of the splitting tensile strength development based on s 2 = 0.25 for all concrete mixes with SCMs. It can be seen that the prediction of timedependent splitting tensile strength was different from that of the prediction of compressive strength. The predicted values were higher than that of the experimental values. As such, the value of s 2 in the prediction model also needs to be reconsidered.
The development of the elastic modulus of concrete in Eurocode 2 and FIB 2010 models [31,32] is expressed as a function of time and elastic modulus at 28 days, as shown in Equation (3): where s 3 depends on the cement type and is equal to 0.2, 0.25, and 0.38 for cement strength classes of 32.5N, 42.5N, and 52.5N, respectively. Figure 10 presented the development of experimental values of elastic modulus and the values from prediction models based on s 3 = 0.25 for concrete mixes with SCMs. It can be seen that the prediction values were close to experimental values for fly ash and GGBFS mixes because the early age elastic modulus results are unknown. However, the prediction results were overestimated for the ternary binder group. Therefore, the value of s 3 is recalculated to fit the experimental results. The development of the splitting tensile strength of concrete in Eurocode 2 and FIB 2010 models [31,32] is expressed as a function of time and splitting tensile strength at 28 days, as shown in Equation (2): where depends on the cement type and is equal to 0.2, 0.25, and 0.38 for cement strength classes of 32.5N, 42.5N, and 52.5N, respectively. Figure 9 displayed the prediction of the splitting tensile strength development based on = 0.25 for all concrete mixes with SCMs. It can be seen that the prediction of timedependent splitting tensile strength was different from that of the prediction of compressive strength. The predicted values were higher than that of the experimental values. As such, the value of in the prediction model also needs to be reconsidered. The incorporation of SCMs can influence prediction results as per [31,32,57]. As such, Table 4 displayed the values of fitting parameters based on the experimental results using the Levenberg-Marquardt least square regression analysis method. It can be seen that the fitting parameters values of s 1 , s 2 , and s 3 for different groups were different. The average values of fitting parameters s 1 , s 2 , and s 3 were 0.28, 0.48, 0.23 and 0.34, 0.39, 0.15 and 0.35, 0.40, 0.45 for fly ash group, GGBFS group, and ternary binder group, respectively. It also obvious that these values were different from the suggested values in standards [31,32]. Therefore, it is recommended to consider the parameter capturing the effect of fly ash and GGBFS on concrete mechanical properties development. The development of the elastic modulus of concrete in Eurocode 2 and FIB 2010 models [31,32] is expressed as a function of time and elastic modulus at 28 days, as shown in Equation (3): where depends on the cement type and is equal to 0.2, 0.25, and 0.38 for cement strength classes of 32.5N, 42.5N, and 52.5N, respectively. Figure 10 presented the development of experimental values of elastic modulus and the values from prediction models based on = 0.25 for concrete mixes with SCMs. It can be seen that the prediction values were close to experimental values for fly ash and GGBFS mixes because the early age elastic modulus results are unknown. However, the prediction results were overestimated for the ternary binder group. Therefore, the value of is recalculated to fit the experimental results. Splitting tensile strength (MPa)

Autogenous Shrinkage of Concrete
Regarding autogenous shrinkage of concrete mixes with SCMs, three empirical models from standards are assessed in this study [31,32,58]. The reason to select these prediction models is due to the simplicity and wide-range application in the industry. The prediction in these models is a strength-based approach rather than focusing on the mix design, which is widely used in design engineering field.
The expression of autogenous shrinkage of concrete in Eurocode 2 [31] is shown as follows: where ε * au is the final autogenous shrinkage which is calculated as Equation (5): where f c,28 is the characteristic compressive strength of concrete. The incorporation of SCMs can influence prediction results as per [31,32,57]. As such, Table 4 displayed the values of fitting parameters based on the experimental results using the Levenberg-Marquardt least square regression analysis method. It can be seen that the fitting parameters values of , , and for different groups were different. The average values of fitting parameters , , and were 0.28, 0.48, 0.23 and 0.34, 0.39, 0.15 and 0.35, 0.40, 0.45 for fly ash group, GGBFS group, and ternary binder group, respectively. It also obvious that these values were different from the suggested values in standards [31,32]. Therefore, it is recommended to consider the parameter capturing the effect of fly ash and GGBFS on concrete mechanical properties development.  The autogenous shrinkage in FIB 2010 [32] is estimated by means of basic notional shrinkage coefficient and the time function, as shown in Equations (6) and (7): where α bs depends on the cement type and is equal to 800, 700 and 600 for cement strength class of 32.5N, 42.5N and 52.5N, respectively. In AS3600-2018 [58], the autogenous shrinkage strain is calculated as follows:

Proposed Model
As mentioned in the previous section, the effect of fly ash and GGBFS on concrete mechanical properties development should be considered in existing models. As shown in Table 4, a linear relationship can be observed for regression coefficients and . However, due to insufficient reliable data of , the modeling of was not presented in this study. Therefore, the model capturing the effect of fly ash and GGBFS was proposed for and as follows: where , and , are the recommended parameters for concrete without fly ash and GGBFS, and the recommended values in this study are 0.146 and 0.389 for , and , , respectively; and are the factor depending on the percentage of fly ash and GGBFS for 1, as shown in Equations (12) and (13); and are the factor depending on the percentage of fly ash and GGBFS for 2, as shown in Equations (14) and (15).

