A bibliometric review on stability and reinforcement of special soil subgrade based on CiteSpace

1.   Introduction

Subgrade stability is a significant factor aff5cting the highway and railway engineering construction quality. Because of geographical environment, climatic condition, and geological origin, there are obvious soil texture differences in different regions, which means that soil subgrade treatments have become a key construction technology. For example, if frozen soil is cemented by ice, it could be subjected to frost heave or could thaw with a change in temperature (Kudryavtsev et al., 2016; Varlamov, 2018). Because loess has a multi-porous, loose structure, poor mechanical strength and low natural water content, it can easily collapse, which can lead to local deformation when the loess subgrade is soaked in water (Hasan et al., 2018; Jiang, 2020; Li, 2016; Zhao et al., 2018). Waste soil is basically subgrade filled with waste material, it is prone to landslides and uneven settlement (De Oliveira et al., 2020; Wang, 2011). Expansive soil is a special kind of clay, it expands when encountering water and shrinks and cracks when losing water, which significantly weakens its strength (Al-Taie et al., 2016; Ashango and Patra, 2016). During subgrade construction and even after construction, the properties of these types of soils could lead to serious problems, such as uneven differential settlements, slope collapses, and landslides, which in turn would aff5ct road safety, transportation efficiency and road service life. Consequently, stability and reinforcement technologies to cater for these types of special soil subgrades have become major research foci; therefore, it is beneficial to reveal the macro development research trends to guide the future research frontier.

There has been a great deal of research into special soil subgrades. For example, Zhao et al. (2015) systematically studied soft soil stress deformations and settlement control based on soft soil subgrade settlement control data from the Beijing-Shanghai Railway, the Wen-Fu Railway and the Hu-Hang Passenger special area, which provided guidance on treatment measures and settlement control. Liu (2018) summarized the influencing factors and treatment measures for the differential settlements caused by subgrade widening in soft soil areas, and proposed research ideas for differential settlement characteristics and municipal road widening project treatments. Xiong et al. (2010) assessed the stability of the permafrost slopes for the Qinghai-Tibet Railway based on the thermal thawing slump and thawing frost mud flow deformation characteristics, and developed a stability evaluation method and engineering treatment measures, from which it was determined that the stability of the thick layer underground ice and the steep slope section subgrade of the artificial slope needed attention. In other related studies, to improve the uneven vertical or horizontal expansion and shrinkage of the gene-expansive soil in the Yunnan-Guangxi Railway road that was causing roadbed displacement, cracking and other damage, Cao and Cao (2018) proposed specific engineering measures to fully close the foundation bed structure, which was then applied to other construction sites to reduce the risk of these hazards. Song et al. (2019) established a thermal-mechanical coupling model and an external force subgrade structure action model based on the crushed stone subgrade in frozen soil areas, and by monitoring the data and calculating the mechanical properties to check whether the model was reliable, established three frost heave subgrade models to compare the mechanical properties and selected the model that was best able to reduce the frost heave subgrade deformation. Qin (2017) proposed adding an ISS soil curing agent into red clay, and through testing found that this method improved the red clay subgrade swelling and shrinkage problems. Deb et al. (2010) established two generalized mathematical models to examine the soil arching effect frequently occurring in soft soil foundations; a stone column geosynthetic material mathematical model, and a non-reinforced embankment soil arching mathematical model; and used a spring damper system to idealize the soft soil subgrade to determine the factors aff5cting the soil arch degree, from which it was found that embankment height, the degree of soft soil consolidation, the stiffness of the stone column materials, and the use of geosynthetic reinforcement all aff5cted the soil arch degree. Oh and Vanapalli (2011) proposed a new method to predict the bearing capacities and settlement of shallow unsaturated sand foundations based on soil parameters obtained under saturated and soil characteristic curve conditions. The settlement was predicted using the relationship between the vertical stress and the surface settlement, which was analyzed using the finite element method and the elastic–plastic model. The results from the two methods were found to be consistent, proving that the proposed method could conveniently and concisely predict foundation bearing capacities and settlements. Kadela et al. (2017) conducted a series of tests on the density, porosity, compressive strength and flexural strength of foam concrete to assess its pavement structure subbase potential. Expansive soil has obvious movement when the water content changes, which can cause serious damage; therefore, to resolve this problem, Hasan et al. (2016) conducted a series of tests with bagasse ash to demonstrate its potential in restraining the adverse effects of expansive soil on roads. Orakoglu et al. (2016) used a Stefan equation and a modified Bergren equation to predict frost depth in a permafrost region of eastern Turkey, compared the results with actual observation data, and gained valuable insights into subgrade design and maintenance in frozen soil areas. From experiments designed to explore expansive soil stabilizers, Pooni et al. (2019) found that adding enzyme based stabilizers could significantly stabilize and strengthen the soil and enhance its bearing capacity. Pancar and Akpinar (2016) proposed several soil improvement methods and determined that lime improved geosynthetics produced a better effect on soft soil subgrade. Tarnocai and Bockheim (2017) studied and classified the causes and properties of permafrost in Canada to assist agricultural and construction projects (such as roads, urban sites and airstrips) and inform farming and construction methods. Saberian and Khabiri (2017) conducted tests on the damage of expansive soil pavement caused by water imbalance and gained the conclusion that expansive soil and alkali activated binder (AAB) were combined with coir and hemp fiber can improve the strength of expansive soil subgrade. Khabiri (2010) used waste building materials for pavement subbase and added lime to stabilize subgrade strength.

With the recent developments in scientific metrology, data visualization, knowledge map construction and application, CiteSpace can be used to visualize the knowledge evolution in a research field (Chang et al., 2016; Chen et al., 2019a, 2020; Gao et al., 2018; Mao et al., 2019; Naveau et al., 2017; Qian et al., 2016; Shi et al., 2019; Wang and Lu, 2019). Therefore, this article examined the WOS and CNKI research literature on special soil subgrade from 2005 to 2019, and then employed CiteSpace software to generate scientific knowledge maps to identify the innovative special soil subgrade research developments.

2.   Data and methods

2.1   Data source

The data were extracted from the Web of Science (WOS) database and the China National Knowledge Infrastructure (CNKI) database. The WOS core collection database advanced retrieval search string was: topical subject (TS) = (subgrade OR roadbed OR embankment OR road OR railway) and TS = (soil OR loess OR cultural relics OR expansive soil OR fissured soil OR frozen soil OR red clay OR saline soil OR bentonite OR residual soil OR garbage soil OR dispersive soil OR ice water accumulation OR red sand soil OR flammable ice OR lunar soil) and TS = (kinetic OR characteristics OR constitutive model OR defect OR lesion OR multi-field Coupling OR Settlement OR deformation OR dynamic buckling OR stabiliz OR reinforce OR static properties OR broaden OR genetic mechanism).

