Destructive and nondestructive characteristics of solidified reservoir sediments incorporating microstructural analyses

Abstract

Reservoir sediments create a range of severe problems for hydropower dams. Although reservoir sediments can be excavated, nonetheless, sizeable dumping lands for such excavated sediments are unavailable at this time. This paper presents an experimental investigation of the destructive and nondestructive properties and microstructural characteristics of reservoir sediments solidified with fly ash–cement blend for reuse as construction materials. The obtained natural sediment was classified as well-graded sand with silt. The destructive experiments comprised unconfined compression, indirect tension, California bearing ratio, resilient modulus, and durability against wet–dry cycle tests, while the nondestructive experiments included a free–free resonance test. Microstructural investigations consisting of X-ray diffraction, thermogravimetric analysis, scanning electron microscopy, and energy-dispersive X-ray spectrometry were performed to verify the macroscale test results. The results showed that fly ash–cement blend exhibited increased strength values on the order of 2 to 9 times that of unsolidified sediments. Using fly ash-blended cement was more effective than using sole cement or sole fly ash, and mixtures with 10% fly ash delivered the best strength and modulus values. Various functional empirical correlations were proposed. Utilizing six wet–dry cycles is acceptable because the strength of the samples subjected to the six wet–dry cycles was lower than the given value. The results of the peak intensities of calcium silicate hydrate, mass losses, calcium contents, and scanning electron microscopy images derived from the microstructural investigations confirmed the macroscale test results.

Introduction

Reservoir sediments represent an amalgamation of various noncompacted soil particles, including gravel, sand, silt, and clay, induced by soil erosion and rock weathering processes (Dubois et al. 2011). Reservoir sediments develop from the transportation and deposition processes of natural sediments in reservoirs contained by dams (Banoune et al. 2016; Chan 2016; Dang et al. 2013; Furlan et al. 2018; Silitonga et al. 2009, 2010; Wang et al. 2012, 2013, 2018; Zeng et al. 2017; Zentar et al. 2021; Zhang et al. 2018). More than twenty hydropower dams operated by the Department of Alternative Energy Development and Efficiency (DAEDE) under the Ministry of Energy of Thailand have faced large amounts of excessive reservoir sediments behind their dams. Such excessive sediments generate many severe problems, including reducing the storage capacity, discharge ability, flood control capability, and electric energy production efficiency of the dams. Furthermore, reservoir sediments also affect dam safety, such as leading to excessive loads on the dam body and dam gates, damaging mechanical equipment, and creating a wide range of environmental impacts. Reservoir sediments are regularly excavated to maintain dam abilities and safety. Therefore, the excavated reservoir sediment is characterized as waste and, as such, is generally discarded in the sea or land dumps (Kamali et al. 2008).

The DAEDE arranges to excavate such excessive reservoir sediments. However, the sizeable dumping land areas needed to support such excavated sediments are unavailable because most of the area around such reservoirs tends to be agricultural and farming regions. Thus, the DAEDE decided to reuse these wastes as construction and building materials, such as compressed earth blocks and road materials, to benefit the community within the vicinity. The first pilot project was the reuse of reservoir sediments in the Huai Mae Phong hydropower dam, located in Phayao Province, Thailand (Fig. 1a and b). However, the reuse of reservoir sediments without treatment would not feasibly deliver appropriate standard construction materials. Solidification of the excavated reservoir sediments using stabilizers that are locally available in Thailand, such as cement and fly ash, is introduced in regard to cost-effectiveness, construction practice, and waste management.

Fig. 1

a Location of Phayao province, Thailand, and b Huai Mae Phong reservoir

Solidification with cement alone is the most conventional method for enhancing the strengths and moduli of problematic soil properties in various applications (Güllü et al. 2017; Horpibulsuk et al. 2010, 2011a, b; Jamsawang et al. 2017a, b; Kamali et al. 2008; Mohammadinia et al. 2015; Voottipruex and Jamsawang 2014; Yoobanpot et al. 2017; Chompoorat et al. 2019). Class C fly ash is a famous byproduct derived from the combustion process of lignite coal in power plants (Jamsawang et al. 2017a, b; Yoobanpot et al. 2017). Class C fly ash is considered cementitious and pozzolanic because it has a suitable self-cementing property and contains active silica and alumina (Jamsawang et al. 2017a, b; Yoobanpot et al. 2017). The fly ash production rate is much higher than its reuse rate. Therefore, recycling of fly ash as road materials in civil engineering applications provides notable environmental advantages, such as relieving air contaminants and reducing water pollution (Banoune et al. 2016; Silitonga et al. 2009; Wang et al. 2012; Zentar et al. 2021). The use of fly ash alone can increase the strength and stiffness and reduce the compressibility and swell potential of poor soils due to cation exchange, flocculation, and microfillers (Horpibulsuk et al. 2009; Silitonga et al. 2009; Tastan et al. 2011; Wang et al. 2012). An amount of 21% fly ash was found to be optimal for obtaining the best strength development with soft Bangkok clay within the active zone (Horpibulsuk et al. 2013). A sole fly ash content of 14% was sufficient to provide the same swell potential of the treated expansive clays as a sole cement content of 3%. However, the exclusive use of fly ash was not able to deliver the same strength as the use of sole cement (Voottipruex and Jamsawang 2014). The strength and modulus of the soft organic soils solidified with 10–20% fly ash were 2 to 6 times greater than those of the unsolidified sample (Horpibulsuk et al. 2009). Class C fly ash can also be blended with cement to increase the performance of the soil stabilization process by increasing the reactive surface area for hydration and pozzolanic reactions (Tastan et al. 2011; Chompoorat et al. 2022).

The important destructive properties for solidified soils intended for use in construction materials include the unconfined compressive strength (q_u), indirect tensile strength (q_t), strain at failure (ε_f), California bear ratio (CBR), secant elastic modulus (E₅₀), and resilient modulus (M_r). In contrast, the main nondestructive characteristics comprise the shear wave (V_s) and compression wave (V_p) velocities, which can be determined by free–free resonance (FFR) tests. The FFR test is extensively employed for characterizing material properties because it is a rapid test that can be carried out on samples of various sizes (Bogas et al. 2013; Biswal et al. 2020; Moon et al. 2020; Chompoorat et al. 2021a). The dynamic moduli, including the dynamic Young’s modulus (E₀), shear modulus (G₀), and dynamic Poisson’s ratio (ν₀), are essential parameters for geotechnical design and predicting the behavior of soil subjected to earthquakes, explosions, and vibrations. The dynamic moduli can also be used as indirect soil parameters for general analyses in civil engineering works (Bogas et al. 2013; Biswal et al. 2020; Moon et al. 2020; Chompoorat et al. 2021a). Microstructural analyses by X-ray diffraction (XRD), thermogravimetric analysis (TGA), energy dispersive X-ray spectrometry (EDX), and scanning electron microscopy (SEM) techniques have commonly been applied for investigating the changes in microstructures of stabilized soils and sediments after the completion of destructive tests (Chan et al. 2016; Güllü et al. 2017; Kang et al. 2014; Jamsawang et al. 2017a, b; Jiang et al. 2016, 2018; Voottipruex and Jamsawang 2014; Wang et al. 2013; Zhang et al. 2018; Chompoorat et al. 2021b).

