Effect of Consistency Limit on the Strength of Cement-Solidified Dredged Sludge: Modelling and Micro-Mechanism

Abstract

The unconfined compressive strength (UCS) of sludge with different consistency limits solidified by cement was investigated. The results showed that under the condition of constant initial water content, a higher liquid index of soil resulted in higher UCS. A novel strength-evaluation model based on the ratio of the liquid index to the cement content was developed, and the prediction deviation of the model was within 30%. The influence mechanism of the consistency limit of sludge on the cement solidification was revealed by scanning electron microscopy, mercury intrusion porosimetry, X-ray diffractometer and thermogravimetric analysis. For the cement-solidified dredged sludge (CDS) with a lower liquid index, a large amount of hydrate was interlaced with each other and wrapped soil particles, promoting the formation of a dense structure. For the CDS with a higher liquid index, hydrates such as C-S-H and ettringite challenged each other to play the role of “cementing particles” and “filling pores”, resulting in the formation of the porous structure. The mineralogical analyses confirmed that more C-S-H gels and ettringites were generated in the CDS with a lower liquid index, but less calcite was formed due to its denser structure. In engineering applications, reducing the liquid index by adjusting the consistency limit can improve the strength performance of CDS.

1. Introduction

Over the last few decades, a large amount of sludge has been generated in water environment regulation, channel dredging and port construction [1,2,3,4,5,6]. This raw dredged sludge (RDS) is characterized by high water content, high compressibility and low shear strength, which are generally unsuitable for geotechnical utilization [7,8,9]. Cement solidification is a widely used soft sludge reinforcement technology to achieve land reclamation, embankment reinforcement and roadbed filling [10,11]. When the cement is mixed into the RDS, a series of physicochemical interactions occur between the pore water, cement and soil particles, as shown in

Table 1. The chemical reaction of cement stabilized dredged sediment [5].

Reaction	Chemical Formulas
The direct hydration reaction between cement and pore water in the sludge	2(3CaO·SiO2) + 6H2O→3CaO·SiO2·3H2O + 3Ca(OH)2 C-S-H
	2(2CaO·SiO2) + 4H2O→3CaO·SiO2·3H2O + Ca(OH)2 C-S-H
	3CaO·Al2O3 + 6H2O + Ca(OH)2 →3CaO·Al2O3·6H2O C-A-H
	4CaO·Al2O3·Fe2O3 + 7H2O→3CaO·Al2O3·6H2O + CaO·Fe2O3·H2O C-F-H
	4CaO·Al2O3·13H2O + 3(CaSO4·2H2O) + 14H2O→3CaO·Al2O3·3CaSO4·32H2O + Ca(OH)2 AFt
The secondary pozzolanic reaction between calcium hydroxide (cement hydration product) and active silicon aluminum in the clay minerals	OH− + SiO2 + Ca2+ + H2O→C-S-H
	OH− + Al2O3 + Ca2+ + H2O→C-A-H

. The main hydrates such as calcium silicate hydrate (C-S-H), calcium carbonate (CaCO₃) and ettringite (AFt) are formed. Then, the geotechnical properties of cement-solidified dredged sludge (CDS) were improved by reducing pore water, cementing soil particles and filling micropores [12,13,14].

Being able to evaluate and predict the strength property of CDS is crucial for reliable engineering design and construction control. Many studies have investigated the key factors influencing the strength properties and established the corresponding empirical evaluation model. Horpibulsuk et al. [15,16] proposed an empirical model with the water/cement ratio as the parameter, which confirmed that cement-stabilized clay with the same water/cement ratio has actually the same strength characteristics. Lorenzo and Bergado [17,18] demonstrated that the ratio of void ratio to cement content yielded a unique relationship of unconfined compression strength. By analyzing many test data, Zhang et al. [19] proposed that when evaluating the unconfined compression strength, the water content and cement content should be taken as independent variables. Based on the water/binder ratio hypothesis, Ma et al. [20,21] proposed a mathematical model for evaluating the strength of cement-based composites in stabilized soil, and the deviation was within 10%. Correia et al. [22] established a strength model to assess the unconfined compressive strength based on normalization by the liquid index. Yao et al. [23] proposed a power function model combining the effects of curing age, cement content and water content on the strength of cement-stabilized clay. Bi and Chian [24] introduced a cumulative distribution function to characterize the complete three-phase strength development.