The advanced professional search was selected in the CNKI with the journal sources set at Science Citation Index, Engineering Index, CORE, Chinese Social Science Citation Information, Chinese Science Citation Database, and the following search string employed: subject (SU) = (subgrade + railway + subgrade + highway) and SU = (soil + loess + expansive soil + fissured soil + frozen soil + red clay + salinized soil + bentonite + residual soil + garbage soil + dispersive soil + ice water accumulation + red sand + coral sand + cultural relic soil + combustible ice + lunar soil) and SU = (dynamic properties + constitutive model + disease + multi coupling action + settlement + instability + stability + reinforcement + static properties + widening + genetic mechanism). The retrieval time was January 19, 2020.

To avoid accuracy problems because of the differences between the two databases, the retrieval years of literature were set at 2005–2019 and irrelevant academic articles such as newspapers articles were excluded. Finally, 2102 documents from the WOS database and 2601 documents from the CNKI database were extracted.

2.2   Methods

CiteSpace software, which was jointly developed by Chen from the School of Information Science and Technology at Drexel University and the WISE Laboratory at Dalian University of Technology, is a information visualization software that measures and analyzes research data and then illustrates this analysis using knowledge maps. Before the analysis project was established in CiteSpace, the CNKI data were transformed; however, the data exported from the WOS core database did not need to be transformed. The CiteSpace information visualization software then developed a keyword network visualization map that indicated the time distribution zone, the high-frequency keywords, the emerging words in the special soil subgrade field, the most popular topics, and the research turning points, thereby revealing the evolution, the research frontier and the development trends. Then, quantitative analyses of the countries/regions, organizations, source journals, and keyword co-occurrences were conducted, from which a comprehensive report on special soil subgrade was developed.

3.   Results and discussion

3.1   Time distribution feature

The number of article published over time reveals the development trends in a field. Therefore, the number of annually published article on special soil subgrade in the WOS and CNKI databases were collated and a distribution map developed to allows for an analysis of the activity in this field.

Fig. 1 shows the WOS core database special soil subgrade published article from 2005 to 2019 and the CNKI database special soil subgrade published article from 2005 to 2019. As it can be seen, from 2005 to 2009, only a small number of special soil subgrade article appeared in the WOS, indicating that this field was only at its frontier development stage. From 2010 to 2019, however, the number of published article increased significantly from only around a handful in 2009 to over 550 in 2019. The CNKI results, however, indicated that there had been special positive development soil subgrade research trend from as early as 2005 until the peak in 2009, after which the research interest stabilized and slightly decreased. The special soil subgrade CNKI research development, therefore, could be divided into two stages: a sharp growth period from 2005 to 2009; and a slow reduction period from 2010 to 2019.

Figure  1.  Published article from 2005 to 2019.

DownLoad: Full-Size Img PowerPoint

3.2   Quantitative analysis of productive countries/regions

The analysis also revealed the countries/regions that were most focused on special soil subgrade research and the macro, meso and micro cooperative relationships. The node type was set to country and the time slice threshold set to the top 50, after which CiteSpace conducted a visualization analysis and developed the network visualization country/region map shown in Fig. 2.

Figure  2.  National cooperative special soil subgrade network atlas.

DownLoad: Full-Size Img PowerPoint

The quantitative analysis revealed that 96 countries had participated in special soil subgrade research; however, there has been few article from West Asia, Africa or South America. In Fig. 2, each circle is a node, with the size of the nodes indicating the frequency of the published special soil subgrade articles, that is, the size of the node is directly proportional to the number of published article. The node color corresponds to the publication time of the article, and the connections between two nodes reflect the strengths of the cooperative relationships; the closer the ties, the closer the cooperation. For example, the larger nodes for China, the United States, Australia, Iran and India indicate that these are the most active special soil subgrade research countries. From 2005 to 2019, China published 740 articles or 35.2% of the total and the United States published 374 (17.8%); however, China's CPP (citations per article) was far lower than that of the United States, which indicated that the US special soil subgrade achievements had a more significant impact. Fig. 2 reveals that both China and the United States were relatively independent as they had few connections with other countries. However, the Netherlands had the most special soil subgrade research network connections.

3.3   Quantitative analysis of the main research organizations

The institutional cooperation analysis revealed the most effective organizational information in the special soil subgrade field. To analyze and classify the institutions, the time spanparameter was set at 2005–2019, the year per slice at 1, title, abstract, author, and keywords set in the text processing function area, Institution selected in the network configuration function area, and top N set as 20 in the selection criteria.

The literature collected in the WOS database was then analyzed and the network visualization map shown in Fig. 3 generated, which shows that there were eight institutions that each had published more than 30 articles in a total of 420 journals, which accounted for 19.98% of all identified studies. The eight most prolific institutions were: the Chinese Academy of Science; Tongji University; Hohai University; Central South University; Zhejiang University; Southwest Jiaotong University; Southeast University; and the Chinese Academy of Science University. The Chinese Academy of Science and Tongji University had the largest number of publications at 157 and 58. The network visualization map revealed that the main scientific research was being conducted in universities and research institutes. Fig. 3 has 425 nodes and 439 connecting lines, a research network density of 0.0055, which indicated that the cooperation between the various institutions was low.

Figure  3.  WOS special soil subgrade research institution network map.

DownLoad: Full-Size Img PowerPoint

The data from the CNKI database was then analyzed and Fig. 4 generated, from which it can be seen that there were eight institutions that had each published more than 30 special soil subgrade articles: the School of Civil Engineering, Southwest Jiaotong University; the China Railway Northwest Research Institute Co., Ltd; the Department of Underground Architecture and Engineering, Tongji University; School of Civil Engineering, Beijing Jiaotong University; School of Highway, Chang'an university; the State Key Permafrost Laboratory, the Institute of Environment and Engineering in Cold and Dry Areas; the Chinese Academy of Sciences; College of Civil Engineering, Lanzhou Jiaotong University; School of Civil Engineering, the Central South University. The School of Civil Engineering of Southwest Jiaotong University, and the China Railway Northwest Research Institute Co., Ltd have the largest number of publications of 64 and 53 articles, respectively. The network visualization map revealed that the main scientific research was being conducted in universities and research institutes. Fig. 4 has 435 nodes and 319 connecting lines, a research network density of 0.0034, which indicates that the cooperation between the various institutions was relatively low.

