On Methods for Determining the Relative Subsidence of Loess-Like Soils

F. P. Nifantov

(Presented by the scientific seminar of the Department of Hydrogeology and Engineering Geology)

Introduction

In the course of engineering-geological investigations and the design of buildings on loess-like soils, the magnitude of subsidence of the stratum of these deposits is usually determined [1, 10, 13]. At the same time, in recent years a method has been developed for determining the actual subsidence of structures being designed, taking into account the collapsibility only of the compressible stratum of loess-like rocks [3, 4, 7, 8, 9]. In the calculation formulas of both methods, one of the principal design characteristics is the relative subsidence of loess-like soils. The correctness with which this characteristic is determined largely governs the accuracy of calculating the probable subsidence of the stratum and the actual subsidence of structures, and consequently the correctness and economy of assigning anti-subsidence measures that ensure the stability of the buildings being designed.

Despite the great influence of relative subsidence on the final results of calculations of the subsidence of structures or of a soil stratum, this characteristic is still determined very approximately, by different methods, in different instruments, under different experimental conditions, etc. Naturally, therefore, the final results of calculations—especially the magnitudes of stratum subsidence determined according to N and TU 137—56—are inaccurate and cannot serve as a basis for prescribing the types and scope of anti-subsidence measures. It is necessary to develop more accurate methods for determining the relative subsidence of loess-like soils.

The solution of the problem posed and the further improvement of methods for determining the actual subsidence of structures being designed should be carried out on the basis of a comprehensive analysis and consideration of the engineering-geological features of construction sites, the nature of the strength and changes in the subsidence properties of loess-like rocks, with allowance for the action of different types of structures. In this connection, along with general standard methods for determining relative subsidence and methods for calculating stratum subsidence, particular methods should be developed for determining the probable subsidence of various types of structures designed on loess-like soils that differ in their properties and stability. At the same time, along with laboratory methods, field methods should be developed.

methods for determining relative subsidence, including by means of plates in pits and boreholes.

The need to improve methods for investigating and evaluating the subsidence of loess-like deposits is especially emphasized in the decisions of the Dnepropetrovsk conference, where the advisability of revising the existing standards for construction on loess-like soils is indicated [6].

Taking into account the above-mentioned recommendation of the conference, some directions for improving the laboratory method for determining the relative subsidence of loess-like deposits are considered below.

Modern methods for determining the relative subsidence of loess-like rocks

In our country, the most widespread at present are laboratory methods for determining relative subsidence. Field methods, for example the method of testing with plates, have proved to be almost forgotten, evidently because of their cumbersomeness, high cost, and, most importantly, the impossibility of directly obtaining the value of the relative subsidence of the rocks under study. Among laboratory methods, the methods of Yu. M. Abelev and N. Ya. Denisov, which are closely similar to one another, have become best known [1, 3, 4]. According to N. Ya. Denisov’s method, samples of the rocks under study, with their natural moisture content and natural structure preserved, are placed in the rings of compression devices and, under specified pressures, are soaked with water. As a result, the additional settlement of the soil upon soaking is determined. From this value the “index of additional settlement” is calculated by the formula

\[ R=\frac{S_2}{h-S_1}\cdot 100, \]

where

\(S_2\)—additional settlement of the rock in the device upon soaking under a specified load \(P\) kg/cm²;

\(S_1\)—total settlement of the rock in the device under the specified load before soaking;

\(h\)—initial height of the specimen.

Despite its sufficient clarity and simplicity, N. Ya. Denisov’s method has not been included in standards and technical specifications as a standard method and has not found wide application in production tests of the subsidence of loess-like rocks at construction sites.

The method of Yu. M. Abelev has become much better known and more widely used in practice. According to this method, relative subsidence (the coefficient of relative subsidence) is calculated from the coefficient of macroporosity. The latter, as in N. Ya. Denisov’s method, is calculated from the additional settlement upon soaking of moist subsiding soil, but compressed in a special or ordinary compression device under a pressure of 3 kg/cm², and not under the actual load. The determination of relative subsidence is carried out by the formula

\[ i_m=\frac{\varepsilon_m}{1+\varepsilon_p};\qquad \varepsilon_m=\varepsilon_n-\varepsilon_p, \]

where \(\varepsilon_m\)—coefficient of macroporosity, determined from the difference in the coefficient of porosity of the soil before and after soaking under a load of 3 kg/cm²;

$\varepsilon_p$ — the porosity coefficient of the soil in its natural moisture state, compressed in the oedometer under a load of \(3\ \text{kg}/\text{cm}^2\);

\(\varepsilon'_p\) — the porosity coefficient of the soil after subsidence under the same load.

