2025 Volume 66 Issue 1 Pages 31-36
As ballast on railway tracks is compacted and fragmented, settlement of the track tends to occur even after tamping, which leads to a need for more frequent maintenance. The basic measure to reduce maintenance frequency of replacing old ballast with new ballast, is costly. This calls for a low-cost method to reduce ballast settlement without need for ballast replacement. Therefore, we developed a low-strength stabilization method for reducing settlement without replacing ballast. This study confirms the effectiveness of the developed method for reducing settlement through laboratory tests. We conducted field tests on a commercial line to verify the effectiveness of this method in reducing settlement.
Repeated train loads on ballasted tracks causes settlement. Sections of track with significant settlement, identified during regular inspections, need tamping using portable tampers. If the ballast does not undergo progressive crushing and granulation, then the ballast settlement after tamping is small, thus resulting in low maintenance frequency. However, if ballast is progressively crushed and granulated (“fouled ballast”) by cyclic train loads and tamping [1, 2], then ballast settlement is likely to occur even after tamping, which consequently increases the maintenance frequency. Therefore, ballast is generally replaced to reduce maintenance frequency. However, the high cost of replacing ballast calls for low-cost settlement control methods that suppress the settlement of fouled ballast. Previously developed methods that improve the strength of fouled ballast and suppress the settlement of ballasted tracks include the granulation method, which fills fouled ballast with grout material and then crushes and granulates it after solidification to increase drainage and strength [3]. Then there is the polymer-stabilization method, which stabilizes fouled ballast using an aqueous biodegradable polyvinyl alcohol (“PVA”) polymer solution and a reaction accelerator to increase strength [4].
However, the granulation method requires the removal of sleepers, mixing of grout, and crushing of the grout-filled layer, which challenges workability. The polymer-stabilization method, on the other hand, requires an aqueous PVA polymer solution that contains a significant amount of water. This reduces the strength of the fouled ballast immediately after it has been added, rendering it susceptible to initial settlement.
Therefore, we developed a low-strength stabilization method [5, 6, 7] that uses a mixture of ultra-fast hardening cement and a polymer solidification material (“stabilizing material”). This method is easy to implement, suppresses the settlement of fouled ballast, and allows normal tamping to be conducted even after hardening.
This study reports on cyclic triaxial compression and cyclic loading tests using a full-scale track model, and test implementation on a commercial line. These tests were conducted to confirm the settlement-suppression effect of the low-strength stabilization method.
The low-strength stabilization method suppresses track settlement by the addition of stabilizing material during tamping using a portable tamper and stabilizing the fouled ballast. Stabilizing material is added at eight locations per sleeper (Fig. 1). The features of this method are as follows:
(1) Applicable using a portable tamper, tamping machine, and backhoe tamper.
(2) Suppresses ballast settlement immediately after application. This means that implementation of the method [does not have to rely on nighttime track closures but] can be completed within daytime track work windows.
(3) Allows tamping to be performed using portable tamper even after application.
The stabilizing material is a 1:1 (by weight) mixture of two types of materials (Fig. 2): ultra-fast hardening cement and a polymer solidification material. The low-strength stabilization method does not require replacement of ballast or the disposal of generated ballast. Therefore, its application incurs only approximately 10% of the cost of ballast replacement.
The procedure for the low-strength stabilization method is as follows:
1) Excavate an area to the depth of the sleeper bottom at the tamping position (Fig. 3(a)).
2) Add 500 g (standard addition amount) of stabilizing material to each tamping position (eight positions per sleeper) (Fig. 3(b)).
3) Backfill using ballast, perform tamping, and stabilize the fouled ballast (Fig. 3(c)).
Even when a tamping machine is used, the stabilizing material is added to the excavated area as per the abovementioned procedure. However, as described later in Section 5.2, even when the stabilizing material was dispersed on the ballast surface during tamping work, a settlement-suppression effect was observed in tests. Considering workability, this therefore suggests when tamping work alone is being carried out, it is also effective to simply spread the stabilizing material on the ballast surface.
Additionally, the area where tamping can be performed using a portable tamper was assumed to be a cylinder with a diameter of 300 mm and a height of 150 mm (Fig. 4), and the added amount of stabilizing material AS (g) was determined using (1).
| (1) |
where V is the cylinder volume (cm3), ρd the dry density of fouled ballast (2.05 g/cm3), and Rs the proportion of added stabilizing material (2%). In this study, Rs was determined based on the results of presented in Sections 4 and 5.
Additionally, we examined whether tamping can be conducted after adding stabilizing material using a portable tamper to tamp the fouled ballast in a circular soil tank (diameter 300 mm) that had been stabilized using four times the standard added amount of stabilizing material (additional amount of 8%). Tamping was conducted three months after air curing, i.e., when the strength was assumed to have developed sufficiently, and ballast tamping was confirmed to be feasible.