It can be concluded that the total water content (C_W) and cement content (C_P) are the dominant factors affecting the strength of CDS. However, according to the traditional theory of soil mechanics, the pore water in sludge can be divided into free water and adsorbed water bound by the surface charge of clay particles [25], and different from the limited reaction activity of adsorbed water, the free water is the main part involved in cement hydration [26]. Furthermore, the difference between total pore water content and liquid limit (L_L) is the free water content. The L_L can be regarded as the boundary between adsorbed water and free water content [27]. Therefore, the consistency limit is a crucial index to evaluate the strength properties of CDS.

To investigate the influence mechanism of the consistency limit of sludge on the cement solidification, a series of laboratory samples were prepared for different cement contents (from 15 to 35%, step 5%) and liquid index (which range from the natural state, 2.69, to 1.58). Based on the results of the unconfined compression tests, the effects of C_P and I_L on the strength development were quantitatively evaluated, and the corresponding mathematical models were set up. In addition, the porous structure and mineral phase of CDS were analyzed by microscopic testing methods, confirming that the macro mechanical properties were highly consistent with its microstructure characteristics.

2. Materials and Methods

2.1. Raw Materials

The RDS used in the laboratory experiments was taken from the bottom sludge of a river in Suzhou city, Anhui Province, China. The basic physicochemical properties of RDS were determined according to the relevant standards [28,29,30,31], and the results are shown in

Table 2. Basic properties of RDS.

Property	Value	Standard
Specific gravity	2.71	ASTM D854-10 [28]
Liquid limit (LL), %	49.1	ASTM D4318-10 [29]
Plastic limit (PL), %	24.9
Plastic index (IP), %	24.2
Clay fraction (d < 0.005 mm), %	21.0	ASTM D422-63 [30]
Silt fraction (0.005 mm < d < 0.075 mm), %	64.0
Sand fraction (d > 0.075 mm), %	15.0
Organic content, %	2.5	ASTM D2974-14 [31]

. According to the Unified Soil Classification System (USCS), the raw dredged sludge (RDS) was classified as a “Low liquid limit clay” (CH), as shown in Figure 1. The sampled RDS was firstly naturally air-dried, then thoroughly dried at 60 °C, and finally mechanically broken into powder for reserve. The binder for solidification of RDS in this study is ordinary Portland cement (PC, ASTM Type I). The sodium bentonite used in this experiment is a fine clay mainly composed of montmorillonite, which has the characteristics of strong water absorption and high specific surface area (54.5 m²/g). The liquid limit L_L and plastic limit P_L of bentonite are 237.5% and 57.9%, respectively. The chemical compositions (determined by X-ray fluorescence test) of RDS, PC and bentonite are listed in

Table 3. Major chemical compositions of RDS, PC and bentonite.

Raw Materials	Chemical Composition, %
	SiO2	Al2O3	CaO	Fe2O3	MgO	Na2O	Loss
RDS	58.1	18.8	4.6	5.8	2.7	2.1	6.4
PC	22.3	4.2	64.8	2.3	2.5	1.2	1.7
bentonite	69.5	15.5	3.0	2.0	3.0	1.0	2.5

. To investigate the influence of the consistency limit on the cement solidification, according to the experience of the pre-experiment, based on the dry weight of RDS, four types of sludge with varying consistency limits were prepared by adding 15%, 25%, 35% and 45% bentonite into RDS, and the corresponding mixtures were denoted as RBI, RBII, RBIII and RBIV, respectively. All five types of sludge were mixed with a certain amount of water to an initial water content of 90%, then sealed for a week to ensure the dry soil particles were fully “hydrated”. The consistency limits of five types of sludge and their corresponding liquid indices are listed in

Table 4. Mix design and testing program.

Dredged Sludge Type	Initial Water Content, CW (%)	Bentonite Content, CB (%)	Liquid Limit, LL (%)	Plastic Limit, PL (%)	Plasticity Index, IP	liquid Index, IL	Curing Age, (Days)	PC Content, CP (%)	Symbol
RDS	90	0%	49.1	24.9	24.2	2.69	7, 28	15	RDS-15
								20	RDS-20
								25	RDS-25
								30	RDS-30
								35	RDS-35
RBI	90	15%	57.8	32.6	25.2	2.28	7, 28	15	RBI-15
								20	RBI-20
								25	RBI-25
								30	RBI-30
								35	RBI-35
RBII	90	25%	63.3	35	28.5	1.94	7, 28	15	RBII-15
								20	RBII-20
								25	RBII-25
								30	RBII-30
								35	RBII-35
RBIII	90	35%	67.1	37	30.1	1.76	7, 28	15	RBIII-15
								20	RBIII-20
								25	RBIII-25
								30	RBIII-30
								35	RBIII-35
RBIV	90	45%	71.0	38.3	32.7	1.58	7, 28	15	RBIV-15
								20	RBIV-20
								25	RBIV-25
								30	RBIV-30
								35	RBIV-35

.