This method for determining relative subsidence has a number of shortcomings. First of all, the magnitude of additional subsidence is determined not under the actual load that will be transmitted to the soil by the structure. Secondly, wetting of the soil in the apparatus is carried out from above, which, on the one hand, leads to the accumulation of air in the pores of the soil and to a reduction in the amount of subsidence, and, on the other hand, with prolonged filtration in the special Abelev–Ozeretskii apparatus, salts may be leached out. Finally, the method of constructing the curve and calculating relative subsidence by the macroporosity coefficient is more complicated than calculating the additional-settlement index by the method of N. Ya. Denisov.

Despite the shortcomings noted, Yu. M. Abelev’s method, with minor changes, has been included in the current building codes and regulations (SNiP) and in N and TU 137—56 [10] as mandatory for broad practical application. The Dnepropetrovsk conference on construction on loess-like soils recommended revising N and TU 137—56, eliminating from them the evaluation of subsidence at a pressure of \(3\ \text{kg}/\text{cm}^2\), and replacing it with the actual pressure that will arise in the zone of subsiding soil beneath the structure. It should be emphasized that this recommendation had repeatedly been proposed in a number of works before the publication of N and TU 137—56, but was not taken into account when the latter were compiled.

Unfortunately, in the resolutions of the conference on the methodology for investigating and evaluating subsidence, the principal directions for the study of loess-like soils were not emphasized. The conference recommended only “broadly developing an in-depth study of the properties of loess-like subsiding soils and improving methods of construction on them.” Such a general recommendation is explained by the fact that the process of additional settlement (subsidence) of structures depends on many factors and often proceeds in a highly peculiar manner. In each specific case, the subsidence of loess-like soil and of the structure depends substantially on the degree of wetting of the soil, the applied load, the nature of the foundation, and the properties of the soil. It is therefore very difficult to establish in advance general regularities in the subsidence process of various structures erected on different subsiding soils.

In works of recent years, a whole series of differing propositions on the question under consideration has been presented, which indicates that the problem remains unresolved. In addition to the obvious dependence of subsidence on composition, structure, moisture content, and load, emphasis is also placed on the influence of lateral thrust, of air trapped in the pores of the soil, and of the influence of methods and apparatus on the course and magnitude of subsidence. Thus, for example, Yu. M. Abelev, in a report read at the Dnepropetrovsk conference, points out that laboratory methods for determining relative subsidence usually give underestimated results in comparison with actual values, and that corrections must be introduced each time into the results of these investigations [2]. Indications of deviations of laboratory data from actual values are given in their works by N. V. Ornatskii [11], L. A. Rubinshtein [12], and other researchers. But what correction coefficients should be introduced into laboratory data has not yet been established. Moreover, the coefficients will obviously not be the same for different soils, for different methods of testing subsidence, and for different structures, which again introduces uncertainty and subjectivity into the resolution of the question.

Taking the foregoing into account, it is first of all necessary to decide whether laboratory methods for determining the relative subsidence of loess-like soils should be retained and improved, or whether one should switch completely to field methods, which have not yet been developed at all. In order to give a fuller answer to the question posed, let us consider some materials clarifying the dependence of relative subsidence on load and moisture content.

Dependence of the relative subsidence of loess-like loams of the industrial regions of Kuzbass on load and moisture content

For the loess-like loams of Kuzbass, just as for loess-like rocks of other regions of the USSR, a change in relative subsidence with the applied load at the moment the soil is wetted is clearly observed. This dependence had previously been established for subsiding loams of the Kemerovo, Prokop’evsk, and Tomsk districts [7, 8, 9]. Below, in Fig. 1, new data are presented on the change in relative subsidence with load, obtained for loess-like soils of the Belovo district of Kuzbass.

Fig. 1. Dependence of relative subsidence on load.
1 — slightly moist loams with a degree of moisture less than 0.5; 2 — moderately moist, with a degree of moisture of 0.5–0.7; 3 — moist and saturated loams with a degree of moisture greater than 0.7.

As can be seen from the graph, relative subsidence increases sharply with increasing load in slightly moist soils, less markedly in moderately moist soils, and increases very weakly or not at all in moist soils. These data indicate the need to determine the relative subsidence of loess-like soils at construction sites under such loads as will be transmitted to the soil at the depth from which monolith samples are taken by future structures.

An even more characteristic dependence proved to be that of relative subsidence on the initial natural moisture content at which the subsiding soil is wetted. This dependence, like the dependence of subsidence on load, has been established for loess-like loams of a number of industrial regions of Kuzbass [7, 8, 9], including for the Belovo district (Fig. 2).

Fig. 2. Dependence of relative subsidence $i_m$ on degree of moisture. Determinations in compression devices under a vertical load of 3 kg/cm².

