Glass fiber reinforced Polyethylene pipes tested at soft ground
Presented at the 19th Plastic Pipes Conference PPXIX
Agricultural water is mainly supplied by waterway etc. laid on the ground. Recently, pipelines laid in the ground have been increasing from the viewpoints of water effective utilization, maintenance and pumping. Selection of the places in which the pipelines are laid is difficult, and the places are determined by planning routes and cost performance, and states of ground structure. Therefore, the pipelines are used in soften ground referred to as the peat soil in several cases. The peat generally means a material mainly formed of naturally accumulating putrefaction of hygrophyte while the decomposition is insufficient for many years under conditions of low temperature and high humid. Most of them are distributed in Hokkaido in Japan, but they are scattered also from Tohoku District to Kyushu District although in a small scale. The peatland as wide as about 2.000 km2 is reputedly distributed in Hokkaido, which corresponds to about 2,4 % of the total area of Hokkaido, or about 6 % of the plain area. In general, such a soften ground reputedly has about 1/4 to 1/6 of ground reaction in comparison with sand or sandy soil, from properties of low shear strength, high compressibility, and a high groundwater level, and differential settlement is caused in the pipelines. The fiberglass reinforced plastic mortar (FRPM) pipes having low density, light weight and inexpensiveness have been utilized in many cases, rather than the pipe materials having high density, such as steel pipes and ductile iron pipes against the differential settlement in Japan. However, it is reported that the FRPM pipes are flattened to the designed value or higher by vertical soil pressure in the pipe circumferential direction, leading to damage in the soften ground in which the ground reaction is unstable and accidents of causing water leak are never ended. Therefore, we conducted the following applicability tests using glass fiber reinforced polyethylene pipes.
1) Evaluation tests using a large soil tank. The differential settlement is forcibly reproduced in the state of applying a load equivalent to T14 and the internal pressure of 0.5 MPa.
2) Evaluation tests using on-site field.
Fig. 1: Large scale sand box
Evaluation using a large soil tank
The pipe was laid in a large soil tank having a width of 635 mm, a length of 1830 mm and a depth of 1080 mm shown in Fig. 1. After backfilling, the upper part was provided with forcible loading by using hydraulic jacks. Then, air springs arranged on the bottom part of the sample pipe were deflated to reproduce ground settlement. As the pipes, a glass fiber reinforced polyethylene pipe (inner diameter: 205 mm, pipe thickness: 8,5 mm) or a high density polyethylene pipe (inner diameter: 205 mm, pipe thickness: 11,5 mm) were applied, and Fig. 2 summarizes the test cases.
As the pipe thickness, ring stiffness values in the pipe circumferential direction as determined from Eq. (1) and Eq. (2) were adjusted to be substantially equivalent, and the tests were conducted. As backfilling soil, No. 6 to 7 mixed silica sand was used, and Table 2 show the properties. Backfilling was performed by setting a spreading thickness to 100 mm and finished to 400 mm in covering of the above of pipe to be 25% in relative density. After backfilling, the sample pipe was evaluated according to the following procedures. Both ends of the sample pipe were connected to flange pipes by using BUTT fused joints to construct the closed pipeline.
b) The whole upper part was provided with a vertical load of 55.9 kN/m2 by using hydraulic jacks.
c) Depressurizing the air springs arranged on the bottom surface of the sample pipe to cause sedimentation of the pipes at a sedimentation speed of 1 mm/min to simulate ground settlement.
d) The operation was performed to a sedimentation level of 30 mm.
Table 3 shows the measurement conditions.
A settlement level of the pipe was measured by forming a structure in which a wire displacement gauge was protected with a cylindrical bar to prevent the wire displacement gauge from being interfered with sand. The displacement of the pipe was measured by attaching strain gauges in positions divided into 18 in the pipe axial direction as shown in Fig. 2 and in positions divided into 24 at a maximum for 5 sections in the pipe circumferential direction as shown in Fig. 3.