Proposed Model
As mentioned in the previous section, the effect of fly ash and GGBFS on concrete mechanical properties development should be considered in existing models. As shown in Table 4, a linear relationship can be observed for regression coefficients s 1 and s 2 . However, due to insufficient reliable data of s 3 , the modeling of s 3 was not presented in this study. Therefore, the model capturing the effect of fly ash and GGBFS was proposed for s 1 and s 2 as follows: where s 1,0 and s 2,0 are the recommended parameters for concrete without fly ash and GGBFS, and the recommended values in this study are 0.146 and 0.389 for s 1,0 and s 2,0 , respectively; m 1 and n 1 are the factor depending on the percentage of fly ash and GGBFS for s 1 , as shown in Equations (12) and (13); m 1 and n 2 are the factor depending on the percentage of fly ash and GGBFS for s 2 , as shown in Equations (14) and (15).
n 1 = 0.009 × GGBFS + 2.08 (13) and m 2 = −0.026 × FA + 2.03 (14) n 2 = −0.006 × GGBFS + 1.19 (15) where FA and GGBFS are the percentage of fly ash and GGBFS used in concrete. It should also be noted that the Equations (10) and (11) can be used not only for concrete with fly ash or GGBFS, but also extend to concrete with ternary binders. Figures 12 and 13 show the comparison of development of compressive strength and splitting tensile strength and the predicted results using the proposed model. It can be observed that the predicted values agreed well with the experimentally measured results. This indicates that the proposed model for calculating parameters s 1 and s 2 , which captured the effect of fly ash and GGBFS, could be used in the existing models.

Proposed Model Comparison with Experimental Results
The evaluation and error values have been widely interpreted in many studies [59][60][61]. The mean value and coefficient of variation (CoV) are typically used to indicate the error and robustness. As shown in Figure 14, the mean value and CoV of the ratio by the proposed model and experimental results are 1.00, 0.10, and 0.95, 0.17 for the s 1 and s 2 prediction models, respectively. As a result, it can be seen that the statical analysis confirms the proposed model has been successfully utilised in the existing models to calculate the time-dependent compressive strength and splitting tensile strength of concrete.
= −0.006 × + 1.19 (15) where and are the percentage of fly ash and GGBFS used in concrete. It should also be noted that the Equations (10) and (11) can be used not only for concrete with fly ash or GGBFS, but also extend to concrete with ternary binders. The evaluation and error values have been widely interpreted in many studies [59][60][61]. The mean value and coefficient of variation (CoV) are typically used to indicate the error and robustness. As shown in Figure 14, the mean value and CoV of the ratio by the proposed model and experimental results are 1.00, 0.10, and 0.95, 0.17 for the and prediction models, respectively. As a result, it can be seen that the statical analysis confirms the proposed model has been successfully utilised in the existing models to calculate the time-dependent compressive strength and splitting tensile strength of concrete.

Conclusions
The results presented in this work provide insight into the mechanical propertie autogenous shrinkage development of concrete mixes with various fly ash and G content. The observed trends and patterns of behaviour could be utilised to refle early-age behaviour of concrete. The main observations are as follows:

Conclusions
The results presented in this work provide insight into the mechanical properties and autogenous shrinkage development of concrete mixes with various fly ash and GGBFS content. The observed trends and patterns of behaviour could be utilised to reflect the early-age behaviour of concrete. The main observations are as follows: • Fly ash and GGBFS reduced the early age mechanical properties of concrete, but the compressive strength and splitting tensile strength of fly ash concrete were similar to OPC concrete at 28 days, while GGBFS increased the compressive strength of concrete at 28 days. Both fly ash and GGBFS decreased the flexural strength and elastic modulus of concrete. When ternary binders were incorporated, the concrete also exhibited low mechanical properties.

•
The model in existing standards overestimated the early age of compressive strength development but underestimated the splitting tensile strength of concrete, leading to an unreliable calculation for design purposes. As such, it is recommended to consider the effect of fly ash and GGBFS on concrete's mechanical properties development.

•
A model capturing the effect of fly ash and GGBFS on the development of compressive and splitting tensile strength was proposed. The statistical analysis of the comparison between predicted and measured compressive strength and splitting tensile strength showed a good agreement. The mean value and CoV are 1.00, 0.10, and 0.95, 0.17 for the s 1 and s 2 prediction models, respectively.

•
The autogenous shrinkage of concrete mixes with fly ash and GGBFS was different from that of OPC concrete. The model in existing standards failed in predicting the early-age autogenous shrinkage development of fly ash and GGBFS concrete. As a result, a broader database of concrete containing fly ash and GGBFS values is recommended to improve and establish optimal parameters in existing models.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: Data sharing is not applicable to this article.