Figure  4.  CNKI cooperative special soil subgrade research institution network map.

DownLoad: Full-Size Img PowerPoint

The research topic of special soil subgrade stability and reinforcement is relatively mature in both China and the world, and there are some highly cited articles, such as the article entitled "undrained deformation behavior of saturated soft clay under long-term cyclic loading" published in Soil Dynamics & Earthquake Engineering in 2013 (Guo et al., 2013). The article has been cited 115 times until March 17, 2021. Based on a series of monotonic triaxial tests and long-term cyclic triaxial tests, the prediction formulas of long-term elastic modulus and permanent strain were established in the article, and a new critical value of CSR was proposed. "summary of foundation treatment technology development" which published by Zheng et al. (2012) in "Building Science and Engineering" in 2012, has been cited 294 times and downloaded 13, 295 times. In this article, based on the detail analysis of the relevant specifications of China's foundation treatment, the foundation treatment technologies developed in recent years at home and abroad were applied comprehensively and a new type of composite reinforcement technology was formed. This article reviewed and summarized the historical development of China's foundation treatment, and promoted the development of this field, making a significant contribution for the future research direction (Tables 1 and 2).

Table  1.  Classic literature on subgrade stability and reinforcement of special soil in WOS.

No.	Document name	Author	Year	Citation time
1	Undrained deformation behavior of saturated soft clay under long-term cyclic loading	Guo et al.	2013	103
2	Calcium carbide residue: Alkaline activator for clay-fly ash geopolymer	Phetchuay et al.	2014	94
3	DEM simulation of the behaviour of geogrid stablished ballast fouled with coal	Ngo et al.	2014	93
4	Investigation of factors influencing behavior of single geocell-reinforced bases under static loading	Pokharel et al.	2010	87
5	Bearing capacity of geocell reinforcement in embankment engineering	Ling et al.	2010	85
6	Swell-shrink and strength behaviors of lime and cement stabilized expansive organic clays	Saride et al.	2013	83
7	Geotechnical influence on existing subway tunnels induced by multi-line tunneling in Shanghai soft soil	Zhang and Huang	2014	82

 | Show Table

DownLoad: CSV

Table  2.  Classic literature on subgrade stability and reinforcement of special soil in CNKI.

No.	Document name	Author	Year	Citation times
1	Summary of foundation treatment technology development	Zheng et al.	2012	293
2	Research progress of foundation treatment	Liu and Zhao	2016	240
3	Basic theoretical research on unsaturated soil and special soil mechanics	Chen	2014	217
4	Impact analysis of subway shield tunnel under Beijing Tianjin Intercity High Speed Railway	Xu et al.	2009	175
5	Analysis of rigid pile embankment interaction on soft soil foundation	Chen et al.	2005	172
6	Large scale field immersion test study on collapsible deformation characteristics of large thickness selfweight collapsible loess site	Huang et al.	2006	170
7	Undisturbed soil tangent modulus method for nonlinear settlement calculation of foundation	Yang	2006	150
8	New method of foundation settlement calculation	Yang	2008	133

 | Show Table

DownLoad: CSV

3.4   Quantitative analysis of the main source journals

As journal is the main method for the dissemination of scientific achievements, research on journal distributions can highlight the theoretical and practical value of a research field. The 2102 special soil subgrade article extracted from the WOS were published in 520 journals, and the most prolific of which are: the Journal of Materials in Civil Engineering (64 article), Transportation Research Record (64 article), Cold Regions Science and Technology (58 article), Construction and Building Materials (49 article), and Transportation Geotechnics (47 article). The quantitative visualization analysis of the retrieved literature from WOS also generated a discipline classification distribution map, which revealed that the engineering discipline journals accounted for 1283 published article or 60.037% of the total, and the geology discipline accounted for 547 published article or 25.597%. It can be seen that the research on the stability and reinforcement of special soil subgrade is a foci in engineering research.

The 2601 special soil subgrade CNKI article were published in 131 journals, with the most article from 2005 to 2019 being published in: Rock and Soil Mechanics (185 article), the Chinese Journal of Geotechical Engineering (114 article), Highway (113 article), Railway Engineering (111 article); and Chinese Journal of Rock Mechanics and Engineering (108 article). The quantitative visualization analysis of the retrieved literature from CNKI also generated a discipline classification distribution map, which revealed that 1455 article or 55.96% of the total were published in the traffic and transportation discipline, and 1024 article or 39.37% were published in architectural science journals. It can be seen that the research on the stability and reinforcement of special soil subgrade in China is a foci in the field of transportation technology.

3.5   Research hotspots

Keywords are used to summarize the content of the article; therefore, analyses of keywords can reveal the research foci across time. The "node type" selected was keywords, the time slice threshold was set to TOP 50, and CiteSpace run again to analyze the keywords and generate the WOS and CNKI keyword visualization maps.

In Fig. 5, which shows the WOS key word distributions, each node represents the keyword, the size of the node represents the keyword occurrence frequencies, and the line thickness between the nodes represents the correlation strength there were 195 network nodes and 301 connecting lines, with the total network density being 0.0159. Fig. 5 indicates that the stability and reinforcement of special soil subgrade research strength. The analysis revealed that the WOS articles were most related to soil, behavior, model, strength, clay, settlement, stability, fly ash, prediction, and resilient modulus. The top 20 high frequency WOS keywords are shown in Table 3.

Figure  5.  Research knowledge map for the WOS special soil subgrade field articles.

DownLoad: Full-Size Img PowerPoint

Table  3.  High frequency keywords for the WOS special soil subgrade field articles.