The destructive properties and corresponding microstructural characteristics of sediments solidified by either cement or fly ash alone have been studied by many researchers (Bogas et al. 2013; Dubois et al. 2011; Furlan et al. 2018; Silitonga et al. 2009; Wang et al. 2012, 2013; Zeng et al. 2017; Yoobanpot et al. 2020a, b; Zentar et al. 2021; Jamsawang et al. 2021). In contrast, few studies have been conducted entirely on sediments solidified with fly ash–cement blend. As mentioned above, most of the previous test results were concentrated on only the q_u and E₅₀ values rather than other destructive properties. Only XRD and SEM analysis results were presented to confirm the increases in the q_u and E₅₀ values of the solidified sediments. Moreover, as previously mentioned, limited relevant nondestructive characteristics have been determined to investigate correlations between destructive and nondestructive properties. This study presents experimental work to evaluate various destructive and nondestructive properties of reservoir sediments solidified with fly ash–cement blend relevant to construction materials. Numerous functional correlations are provided. Several microstructural analyses are employed to investigate changes in the microstructures and support the macroscale test results.

Research methodology

Materials

The primary raw materials used in the current study included reservoir sediment, type I ordinary Portland cement, and fly ash, as shown in Fig. 2. The natural reservoir sediments were excavated with an excavator from the Huai Mae Phong Reservoir, Phayao Province, Thailand. They were then delivered to the geotechnical engineering laboratory at King Mongkut’s University of Technology North Bangkok, Bangkok, Thailand. The fly ash, a byproduct of coal ash lignite, was delivered from the Mae Moh electric power plant, located in Thailand’s Lampang province. SEM photos of the primary raw materials, as shown in Fig. 2, reveal that the sediments contained various particle sizes, such as coarse-grained and fine-grained particles with irregular shapes. The cement particles exhibited rough surfaces and nonuniform angular shapes, whereas the fly ash particles showed spherical shapes with uneven surfaces.

Fig. 2

Real and SEM photos of the raw materials used in this study

Figure 3 presents a grain size distribution curve and the geotechnical properties of the raw sediment. The raw sediment was comprised of 89% sand, 11% silt, and 4% clay contents, classified as group symbols of SW-SM and group names of well-graded sand with silt following the Unified Soil Classification System (Table 1). The maximum dry unit weight was 16.5 kN/m³, falling in the backfill sand’s typical value for use in road construction works (Jamsawang et al. 2015, 2018). Comparisons of the grain size distribution curves of the raw materials in Fig. 2, obtained from sieve and laser analyses, indicate that the mean particle size values of the sediment, cement, and fly ash were 0.85, 0.014, and 0.019 mm, respectively. Thus, the sediment’s mean particle size was 50 times greater than that of cement and fly ash. The fly ash acts as both a pozzolan and a filter in the sediment pore space.

Fig. 3

Grain size distribution curves of the raw materials used in this study

Table 1 Geotechnical properties of raw reservoir sediment

Table 2 shows the chemical compounds of raw sediment, cement, and fly ash derived from XRF analyses. The fly ash is characterized as class C because the sum of its SiO₂, Al₂O_3, and Fe₂O₃ fractions is 69.14% (ranging from 50 to 70%), based on ASTM C 618. The fly ash can also contribute cementitious and pozzolanic properties since it is comprised of both 17.85% CaO and 37.34% SiO₂ contents. As a result, fly ash alone can be an admixture for enhancing sediment properties. Cement contains an excellent CaO content of 66.35%, which is more significant than the CaO content in fly ash. Therefore, cement alone is a primary source of cementitious properties and can effectively stabilize sediments in short-term conditions. In this study, fly ash-blended cement is more useful than sole cement due to the high SiO₂ content in the fly ash, which actively reacts with the CaO in cement to deliver long-term strength via the pozzolanic reaction.

Table 2 Chemical composition of the reservoir sediment, cement, and class C fly ash

Preparing and molding samples

The raw sediments, which have a high natural moisture content, were air-dried for 24 h, shaken through a sieve with a 4.75-mm opening size to remove the gravel particles, and later oven-dried for 24 h to obtain a zero initial moisture content before molding. All mechanical tests were performed utilizing cement alone, fly ash alone, and fly ash-blended cement. Cement contents of 1.5, 2.5, 5, and 7.5% (by dry sediment weight) and fly ash contents of 5, 10, 15, and 20% (by dry sediment weight) were employed to mix with the prepared sediments. Notably, the fly ash and cement contents used in this study were selected based on previous similar studies and several prior attempts by the authors during preliminary inquiry (Yoobanpot et al., 2020a, b; Jamsawang et al., 2021). The mixture’s water content attains the optimum moisture content of 18.5%, derived from modified Proctor compaction tests. All mixture proportions were tested in triplicate for each testing condition, as listed in Table 3. The cylindrical molding samples for the unconfined compression (UC), indirect tensile (IT), wetting and drying cycles (WDC), resilient modulus (M_r), and free–free resonance (FFR) tests had diameters of 50 mm and heights of 100 mm, whereas those for the CBR tests had diameters of 152 mm and heights of 116 mm. All dry mixtures were prepared by hand mixing of dry sediment, cement, and fly ash for 5 min, and water was later gradually added to the dry mixtures by spraying. The dry mixes and water were hand mixed again for 5 min until uniform and homogeneous mixtures were visually observed.

Table 3 Mixture proportion

Each sample for the UC, IT, WDC, M_r, and FFR tests was molded by containing each prepared mixture in a cylindrical split stainless mold lubricated inside with grease. The sample was subsequently statically compressed in three layers using a compression machine until each layer achieved the maximum dry unit weight of 16.5 kN/m³, derived from modified Proctor compaction tests. After completing the sample molding process, the sample was carefully removed by splitting the mold into two pieces to minimize sample damage. Once removed, each specimen was instantly wrapped with plastic wrap to prevent sample moisture loss. The covered sample was then cured for 7, 14, 28, 60, and 120 days inside a storage cabinet under a controlled room temperature of 40 ± 2 °C to emulate Thailand’s weather.

