Effect of fibre content on the geotechnical properties of peat

The peat soil tested in this study was collected from the peatland in the north-eastern Poland, 250 km north-east of Warsaw. The area under consideration is located in the lake region within the river basin. Figure 1 shows the occurrence of peatlands in Poland and the location of the study area. The subsoil profile of considered peatland consists of peat with a thickness of about 1–7 m. Below the peat layer, glacial sediments, mainly in the form of silt and sand, are deposited. The water level is at a depth from 0.1 to 1.5 m below the ground surface.

Figure 1

The peat was sampled with thin wall cylinders with a diameter of 70 mm. The cylinders were pressed into the subsoil in a vertical direction. Samples were collected from eight different depths, 1.0, 1.4, 1.9, 2.3, 3.1, 3.6, 4.7 and 5.0 m, below the ground surface, over an area of approximately 2500 m². A total of 40 samples were tested. The results presented in the study are the average test results for each group of peat samples collected from a certain depth.

The bulk and particle densities, natural water content, organic content, initial void ratio and the degree of decomposition were investigated as the physical properties of peat. Tests were performed on the samples with natural water content and natural structure. The bulk and particle densities and the natural water content were determined in accordance with the international standards (ISO 17892-1:2014, ISO 17892-2:2014, ISO 17892-3:2015). The determination of the water content was carried out according to the procedure used commonly in soil testing. The organic content of peat was determined using the method of loss on ignition (LOI) in accordance with the European standard (EN 15935:2012). The method is widely used by researchers and consists of strongly heating the oven-dried soil sample at a specified temperature (Heiri et al., 2001; Hoogsteen et al., 2015). The temperature of 550°C causes combustion of organic matter to ash (Heiri et al., 2001). When the natural water content of peat w_n is known, the organic content I_om can be predicted using the equation given below (Huat et al., 2009): (1) Iom=0.0592wn+54.34 {I_{om}} = 0.0592{w_n} + 54.34 where w_n is the natural water content in percentage.

The degree of decomposition of peat was determined using the von Post scale. Table 1 shows the categories of peat based on the squeezing and visual tests results (Landva and Pheeney, 1980).

The consolidation and compressibility characteristics of peat were determined in an oedometer by incremental loading. Tests were carried out on cylindrical samples with an initial height of 20 mm and a diameter 63.5 mm using a set of oedometers with automatic registration of displacement sensor readings presented in Figure 2. The peat samples were tested in non-deformable rings with top and bottom porous stones. Tests were performed in accordance with the international standard (ISO 17892-5:2017) at different effective vertical stresses σ_v′: 15, 32, 64, 96, 128 and 256 kPa. The maximum stress applied to the sample depended on the peat compressibility and the technical capabilities of the laboratory equipment.

Figure 2

With the vertical stress increment, the height of the sample and the void ratio of soil decrease. When the initial void ratio e₀ is known, the void ratio e at any stage of the oedometer test can be calculated from the equation (Head, 1994) (2) e=e0−(ΔH/H0)⋅(1+e0)=e0−εv⋅(1+e0) e = {e_0} - \left( {\Delta H/{H_0}} \right) \cdot \left( {1 + {e_0}} \right) = {e_0} - {\varepsilon _v} \cdot \left( {1 + {e_0}} \right) where ΔH is the change in the sample height as a result of the application of stress, H₀ is the initial height of the sample and ɛ_v is the vertical strain.

Based on the oedometer tests, the constrained modulus, compression and secondary compression indexes of peat were determined. The constrained modulus was determined by measuring the sample height changes under an applied stress. The constrained modulus E_oed can, therefore, be calculated from the formula (3) Eoed=Δσv′/εv=(Δσv′H)/ΔH {E_{oed}} = \Delta \sigma _v^\prime /{\varepsilon _v} = \left( {\Delta \sigma _v^\prime H} \right)/\Delta H where Δσ_v′ is the effective vertical stress increment and H is the height of the sample before the stress increment.

The compression index C_c is one of the most important parameters in the analysis of primary consolidation of soil. It is used to predict the foundation settlements. The parameter describes the change in the void ratio Δe as a function of the change in vertical effective stress plotted in the logarithmic scale Δlog σ_v′, C_c = Δe/Δlog σ_v′.