The internal pressure was measured thereon in a state of holding the water pressure in the sample pipe by using a pressure gauge attached at the end of the sample pipe. The internal water pressure was adjusted to 0.5 MPa. As the load, the magnitude of the vertical load by the hydraulic jacks was measured. The loads in three places in the upper part by the hydraulic jacks were regarded as uniformly distributed loads and adjusted to 55.9 kN/m2. The magnitude of the load was expressed in terms of values corresponding to the prescribed standard strength values (T-14) for pavements and agricultural roads by the Road Structure Ordinance.Fig. 2: Strain gauge for axial direction
Field test 2 (Evaluation
using on-site field)
Evaluation using the test field was performed in Nishi-bibaicho, Bibai, Hokkaido, Japan, in which in the peat soil is spread. According to the boring exploration conducted in the test field, the N-value was 0 to 1 down to a depth of 6 m.
As the sample pipes, three glass fiber reinforced polyethylene pipes (inner diameter: 610 mm, pipe thickness: 20 mm) having a length of 11 m were arranged and jointed by electrofusion (EF joint.) to form an integrated pipeline having a length of 33 m. Both ends were connected to flange pipes to construct the closed pipeline. BUTT fused joints were arranged in the central part of the pipe and EF joint parts were provided in positions of 5 m from the both ends, respectively. Wooden piers each having a length of 7.2 m were buried below the EF joint parts, and the supporting points for suppressing the vertical displacement were provided. The central part of the pipeline was provided with the forcible loading by using the hydraulic jacks in simulation of the flat plate loading test method. Forcible loading, unloading and leaving to stand for 12 hours were repeated three times to verify the behavior of the respective pipes. Fig. 4 shows a schematic diagram of the test field.
Design conditions of the pipes
The calculation expression for the pipe thickness determined from the internal and external pressure acting on the sample pipe was determined based on Eq. (3) according to the Land Improvement Business Plan Design Criterion and Commentary, issued by the Japanese Society of Irrigation, Drainage and Rural Engineering.
Excavation and backfilling conditions
As excavation, the peat ground was excavated in sections shown in Fig. 5, and the sample pipe was laid. Table.4 shows material properties of the on-site ground. An excavation gradient was adjusted to 1 : 0.3 according to the properties of cohesive soil. Backfilling was finished from the above of pipe up to 1200 mm for every 300 mm by using the peat soil by using a wooden ram and the final layer was finished to be 1400 mm in height from a height of 200 mm. The ground density and the ground reaction coefficient on the side surface of pipe were 1.06 g/cm3 and 250 kN/m2, respectively. After backfilling, the test was performed according to the following procedures. Both ends of the sample pipe were connected to flange pipes by using BUTT fusion to construct the closed pipeline.
a) Conditions after backfilling
b) Loading of an extra load of 200 kN (loading speed: 50 mm/min)
c) Unloading (unloading speed: 1000 mm/min)
d) Repeating of loading and unloading (repeating number of times: three times)
e) Leaving to stand for 12 hours Airtight test.
Fig. 4: Field test 2
The sedimentation levels were measured by installing the gauges vertically in 5 points in A to E sections, the pipe axial strain was measured by attaching uniaxial strain gauges in 40 points for every 250 mm from the center of the bottom part of pipe on the pipe outer surface in the pipe axial direction. On the other hand, the pipe circumferential strain was measured by attaching pipe circumferential strain gauges to 48 points formed by uniformly dividing A to C sections into 16 (× 3 sections) at 22.5° in the pipe circumferential direction. The soil pressure was measured by installing soil pressure gauges on the above of pipe, the side of pipe and the bottom of pipe in A to C sections. However, no gauges were installed on the top part of pipe in the C section from influence of loading jigs. As the internal pressure, water was injected from the flange at the pipe end, and pressurized up to an internal water pressure of 0.5 MPa by using the plunger pump. Presence or absence of leak was confirmed using the pressure gauge.
