The Influence of Water Content on the Settlement Behaviour of Polypropylene Fibre-Reinforced Dredged Marine Soil (DMS)

The use of fiber as soil reinforcement is not new in civil engineering field. In the earlier times, rice straw was mixed together with mud or clay to produce construction materials such as brick and concrete. Conventional concrete mix without fiber tends to exhibit brittleness behaviour. Hence, there is a growing attention on using current reinforcement materials such as steel, polypropylene and glass fibers. It is reported that fiber in concrete provide bridging effect, which transfer and distribute load evenly, thus increasing ductility. Now, similar concept of fiber inclusion in concrete can be applied to the case of problematic soil. The addition of chemical additives such as cement in soil resulted with stiffness and brittleness. As solution, numerous studies have shown that the fiber inclusion in soil have increased strength, permeability and ductility. Due to the many studies of fibre-reinforced soil related to its shear strength, the present study will investigate the compressibility behaviour of the fibre-reinforced soil through oedometer test. In this study, the dredged marine soil (DMS) was mixed together with 0.25, 0.5, 0.75 and 1 % of polypropylene (PP) fiber. Two conditions of soil, namely high water content (1.40LL) and low water content (0.90LL) were tested. Samples with 0.90LL water content show great reduction of settlement than samples with 1.40LL water content. The outcome of this study will suggest the beneficial reuse of DMS for engineering application such as backfill material, land reclamation or clay liner for landfills.


Introduction
Maintenance dredging is necessary to be carried out constantly due to the reoccurring siltation which may affect the waterways for ships to the jetty. This type of dredging has the most volume of dredged marine soil (DMS) as compared to the other types of dredging projects, namely capital and remedial dredging. In Malaysia, dumping DMS in the open sea is a common practice. This activity has caused short and long terms of negative impacts, such as turbidity and disruption of marine habitats respectively [1,2]. Numerous researchers have extensively studied the beneficial reuse of DMS mainly in construction industry such as capping material for landfill, paving block, blended cement, construction brick and pavement base material [3][4][5][6][7][8]. By reusing DMS, it will provide an alternate management of DMS and act as preventive measure to resolve the aforementioned negative impacts. However as reported by many researchers, DMS has high water content, high compressibility, low permeability, low bearing capacity and susceptible to significant consolidation settlement [5,[9][10][11][12]. Hence, soil improvement is required to enhance the poor soil properties of DMS. Among the soil improvement technologies outlined by [13] are soil replacement, densification, reinforcement, admixture stabilization and thermal stabilization. Soil reinforcement by using fiber is gaining attention nowadays, although the use of fiber in engineering application is not new. Over the years, natural organic fiber such as rice straw was mixed together in mud or clay brick for reinforcement [14]. As the technology advances, synthetic fibers such as steel, polypropylene and glass fibers were regularly used in concrete mixture [15][16][17]. It is reported that the application of fiber in concrete has provided the bridging effect which distribute the load evenly and increase the ductility of the concrete. The similar concept of fiber inclusion in concrete can also be apply to the case of problematic soil. Many studies have proven the increment of strength, permeability, bearing capacity, and reduction of crack and settlement of fiber-reinforced soil [18][19][20][21][22][23]. Despite the various types of material and experimental procedure used in these studies, polypropylene fiber (PP) was most favored in soil reinforcement. According to [24][25][26][27], reinforcing soil using PP fiber is an effective approach in term of cost reduction, whereby only small amount of PP fiber is required in the soil mixture (less than 5 % by dry weight of soil). Most of the currently published and available literature deals with soil with water content below its liquid limit [28][29][30][31][32]. However, only a limited number of studies have been carried out on soil with high water content that exceed its liquid limit, particularly DMS [33,34]. It should also be pointed out that there has not been a general explanation on the influence of water content on the consolidation settlement behavior of the PP fiber-reinforced soil through oedometer test. Hence, the objective of the current study was to observe the behavior of fiber-reinforced DMS with various water content in terms of consolidation settlement.