Similar relationships are confirmed for loess-like deposits in other regions of the USSR as well. However, with regard to the influence of natural moisture content on subsidence, it is always necessary to bear in mind that in the aeration zone, i.e., usually at the depth at which foundations are placed, the natural moisture content changes over time; and during construction the soils in foundation pits prepared for foundations often become saturated with water to the state of full moisture capacity. In such cases, loess-like rocks sharply reduce their structural bonds and pass from the category of subsiding soils into strongly compressible soils. Thus, the value of relative subsidence determined for slightly moist soils during the investigation period may prove unrealistic for the construction period and for the moment when the soil beneath the structure is wetted.

The question arises: at what initial natural moisture content should the relative subsidence of loess-like rocks be determined? Should one adopt the natural moisture content for the period of investigations, as is done at present, or take soils with natural moisture contents characteristic of the dry and wet periods of the year? How should one account in advance for the possible moistening of the soil during the construction period, when various trenches and foundation pits will be excavated on the construction site, periodically flooded with water, and the soils of the compressible zone in a number of areas will pass into the state of a two-phase system, changing their properties and resistance to load?

It is quite obvious that investigations of the subsidence and stability of loess-like rocks should be carried out at the natural moisture content and density that the soil will have by the time of construction and wetting. But it does not appear possible to predict this moisture content during investigations of construction sites. Therefore, in the course of investigations it is wholly insufficient to determine the relative subsidence only at some single accidental natural moisture content and load, as is recommended by modern methods and standards; rather, broader and more comprehensive investigations of subsidence must be undertaken under changes in moisture content, load, and degree of wetting. The latter—that is, the degree of wetting of the soil—has no less significant an effect on the amount of subsidence [1, 2].

For calculating the maximum subsidence of planned structures, it is required, at a minimum, to determine the following relationships for each construction site with subsiding soils: the variability of the relative subsidence of soils under different loads and with complete soaking of the soil with water during the tests; and, in addition, the dependence of subsidence on the initial natural moisture content. At the same time, changes in structure and physical properties should be studied in order to ascertain the causes of soil subsidence.

The relationships listed above are readily established by laboratory investigations in ordinary compression apparatuses, provided that the methodology of these investigations is changed.

Principal directions for improving the laboratory method of determining the relative subsidence of loess-like rocks

The investigations we have carried out showed that, among the existing laboratory methods, the best results in determining the relative subsidence of loess-like rocks are obtained by the method of plotting two compression curves, published by R. A. Tokar as early as 1937. Later this method was described in a special instruction [5]. According to this method, for each monolith of subsiding soil taken at a construction site, two compression curves are obtained:

curves: the first at the minimum moisture content of the soil, for which purpose the wet samples are dried; the second at full saturation of the soil with water before its compaction.

Taking into account the present-day requirements for the methodology of determining the relative subsidence of loess-like deposits, this method requires modifications, which are as follows. The first compression curve should be obtained for naturally moist, undisturbed soil having the minimum natural moisture content. For this purpose, monoliths for investigation should be taken during the driest periods of the year. While obtaining the first curve, given the low moisture content of the soil, the load increments may be increased to one kg/cm², which will speed up the process of obtaining the curve. At a load of 4 or 5 kg/cm², when stabilization has occurred, the soil is soaked with water from below and the course and magnitude of the actual subsidence are determined. The latter is necessary for comparison with the data of the second curve. After completion of the subsidence, one more load increment is applied. When stabilization of settlement occurs, the water is completely removed from the device and the elastic rebound curve of the soil is obtained. Then the device is unloaded and the bulk density, moisture content, and degree of closure of the soil macropores are determined. The latter is done by comparing the surface of the compacted specimen with its photograph taken before compression.

Following the first curve, on the same compression apparatus an ordinary compression curve is obtained for the same monolith, with the soil soaked before the load is applied, so that no swelling of the soil occurs. This curve should be obtained in exactly the same way as ordinary compression curves are obtained when studying the compressibility of other soils. The only difference is that soaking of the soil, as when obtaining the first curve, should be carried out from below, in order better to displace air from the soil pores without carrying out fine particles and leaching soluble salts. In both cases, soaking should be done with cold water containing a minimal amount of adsorbed air and dissolved salts. After the second curve has been obtained, the water is removed from the device, the elastic curve of the soil is obtained, and the bulk density, moisture content, and degree of closure of the macropores are determined. All these operations are performed as in the first case. The obtained coefficients of porosity and moisture content are compared with the first ones, and if the discrepancy proves to be no more than 3–5%, the tests are concluded and the initial data are sent for processing. In the course of processing the materials, two graphs are constructed (Fig. 3), each with two compression curves, from which the relative subsidence of the soil is easily determined for any given vertical load.