No.	Keyword	Frequency	Centrality
1	Soil	392	0.14
2	Behavior	259	0.08
3	Model	186	0.13
4	Strength	137	0.06
5	Clay	114	0.04
6	Deformation	114	0.06
7	Stabilization	111	0.08
8	Fly ash	110	0.06
9	Performance	98	0.05
10	Settlement	97	0.05
11	Lime	95	0.06
12	Subgrade	87	0.03
13	Prediction	80	0.05
14	Resilient modulus	74	0.04
15	Sand	65	0.07
16	Geosynthetics	61	0.06
17	Design	59	0.03
18	Soil stabilization	59	0.03
19	Impact	59	0.04
20	Bearing capacity	42	0.03

 | Show Table

DownLoad: CSV

In Fig. 6, which shows the CNKI key word distributions, each node represents the keyword, the size of the node represents the keyword occurrence frequencies, and the line thickness between the nodes represents the correlation strength. The analysis revealed that there were 450 network nodes and 660 connecting lines, with the total network density being 0.0065. Fig. 6 indicates that the special soil subgrade CNKI articles were most related to settlement, subgrade treatment, cement fly-ash gravel (CFG) pile, collapsible loess, soft soil, geogrids, numerical simulation, dynamic compaction, soil mechanics, and frozen soil. The top 20 high frequency CNKI keywords are shown in Table 4.

Figure  6.  Research knowledge map for the CNKI special soil subgrade field articles.

DownLoad: Full-Size Img PowerPoint

Table  4.  High frequency keywords for the CNKI special soil subgrade field articles.

No.	Keyword	Frequency	Centrality
1	Settlement	223	0.22
2	Subgrade treatment	221	0.20
3	CFG	166	0.17
4	Collapsible loess	137	0.06
5	Suicide	108	0.10
6	Geogrid	103	0.09
7	Numerical simulation	102	0.17
8	Dynamic compaction	94	0.05
9	Soil mechanics	88	0.13
10	Frozen soil	86	0.06
11	Stress ratio	72	0.03
12	Model	68	0.05
13	Bearing capacity	61	0.05
14	Expressway	61	0.06
15	Qinghai-Tibet Railway	57	0.03
16	Stabilization	48	0.01
17	Reinforce	45	0.02
18	Finite element	42	0.03
19	Deformation	39	0.03
20	Pore water pressure	34	0.03

 | Show Table

DownLoad: CSV

Centrality means betweenness centrality, which is an index to measure the importance of nodes in the network. CiteSpace uses this index to find and measure the importance of literature, and uses purple circle to analyze this kind of literature (or keywords, institutions, authors and journals, etc.), and the node with purple circle has intermediary centrality of no less than 0.1.

The comparison of the WOS and CNKI special soil subgrade field high frequency special soil subgrade keywords revealed that the basic research directions had been much the same.

Table 3 indicates that the WOS high-frequency special soil subgrade keyword distribution was uneven and the research focus relatively scattered because there were many research directions and many cross research fields. However, generally, the WOS article special soil subgrade research was focused on three main aspects.

(1) The first main research area was focused on the mechanisms and causes of subgrade problems, such as settlement deformations, longitudinal cracks, highway widening, highway slopes, and subgrade mud pouring. Table 1 shows that settlement deformation was the most focused on subgrade problem, with 211 research article. For example, Ma et al. (2011) summarized the subgrade deformation characteristics under different annual mean ground temperatures (MAGTs) in four permafrost regions along the Qinghai-Tibet Railway. Based on further analysis on some typical monitoring sections, the deformation mechanism of these embankments was discussed according to the details of the temperature and underground conditions. Batenipour et al. (2014) studied the nature and cause of embankment deformations by monitoring highway embankment settlements and temperatures in a discontinuous permafrost region about 18 km northwest of Thompson, Manitoba for three years. Benmebarek et al. (2015) carried out numerical simulation of geosynthetics-reinforced embankment on local weak zone, and analyzed the influence of compression parameters, stiffness of geosynthetics, geometric parameters and friction angle of embankment fill on soil. The results showed that the combination of membrane strength effect of geosynthetics and arch effect in embankment filling can improve the differential settlement of embankment. Tafreshi and Norouzi (2015) conducted a series of repeated load tests on the 150 mm slab simulating vehicle traffic and demonstrated the role of soil rubber mixture in reducing the settlement of embankment surface. The result showed that the rubber reinforcement effect is related to the rubber soil mixture and cap on the mixture, which encourage the use of rubber debris from non reusable tires as viable materials in road engineering. Soeung et al. (2018) elaborated the factors causing the additional settlement of high-speed railway in the article. The potential influencing factors of additional settlement includes geological characteristics, embankment height, layer thickness, foundation treatment method, groundwater level drawdown, frozen soil compaction and dynamic load of train operation. It showed that the settlement of concrete track built on alluvium and biotite granodiorite is large, and the settlement increases with the increase of embankment height and subgrade thickness. In the high settlement site, the settlement will increases suddenly when the water level decreases in dry season. In addition, the compaction of embankment in winter and the dynamic load of train operation also caused the settlement increased to a certain extent. Bowman and Haigh (2019) combined the Cambridge accelerated pavement test machine (APT) and digital image correlation (DIC) technology to study the progressive failure of flexible pavement structures under repeated wheel loads, which helped reduce the costs of road and airport construction by identifying the correct sized structural layer. Anand et al. (2019) introduced the application of geocell in desert sand and a series of plate load tests were carried out on unreinforced soil and geocell reforced homogeneous sand layer. The field test results of heavy military vehicles and other vehicles with different load levels on geocell reinforced soil are compared with the laboratory test results. The results show that the strength and stiffness of desert sand are significantly improved by using geocell, which provides a solution to prevent the settlement of deserts and subgrade.