When the assigned curing time of the specimen for each test was achieved, the test was then immediately performed. The acceptance criterion was designated that the individual test values of three samples, molded with the same characteristics, must deviate from the mean test value by less than 10% to avoid the error caused by the discreteness of the sample. For all tests in this study, an average value of the test results based on three specimens was reported.

Experiments

This section describes the UC, IT, WDC, M_r, and FFR experimental details and procedures. After the assigned curing time of the sample was achieved, the experiments were immediately performed. The test results’ average value based on three replicates was reported for all experiments in this study. All the tests were performed using an automatic loading machine with a capacity of 50 kN, brand “Controls with V1072 series multispeed digital version.” The machine consisted of a stand-alone automatic digital load frame with a 50-kN capacity. Electric motors with closed-loop speed control were used to control the deformation rate during each test and provided a fully variable test speed of 0.2 to 51 mm/min. The machine can be equipped with analog or digital load/displacement measurement systems and specific accessories to suit either the field or central laboratory requirements. Electric load cells of brand “Kyowa” and linear variable differential transducers (LVDTs) by “Kyowa” were utilized to measure the applied loads and deformations of the specimens. The electric data logger system, brand “Kyowa” model UCAM-550A, was used because it is a fast data logger that repeatedly measures a maximum of 1000 channels at an interval of 0.02 s and supports load and displacement transducers. Measuring channels are for 1 unit, a maximum of 50 channels, and with 20 cascaded units, a maximum of 1000 channels, and these are suited for small-scale to large-scale measurements. Therefore, the electric data logger used in this study can provide sufficient measurement data to create precise stress–deformation curves for assessing the reliability of the mechanical performance of the tested specimen and for the interpretation of the accuracy of the results.

Unconfined compression tests

The UC tests were performed based on ASTM D 2166 (2016a) to generate unconfined compressive stress–axial strain curves for determining the q_u and the secant modulus (E₅₀) values of the specimens. A 50-kN-capacity load cell and a 50-mm-capacity LVDT were used. A vertical displacement rate of 1 mm/min was applied to the test sample until 15% axial strain was attained. The q_u is defined as the maximum compressive stress when the stress–strain curve exhibits apparent peak stress. Otherwise, q_u is examined as the compressive stress at 15% axial strain. The E₅₀ is estimated from the stress–strain curve slope and is defined as the secant modulus at 50% q_u, according to a safety factor of 2.

Indirect tensile tests

According to ASTM D 3967 (2016b), IT tests were performed to determine the q_t value of the solidified samples. Each specimen was placed in direct contact with two steel-bearing strips. A 20-kN-capacity load cell with a vertical displacement rate of 1.00 mm/min was utilized to obtain the accurate failure load. The load was applied to the specimen from the top to produce a vertical crack along with the sample diameter induced by splitting tension. q_t was defined as follows: ${\text{q}}_{t}=2P/\pi {\text{dl}}$, where q_t is the indirect tensile strength, P is the compressive load at failure, d is the sample diameter, and l is the sample length.

Resilient modulus tests

The M_r tests were performed under the AASHTO T307 (2012) standard test method to determine the M_r of both unsolidified and solidified specimens. The M_r test apparatus was comprised of two 20-mm-capacity LVDTs placed symmetrically on the sample’s top to detect the vertical displacements. A 10-kN-capacity load cell mounted on the loading piston with a vertical displacement rate of 1.00 mm/min was utilized to create the deviatoric stress–axial strain curve. The test was finished when a total vertical permanent strain of 5% was obtained. A pressurized air chamber was used to generate confining pressures for accommodating the specimen throughout testing. Axial cyclic stress with a constant magnitude of 0.1 s, followed by a 0.9-s rest period, was applied to the test sample. A loading cycle of 1 s is composed of a stress and rest period, representing a 1-Hz frequency. Each specimen was subjected to preconditioning, including a 103.4-kPa confining pressure, 93.1-kPa deviatoric stress, and 10.3-kPa contact stress with 500-cycle stress repetitions. The specimen was subsequently subjected to 15 sequences of applied stress conditions, consisting of five distinct confining pressures and three different deviatoric stresses for each confining pressure with 100-cycle stress repetitions, as listed in Table 4. The M_r value for each sequence was determined by considering the average of the previous five records. In contrast, the M_r value reported for the designated specimen was based on the average value derived from 15 sequences (Mamatha and Dinesh 2017).

Table 4 Resilient modulus testing sequences

California bearing ratio tests

The CBR tests were operated per ASTM D1883 (2016c). Each mix was dynamically compacted in a standard compaction steel mold, following a modified Proctor compaction test. The compacted sample in the mold with a 45-kN surcharge was soaked for 96 h to emulate the road material characteristics when subjected to rainfall or flooding conditions. After the specimen was completely soaked, the penetration piston load was applied to the soaked sample with a 1.27 mm/min penetration rate using a 20-mm-capacity LVDT to create a penetration stress–penetration curve. The CBR is defined as the test stress at either a 2.54 mm or 5.08 penetration divided by the corresponding standard stress of either 6900 or 10,300 kPa multiplied by 100%, respectively. The CBR is usually reported based on the value at a 2.54-mm penetration. If the CBR at 5.08 mm penetration is larger than that at 2.54-mm penetration, retesting will be needed. However, the CBR at 5.08-mm penetration will be reported if the retest shows similar results.

Wetting and drying cycle tests

The ASTM D559 (2015) standard test method was adapted for the specimen preparation for WDC tests. In a cycle of the wetting and drying process, each solidified sample was preliminarily soaked at room temperature for 5 h, then oven-dried at 71 °C for 42 h, and subsequently cooled to room temperature for 3 h. A total number of three samples were required for each mix to complete the tests. The first, second, and third samples were subjected to 1, 3, and 6 WDCs. After each assigned cycle was attained, the prepared specimen was instantly soaked in water at 25 °C for 1 h. The UC test was subsequently conducted on the soaked sample following “Unconfined compression tests” to determine the residual q_u under 1, 3, and 6 WDCs, designated as q_u(WDC).