2.2. Mix Design and Samples Preparation

The mix design is shown in Table 4. The contents of bentonite (C_B), water (C_W) and PC (C_P) were defined as their respective mass ratios to dry DS. Although the initial water content of five types of sludge was fixed at 90%, the consistency limits were different, and the liquid index (I_L) changed accordingly. The prepared sludge was homogeneously mixed with a predetermined amount of cement for 5 min. Then, the solidified sludge was put into cylindrical PVC split molds with a diameter of 50 mm and a height of 50 mm. The prepared specimens were then transferred into a standard curing room (20 ± 2 °C, relative humidity ≥ 95%) for 1 day until demolding. Finally, the specimens were sealed with a plastic membrane and cured for 7 and 28 days before subsequent analyses. It is noted that although the strength development of CDS is a time-dependent process, the time factor is not the focus of this study. Four parallel samples were prepared for each solidified sludge mixture, three of which were subjected to an unconfined compression test. The remaining parallel specimen was used as a backup.

2.3. Testing Methods

According to the standard ASTM D4219-08 [32], the UCS of CDS was measured by a hydraulic servo testing machine. The interface micromorphology of CDS was observed by using scanning electron microscopy (SEM). The pore structure was characterized by performing mercury intrusion porosimetry (MIP). The crystalline-phase mineralogy of broken samples was examined by an X-ray diffractometer (XRD) over the 2θ range of 10–80° at a rate of 5°/min. The weight change between 30 and 1000 ℃ was assessed by performing thermogravimetric analysis (TGA) at the heating rate of 10 ℃/min with argon as stripping gas. The crushed sample pieces after the UCS test were immediately dried by the vacuum freeze-drying method and ground into powder for XRD and TGA testing. However, the specimen used for MIP and SEM analysis was taken from the undisturbed parallel sample to exclude the influence of microscopic damage caused by external factors.

3. Results and Discussion

3.1. Curing Strength Development of Different Types of Sludge with Cement Content

Figure 2 presents the UCS change in CDS with cement content for different types of sludge. The UCS increased with increasing C_P for a given kind of sludge and curing age, and the strength of CDS at 28 days of curing was greater than that of 7 days of curing. This result is consistent with the findings reported by [11,23,33,34], which was attributed to the formation of more cement hydration products. It is also evident from Figure 2 that the sludge type had a significant influence on the strength of CDS. With the same cement content (C_P) and water content (C_W), the strength of CDS increased significantly with the increase in the plasticity index (I_P) at various curing ages. For example, when the C_P was 35%, the 28-day curing strength of the cement-solidified raw dredged sludge (RDS-35, I_P = 24.2) was 760 kPa, while that of RBIV-35 (I_P = 32.7) was 2077 kPa, increasing by 173%.

By arranging the data in the form of Figure 3, it is clear that there is a linear relationship between C_P and the UCS of CDS. The increase rate of UCS with C_P was more remarkable for the sludge with a higher consistency limit: for example, the fitting line slope for sludge RDS, RBI, RBII, RBIII and RBIV in Figure 3b was 24.8, 35.7, 47.5, 60.4 and 74.9, respectively. This denoted that under the same C_W, the effectiveness of cement was more significant in the sludge with a higher consistency limit. The main reason is that under the same C_W, the free water content and soil particle spacing in the sludge decreased with the increase in the consistency limit, resulting in the formation of denser microstructure and higher strength of CDS with the same cement content. This indicated that the pore water state and fluidity of sludge were the key factors affecting the solidification effect of the cement. Therefore, it is necessary to further explore the quantitative relationship between the strength and water characteristics. It is worth noting that many studies showed that there are typical four-stage evolution characteristics between the UCS of stabilized soft soil and the cement content [1]. In this study, when the cement content varied from 15% to 35%, the UCS of different types of CDS all changed within the active zone II (i.e., clay–cement interaction zone), thus increasing the cement content can further strengthen the inter-linkages amongst clay particle clusters. Therefore, considering technical and economic factors, the cement solidification of sludge is preferentially designed in Zone II, where the increase rate of UCS with C_P is highest.