(2) As shown in Table 1, the keyword "model" appeared 186 times. Subgrade prediction and analysis models have been widely used to detect subgrade bearing capacities, study the dynamic characteristics, calculate the elastic modulus, run simulations, and conduct finite element analysis. Various methods, such as stress–strain relationships and P-wave velocity, have been applied to obtain the elastic modulus data to measure road performances (Abu-Farsakh et al., 2015; Jamei et al., 2015). For example, Hirain (2014) used a Winkler model to analyze rectangular tree pile settlements under a vertical load in non-uniform soil, and developed a recurrence settlement and vertical load equation to ensure the safety and availability of road subgrade. Yesuf and Hoff (2015) established a plastic strain subgrade soil model to study failure modes and predict subgrade deformations. Zhang et al., 2015, Zhang and Huang, 2014 proposed the use of civil composite materials to concentrate embankment stress and conducted field tests to analyze the lateral displacement of clay. Premekumar et al. (2016) conducted a contact scour experiment on dispersive soil subgrade using groundwater to ensure that the dispersive clay subgrade was treated to maintain the compactness of each layer. Peduto et al. (2016) used a Dutch railway embankment as an example to propose a new subgrade settlement prediction model based on the secondary settlement of soft soil subgrade. Indraratna et al. (2016) took the problem of gradual deterioration of track, excessive deformation and deterioration of ballast and unacceptable uneven settlement of track caused by large and frequent load of heavy freight train and passenger train as the research background. The stress–strain and degradation response of ballast bed were analyzed by discrete element method (DEM) and finite element method (FEM). In this article, the advantages of the proposed DEM and FEM models in capturing the correct stress–strain and degradation response of ballast were summarized, with particular emphasis on particle breakage and fouling, as well as the application of geosynthetics grids and seismic pads. Jamsawang et al. (2016) analyzed the slope failure during the construction of Suvarnabhumi drainage channel. To find the reason of channel failure, laboratory study on undrained creep characteristics and finite element analysis of soft soil creep (SSC) model were carried out. The creep failure time of soft clay is analyzed when the peak strength deviating from creep stress is 70% and 100%. In addition, through the finite element analysis of the channel slope safety factor, this article introduced a new method to solve the channel section. The field monitoring and finite element analysis of the new test section verified the effectiveness of the solution method. Arghadeep et al. (2016) studied the mechanical behavior of rigid clay foundation strengthened with geosynthetics under cyclic loading through a series of indoor model tests. Five different series of tests were carried out in homogeneous (clay or sand) and layered structures. The test results show that the performance of the rigid clay foundation bed was significantly improved by reinforcement of plane geogrid and three-dimensional geocell. To make the applicability behavior and deformation of geosynthetics reinforced column supporting embankment to control its design applicability, King et al. (2017) emphasized the need to couple arch stress deformation model to describe the accurate usability behavior and the relationship between soft soil parameters and usability behavior, as well as the conditions leading to progressive collapse of grcse were described. At the same time, in some field case studies, the problems of long-term performance and the way to achieve acceptable performance in the short term were proposed. Roy and Deb (2017) carried out model plate loading tests on rigid rectangular footings with different length diameter ratios on sand overlying soft soil. The effects of the length width ratio of rectangular footing and the thickness of sand layer on the bearing capacity, settlement characteristics, load distribution angle and size (the optimal thickness of sand layer) of geogrid layer were studied. Based on the test results, an analytical method for calculating the bearing capacity of rectangular foundation on soft soil with or without reinforcement was proposed. To establish a deformation prediction model for expansive subgrade pavement, Ahmed et al. (2019) used humidity, temperature sensors and rain gauges to regularly monitor two pavement sites in Texas. The influence of temperature and resistivity was also considered in the model. In addition, the relationship between pavement deformation and rainfall can be predicted at any time of the year.

(3) There has been an increased research focus on subgrade reinforcement optimization and technology, such as grouting reinforcement technology, dynamic compaction methods, powder spraying pile reinforcements, and adding geogrid to stabilize and strengthen the subgrade, as well as research into the mechanical properties of subgrade soil materials. For example, Ngo et al. (2014) used discrete element modelling (DEM) to simulate a direct shear test for coal gangue ballast that had been reinforced using geogrid, finding that DEM could effectively reflect the geogrid coal gangue stabilizing effect. Biabani and Indraratna (2015) conducted a series of experiments to study the interface shear strength of geogrid and geomembrane stabilized ballast. Dessouky et al. (2015) conducted visual inspection, field and laboratory tests, ground-exploration radar scanning and structural design evaluation on three project sites to evaluate the existing pavement treatment options. The study showed that geogrid reinforcement combined with lime treatment is an effective way to repair low to medium plastic soil area. In the area of high plastic expansion soil, cementitious was an effective treatment method. Hussaini et al. (2016) used a series of materials to conduct comparative tests on the railway track bed with or without geogrid, finding that geogrid could reduce subgrade settlement harm. Yazdani et al. (2016) monitored soil structure using a geogrid application with sensors, and its application and advantages are introduced. Keller (2016) comprehensively described a geosynthetics application for road engineering. To solve the problem of the rapid deterioration of base material and gradual permanent surface deformation which caused by the static and dynamic traffic load on the weak compressibility and strength of soft soil subgrade, Kiptoo et al. (2016) reinforced geosynthetics (geogrids and geotextiles) were used as reinforced inclusions in granular base under soft subgrade. The results showed that it can significantly prolong the service life of the pavement and effectively control the settlement of the subgrade. Zornberg et al. (2017) established a soil-geosynthetic material interaction framework, and gained the stiffness of geosynthetics under small displacement by using a single and repeatable parameter. In addition, the results are introduced. These tests results of the geosynthetics interaction device confirmed the hypothesis of the model and the applicability of the results. Rasul et al. (2018) established a performance model to determine the elastic deformation characteristics of dry and wet soil. Sharma et al. (2018) studied the subgrade stabilization effect of lime and cement on soil. Singh et al. (2019) conducted an experimental study on the performance of subgrade soil by placing single-layer and double-layer geosynthetic reinforcement (GLASGRID, Tenax 3D mesh and Tenax multimat) at different depths on the top surface of subgrade soil. Through the indoor CBR test, the optimal depth of reinforcement layer was determined. The results showed that the increase of CBR value of soil would lead to the decrease of design thickness of pavement layer above subgrade soil, which was a significant contribution.

The nodes in CiteSpace that have a centrality of more than 0.1 are key nodes. As shown in Table 4, the centralities of only five keywords in the Top 20 CNKI special soil subgrade field articles were greater than 0.1; settlement, subgrade treatment, CFG, collapsible loess, and numerical simulation; with the centralities of all other keywords being less than 0.1, indicating that the research areas were not yet fully developed. The node sizes for each keyword in Fig. 6 varied greatly, with each node being loosely connected with small tightness. The CNKI research on special soil subgrade from 2005 to 2019 was focused on the following three areas.

(1) As the subgrade characteristics of different soils in different regions vary significantly, subgrade is generally classified by soil type, such as loess, soft soil, frozen soil and red clay. For example, Liu et al. (2006) conducted eight subgrade simulation tests on an expansive soil section in Guangxi and Hunan, The damage mechanism of water on expansive soil subgrade under the action of dry wet cycle is studied. Chen (2017) studied the variation of resilient modulus of red clay with the number of drying and wetting cycles by using the indoor bearing plate method and the resilient modulus tests under different drying and wetting cycles. The results show that the resilient modulus of remolded red clay decreases with the increase of drying and wetting cycles. The research results can provide reference for the long-term stability evaluation of red clay subgrade. Wu and Zhang (2019) studied the subgrade in a permafrost area of the Gonghe-Yushu (Jiegu) Highway and analyzed the relationships between the subgrade and permafrost type, subgrade treatment type, subgrade height and subgrade flatness to resolve specific subgrade problems and propose relevant treatment technologies, Zhang (2019) analyzed the causes of loess plateau subgrade from various soil characteristics and proposed corresponding maintenance schemes. Based on the crushed stone subgrade in frozen soil areas, Song et al. (2019) established a thermal-mechanical coupling model and an external force subgrade structure action model. The model was proved to be reliable by the monitoring data and the mechanical properties calculating data.