Thus, this method makes it possible to determine the change in the degree of relative subsidence of the soil as a function of load and to obtain the ordinary compression curve required for calculating the settlements of structures. By contrast, the currently used standard method makes it possible to determine only one purely accidental value of relative subsidence under one specified load. As for N. Ya. Denisov’s opinion regarding the impossibility of compacting the soil when obtaining the second compression curve, our investigations did not confirm this. On the contrary, as is clearly seen from Figure 3, the compaction of the soil when obtaining the second curve almost always proved to be greater than after subsidence under the specified load. The exceptions were individual determinations, evidently with errors in the tests or with different macroporosity of the specimens taken. The latter was also noted by N. Ya. Denisov as a major defect of this method. In order to reduce it to a minimum, the specimens in the rings of the apparatus...

samples should be taken from one section of the monolith successively one above another, that is, along the direction of development of macropores, which are almost always oriented vertically.

Should the method recommended by SNiP and Pom and N and TU 137—56 [10] be retained in practice? It seems to us that it should not, since the results obtained by this method, as emphasized above, very often turn out to be accidental, usually understated, and not corresponding to the actual maximum soil subsidences under the specified pressures. Only in isolated cases, when it is necessary to carry out investigations very quickly and to determine the amount of subsidence approximately, may the use of this method be allowed, with soaking of the soil being carried out under actual loads.

Fig. 3. Graphs for determining the coefficient of macroporosity and relative subsidence from data of testing loess-like soils in compression apparatuses. 1 — compression curve of soil with minimum natural moisture content; 2 — compression curve of soil saturated with water before soaking.

In those cases in which the relative subsidence of the soil must be determined for foundations transmitting dynamic loads, tests in compression apparatuses must be carried out by this and other methods with the application of the corresponding dynamic loads.

Conclusion

The observations noted above concerning the application of the two-curve method for determining relative subsidence are regarded by us as the first stage in the necessary fundamental improvement of investigations of the properties and collapsibility of loess-like rocks. Simultaneously with the development of laboratory methods, simple field methods that provide the necessary accuracy should be more widely developed for investigating the stability and collapsibility of loess-like rocks, including methods using small plates in boreholes.

In the course of laboratory investigations, special attention must be paid to clarifying the nature of the strength and collapsibility of loess-like rocks and to determining the dependence of the amount of subsidence on various factors. Investigations of this character will make it possible more quickly to develop the most accurate methods for assessing the stability and collapsibility of structures designed on loess-like soils and to devise effective and simple measures for combating subsidence.

LITERATURE

1. Abelev Yu. M. Fundamentals of Design and Construction on Macroporous Soils. Stroyvoenmorizdat, 1948.

2. Abelev Yu. M. Subsidence phenomena and their regularities for clayey soils. Abstracts of a report at the Dnieper Conference, 1957.

3. Denisov N. Ya. On the Nature of Deformations of Clayey Rocks. Publishing House of the Ministry of the River Fleet, 1951.

4. Denisov N. Ya. Construction Properties of Loess and Loess-Like Loams. Moscow, 1953.

1. Instructions for Determining the Subsidence Properties of Macroporous (Loess-Like) Soils (I-111-48), Stroyizdat, 1950.

2. Materials of the Dnepr Conference. Dnepropetrovsk, 1957.

3. Nifantov, F. P. Geology of loose deposits and the physical properties of surficial loess-like rocks of the Kemerovo district of Kuzbass. Tomsk Polytechnic Institute, 1951.

4. Nifantov, F. P. Engineering-geological characteristics of loess-like deposits in certain industrial districts of Kuzbass. Proceedings of the Conference on Engineering Geology, vol. I, Academy of Sciences of the USSR, Moscow, 1956.

5. Nifantov, F. P. Determination of the subsidence of structures on loess-like loams of Kuzbass. Bulletin of the Tomsk Polytechnic Institute, 1958.

6. Standards and Technical Specifications for the Design and Construction of Buildings and Industrial Structures on Macroporous Subsiding Soils (N i TU 137—56), Moscow, 1956.

7. Ornatsky, N. V. Underestimation of settlements in determining soil subsidence by the method of soaking under load. Abstracts of a report at the Dnepropetrovsk Conference, Dnepropetrovsk, 1957.

8. Rubinshtein, A. L. Engineering-geological properties of loess-like soils and their deformations beneath irrigation structures. Proceedings of the Conference on Engineering Geology, vol. I, Academy of Sciences of the USSR, Moscow, 1956.

9. Construction Norms and Rules, Part II. State Publishing House for Literature on Construction and Architecture, Moscow, 1954.

10. Tokar, R. A. Quantitative characterization of macroporous (loess-like) soils. Collection of Fundamentstroy, No. 7, 1937.