The number of correlations between the compression index and other physical parameters of soil (water content, void ratio, liquid or plastic limit) is known. The following dependencies can be used to predict the compression index in peat and other organic soils: C_c = 0.01w_n (Mesri and Ajlouni, 2007), C_c = 0.0115w_n (Moran et al., 1958), and C_c = 0.35·(e₀ − 0.5) (Hough, 1957).

The secondary compression index C_α is often more significant in peat than in other soils because of the relatively short duration of the primary consolidation and the long-term secondary consolidation of peat deposits (Mesri et al., 1997). It can be defined as the tangential slope of the e □ log t curve at any particular time t after the end of primary consolidation. It is directly related to the compression index (Mesri and Castro, 1987; Mesri and Godlewski, 1977) and stress (Jiang et al., 2020). The secondary compression index can be determined from the equation (4) Cα=(Δe/Δlogt)=(Δε/Δlogt)(1+e0) {C_\alpha } = \left( {\Delta e/\Delta \log t} \right) = \left( {\Delta \varepsilon /\Delta \log t} \right)\left( {1 + {e_0}} \right)

Table 2 shows the physical properties of the tested peat. Based on the degree of decomposition, peat from the peatland in north-eastern Poland was divided into fibric, hemic and sapric. Peat from a depth of 3.1 m, designated as P1, was classified as fibric, peat from the depths 1.9, 2.3, 3.6, 4.7 and 5.0 m below the ground surface (peat P2–P6) was classified as as hemic and peat designated as P7 and P8 from the depths 1.0 and 1.4 m was classified as sapric.

It can be observed in Table 2 that the lowest bulk density (ρ = 1.010 g/cm³) was estimated for fibric peat, whereas it was the highest (ρ = 1.112 and 1.127 g/cm³) for sapric peat. The bulk density of hemic peat varied from 1.029 to 1.039 g/cm³. The estimated values were within the range reported in the literature (ρ = 0.80–1.43 g/cm³) (Huat et al., 2011b; Rahgozar and Saberian, 2016; Wong et al., 2008). However, even lower values of bulk density have been reported by Zaccone et al. (2017, 2018). The particle density of the tested peat was determined to be 1.443–1.530 g/cm³, whereas the particle density presented in the literature varies from 1.23 to 1.64 g/cm³ (Boylan et al., 2008; Rahgozar and Saberian, 2016; Wong et al., 2008). As expected, the fibric peat had the highest natural water content (w_n = 828.3%), whereas sapric had the lowest water content (w_n = 236.5% and 320.2%). The water content of hemic peat was in the range of 335.7%–465.4%. The water content of peat presented in the literature has a wide range of 200%–1950% (Mesri and Ajlouni, 2007; Rahgozar and Saberian, 2016; Wong et al., 2008, 2009). The organic content of the tested peat varied from 83.42% to 94.86% and was within the range presented in the literature (I_om = 65.00%–99.06%) (Lechowicz et al., 2018; Mesri and Ajlouni, 2007; Rahgozar and Saberian, 2016; Wong et al., 2008). The highest value of organic content was determined for fibric peat, which had the highest natural water content. Figure 3 shows the I_om–w_n dependence.

Figure 3

It can be seen in Figure 3 that the organic content increased with the increase in the natural water content. However, the tested peat was characterised by a lower variability of organic content depending on the natural water content than the results obtained from Equation (1) (Huat et al., 2009). An empirical correlation between the organic and water contents of the tested peat was established (Fig. 3). The coefficient of determination R² of the linear regression fit was relatively low (R² = 0.442). The prediction of the organic content based on the water content using empirical correlations should be made cautiously.

It can be observed in Table 2 that the highest initial void ratio (e₀ = 12.26) was determined for fibric peat, whereas it was the lowest (e₀ = 3.399 and 4.638) for sapric peat. For hemic peat, the e₀ values ranged from 5.156 to 6.969. The initial void ratio of hemic and sapric peat was lower than that reported by researchers (e₀ = 7.28–14.2) (Mesri and Ajlouni, 2007; Rahgozar and Saberian, 2016; Wong et al., 2008). This may be connected with the high degree of decomposition of the peat soil investigated in this study.