Evaluation results using a large soil tank
1) Pipe axial strain
Fig. 6 shows changes in strain of PE-sGF pipes caused in the pipe axial direction, in which a black line in the figure indicates strain immediately after backfilling to the above of pipe, red lines indicate strain at starting settlement, and blue lines indicate strain when settlement progressed to 30 mm. Similarly, Fig. 7 shows changes in strain of PE pipes caused in the pipe axial direction. In all the cases, the pipes are found to be deflected downward because compressive strain is caused on the above of pipe and tensile strain is caused in the bottom of pipe. Moreover, the strain caused in the bottom of pipe is larger than the strain caused on the above of pipe. The reason is considered that compressive strength is larger than tensile strength generally in the polyethylene material. PE-sGF reinforced with glass fibers is found to have the same trend. When unloading was performed after completion of measurement, the sample pipes were confirmed to be restored and returned to the original states in all the cases, which is considered to be resulted from exhibiting elastic response. On the other hand, as a result of comparison of PE-sGF with PE, although the ring stiffness in the pipe axial direction is substantially equivalent to each other, the tensile strain and also the compressive strain are found to be larger in PE.
Table 6: Comparison of deflection ratio and maximum strain
2) Pipe circumferential strain
Fig. 8 shows changes in strain of PE-sGF pipes caused in the pipe circumferential direction, in which a black line in the figure indicates strain immediately after backfilling to the above of pipe, red lines indicate strain at starting settlement, and blue lines indicates strain when settlement progressed to 30 mm. Similarly, Fig. 9 shows changes in strain of PE pipes caused in the pipe circumferential direction. In all the cases, it is found that the compression is caused on the above of pipe and tension is caused on the side of pipe. The trend in which the compression on the above of pipe becomes larger than the tension in the bottom of pipe was found in PE, which is considered to be resulted from causing further significant elliptic deformation of PE in comparison with PE-sGF.
Table.6 summarizes comparison of deflection rates and the maximum strain depending on differences in pipe types and loading states of internal water pressure. Although the ring stiffness is equivalent to each other, both the deflection rates and the maximum strain are found to be larger in PE pipes, which is assumed that the deformation (ground followability) in the axial direction influences the pipe circumferential strain. The deflection rate of PE-GF under no pressure loading was about a half of the rate of PE pipes, and PE-GF is found to keep a circular shape even though PE-GF is deformed in the pipe axis.
Field test 2 (Evaluation
results using on-site field)
1) Vertical settlement level
Fig. 11 shows transitions of vertical settlement levels after backfilling. A line of ◊ shown in Fig. 10 shows a state immediately after backfilling, a line of □ shows a state after elapse of 10 hours from backfilling, a line of Δ shows a state after further 2 days, and a line of ○ shows the settlement levels when providing the pipe with the extra load of 200 kN. A to E shown in the X axis indicate each section in Fig. 4, in which the wooden piers are fixed in A and E sections in the vertical direction from the bottom of pipe. The reason why the sedimentation progresses in A and E sections after 10 hours from backfilling is considered that the ground is softer than the circumference under an influence of performing joint excavation of the vicinity ground because the sections are connected with the Electrofusion joints.
The bottom part of pipe is found to reach the wooden piers because the sedimentation of A and E sections stopped at about -60 mm. An aspect in which the C section in the central part causes sedimentation by a concentrated load can be observed.
Fig.11 shows vertical settlement levels under forcible loading, in which a line of ◊ in the figure shows a state before loading, a line of ○ line shows the settlement levels when the extra load of 200 kN was loaded, and a state immediately after unloading thereafter was shown by a line of ×, and an aspect when 12 hours further elapsed from unloading is shown by a line of *. No change was observed in the displacement in A and E sections because the wooden piers are installed in the bottom part in each section. The pipes are found to be substantially restored when 12 hours elapsed after unloading. The results obtained by repeating this process 3 times were confirmed to be the same for all.