Free–free resonance tests

The FFR tests were performed based on an adaptation of the ASTM C215 (2019) standard test method for the determination of the V_p and V_s of both unsolidified and solidified samples. Before the destructive UC and IT tests were performed, the same specimen was employed for the nondestructive FFR tests to ensure that the destructive and nondestructive test results were based on similar specimen conditions. Thus, the relationships among them were reliable. The specimen was freely mounted on a rigid frame in the horizontal direction and suspended with a tendon to approach free boundary conditions. An accelerometer was set in contact with one end of the sample, while another end was knocked with a light hammer. The following relationships can be used to calculate V_p and V_s: ${\text{V}}_{p}=2{\text{lf}}_{p}$ and ${\text{V}}_{s}=2{\text{lf}}_{s}$, where l is the specimen length, f_p is the resonant frequency induced by a compressive wave, and f_s is the resonant frequency caused by a shear compressive wave. The wavelength is two times the sample length when the sample with a length-diameter ratio of 2 is subjected to a free vibration with both ends free (Ryden et al. 2006). The theory of one-dimensional wave propagation in an elastic rod was adapted to estimate the dynamic Young’s modulus (E₀) and shear modulus (G₀) of the test specimen. E₀ and G₀ can be evaluated based on the sample’s measured V_p and V_s with the sample’s density (ρ) using ${E}_{0}=\rho {V}_{\text{p}}^{2}$ and ${G}_{0}=\rho {V}_{\text{s}}^{2}$ (Nazarian et al. 1999). The dynamic Poisson’s ratio (ν₀) can also be derived ${\nu }_{0}=\text{(}{E}_{0}/{2}{G}_{0}\text{)}-{1}$ based on the assumption of a linear-elastic, homogeneous, and isotropic material.

Microstructural analyses

X-ray diffraction analyses

XRD analyses were performed on both unsolidified and solidified samples. The small samples for XRD analyses were extracted from the test specimens’ failure surfaces after the UC tests were accomplished. The retrieved piece was then placed in a desiccator on silica gel to be air-dried for 24 h. The dried sample was then ground with a medicine grinder to prepare a powder material, and the powder material was then sieved through a 0.075-mm sieve (no. 200 sieve) until 20 g of powder finer than 0.075 mm was obtained. XRD analysis was performed on the fine powder by a Siemens D-500 XRD instrument with Stirling equipment and Cu Kα radiation at 40 kV and 30 mA. Intensity data were recorded every 0.02° from 2 to 80°.

Thermogravimetric analyzer analyses

TGA analysis was performed on the solidified samples utilizing the STA 449F3–Jupiter equipment. The specimen preparation was similar to the XRD analyses. The 100-mg powdered sample was placed on a balance located in the furnace. The heating process was subsequently applied over the temperature range from 25 to 750 °C under a 10 °C/min heating rate in a flowing nitrogen environment.

Scanning electron microscopy–energy-dispersive X-ray spectometry analyses

SEM–EDX analyses were conducted on both unsolidified and solidified samples to investigate the change in elemental composition/distribution induced by hydration and pozzolanic reactions. A small rectangular piece of the sample with approximate dimensions of 5 mm × 5 mm × 2 mm was collected from the same specimens as the XRD analyses. The acquired sample was placed in a desiccator on silica gel to be air-dried for 24 h and was then coated with platinum over 30 s at a current of 50 mA. SEM micrographs with 3500 times magnification were derived with Do SEM JSM-5410 LV equipment. EDX analysis was then conducted on the corresponding SEM micrograph regions with a Do SEM Link ISIS300 instrument.

Test results

Destructive properties

Unconfined compressive strength

A strength increase in sediments solidified with fly ash and cement is expressed through the primary cement hydration reaction (Eq. 1) and secondary pozzolanic reaction (Eq. 2). The cement hydration reaction develops early after fly ash–cement blend and sediment are mixed with the optimum water content, produced by the CaO segregation from cement and fly ash particles. SiO₂ in the cement is disintegrated by high CaO concentrations. The cement hydration products subsequently create calcium silicate hydrate (C-S–H) gel and calcium hydroxide (Ca(OH)₂). The C-S–H gel is a cementitious substance that gives adhesive properties to the sediment particles, while Ca(OH)₂ is a byproduct of the cement hydration process. The pozzolanic reaction later develops since the SiO₂ in fly ash continuously adsorbs the Ca(OH)₂ obtained from the cement hydration reaction. The pozzolanic reaction products (additional C-S–H gels) are consequently generated.

$$\begin{aligned}\mathrm{Cement\ hydration\ reaction:{CaO+H}}_{2}\text{O} =& \underset{\mathrm{Hydration\ product}}{\mathrm{C-S-H}}\\&{ + \underset{\mathrm{Byproduct\ from\ hydration}}{\mathrm{Ca(OH)}}_{2}}\end{aligned}$$

(1)

$$\begin{aligned}\mathrm{Pozzolanic\ reaction:}\underset{\mathrm{Byproduct\ from\ hydration}}{{\mathrm{Ca(OH)}}_{2}}&{ + }\underset{\mathrm{Compound\ in\ pozzolan}}{{\mathrm{SiO}}_{2}}\\&{ = \underset{\mathrm{{P}}\mathrm{ozzolanic\ product}}{\mathrm{C-S-H}}}\end{aligned}$$

(2)

The strength increase characteristics are highly dependent on the C-S–H product amounts derived from either the cement hydration reaction or pozzolanic reaction. The C-S–H products induced by the cement hydration reaction provide a short-term strength increase. In contrast, the C-S–H products from the pozzolanic reaction give a long-term strength increase (Horpibulsuk et al. 2009; Kang et al. 2014; Shon et al. 2010; Tastan et al. 2011; Wang et al. 2013). The C-S–H gels bind the reservoir sediment particles together, providing strong cohesiveness to the sediment matrix. This phenomenon results in a strength increase with increasing C-S–H products.

In summary, fly ash–cement blend was most effective in solidifying the sediment, followed by sole cement and sole fly ash. Based on Eqs. 1 and 2, the fly ash-blended cement provided the most extensive C-S–H contents because the C-S–H contents were derived from a combination of C-S–H contents produced from both cement hydration and pozzolanic reactions. The 10% fly ash revealed the best q_u values for all mixes, including F10-C1.5, F10-C2.5, F10-C3.5, and F10-C5, because the 10% fly ash gave the most suitable SiO₂ and CaOH₂ portions obtained from the fly ash and cement hydration reaction, respectively. Therefore, the 10% fly ash caused the highest pozzolanic reaction level, creating the most extensive C-S–H products. However, fly ash over 10% can block cement particles, preventing water and cement chemical reactions. This occurrence reduces the C-S–H products, and q_u subsequently decreases (Horpibulsuk et al. 2011a, b). The use of sole cement yielded a smaller q_u than the fly ash–cement blend because most C-S–H products were derived from the cement hydration reaction only due to the absence of SiO₂ in FA. The sole fly ash gave the smallest q_u because of the lack of CaO from cement to produce the C-S–H products in the cement hydration reaction, resulting in the lowest C-S–H amounts.