(2) Table 4 indicated that settlement deformation was studied 262 times and was a major research area. For example, Fu et al. (2007) established a differential settlement calculation model to analyze the differential settlement characteristics of widened subgrade and examined the influence of subgrade height, soil foundation compression modulus, widening and width on the differential settlement of subgrade and the settlement curve shape to determine the influence of differential settlements on widened roads. Xie et al. (2009) studied deformation type, slope failure, and the characteristics of different slope types using a centrifugal model test for different slope types, and Niu et al. (2011) gave a comprehensive description of the road engineering and subgrade disasters on the Qinghai-Tibet Plateau. Wei et al. (2017) established a simplified discrete pile net subgrade element model using PFC3D software to analyze the contact force chain between soil particles, the lateral and vertical displacement of particles, and the vertical distribution of particle stress in the subgrade to determine a method for weakening the geotechnical effect. Taking the red sandstone subgrade in northwest of Guizhou Province as research object, Guo et al. (2019) analyzed the stability and settlement deformation of subgrade with different filling height in different foundation environment by using FLAC 3D software. Combined with the tracking monitoring data of embedded test instruments in the construction site, the theoretical analysis results were verified. The research results provided reliable technical support for the construction of expressways in the red sandstone area of northwest of Guizhou Province. Taking high fill embankment as prototype, by using geotechnical centrifugal model test, Sun and Sun (2019) studied the settlement and deformation law of soil rock mixture filling in the subgrade under different soil rock ratio, different filling height and different filling technology. The results showed that the settlement deformation of the filling body increases with the increase of filling height, and the layered settlement law was significant. Focused on the structural characteristics of CRTs I and CRTs II slab ballastless track on subgrade, Xiang et al. (2019) established the finite element models to study the transfer law and influence of stress and deformation of different slab ballastless track under the action of uneven settlement of subgrade.

(3) The third main research area has been focused on subgrade reinforcement and stabilization, such as dynamic compaction, CFG piles, and geogrids. For example, Chen et al. (2006) used a dimensional analysis method for a model experiment on geocell gravel cushion gravel pile composite foundation to assess whether the interaction between the geocell gravel cushion and the gravel pile composite foundation could improve its bearing capacity and reduce the differential settlement. Guo et al. (2008) calculated the theoretical compactness, void ratios and effective soil influence depths during compaction based on the principle that the compaction pit volume was equivalent to a cylinder, and also determined the theoretical calculation formula for the compaction amount, the void ratio, compactness, compaction energy, and compaction amount by changing the compaction times. Ma et al. (2014) analyzed Beijing-Shanghai high speed railway road sections using numerical simulations to assess the CFG pile composite foundation settlement method. Sun (2019) geogrid reinforcement of a soft soil subgrade widening project as the background. Midas/GTS finite element software is used to establish the three-dimensional model of subgrade. The displacement, stress and stability of soft soil subgrade with different geogrid reinforcement layers are analyzed. The best geogrid reinforcement layer is selected. It can effectively reduce the displacement and stress of subgrade and improve the stability coefficient. It reduces the construction difficulty and engineering cost. Taking Shawan Project of S101 line as the test section, Duan et al. (2019) established a two-dimensional seismic dynamic analysis slope model. The dynamic characteristics of geogrid reinforced soil slope embankment in high intensity area were studied, and the displacement response, acceleration response and stress response of the external and internal slope under seismic acceleration were analyzed. The results showed the stability effect of geogrid on slope embankment. Li et al. (2019) used hard shell layer to replace the surface plain fill. The settlement and vertical stress of plain fill foundation, hard shell foundation and hard shell foundation with geogrid were compared and analyzed. The results showed that all of the settlement deformation of the foundation and the vertical stress of the underlying soil layer can be decreased by using hard shell layer and geogrid hard shell layer to replace plain fill either in sunny day or in rainy day condition. In the case of rainfall, the foundation with geogrid hard shell layer would play a more effective role in reducing foundation settlement deformation and controlling vertical stress than in the case of sunny day.

3.6   Research frontier identification

The CiteSpace timezone view reveals the knowledge evolution and development trends over time. Therefore, the CiteSpace timezone view was employed to examine the evolutionary path for the most popular special soil subgrade research areas in the WOS and CNKI from 2005 to 2019. The node type was set to "keyword" in CiteSpace, the "time interval" was set to 2005–2019, and the "years per slice" was set to 3. After the analysis, the visualization was set to "timezone view" in the control panel, which revealed the emerging keywords for the different research hotspots over time, as shown in Figs. 7 and 8. It was found from the citation burst history in the visualization for the WOS articles that there were 19 keywords with high mutation intensity between 2005 and 2019, which are shown in Table 5. It was found from the citation burst history in the visualization of the CNKI articles that there were 25 keywords with high mutation intensity between 2005 and 2019, which are shown in Table 6.

Figure  7.  Time zone view map for the special soil subgrade WOS articles.

DownLoad: Full-Size Img PowerPoint

Figure  8.  Time zone map for the special soil subgrade CNKI articles.

DownLoad: Full-Size Img PowerPoint

Table  5.  Emerging words in the special soil subgrade WOS articles.