2) Pipe axial strain
Fig.12 shows changes in pipe axial strain in the bottom part of pipe by forcible loading, in which a line of × in the figure shows a state before loading, and a line of ○ shows strain levels when the extra load of 200 kN was loaded, and an aspect after 10 hours from unloading thereafter was shown by a line of □, and an aspect when two days further elapsed from unloading was shown by a line of Δ. The strain in the bottom part of pipe was output on the tensile side in the central part forcibly loaded, and the compressive strain was confirmed in A and E sections served as the supporting point. The strain was 0.2 % at a maximum in the central part of the bottom of pipe.
3) Pipe circumferential strain
Fig. 13 to Fig. 15 show changes in the pipe circumferential strain by forcible loading. Fig. 13 shows the A section and Fig. 14 shows the B section, but no influence of forcible loading is found. On the other hand, Fig. 15 shows the C section, in which a blue line in the figure shows a state before loading, a red line shows strain levels when the extra load of 200 kN was loaded, and a state immediately after unloading thereafter was shown by a green line, and an aspect when 12 hours further elapsed after unloading was shown by a purple line. The pipes are deformed in such a manner that the deformation in the places in contact with the loading jig as shown in Fig. 16 is transferred in the C section, but the pipes are found to be substantially restored after unloading.
Fig. 12: Actual strain of axial direction
4) Soil pressure
Fig. 17 to Fig. 19 show measurement results of soil pressure. Theoretical values of soil pressure on the above of pipe were determined by the vertical soil pressure equation (Eq. (4)). Theoretical values of soil pressure on the side of pipe were determined by Eq. (5) based on horizontal deflection levels determined by Marston-Spangler's Equation (Eq. (6)). Table.7 shows the properties of the circumferential ground. Fig. 17 shows actual values of soil pressure on the above of pipe and on the bottom of pipe in A section and theoretical values of soil pressure on the above of pipe in A section. The actual values of the initial soil pressure are consistent with the theoretical values, and no particular displacement is found. In and after the third times, the actual values are found to be reduced in comparison with the theoretical values. The reason is considered that only deformation of the pipe was restored while the circumferential ground was settled in and after the first time loading. Fig. 18 shows actual values and theoretical values of soil pressure on the side of pipe in A section. The soil pressure caused on the side of pipe is found to be substantially consistent with the theoretical values. The reason is considered that no local deformation is caused on the pipe. Fig. 19 shows actual values of soil pressure on the side of pipe and on the bottom of pipe in C section. No gauges are installed on the above of pipe under an influence from the loading jig shown in Fig. 16. The soil pressure on the bottom part of pipe was confirmed to be 70 to 80 kPa by the forcible loading. Meanwhile, the soil pressure on the side part of pipe was confirmed to be constant, irrespective of the change. These results are considered to be caused by a small change of the pipe in the sectional direction. From the phenomenon in which the soil pressure on the bottom of pipe returns to the original in 24 hour, it is estimated that voids are gradually lost in the circumferential ground, while the sample pipe was quickly restored on the contrary.
(1) Evaluation results using large soil tank
The following results were found:
• Although the ring stiffness was equivalent, both the deflection rate and the maximum strain were larger in the PE pipes.
• The pipe circumferential strain was output in a level smaller in PE-sGF than the pipe axial strain.
• The tensile strains was larger than the compressive strain in PE-sGF.
• PE-sGF was found to follow the ground displacement, while the shape was maintained in the pipe circumferential direction.
(2) Evaluation results using the test field
The following results were found.
• The pipes followed the change in the ground in the state of being buried, and even if bending was caused in the pipe axial direction, the shape in the pipe circumferential direction was maintained.
• The pipes exhibited the elastic response even in the peat ground.
• With regard to the restoration speed of the PE-sGF pipes, the restoration was responded at a speed faster than the restoration in the peat ground.
Mitsuaki TOKIYOSHI High Stiffness Polyethylene Pipes Association Tokyo, Japan; Gentaro TAKAHARA Dainippon plastics Osaka, Japan; Joji HINOBAYASHI Dainippon plastics Osaka, Japan; Toshinori KAWABATA Kobe University Hyogo, Japan
Takashi KURIYAMA Yamagata University Yamagata, Japan.
The full report can be requested from Dainippon Plastics, Japan.