Figure 4a shows the effect of fly ash and cement contents on the q_u of solidified samples for various curing periods of 7 to 120 days. The q_u values for all mixes regularly increased with the curing period. Notably, the q_u values of the specimens solidified with sole cement grew with the cement contents. In contrast, sole fly ash and fly ash–cement blend solidifications were characterized as two principal regions, including active and degradable zones. The strengths and moduli increased clearly with increases in the fly ash content in the active zone and attained maximum strength and modulus values at 10% fly ash. However, the strengths and moduli continuously decreased as the fly ash content exceeded 10% in the degradable region. Although the strengths and moduli increased compared to the unsolidified sample, in which the rate of increase was relatively low. Thus, the solidification in the degradable zone was uneconomical compared to that in the active region. For the short-term q_u of the 28-day curing time, the q_u values of the samples solidified with sole 10% fly ash (F10), sole 5% cement (C5), and 10% fly ash–cement blend (F10-C5) were 0.8, 3.2, and 3.6 MPa, respectively. As the curing period increased to 120 days, the long-term q_u increased to 3.1, 4.4, and 6.8 MPa for F10, C5, and F10-C5, respectively.

Fig. 4

Destructive properties versus curing times

Indirect tensile strength

Figure 4b shows that the q_t and q_u values of the solidified specimens exhibited similar strength evolution trends. The q_t of the unsolidified sample was 0.09 MPa. In contrast, the 28-day q_t values were found to be 0.25, 0.46, and 0.57 MPa, and the 120-day q_t values were equal to 0.50, 0.74, and 1.09 MPa for F10, C5, and F10-C5, respectively. Figure 5a presents the relationship between the q_t and q_u of solidified sediments for q_t variations between 0.12 and 1.09 MPa and q_u between 0.82 and 6.84 MPa. The q_t value was 1.3 to 12.1 times greater than that of the unsolidified sample. The q_t–q_u correlation for the solidified reservoir sediments was expressed as q_t = 0.16q_u, which was within the range of q_t = 0.1–0.2q_u for solidified soils with cementitious materials, as reported in previous studies (Baldovino et al. 2018; Consoli et al. 2014; Kolias et al. 2005).

Fig. 5

Correlation between destructive properties and q_u

California bearing ratio

The CBR results in Fig. 4c show similar trends to the q_u results of the solidified sediments. The CBR value of the unsolidified sample was 6%. In contrast, the solidified sediments with sole fly ash, sole cement, and fly ash–cement blend had variations in CBR values of 16–50%, 30–106%, and 32–136%, respectively. The empirical relationship between the CBR and the q_u value derived from the current study is shown in Fig. 5a. The correlation varies from CBR(%) = 10–30q_u(MPa), and the best fit is CBR(%) = 20q_u(MPa), which falls in the range of generally treated soils as proposed by Agapitus (2014); Jiang et al. (2016); Voottipruex and Jamsawang (2014); Yoobanpot et al. (2018).

Strain at failure

Figure 4d shows that the strain at failure (ε_f) values of solidified samples decreased as the curing time increased, indicating that the degree of the specimens’ brittleness increased with increasing time. Most values of ε_f were distributed in the range of 2 to 3.2%. The specimens’ brittleness significantly depends on their q_u magnitudes and exhibits inverse variation to the q_u values. Thus, an increase in q_u induced brittleness in solidified samples, decreasing ε_f. The correlation between the ε_f and q_u values found in the current is presented in Fig. 5b. The ε_f–q_u correlation of the solidified sediment obtained from this study is well represented by a power function, expressed by ${\varepsilon }_{\text{f}}=2.95{q}_{\text{u}}^{-0.22}$. This relationship corresponds to that reported by Al-Mukhtar et al. (2010b) for cement-stabilized zinc-contaminated clays.

Secant elastic modulus

The evolution of the E₅₀ of solidified sediments in Fig. 4e shows that the variations in E₅₀ values were consistent with the q_u values for the designed samples. The variations in E₅₀ values were 56–628 MPa for q_u values of 0.82–6.84 MPa. The E₅₀–q_u linear correlation in the current study is presented in Fig. 5c, represented by E₅₀ = 81q_u. Solidified marine clay with cement-blended metakaolin showed a brittle response with a relation of E₅₀ = 160q_u (Tongwei et al. 2014), overestimating the linear correlation derived from the current study by a factor of two. In contrast, marine sediment solidified with cement-blended lime and fly ash exhibited semiductile behavior with a correlation of E₅₀ = 60q_u (Wang et al. 2013), underestimating the current study result by a factor of 1.4. However, a similar correlation was reported for soft clay solidified with cement replaced with bagasse ash (Jamsawang et al. 2017a). The greater-strength sample exhibits a greater E₅₀ level due to its high degree of brittleness. Typically, E₅₀ depends on various factors, including the soil property, stress history, soil fabric, interparticle contact stiffness, effective stress particle roughness, and particle angularity (Verástegui–Flores et al. 2015).

Resilience modulus

Figure 4f reports the M_r values after 15 cycles, revealing the same characteristics as those of the E₅₀ values. The M_r value of the unsolidified sample was 130 MPa. After solidification with fly ash alone, cement alone, and fly ash–cement blend, the M_r values increased to 140–430, 200–660, and 240–980 MPa, respectively. The M_r–q_u correlation of the current study in Fig. 5c shows that M_r increased proportionally with q_u, similar to the E₅₀–q_u correlation. A linear correlation can be reasonably used for modeling M_r–q_u relationships (Austroads 2017), which is expressed as M_r = 151q_u. The linear regressions found by other researchers were M_r = 70–270q_u for fly ash-treated organic soils, quick lime-stabilized silty soils, and clayey soils solidified with cement, lime, and fly ash (Ardah et al. 2017; Arora and Aydilek 2005; Tastan et al. 2011), which are the lower and upper limits of the linear correlation derived from the current study.

Nondestructive properties

Compression wave velocity and shear wave velocity

Figure 6a and b present the V_p and V_s assessed from the FFR tests versus curing periods, respectively. In general, V_p and V_s increased with the curing time due to the strength development over time of the solidified sediments (Yesiller et al. 2001). The values of V_p and V_s were between 375–1330 m/s and between 195–730 m/s, respectively, for the q_u values of 0.82–6.84 MPa. The V_p and V_s propagations occur along the fastest pathway in the solidified samples. Therefore, they depend on the static stiffnesses, which are affected by similar factors, including particle bonding, porosity, and microstructure (Larsson 2005). The variations in V_p and V_s of the solidified samples are possibly related to the sediment homogeneity, uniformity of admixture distribution, and the existence of specific unmixed sediment inclusions (Guimond–Barrett et al. 2013). The V_p–q_u and V_s–q_u relationships from the present research in Fig. 7a illustrate the power–law functions as follows: ${q}_{\text{u}}({\text{MPa}})\text{ = 0.00007}{V}_{\text{p}}{}^{1.92}\text{(m/s)}$ and ${q}_{\text{u}}({\text{MPa}})\text{ = 0.00002}{V}_{\text{p}}{}^{1.92}\text{(m/s)}$, respectively.