Cyclic triaxial compression tests were conducted to examine the added amount of stabilizing material to the fouled ballast. The fouled ballast used for the test (“sample”) had a fouling index of ballast “FI” (sum of passing mass percentages at 0.075 mm sieve and 4.75 mm sieve) [8] of 36% (Fig. 5). Additionally, Fig. 5 shows the maximum dry density ρdmax and optimal moisture content was obtained from a compaction test performed using the E method specified in JIS A 1210. When the FI is 20% or higher, the ballast is evaluated as “fouled,” which signifies that settlement is likely to increase [7]. The samples were prepared by mixing new ballast with cobble stone and kaolin clay such that the particle-size distribution would be similar to that of the ballast obtained on-site. In triaxial compression tests, the specimens used were cylindrical, with a diameter of 100 mm and height of 200 mm (which complies with JGS 0524 (Method for consolidated-drained triaxial compression test on soils)), and the peak particle size of the sample was adjusted by considering the specimen dimensions and maximum particle size. The sample was prepared by mixing a sample with a moisture content of 8% (saturation degree Sr = 68%) with the added amount of stabilizing material shown in Table 1, which shows the relationship between the permanent vertical strain and the number of loadings. Subsequently, the sample was divided into five portions and placed in a cylindrical mold to be compacted to the target compaction degree Dc = 92% [9]. The moisture content was set to the wetter side of the optimal moisture content by approximately 2%, which was the condition for obtaining a significant amount of settlement. The added amount of stabilizing material is shown herein in mass% relative to the dry sample (Table 1).
| FI | added amount of stabilizing material |
| 36% | 0% (without stabilizing), 0.02%, 0.2%, 2% |
| 52% | 0% (without stabilizing), 0.2%, 0.3%, 3% |
The test involved conducting isotropic consolidation at a negative pressure of 20 kPa at first, followed by cyclically loading the sample with a haversine wave. The loading conditions were 5,000 loads, a loading frequency of 0.5 Hz, and a confining pressure of 20 kPa. The vertical-stress amplitude was determined by setting the load distribution rate of the rails for a train load of 160 kN to 0.4 [10] and then dividing it by the base area of the sleeper to be used in the full-scale test in the next section. Here, the train load of 160 kN was determined based on the actual axle loads: 90 kN [11] for conventional railways and 110 kN [12] for Shinkansen. For Shinkansen, a dynamic load factor accounting for high-speed impact [10] was applied, whereas for conventional railways, the load was adjusted to reflect the impact factor at rail joints [10]. In the next section, tests in case of FI=36% were conducted using PC sleepers (3PR specified in JIS E 1201), which is the bottom area of the sleeper of 0.48 m², installed on conventional lines. And tests in case of FI=52% were conducted using PC sleepers (3H specified in JIS E 1201), which is the bottom area of the sleeper of 0.79 m², installed on Shinkansen lines. Therefore, the vertical stress amplitude in case of FI=36% was set to 133 kPa and the vertical stress amplitude in case of FI=36% was set to 88.6 kPa. The curing time for this method was set to 2 hours after accounting for the time between nighttime work and the first train.
Figure 6 shows that the permanent vertical strain was reduced to approximately 1/8 by adding at least 0.2% of stabilizing material as compared with the case without any measures implemented.
Cyclic loading tests were conducted on a full-scale track model to examine the appropriate added amount of stabilizing material during actual construction. The test cases used a full-scale model with PC sleeper (3PR) in case of FI=36% (Case 1) and PC sleeper (3H) in case of FI=52% (Case 2), and the same ballast material as in the cyclic triaxial compression test. The full-scale track model was a ballasted track with one sleeper. Figure 5 shows the particle-size distribution of the ballast. The loading conditions were 300,000 loading cycles, a loading frequency of 5 Hz, a minimum load of 5 kN, and a maximum load of 85 kN. Figure 7 shows the test conditions. The construction procedure for the low-strength stabilization method is as presented in Section 2. Additionally, a cyclic loading test was conducted separately for the new ballast using one PC sleeper (3PR) under the same conditions as this test.
Figure 8(a) shows the amount of sleeper settlement in Case 1. In the case without any measures implemented, settlement progressed rapidly from the start of loading and continued to increase gradually, with approximately 32 mm of settlement accumulated after 300,000 cyclic loadings. By contrast, in the cyclic triaxial compression test, the addition of 0.2% stabilizing material, which provided the settlement-suppression effect, resulted in a settlement amount of approximately 26 mm after 300,000 loadings, which corresponded to a reduction by approximately 20% as compared to the case without measures implemented. The less prominent settlement-suppression effect in the cyclic triaxial compression test is attributable to the inability of the stabilizing material to mix uniformly with the ballast when it is mixed with the ballast using a portable tamper in actual construction. Therefore, we considered the variation in the on-site strength and added 2% stabilizing material, i.e., 10 times of the 0.2% stabilizing material. Results showed that the settlement amount after 300,000 loadings was approximately 5.6 mm, which was approximately 1/6 of the amount for the cases without measures implemented. In fact, this amount is similar to the settlement amount yielded when using the new ballast.
Tests using portable tamper were conducted in a straight section with an annual passing tonnage of 17 million tons (Fig. 9). The FI of the ballast was 24%, significant amount of settlement was indicated and the amount of maintenance required had increased in this area. Excavation was manually conducted for this intervention (Fig. 9(a)), and 2% stabilizing material was added to 15 sleepers before and after the joint where mud pumping had occurred.