Based on the results presented in Table 2, it can be concluded that the initial void ratio had an influence on the bulk density and the natural water content of peat. The ρ □ e₀ and w_n □ e₀ dependences are shown in Figures 4 and 5.

Figure 4

It can be observed in Figure 4 that with the increase in void ratio (volume of voids), the bulk density of peat decreased. An empirical correlation between the bulk density and the initial void ratio was established (Fig. 4). The R² coefficient obtained for the regression fit line was 0.518. As expected, the natural water content of saturated peat increased with the increase in initial void ratio. The correlation between the w_n and e₀ parameters was determined and is shown in Figure 5. The coefficient of determination of the regression line estimated for w_n □ e₀ dependence was very high (R² = 0.993); thus, the prediction of the natural water content based on the initial void ratio of the tested peat would be possible with a high accuracy.

Figure 5

Figure 6 shows the relation between vertical strain and effective vertical stress for fibric, hemic and sapric peat for the primary loading and unloading phases. Due to the high compressibility of the tested soil and the technical capabilities of the laboratory equipment, the maximum vertical stress applied to fibric, hemic and sapric peat was 96, 128 and 256 kPa, respectively.

Figure 6

The biggest change in height of the sample caused by stress increment was determined for fibric peat, whereas it was the lowest for sapric peat (Fig. 6). It can be observed in Figure 6 that the vertical strain during the primary loading phase obtained for fibric peat, depending on the stress level, was approximately two to four times greater than it was determined for sapric peat. For hemic peat, the difference was not so significant.

Based on the results of the oedometer tests, the constrained modulus at different vertical stresses was obtained (Table 3).

The highest values of the constrained modulus were determined for sapric peat, whereas they were the lowest in fibric peat. It can be observed in Table 3 that the constrained modulus of sapric peat was even several times higher than the constrained modulus of fibric and sapric peat. The difference decreased with the increase in vertical stress.

For a complete consolidation analysis, the vertical strain should be related to time. The ɛ_v □ log t dependencies used for determining the secondary compression index are shown in Figure 7.

Figure 7

Figure 7 also shows the highest vertical strain and, therefore, the biggest compressibility of fibric peat in comparison to hemic and sapric peat.

The relations between void ratio and effective vertical stress are shown in Figure 8. The e values were calculated from Equation (2). Based on the results presented in Figure 8a, the compression index of fibric, sapric and hemic peat was determined.

Figure 8

The biggest change in the void ratio with the stress increment was determined for fibric peat, whereas it was the lowest for sapric peat (Fig. 8). The percentage decrease in the e values in relation to e₀ for the primary loading phase at vertical stress up to 64 kPa was estimated to be 38% for fibric peat, 32% for hemic peat and 22% for sapric peat. The σ_v′ □ e dependence for fibric peat shown in Figure 8b is very close to that reported in the literature, σ_v′ = (40/e)^2.7 (Mesri and Ajlouni, 2007). Due to the low void ratio of hemic and sapric peat, the relation σ_v′ □ e estimated for them is slightly different, but still comparable to the values presented by researchers. The empirical correlations between vertical stress and void ratio along with R² coefficients were established and are presented in Figure 8b.

The values of compression index are shown in Figure 9. They were related to the natural water content and initial void ratio.

Figure 9

It can be observed in Figure 9 that C_c □ w_n and C_c □ e₀ correlations were established with a high reliability (R² = 0.973 and 0.966, respectively). The C_c value increased with the increase in w_n and e₀ values. The compression indexes of the tested peat could be predicted based on the natural water content or the initial void ratio from the equations: C_c = 0.007w_n and C_c = 0.67·(e₀ – 0.5). The obtained compression index is comparable to the values presented by researchers (C_c = 0.2–10) (Hough, 1957; Mesri and Ajlouni, 2007; Moran et al., 1958).

The secondary compression index of fibric, hemic and sapric peat and the C_α/C_c correlation are shown in Figure 10. The C_α values were determined for each level of vertical stress.

Figure 10

With a medium reliability (R² = 0.697), a relation between the secondary compression index and the compression index was established, C_α/C_c = 0.07. The determined C_α/C_c ratio for the peat is the same as that reported in the literature (C_α/C_c = 0.06 ± 0.01) (Mesri, 1986; Mesri et al., 1997; Samson and La Rochelle, 1972).