Fig. 6

Nondestructive properties versus curing times

Fig. 7

Correlation between nondestructive properties and q_u

Dynamic Young’s modulus and shear modulus

Figure 6c and d illustrate the influences of the cement and fly ash contents and the curing periods on the E₀ and G₀ values of the solidified specimens, respectively. The E₀ and G₀ values varied from 32 to 227 MPa and 9 to 68 MPa, respectively. The unsolidified sediment had E₀ and G₀ values of 29 and 8 MPa, respectively. For all mixes, E₀ and G₀ increased as the particular cement and fly ash contents and curing period increased, and E₀ and G₀ achieved maximum values at a fly ash content of 10%. E₀ and G₀ gradually decreased when the fly ash content exceeded 10%, which exhibited similar characteristics as those of the E₅₀ values.

Consequently, F10-C5 provided maximum E₀ and G₀ values of 97 and 227 MPa and 28 and 68 MPa for curing times of 7 and 120 days, respectively. As the curing periods increased, the bonding development between the sediment particles was created due to the increasing hydration and pozzolanic reactions. This phenomenon increased the number of contact points between the sediment particles and reduced the number of voids in a solidified sample (Aldaood et al. 2014; Al-Mukhtar et al. 2010a, b), which affected V_p and E₀. Similarly, V_s and G₀ also increased since the solidified samples’ skeletal stiffnesses formed after a curing time of 28 days (Meei–Hoan and Chee–Ming 2014). E₀ and G₀ are plotted versus q_u in Fig. 7b. In the range of the unconfined compressive strength tests, E₀ and G₀ increased linearly with q_u. The ratios of E₀ and G₀ to q_u for the present study were found to be 33 and 10, respectively.

Dynamic Poisson’s ratio

Figure 6e shows the relationship between ν₀ and q_u for all mixes at curing periods of 7, 14, 28, 60, and 120 days. The values of ν₀ for the solidified samples were between 0.255 and 0.346, corresponding to an average of 0.297. The unsolidified sediment had a ν₀ of 0.312, whereas the unsolidified specimens had ν₀ values of 0.255 to 0.346, corresponding to an average value of 0.297. The typical ν₀ values for sandy clay, silt, and sand were found to be 0.25, 0.325, and 0.375, respectively, as Bowles (1997) reported. Thus, the ν₀ value of unsolidified sediment was close to that of silt. However, the ν₀ values decreased in the sandy clay and silt ranges after solidifying with the fly ash–cement blend. No clear correlation between ν₀ and q_u was observed in this study, unlike the E₀–q_u and G₀–q_u correlations. Thus, ν₀ depends on the property of the base sediment rather than that of the solidified sediment.

Wetting and drying cycles

Figure 8a illustrates the q_u(WDC) of all solidified samples subjected to 1, 3, and 6 wet–dry cycles for curing periods of 7, 28, and 60 days, which is compared to the q_u of corresponding solidified specimens without wet–dry processes (0 wet–dry cycles, derived from Fig. 4a). Regarding the samples with a 28-day curing time, the q_u(WDC) values of F10, C5, and F10-C5 subjected to 1 wet–dry cycle were 1.5, 1.9, and 2.2 MPa, respectively. As the sample curing time increased to 60 days, q_u(WDC) increased to 2.3, 2.5, and 3.6 MPa for F10, C5, and F10-C5, respectively. This phenomenon indicates that q_u(WDC) and q_u had similar strength development characteristics. The q_u(WDC) for all samples decreased significantly with the number of wet–dry cycles (C_WD). For example, the q_u(WDC) values of F10, C5, and F10-C5 subjected to 6 wet–dry cycles were 0.29, 0.36, and 0.49 MPa (28-day curing time), and 0.58, 0.59, and 1.07 MPa (60-day curing time), respectively. Therefore, the q_u(WDC) values significantly depend on the q_u values. The wetting causes clay particle swelling due to the expansion of the diffuse double layer (Kampala and Horpibulsuk 2013; Kampala et al. 2014; Tang et al. 2011). In contrast, the drying process reduces the pore volumes due to the water loss in voids between the solid particles, resulting in shrinkage and surface cracks on the solidified samples (Kampala and Horpibulsuk 2013; Kampala et al. 2014; Tang et al. 2011). Thus, the expansion, shrinkage, and surface cracks caused by the wet–dry cycles reduced the strength of the solidified sediment.

Fig. 8

a Effect of the fly ash-blended cement on the q_u(WDC) values, b crack propagation of the samples subjected to 1, 3, and 6 WDC, and c relationship between the normalized strength and N_WDC

The number of cracks increased with an increasing C_WD, as shown in Fig. 8b, resulting in a decrease in strength with growing wet–dry processes. The correlation between the normalized strength, q_u(WDC)/q_u, and C_WD is presented in Fig. 8c and given by the following equation: q_u(WDC)/q_u = 0.694 − 0.1C_WD. The C_WD value may be different, depending on the soil type or road materials. For example, the use of 1 to 12 C_WD is suitable for cement-treated soils used as a base course (PCA 1992), or the use of 1 to 4 C_WD is recommended for silty and clayey soils (Tripathy et al. 2009). COE (2004) suggests that q_u(WDC) values must be more than 1,723 and 2,068 kPa at curing periods of 7 and 28 days, respectively. Thus, the use of 6 C_WD samples was sufficient for solidified sediments in this study because the q_u(WDC) values of the samples subjected to 6 C_WD samples were lower than the values specified by COE (2004).

Normalized destructive and nondestructive characteristic development with time

Based on the results in “Destructive properties” and “Nondestructive properties” the destructive and nondestructive properties of reservoir sediments solidified with fly ash alone, cement alone, and fly ash–cement blend characteristics developed with curing time caused by the results of hydration and pozzolanic reactions. The destructive and nondestructive strength and modulus properties of the solidified sediments at 28 days, strength_(28D) and modulus_(28D), including q_u(28D), q_t(28D), ε_f(28D), CBR_(28D), E_50(28D), M_r(28D), E_0(28D), and G_0(28D), were employed to normalize the development of the properties of the solidified sediments after a curing period of D days, strength_(D) and modulus_(D), including q_u(D), q_t(D), ε_f(D), CBR_(D), E_50(D), M_r(D), E_0(D), and G_0(D). The normalized development of such properties for the solidified sediments is illustrated on a natural logarithmic scale by linear correlations (Horpibulsuk et al. 2009, 2011a, b, 2012; Yoobanpot et al. 2017, 2018), as given in Fig. 9a–c.