No.	Keyword	Strength	Begin	End	2005–2019
1	Geotextile	2.8368	2007	2010
2	Forest road	2.9544	2010	2016
3	Unpaved road	3.9533	2010	2014
4	Reinforcement	3.3246	2010	2014
5	Stiffness	2.9248	2010	2012
6	Permafrost	6.8047	2010	2011
7	Cyclic loading	2.9248	2010	2012
8	Ground vibration	2.9496	2010	2012
9	Highway	4.0284	2011	2013
10	Infiltration	3.3097	2013	2015
11	Geocell	3.0223	2013	2015
12	Geogrid	4.3048	2013	2016
13	Unsaturated soil	2.8500	2015	2017
14	Constitutive model	5.3059	2014	2016
15	Simulation	6.0549	2015	2016
16	Permanent deformation	5.1113	2015	2017
17	Load	4.8170	2016	2017
18	Numerical analysis	3.1675	2017	2018
19	Soft soil	4.9800	2017	2018
20	Mechanism	4.9852	2017	2018
21	Dynamic response	5.9803	2018	2019
22	Granular material	3.1845	2017	2019
Note: "Begin" and "End" in the tables are the starting time and ending time of keyword emergence.

 | Show Table

DownLoad: CSV

Table  6.  Emerging words in the special soil subgrade CNKI articles.

No.	Keyword	Strength	Begin	End	2005–2019
1	Powder jet pile	4.5170	2005	2008
2	Frozen soil	3.2822	2005	2008
3	Geosynthetics	5.3359	2005	2007
4	Geocell	3.6367	2005	2006
5	Subgrade diseases	3.9849	2005	2006
6	Geotextile	4.2882	2006	2008
7	Soft foundation	5.5516	2006	2007
8	Reinforce	8.3726	2006	2007
9	Nonlinear settlement	3.2461	2007	2009
10	Vacuum preloading	3.9205	2008	2011
11	CFG pile	3.8078	2009	2010
12	Pore water pressure	4.6211	2009	2010
13	Embankment load	2.9588	2009	2013
14	Settlement prediction	4.9915	2011	2014
15	Static load test	2.9267	2011	2015
16	Silt	2.9397	2012	2014
17	Settlement monitoring	3.0848	2013	2016
18	Cyclic loading	3.0181	2013	2017
19	Excess pore water pressure	3.1342	2014	2017
20	Soil arching effect	3.6097	2015	2017
21	Numerical simulation	5.7990	2016	2019
22	Foamed lightweight soil	4.1180	2016	2019
23	High fill	3.1842	2016	2018
24	Unconfined compressive strength	4.1180	2016	2019
Note: "Begin" and "End" in the tables are the starting time and ending time of keyword emergence.

 | Show Table

DownLoad: CSV

The analysis of the WOS research trends in Fig. 7 and Table 5 identified several key areas over three different time periods, as detailed in the following.

(1) From 2007 to 2010, the key emerging word was geotextile. During this period, geotextile was widely used as a new type of engineering method in subgrade engineering. For example, in 2007, Sarsby (2007) added high-specification geotextile to soft soil subgrade and successfully played a role in improving the shear strength of subgrade soil.

(2) From 2010 to 2012, the key emerging words were forest roads and permafrost. It shows that during this period, the international community attaches great importance to the operation of forest roads. As forest areas often encounter roads located in swamps, to ensure the efficient operation of forest roads and the strength of forest subgrade, researches on strengthening the stability of subgrade are relatively available at this stage. Meanwhile, the problem of road construction in permafrost area has also been widely concerned. With the thawing settlement of frozen soil caused by atmospheric warming and the deformation of subgrade in frozen soil area caused by accelerated thawing of frozen soil, many scholars put forward effective remedial and preventive measures. For example, to address the impact of the expansion characteristics of soils with high clay content, Mousavi et al. (2014) proposed a method by using the RPP as a polymer stabilizer and studied it's controlling effect on expansive soil. Through a series of tests, the results show that the use of RPP can improve CBR and maximum dry density, and reduce the expansion potential and expansion pressure. In 2010, Tarnocai and Bockheim (2017) conducted a study on frozen soil in Canada to determine the origin and nature of frozen soil and classify it for agricultural and construction projects (such as roads, urban sites and airstrips) to ensure the necessary negative heat balance was maintained.

(3) From 2013 to 2019, the key WOS article research areas were numerical analysis, soft soil mechanism, and dynamic response, which indicated that greater attention was being paid to soft soil subgrade. For example, Basack and Nimbalkar (2017) established a numerical model using free strain analysis to understand the strengthening performances of vertical drainage preloading on soft soil foundations, with the comparisons with existing field test resulting in verifying the correctness of the model and revealing that soft clay foundations could be better treated by shortening the drainage path and accelerating the consolidation speed. Lee et al. (2018) took the Korean high-speed railway as the research object, and through field measurements and numerical analyses proved that a decline in groundwater levels was one of the factors causing soft soil foundation settlement. Lu et al. (2018) used Fast Lagrangian Analysis of Continua in 3 Dimensions (FLAC) (3D) software to simulate the soft soil subgrade in Xinxicheng, established a single pile model, and then conducted load tests to compare the cement soil pile and cement sand pile subgrade settlements. Qiu et al. (2017) conducted model tests on an expansive soil section of the newly built Yunnan-Guizhou High Speed Railway to analyze the dynamic characteristics of the section under the cyclic load of trains in an extreme environment, and found that the dynamic subgrade response tended to gradual stability with an increase in the number of vibrations.

From the perspective of evolution trend, international researchers focused on frozen soil and road engineering stabilization in the early stage, and recently years, they paid more attention to soft soil subgrade diseases. Table 5 shows that the initial structure of strengthening and stabilizing soil subgrade during 2005–2019 was mainly geotextile, and the geocell and geogrid structure were used during 2013–2014. By 2017, the granular material as the subgrade base has a significant impact on the deformation behavior of the subgrade. In the aspect of method research, WOS scholars mostly use the numerical analysis method to understand and analyze the foundation performance and to deal with the influence of parameter changes on the foundation. The constitutive model and numerical calculation method are most widely used in the field of special soil subgrade.

The analysis of the CNKI research trends in Fig. 8 and Table 6 also identified several key areas over three different time periods, as detailed in the followings.

(1) From 2005 to 2010, subgrade deformation and settlement problems were key foci. Permafrost can be easily aff5cted by temperature changes which leads to subgrade deformations because of the collapsibility of the loess, the high moisture content, the low density, and the high compressibility of the soft soil, all of which can occur in post construction settlement, thereby aff5cting the normal use of the road. To improve the road subgrade quality and increase the road service life, the CNKI research was focused on subgrade filling and compaction, pavement drainage, subgrade protection, soft soil foundation treatments, loess collapse treatments and other aspects. For example, Cui and Xu (2007) used a powder jet pile method combined with geosynthetics to strengthen soft soil subgrade widening. Zhang et al. (2005) found that a flexible wedge-shaped geocell slab approach could reduce the fill settlement behind abutments and the bump on the bridge head. Sun et al. (2010) used a vacuum preloading method to strengthen foundations, and introduced a nonlinear constitutive relation (Duncan-Chang model) into Biot consolidation theory, and compiled a finite element program to analyze the engineering settlement and pore water pressure development trends. Peng et al. (2015) used monitoring data from 2004 to 2011 to analyze the Qinghai-Tibet thawing settlement deformations in the permafrost subgrade, and divided it into uniform and uneven deformation rates. Ma et al. (2009) used measures such as hot rods, heat preservation boards and heat preservation berms to treat permafrost subgrade problems along the Northeast Railway.