Materials and Methods
The soil used in this study was retrieved from Kuala Mersing, Johor. As displayed in Figure 1, the sample was obtained using a clamshell dredger from a depth of 4 m at sea level and then sealed securely inside air-tight containers to avoid any moisture loss. Based on the standard [35], the DMS was classified as low plasticity silt (ML). The geotechnical properties of DMS are listed in Table 1. Commercially available fibrillated PP fiber was used in the experiment as seen in Figure 2. Tabulated in Table 2 is the main properties of PP fiber as provided by the manufacturer.    The particle size distribution of DMS is given in Figure 3. Based on the distribution, the DMS was composed of 13.12 % of sand, 68.72 % of silt and 18.04 % of clay, by which categorized the DMS as silty clay. As mentioned from the previous studies, the percentage of PP fiber was kept low (< 1 % by dry weight of soil) to ensure better workability, increase effectiveness of pulling resistance, prevent clumping of fiber and disallow the formation of pocket density which would cause reduction of strength [30,33,34,36]. The index value of water content and liquid limit (WC/LL) of DMS in this study is 1.40. By referring to Table 3, most of the soil tested were below 1.00 times its liquid limit (<1.00LL). Only few researchers dealt with soil 1.00 times more than its liquid limit (>1.00LL), which is essential for DMS that is known to contain high amount of water than its liquid limit. Furthermore, an excessive amount of water in soil tends to delay the effectiveness of soil improvement and unable to withstand any significant load [9]. Hence, the soil samples were mixed together with 0.25, 0.50, 0.75 and 1.00 % of fibrillated PP fiber by dry weight of soil in two water content conditions, namely 0.90LL and 1.40LL. All of the samples were mixed homogeneously using a conventional kitchen mixer for 3-5 minutes. The samples were then placed inside the oedometer cell rings and compacted layer by layer to prevent the fibers to clump together. The experimental procedures of oedometer test were accordance to the standard [38]. The sequential loads were 12.5, 25, 50, 100, 200, 400 and 800 kPa.