Fig. 9

a Generalized strength development, b generalized modulus development, and c generalized strain at failure development for the solidified sediment

Microscopic analysis results

X-ray diffraction

Figure 10a and b show the XRD patterns of the unsolidified sample (F0-C0), solidified specimens with cement only (C5), and fly ash–cement blend (F10-C5 and F20-C5) after short-term curing for 28 days and long-term curing for 120 days, respectively. The C-S–H peaks of solidified samples at 28 and 120 days of curing. The increases in strength and moduli due to increasing C-S–H product amounts, derived from cement hydration reaction and pozzolanic reactions, as discussed in “Destructive properties” are verified by observing the C-S–H peak intensities at the reference reflections (Jamsawang et al. 2017b; Mutaz and Dafalla 2014). The increases in the C-S–H peak intensities of solidified samples compared to the unsolidified specimen are presented in Table 5. The improvement ratios based on the reflection intensity are highest for F10-C5, followed by F10-C20, and are lowest for C5.

Fig. 10

XRD diffraction patterns of the solidified sediments at a 28 days and b 120 days, and c the relationship between the strength and CSH content

Table 5 Influence of fly ash-blended cement solidification on C-S–H peak intensities

The C-S–H peak intensities of C5 are lower than those of F10-C5 and F20-C5, confirming that the C-S–H products of C5 were induced from only the cement hydration reaction because of the lack of SiO₂ in fly ash. In contrast, the C-S–H peak intensities of F10-C5 are greater than those of C5 because additional C-S–H products were created by pozzolanic reactions. The 10% fly ash exhibited appropriate amounts between Ca(OH)₂ (from the cement hydration reaction) and SiO₂ (from fly ash), providing the best C-S–H contents during the pozzolanic response. However, the C-S–H peak intensities of F20-C5 are smaller than those of F10-C5, indicating unsuitable proportions of Ca(OH)₂ and SiO₂. Therefore, the strengths and moduli of the solidified sediments are highest for F10-C5, followed by F10-C20, and are lowest for C5, as discussed in “Destructive properties” The relationships between the average C-S–H peak intensity and q_u and q_t based on three reference reflections are illustrated in Fig. 10c, which can be expressed using exponential growth equations. The increases in q_u and q_t significantly depend on the C-S–H amounts.

Thermogravimetric analyzer

The relationship between the percentage of mass loss (m) versus the temperature (T) obtained from the TGA results for C5, F10-C5, and F20-C5 after curing periods of 28 and 120 days are presented in Fig. 11a and b, respectively. Significant mass losses are clearly observed for C-S–H products (primary m between 50 and 200 °C), calcium aluminum hydrate (CAH), and CASH (primary m between 200 and 300 °C) (Haha et al. 2011). The m increase is produced by releasing CO₂ and CaO induced by decomposing the unreacted hydrated calcium carbonate (Jiang et al. 2016; Sharma et al. 2012). The C-S–H and CAH/CASH contents in F10-C5 are higher than those in C5 and F20-C5, indicating a higher pozzolanic interaction. The correlations between m and q_u and q_t are presented in Fig. 11b. The m appeared to be a direct function of the behavior’s strength, which can be modeled as an exponential growth function.

Fig. 11

TGA analysis results for a C5, b F10-C5, and c F20-C5 at 28 days and 120 days, and d relationships between the strengths and mass loss

Energy-dispersive X-ray analysis

Figure 12a and b show the EDX analysis results of the C5, F10-C5, and F20-C5 samples at curing periods of 28 and 120 days, respectively. Primary elements included silicon (Si), aluminum (Al), and calcium (Ca) (Jamsawang et al. 2017a; Tongwei et al. 2014; Wang et al. 2013). The amounts of Al, Si, and Ca in F10-C5 are higher than those in C5 and F20-C5. Ca and Si’s presence indicates appropriate conditions for forming C-S–H gels (Al-Homidy et al. 2017; Bilondi et al. 2018; Li and Poon 2017). This characteristic reveals that F10-C5 exhibited the best C-S–H gel formation and highest bond strength. These results confirm the finding of the highest strengths and moduli for solidified sediment with fly ash–cement blend at an optimal fly ash content of 10% in “Destructive properties” The correlations of q_u versus Ca content and q_t versus Ca content are presented in Fig. 8, which can be expressed using a modified power equation. The strength generally increased in proportion with the Ca content. The correlations obtained from the microstructural analyses show that q_t ≈ 0.16q_u, which indicated that the destructive and microstructural experiments provide similar results.

Fig. 12

EDX analysis results for C5, F10-C5, and F20-C5 C5-FA0 at a 28 days and b 120 days, and c the relationship between the strengths and Ca content

Scanning electron microscopy

SEM was used to study the microstructure of the mixtures to determine the particulate morphology at select sampling points. Sampling was performed at the failure surface of UC test samples of each mix because such surfaces can accurately represent any local specificities in physicochemical characteristics within the region of interest. At least three points on the failure surface were selected for SEM analysis. The magnification of the SEM micrographs used in this study is 3500 ×.

Figure 13 illustrates the microstructural morphologies of specimens C5, F10-C5, and F20-C5. After a curing period of 28 days, the cement hydration products, e.g., C-S–H gels, were detected in C5, and rough surfaces of the sample were clearly noted, but some voids between sediment particles were also observed (Jamsawang et al. 2017a; Yoobanpot et al. 2017, 2018). After a curing period of 120 days, the additional C-S–H gels derived from the pozzolanic reaction were created. The C-S–H products filled the voids between sediment particles and formed a framework on the sediment aggregates. This C-S–H formation pattern contributes to densifying the sediment aggregates and enhancing interparticle bonds, increasing the overall strength.

Fig. 13

SEM images with 3500 times magnification for samples C5, F10-C5, and F20-C5 at 28 days and 120 days

The SEM images of F10-C5 and F20-C5 reveal three forms of fly ash particles: particles with a disintegrated surface, a smooth surface, and a covering surface with the C-S–H products. The first form represented reacted fly ash particles with Ca(OH)₂ in the pozzolanic reaction, creating additional C-S–H gels (Chindaprasirt et al. 2007; Yoobanpot et al. 2017). The second form showed unreacted fly ash particles, acting as filler materials to fill voids in the sediment aggregates. The third form indicated the presence of chemical reaction products surrounding the fly ash particles. The fly ash particles with smooth surfaces and with covering surfaces with the hydration products were spotted in the SEM images of F10-C5 and F20-C5 at 28 days of curing. The fly ash particles spread in the C-S–H framework and filled the voids, decreasing the porosity of the whole specimen.

The SEM photos of F10-C5 at a curing time of 120 days illustrate the fly ash particles with disintegrated surfaces and fully developed covering surfaces of C-S–H products derived from hydration and pozzolanic reactions. This configuration indicates a high level of hydration and pozzolanic responses. A hydration gel crust was routinely created as the curing period increased, tightly connecting sediment and adjacent fly ash particles. The C-S–H products filled the specimen pore voids, resulting in the integration and connection of sediment particles. Thus, the solidified samples exhibited smaller pores and denser microstructures than the unsolidified sediments (Fig. 1). Thus, the integration and connection of the solidified samples detected at the microscale improved the destructive and nondestructive properties of the unsolidified sediments at the macroscale (Kang et al. 2014).