3.2. Effect of Liquid Index on the Strength Development of CDS

As mentioned above, under the same C_W, the consistency limit of sludge is an important factor affecting its curing strength. Moreover, since the liquid index I_L can comprehensively reflect the relative relationship between the C_W and the consistency limit of sludge, it can be used as a critical parameter to evaluate the strength of CDS. Correia et al. [22,35] proposed that the UCS of CDS could be well normalized by the liquid index of soil, but in their studies, the change in the liquid index was realized by adjusting the initial water content, and the influence of soil properties on the liquid index was not considered. In this study, the effect of I_L on the strength was investigated by changing C_B from 15% to 45%, then the I_L decreased from 2.6 to 1.5, as can be seen in Table 4. The change in UCS with I_L was presented in Figure 4, and the UCS showed a similar evolution trend with I_L for different C_P. Moreover, a power function of y = a∗x^b can be used to characterize the relationship between I_L and UCS development. Although the C_P varied from 10% to 30%, the determination coefficient (R²) indicated that the fitting results are of good quality. The liquid index, rather than the water content, was used to characterize the strength of CDS, because the former can reflect the containing water state and consistency of sludge more accurately by integrating the initial water content, liquid limit, and plastic limit of sludge. Therefore, the I_L is potentially more powerful in developing a generalized strength equation reasonably applicable to a wider range of CDS. The purpose of designing two curing ages was mainly to prove that the model established in this study is suitable for both short-term (7 days) and long-term (28 days).

3.3. Strength Evaluation Model Based on the Ratio of Liquid Index to Cement Content

The above proved that both cement content C_P and liquid index I_L play a significant role in the strength development of CDS. It is necessary to develop a model combing together both the two independent governing parameters (i.e., C_P and I_L) to evaluate the strength of CDS with high water content (I_L ≥ 1). In the last few decades, the water/cement ratio (C_W/C_P) has been the prime parameter to characterize the strength of CDS. Moreover, many studies showed that C_W/C_P is the only factor controlling geotechnical characteristics of CDS: that is, if this ratio remained the same, the strength would be identical even under different combinations of C_W and C_P [3]. However, this evolution law is not completely consistent with the experimental results of this study. For example, the values of C_W/C_P of RDS-15, RBI-15, RBII-15, RBIII-15 and RBIV-15 were all 6.0 (C_W = 90%, C_P = 15%), but the corresponding 28-day UCS was 275.2 kPa, 310.8 kPa, 372.0 kPa, 470.5 kPa and 604.7 kPa, respectively. This indicated that the strength also depended largely on the consistency of the sludge. Therefore, in this section, the parameter I_L/C_P, which can simultaneously reflect the influence of initial water content, cement content, liquid limit and plastic limit, was chosen for characterizing the effect of “pore water” and “curing agent” on the strength development of CDS. The correlation of the unconfined compressive strength q_u (kPa) with parameter I_L/C_P is shown in Figure 5. It can be observed that the q_u of CDS can be well normalized by the parameter I_L/C_P when the curing age is determined. An empirical equation revealing the relationship between q_u and I_L/C_P can be drawn as follows:

$$q_u=\frac{A}{{(I_L/C_P)}^B}$$

(1)

where A and B are fitting parameters whose values are mainly affected by the curing age. The mathematical models for different curing times can be obtained from the two-dimensional q_u-I_L/C_P plot with a very high determination coefficient (R² > 0.96) expressed as follows:

$$q_u{}_{(7\mathrm{days})}=\frac{\mathrm{14,843}}{{(I_L/C_P)}^{1.51}} {\mathrm{R}}^2=0.96$$

(2)

$$q_u{}_{(28\mathrm{days})}=\frac{\mathrm{24,648}}{{(I_L/C_P)}^{1.60}} {\mathrm{R}}^2=0.98$$

(3)

The fitted q_u-I_L/C_P curve is unique at a given curing age, regardless of the liquid index or cement content. To verify the applicability of the proposed framework to other types of clays, additional data from Horpibulsuk et al. [16] and Correia et al. [22] were used to verify the empirical expression proposed in Equation (1). and the predicted strength of specimens at 7 and 28 days of curing was calculated and is presented in

Table 5. UCS of cement-based solidified other soft soils [16,22].