Figure 10 shows the longitudinal level irregularity (10 m-chord versine) before and after construction. The maximum longitudinal level irregularity after two months of normal tamping was approximately -11 mm. Meanwhile, the longitudinal level irregularity after 16 months of the intervention using the low-strength stabilization method was approximately -5 mm. The longitudinal level irregularity within the intervention area after 16 months of the intervention using the low-strength stabilization method was reduced to approximately half of what it was before, and a greater settlement-suppression effect compared to conventional tamping was confirmed.
The increase in longitudinal sectional irregularity directly below the joint was approximately -7 mm between one week and 16 months after intervention. Meanwhile, a maximum longitudinal level irregularity of approximately -6 mm was recorded after 16 months of intervention outside the boundary of the intervention area (horizontal position of approximately 7 m). This is attributable to the fact that the overlift amount within the intervention area was relatively large (i.e., approximately 10 mm) and that the intervention area was relatively larger compared to the area outside it.
5.2 Test using tamping machineA test was conducted using a tamping machine in a straight section with an annual passing tonnage of 5.2 million tons. In this intervention, 2% stabilizing material was added to six sleepers before and after the joint where mud pumping had occurred. The FI of the ballast in this area was 46.5% and a significant amount of maintenance was required.
In this test, no excavation was conducted, and the stabilizing material was dispersed on the ballast surface. In the normal tamping using tamping machine, the tamping tool is opened wide and inserted, whereas the ballast is forcibly pushed from the outside of the sleeper directly below it. Therefore, even if the stabilizing material is dispersed around the sleeper, the stabilizing material might not be sufficiently mixed with the ballast directly below the sleeper. Therefore, we eased the mixing of the stabilizing material with the ballast below the sleeper by narrowing the tamping tool on tamping machine in advance to make it fit along the sleeper and then inserting it into the tamping position, as well as pushing the stabilizing material down to the sleeper bottom (Fig. 11). Finally, we opened the tamping tool and performed tamping in the same manner as in the normal tamping machine construction.
Figure 12 shows the longitudinal level irregularity before and after construction. The maximum longitudinal level irregularity after three months of normal tamping machine repair was approximately -20 mm. Meanwhile, the longitudinal level irregularity after eight months of the intervention using this method was approximately -3 mm, which was approximately 1/7 of the displacement before the intervention.
Additionally, we confirmed that only a trace amount of settlement had progressed between four and eight months after the intervention and that a good track condition was maintained.
We developed a low-strength stabilization method for fouled ballast that suppresses settlement by mixing a stabilizing material composed of ultra-fast hardening cement and a polymer solidification material during tamping. We confirmed the settlement-suppression effect by conducting cyclic triaxial compression and cyclic loading tests using a full-scale track model. Furthermore, we conducted tests using the method on a commercial line to verify the settlement-suppression effect. The main findings obtained are as follows:
a) The fouled ballast with an FI of 20% or more was stabilized using stabilizing material, and a cyclic triaxial compression test was conducted. The settlement amount under FI=36%, which involved the addition of at least 0.2% stabilizing material, was approximately 1/8 of that for the case without any measures implemented.
b) The fouled ballast with an FI of 20% or more was stabilized using stabilizing material, and a full-scale test was conducted. The result of the cyclic triaxial compression test showed that by adding 0.2% stabilizing material by cyclic triaxial compression test, which yielded the settlement-suppression effect, settlement was reduced by approximately 20% compared when no measures were implemented. The settlement-suppression effect was speculated to have been weakened because of variations in strength caused by the use of a portable tamper.
c) A test was conducted in an on-site fouled ballast with an FI of 20% or more, in which 2% stabilizing material was added using a portable tamper. The results showed that the longitudinal level irregularity within the intervention area after 16 months of the intervention reduced to approximately half the longitudinal level irregularity after two months of normal tamping. Additionally, a greater settlement-suppression effect was achieved compared with conventional tamping.
d) A test was conducted in an on-site fouled ballast with an FI of 20% or more, in which 2% stabilizing material was added using a tamping machine. The results showed that the longitudinal level irregularity within the intervention area after eight months of the intervention reduced to approximately 1/7 of the longitudinal level irregularity after three months of normal tamping. Additionally, a greater settlement-suppression effect was achieved compared with conventional tamping.
We would like to express our gratitude to the West Japan Railway Company and the Shikoku Railway Company for their cooperation in this test construction.
|
Takahiro KAGEYAMA, Ph.D. Researcher, Track Structures & Geotechnology Laboratory, Track Technology Division Research Areas: Ballasted Track |
|
Takahisa NAKAMURA, Ph.D. Senior Chief Researcher, Track Structures & Geotechnology Laboratory, Track Technology Division Research Areas: Ballasted Track |
|
Masaru HOJO
Researcher, Track Structures & Geotechnology Laboratory, Track Technology Division Research Areas: Ballasted Track, Ballastless Track |
|
Fumika TAJIMA
Researcher, Track Structures & Geotechnology Laboratory, Track Technology Division Research Areas: Ballasted Track |