The SEM photos for F20-C5 indicate that the fly ash particles were partly encased by C-S–H shells; some fly ash particles had smooth surfaces. These characteristics indicated that the hydration and pozzolanic reaction levels of F20-C5 were lower than those of F10-C5. The decrease in the number of pores in the solidified samples with curing time was induced by the development of C-S–H products, which caused increased internal bonds and filling of the void spaces between sediment particles (Yoobanpot et al. 2018). This phenomenon is the primary basis for destructive and nondestructive property developments, as discussed in Sects. 3.1 and 3.2. An enormous amount of C-S–H products was detected in F10-C5, followed by F20-C5 and C5. In contrast, the smallest pores were also observed in F10-C5, followed by F20-C5 and C5, for the described reasons. The SEM results proved that 10% fly ash–cement blend exhibited the most significant strengths and moduli and the lowest porosity (Jamsawang et al. 2017a).

Proposed strength prediction model

Based on microstructural analysis results, parameters affecting the q_u,D of the solidified sediment with fly ash-blended cement samples included C-S–H peak intensity, mass loss (m), Ca amount, and D (curing period). All influencing factors showed nonlinear correlations with the q_u values of the solidified sediments. General multiple linear regression analyses were performed to integrate these parameters. Multiple linear regression, also known simply as multiple regression, is a statistical technique that uses several explanatory variables to predict the outcome of a response variable. Multiple linear regression aims to model the linear relationship between the explanatory (independent) variables and response (dependent) variables. The formula for multiple linear regression is:

$$y={\beta }_{0}+{\beta }_{1}{x}_{1}+...+{\beta }_{\text{n}}{x}_{\text{n}}+\in$$

(3)

where:

y	=	the predicted value of the dependent variable
x1	=	the first independent variable
x2	=	the last independent variable
β0	=	the y-intercept (value of y when all other parameters are set to zero)
β1	=	the regression coefficient of the first independent variable
βn	=	the regression coefficient of the last independent variable
∈	=	model error

In multiple linear regression, three calculations are carried out to determine the best-fit line for each independent variable, including (1) the regression coefficients leading to the smallest overall model error, (2) the t-statistic of the overall model, and (3) the associated p-value. Then, the t-statistic and p-value are determined for each regression coefficient in the model. The regression results for reservoir sediment solidified with 0–20% fly ash blended with 5% cement are as follows:

$$\begin{aligned}q_{u,D} =& 0.3889(Ca) - 0.0005098 (C-S-H)\\& + 0.903(m) - 0.003814(D) + 3.2146\end{aligned}$$

(4)

where q_u,D is the solidified sediment’s strength with cement blend FA after D days of curing time. A comparison between the experimental observation and predicted values obtained from Eq. 4 is presented in Fig. 14. The observed q_u,D was very close to the proposed predicted q_u,D, with R² = 0.9993. The absolute error percentages of the q_u,D values were 0.8–2% and 0.07–0.19% for the curing periods of 28 and 120 days, respectively. The error for samples subjected to a short period was more significant than that for specimens under a long period since the short-term strengths of solidified samples were unstable and changed significantly among samples. This finding is in contrast to that of the long-term strengths, which developed consistently, reducing the associated prediction error of the model.

Fig. 14

Comparisons between the measured and predicted q_u(D) values based on multiple regression analysis

Role of fly ash in the cement stabilization mechanism

Using fly ash-blended cement was more effective than using sole cement or sole fly ash because of the tremendous amounts of C-S–H produced by both hydration and pozzolanic reactions owing to the presence of SiO₂ in fly ash. Employing sole cement creates C-S–H from the hydration reaction because of the absence of SiO₂ from the fly ash. Utilizing fly ash alone is the least effective approach due to the lack of CaO from cement, which caused the lowest CSH amounts. In this study, a fly ash content of 10% was verified as the most effective because it delivered the best strengths and moduli of the solidified sediments for all cement contents and curing periods. This content provided the most appropriate SiO₂ and CaOH₂ proportions to create the most extensive C-S–H products. Excess fly ash contents of 10% caused decreases in strengths and moduli. They blocked the cement grains and prevented the interaction between the cement and water, reducing the hydration reaction level and reducing the C-S–H products.

Conclusions

This work exhibits the destructive and nondestructive properties and microstructural characteristics of reservoir sediments solidified with a fly ash–cement blend. The raw sediment, which was classified as well-graded sand with silt, was stabilized with class C fly ash and type I ordinary Portland cement. Fly ash (5–20%) and cement (1.5–5%) by weight of dry sediment were mixed to improve the mechanical properties of the mixtures. Based on the destructive and nondestructive test results as well as microstructural characteristics, the following conclusions can be drawn:

• 1. The evolution of the destructive and nondestructive properties of the solidified sediments was similar for all mix proportions. The use of fly ash alone can improve the q_u values approximately 1.25–3.76 times greater than those of the unsolidified sediments. Exclusive cement and fly ash–cement blend can enhance the q_u values by approximately 1.84–9.13 and 1.38–8.43 times, respectively, relative to unsolidified sediments.

• 2. The ε_f and ν₀ values for the solidified sediments ranged from 2–3.2% and 0.255–0.346, respectively. High q_u values caused an increase in the brittleness level of the solidified samples, resulting in decreasing ε_f. However, the relationship between ν₀ and q_u was independent. ν₀ depended on the property of the natural sediment rather than that of the solidified sediments.

• 3. The q_u of the solidified sediments was used as a primary indicator to create the linear empirical relationships, which can be quickly utilized for estimating the other destructive and nondestructive properties. Relations of E₅₀ = 81q_u, M_r = 151q_u, E₀ = 33q_u, and G₀ = 10q_u for the solidified sediment were obtained in the current study.

• 4. The durability of solidified samples against WDC depended on the initial q_u. Using 6 WDCs is sufficient for solidified sediments because of the lower q_u(WDC) than the given values. The correlation between the strength and the number of WDCs can be utilized to evaluate the pavement lifetime using solidified sediments as road materials.

• 5. All microstructural analysis results prove that F10-C5 provided the best C-S–H contents compared to F20-C5 and C5, through various interpretations. The XRD and TGA results found the highest C-S–H peak intensities and maximum mass losses in F10-C5, respectively. The EDX analyses indicated the most significant weight proportion of Ca for F10-C5. The SEM images illustrate the enormous amounts of C-S–H products. The microscale findings confirm that the 10% fly ash–cement blend exhibited the best destructive and nondestructive characteristics at the macroscale.