Results Analysis and Discussions
Note that log time-strain plot is the preliminary result from oedometer test. Figure 4   [40] has studied the influence of water content on dredged soil which contained high water content at low effective stress. In his study, he stated that soil has its own and unique structure. The differences of soil structure were affected by initial water content. Hence, each soil was able to sustain certain extent of external load. If the external load is less than the maximum load of soil, the soil will still maintain its original structure. However, if the external load is greater than the maximum load, the structure will deconstruct and the water content decreases as the water will dissipate out from the soil under acting pressure. From the figures, soil samples with high water content have high strain values. A possible explanation for this is that high water content soil contains high void ratio. There is similarity between the influence of water content on void ratio in this study and that described by [41]. As for low-void soil samples, the strain values are less than the highvoid samples. This is due to the formation of soil structure as the induced stress begin to compressed the soil. The addition of fiber in the soil fabric provide more structure in soil. Hence, the reduction of settlement in fiber-reinforced soil samples. Figure 6 shows a simple diagram of 0.90LL and 1.40LL soil samples with and without 0.50 PP fiber. However, further research should be done to investigate the microstructure of the soil samples. There are two types of compression curves plot, namely log stress (σ' v )-void ratio (e) and log stress (σ' v )-strain (ε) plots. The most preferable method is the log stress-strain plot since the purpose of this plot is to determine the stress-strain properties of the soil [42]. By referring to the compression curves in Figure 7, all of the soil samples exhibited direct normal consolidation line (NCL) which is normal for alluvial soil deposit particularly silt and clay sized particles soil. The present findings seem to be consistent with other research which found that clay or silt with high water content tend to produce direct NCL [43,44]. It is also possible that the direct NCL of the soil samples were influenced by the initial load (12.5 kPa), which was too heavy for soil samples to support. A further study using lower initial load (5 kPa) for all soil samples is therefore suggested. The pre-consolidation pressure (σ p ) can be determine if the initial curve line is visible by using lower initial load of 5 kPa.
What is interesting in this data is that most of the compression curves are almost linear with each other. This finding indicates that all of the soil samples has similar value of compression index (c c ). The only difference in this data is the magnitude of settlement. It is encouraging to compare this figure with that found by [44][45][46] who reported that the compression curves (highlighted at NCL) were parallel to the untreated soil. It can therefore be assumed that the addition of PP fiber in DMS has yet to form a firm soil structure, unlike chemically-treated soils that can develop firm soil structure once the soil and chemical reacted. Hence, future work by adding chemical additive such as cement in fiber-reinforced DMS is suggested. At water content of 0.90LL, fiber-reinforced DMS has reduced the settlement effectively than untreated DMS. Samples 0.25PP and 0.75PP at water content of 1.40LL show increasing settlements. The observed increase in settlement could be attributed by the high water content in the admixed soil. This theory is also depicted in Figure 6. It can thus be suggested that high water content in DMS tend to lubricate the interfacial shear strength of the materials and also decrease the effectiveness of the bridging effect between fiber and soil. This result is consistent with [24], whereby high water content has affected the tensile strength of reinforced DMS. Moreover, samples of 0.25PP, 0.75PP and 1.00LL are less effective than 0.50LL regardless of the water content conditions. This result may be explained by the fact that the fiber content is too low and too high to produce homogeneous mixtures. Without the right consistency and homogeneity, the bridging effect between fiber and soil could not be initiated [33,36]. Interestingly, sample 0.50PP provided the lowest settlement than the other percentages of PP fiber in 0.90LL and 1.40LL water content conditions. A possible explanation for this might be that 0.50 % of PP fiber is the threshold content for DMS in this study. This finding corroborates the results of [29,31,33,34], by which 0.50 % of PP fiber in soil (in the range of 0.70LL-1.60LL water content) has improved the unconfined compressive strength (UCS) effectively. These researchers also agreed that 0.50 % of PP is the optimum fiber content and soil samples with PP fiber more than 0.50 % would reduce the effectiveness of compressive strength.
As mentioned, the compression curves of all samples in both water content conditions display immediate NCL. It is possible that the initial loading of 12.5 kPa is too much for the DMS to sustain and to obtain the pre-yield curve, thus incapable to estimate the effective yield stress (σ y '). It can be suggested the use of load below 12.5 kPa for high water content DMS in the future study. Constraint modulus (D) is an inverted value of volume of compressibility (m v ), which measures the soil stiffness and provides simple estimation of soil settlement [47]. Figure 8 shows the log stress (σ' v )-constraint modulus (D) plot. This plot is located in the elastic region. Elastic is a condition when the load is released, the soil will be back in its origin form. The plot also shows a strain-hardening feature which may be due to the loosely packed soil begin to move closer together as the stress applied. The trend of D value is increasing along with the incremental stress. The higher the D value, the stiffer the soil samples. From the plot, soil sample with 0.50 % of PP fiber at water content of 0.90LL is stiffer than the other samples. This result corresponds well with the compressibility results. Therefore, it is plausible to remark that 0.50 % of PP fiber is the fitting percentage for this type of soil in both 0.90LL and 1.40LL water content conditions.

Conclusions
This study has investigated the influence of water content on the settlement behavior of fiber-reinforced DMS with various fiber contents. Series of samples were tested using oedometer test and the following conclusions can be drawn. Fiber-reinforced DMS with 1.40LL of water content shows higher settlement as compared to fiber-reinforced DMS with 0.90LL of water content. High water content in DMS tend to lubricate the interfacial surface of the materials and reduce the effectiveness of bridging effect between DMS and fiber. The optimum PP fiber in DMS is 0.50 %, regardless of the water content conditions (0.90LL and 1.40LL). DMS admixed with 0.25, 0.75 and 1.00 % of PP fibers in both water content conditions have resulted with less effective settlement reduction than admixed DMS with 0.50 % PP which probably affected by the homogeneity of the soil samples. The constraint modulus (D) or soil stiffness is observable to the soil sample with water content of 0.90LL.