(2) 2011–2016 was a transitional special soil subgrade research stage at which time the CNKI studies began using subgrade design and construction information systems to predict and analyze constructed subgrade settlement deformations, to establish various models to reduce subgrade disasters, and to study the engineering properties of silt subgrade. For example, Ye et al. (2012) developed and applied a cyclic constitutive model to describe the cumulative deformation development process of silt, and then conducted a numerical long-term traffic load dynamic response simulation analysis of the silt subgrade in the Huaibei area of Anhui Province, with the aim of avoiding long-term load subgrade disasters. Wang and Gong (2014) monitored the compaction pit settlement, the amount of ground uplift, and pore water pressure in the reinforcement of saturated silt foundation to obtain the dynamic compaction parameters for the project's silt foundation, and then conducted field tests on the soil layer static penetration and bearing capacity to finally obtain the dynamic compaction parameters suitable for the project.

(3) Researches of 2017–2019 on CNKI harnessed the power of advanced science and technology and new high-tech equipment to promote subgrade construction and select equipment based on the subgrade situation. Subgrade construction technology and schemes that can analyze soil deformation and stress distribution characteristics using numerical simulation methods and determine the influence of the construction parameters on subgrade stability for subgrade widening and reconstruction using numerical analysis have become increasingly mature in China (Feng et al., 2019; Liu et al., 2019a). Because of its high strength, stability and freeze-thaw stability, lightweight foamed soil has been widely used in subgrade construction as a filling material; however, the control of the technology and quality is not perfect. Therefore, Chinese scholars have begun to focus on the application of foamed lightweight soil to roadbed filling (Chen et al., 2019b; Deng, 2018; Jiang et al., 2019; Liu et al., 2019b; Zhao et al., 2019).

In the early stage, Chinese scholars focused on subgrade filling, pavement drainage, subgrade protection, soft soil foundation treatment and loess collapsibility treatment. Recently years, due to the high strength, stability and freeze-thaw stability of lightweight foamed soil, the lightweight foamed soil has been widely applied in roadbed construction by used as filling material in construction. However, because of the inadequate technology and quality control of foam lightweight soil filling roadbed, Chinese scholars began to study the compression and deformation characteristics and to explore the construction technology. Table 6 shows that in 2005, the main methods of strengthening subgrade in China were powder jet grouting pile, geosynthetics, geocell and geotextile. In 2009, CFG pile was used to stabilize the deformation behavior of subgrade.

To sum up, we can see that in the past 2017–2019 years, WOS and CNKI appeared to be more sensitive to stability and reinforcement technology than soft soil and foamed lightweight soil. Therefore, in future research on the field of special soil stability and reinforcement, WOS and CNKI scholars will pay more attention to the soil subgrade stability and reinforcement technology of specific types of soil which were classified by the national policies and relevant standards.

4.   Conclusions

In this article, the WOS core collection and CNKI databases from 2005 to 2019 were interrogated to identify the special soil subgrade research developments. The identified article were loaded in the CiteSpace Visualization software and knowledge maps developed to identify the research development and evolutionary trends, from which the following conclusions were made.

(1) The WOS special soil subgrade research prior to 2009 was relatively low at less than 50 published articles; however, after 2010, there was a significant rise in research interest to over 500 article in 2019. The CNKI special soil subgrade research profile, however, could be divided into two distinct stages: a rapid growth period from 2005 to 2009 and a slow reduction period from 2010 to 2019. The quantitative analysis of the country, institutions and journals most involved in special soil subgrade research found that China was the most active, followed by the United States; however the CPP in the United States was much higher. The quantitative analysis of the WOS and CNKI database data revealed that the co-occurrence network density was low, indicating that there had not been little close research cooperation between different institutions. The research area of most interest in the WOS was the stability and reinforcement of special soil subgrades in the field of transportation technology and in the CNKI was the stability and reinforcement of special soil subgrades in the field of engineering.

(2) The research special soil subgrade foci in the WOS and CNKI were similar. The WOS research was focused on subgrade disaster research, disaster prediction, subgrade stability and reinforcement research, and the CNKI research was focused on soil characteristics, subgrade disasters, stability and reinforcement. The CNKI research tended to pay greater attention to dynamic compaction, CFG piles, and geogrid methods to strengthen and optimize the subgrade, while the WOS research tended to focus more on the use of geosynthetics to strengthen the subgrade.

(3) The special soil subgrade research evolutionary trends also varied across the WOS and the CNKI. The WOS research over the years respectively focusing more on frozen soil, stage road engineering in the early years, and in recent years, they have paid more attention to the study of soft soil subgrade problems, and numerical analysis methods to understand foundation performances and deal with parameter change impacts on the foundation. The application of constitutive model and numerical calculation method has been mostly used in the field of special soil subgrade. The application of constitutive model has increased sharply from 2013 to 2016, and the numerical calculation method is used to determine the construction period, operation period and subgrade settlement during 2017–2019. The CNKI research over the years focused on subgrade filling, pavement drainage, roadbed protection, soft soil subgrade treatment and collapsible loess treatment, and in recent years, on the compression and deformation characteristics of lightweight foam soil, construction technology, and the use of advanced science and technology to improve subgrade construction. In addition, specific types of soil will be classified by the national policies and relevant standards.

(4) After decades of development, the research on the stability and reinforcement of special soil subgrade has matured, with the research directions now being more detailed. The time zone interrogation of the WOS and CNKI data revealed that from 2017 to 2019, besides the focus on soft soil and lightweight foam soil, the research tended to focus more on stability and reinforcement technologies. Therefore, it is predicted that this trend will continue to further improve the application of soft and lightweight foamed soil and stability and reinforcement technologies.

However, to promote special soil subgrade development in China, it is necessary to strengthen both the technological and application innovations, and internationally, it is important that engineering focus more on improving the special soil subgrade systems.