Soft Soil	Cw (%)	IL	CP (%)	IL/CP	Curing Age (Days)	Predicted Strength, qup (kPa)	Measured Strength, qum (kPa)
Bangkok clay (cement + fly ash)	89	1.0	11.18	8.94	7	555	520
	89	1.0	11.56	8.65	7	583	547
	118	1.5	11.14	13.38	7	303	332
	118	1.5	11.51	12.90	7	319	332
	118	1.5	33.42	4.46	7	1577	1371
	118	1.5	34.5	4.32	7	1654	1467
	148	2.0	11.19	17.87	7	196	257
	148	2.0	11.57	17.29	7	207	273
	148	2.0	33.56	5.96	7	1020	844
	148	2.0	34.74	5.76	7	1075	865
	89	1.0	11.18	8.65	28	740	781
	89	1.0	11.56	13.38	28	781	820
	118	1.5	11.14	12.90	28	389	514
	118	1.5	11.51	4.46	28	410	547
	118	1.5	33.42	4.32	28	2255	2056
	118	1.5	34.5	17.87	28	2372	2150
	148	2.0	11.19	17.29	28	244	320
	148	2.0	11.57	5.96	28	258	366
	148	2.0	33.56	5.76	28	1417	1409
	148	2.0	34.74	8.65	28	1498	1549
Bangkok Clay (cement + biomass ash)	89	1.0	22.2	4.50	28	2218	1786
	89	1.0	12.7	7.87	28	908	841
	89	1.0	7.4	13.51	28	382	467
	118	1.5	29.5	5.08	28	1827	1800
	118	1.5	16.8	8.93	28	742	845
	118	1.5	9.8	15.31	28	313	450
	148	2.0	36.8	5.43	28	1642	1800
	148	2.0	21	9.52	28	669	800
	148	2.0	12.3	16.26	28	284	449

. Moreover, the comparison between measured and predicted values is shown in Figure 6. It can be seen that almost all of the measured values of q_u differ within 30% from the predicted values, which is valid and acceptable for geotechnical engineering.

3.4. Microscopic Test Results Related to the Strength

3.4.1. SEM Analysis

This investigation focused on the effect of the consistency limit on the microstructure characteristics of CDS. The comparison of the microstructure of RDS-25 and RBIV-25 cured for 28 days is shown in Figure 7, including a series of SEM images with a magnification of 1000, 5000 and 10,000 times. In particular, the water/cement ratio of RDS-25 and RBII-25 was both 3.6 (C_W = 90%, C_P = 25%), but their liquid index I_L was 2.69 and 1.94, respectively, due to different consistency limits. As shown in Figure 7a,b (low magnification, ×1000), the cross-section of RDS-25 covered with irregular soil particles was rough and uneven. However, the internal section of RBII-25 was relatively uniform and smooth with no apparent distribution of dispersed particles. Figure 7c,d present the micromorphology at 5000 times magnification. For RDS-25 (Figure 7c), the gelatinous cement hydrates such as C-S-H-bonded dispersed soil particles of different sizes in a point-to-point manner, forming the porous network framework. A large number of hydrates interlaced with each other, which contributed to a dense matrix, as shown in Figure 7d. Furthermore, only a small amount of acicular ettringite (AFt) was observed in the pores of RDS-25 at 10,000 times magnification (Figure 7e), and gel hydrates were still not observed clearly, resulting in loose soil aggregates with many micropores. However, no micropores were observed in the RBII-25 sample with a denser matrix even at 10,000 times magnification (Figure 7f). The above microstructural differences caused by the consistency limit can be analyzed from the perspective of the pore water state in sludge: under the same water content and cement content, the smaller liquid index indicated that the free water content decreased and the adsorbed water content increased, which led to closer soil particle spacing, smaller internal pores and better bridging effect of hydrates. Accordingly, the microstructure of CDS transformed from a honeycomb skeleton structure (such as RDS-25, 494 kPa) to a dense integral structure with greater strength properties (such as RBII-25, 780 kPa).

3.4.2. MIP

Researchers have shown that pore structure plays an important role in controlling the geotechnical strength of CDS [36,37]. Figure 8 presents the MIP results for the RDS-25 and RBII-25 cured for 28 days. As shown in Figure 8a, the total pore volume of the RBII-25 (0.57 mL/g) was approximately 0.061 mL/g lower than that of RDS-25 (0.509 mL/g), indicating the denser structure of the former [38]. Figure 8b shows the differential (incremental) aperture distributions derived from Figure 8a. Moura et al. [39] and Wang et al. [40] reported that the aperture distribution can be categorized into three ranges, namely, “capillary pores” (10–100 nm), “mesopores” (100–10,000 nm) and “air pores” (>10,000 nm). It is clear that RBII-25 was mainly characterized by mesopores between hydration products and soil particles, indicating that the internal pores are evenly distributed and contribute to higher compressive strength. However, the inner pores of RDS-25 were mainly composed of capillary pores inside the hydrates and large air pores between the interaggregates, and this internal heterogeneity can easily lead to stress concentration and microcrack generation, resulting in a reduction in strength. This result coincided with the higher q_u value (i.e., 780 kPa) of RBII-25 over that of RDS-25 (i.e., 494 kPa). The MIP test results corresponded to the above UCS and SEM analysis; that is, when the water/cement ratio C_W/C_P is the same, the soil particle spacing decreases with the decrease in the liquid index I_L, so the overall porosity of CDS further reduced, and the internal uniformity of CDS is also improved. As a result, the macro mechanical properties were highly consistent with the microstructure characteristics.

3.4.3. XRD Analysis

Figure 9 presents the XRD diffractograms of 28-day cured typical CDS samples. The minerals in the RDS (unsolidified sludge) were predominantly quartz, montmorillonite and illite. Compared with RDS, the XRD spectra revealed peaks of typical hydrates such as ettringite at 22.9°, calcite at 29.5° and portlandite at 36.5° in RDS-25 (CDS sample) [37,41]. It is worth noting that the diffraction peaks of the amorphous C-S-H and C-A-H gel are not detected, but they can be clearly observed from the SEM images (Figure 7). The calcite formation was attributed to the air exposure of CDS and subsequent carbonation of calcium hydroxide (cement hydration product) during the curing period [36,42]. For RBII-25, the peak corresponding to calcite reduced significantly, which was attributed to its denser matrix (which has been demonstrated from the SEM and MIP tests) and hence a smaller exposure area to carbon dioxide (CO₂) in the air [43,44]. This also showed that the effect of the consistency limit on the strength property was realized to some extent by changing the agglomeration structure of the solidified sludge matrix. The bentonite used to adjust the consistency limit of sludge only showed sharp peaks of montmorillonite at 5.8°, 19.8°, 21.9°, 34.3 and 61.9°, indicating that its composition is relatively pure [39]. The cement-treated sample RDS-25 showed a montmorillonite peak similar to RDS at 19.8°, which was also helpful to confirm the nonreactivity between cement and montmorillonite.

3.4.4. TGA

The results of TGA are presented as a mass loss curve or the first derivative curve of mass loss with temperature, as shown in Figure 10. The weight of RDS-25 and RBII-25 continuously decreased with elevating temperature until 800 °C. The peaks in DTG curves (or mass losses in TG curves) correspond to the presence of cement hydration products during their thermal decompositions. In this study, the significant weight loss at 50–200 °C mainly corresponded to dehydration of C-S-H gel and AFt (loss of crystal water), respectively [40]. The total weight loss rates of RDS-25 and RBII-25 at 50–200 °C were 7.25% and 6.33%, respectively, indicating that more C-S-H and AFt were generated in RBII-25 (corresponding to the higher strength shown in Figure 2), which was consistent with the SEM results. Subsequent weight loss at 400–500 °C was due to the decomposition of portlandite [45]. The loss between 500 and 650 °C was related to the decarbonization of calcium carbonate or silicates/aluminates [43]. The substantial loss between 650 and 750 °C was due to the decomposition of crystalline calcite [46,47], and the weight loss ratio of RBII-25 (4.76%) was significantly lower than that of RDS-25 (3.63%), indicating that the RBII-25 sample contained less calcium carbonate, which was also confirmed by XRD results. The similar evolution characteristics of the two thermogravimetric curves also indicated no significant difference in the new mineral types between RDS-25 and RBII-25.

4. Conclusions

The main findings can be drawn as follows:

(1)

The consistency limit was a key factor affecting the strength development. The relationship between the UCS and liquid index I_L can be characterized by the power function y = a∗x^b. A valid empirical model, q_u = A/(I_L/C_p)^B, was proposed based on the experiment data.

(2)

A large number of hydrates in the CDS with lower I_L interlaced with each other, which promoted the formation of dense structure. However, for the CDS with higher I_L, hydrates cannot play the role of cementing particles, resulting in the formation of the porous structure.

(3)

In engineering applications, under the same initial water content, if the liquid index can be reduced by adjusting the consistency limit, the strength performance of CDS can be improved.