This is the first in a sequence of five papers dealing with engineering aspects of the surface‐wave magnitude $M_s\approx 6.8$, 1988 Armenia earthquake, the devastating effects of which are still being felt in Armenia. Although the central theme of our three‐year study has been to assess the role of local soil conditions on the magnitude and geographic distribution of damage, we have come to recognize that a close link exists among seismological, geological, geotechnical, and structural aspects of the Armenia earthquake. Our field investigations and analysis of various geologic and engineering aspects of the earthquake have culminated in a series of five papers. This and the companion paper present mostly factual information pertaining to: (1) The mechanics of the fault and its surface breakout; (2) some potentially significant topographic, geologic, and geotechnical features; and (3) the statistics of damage in cities of Spitak, Leninakan (now called Kumayri), and Kirovakan, Armenia, correlated to geologic and soil profiles. In the third and fourth papers of this series, which will appear in a subsequent issue of the journal, theoretical analysis of soil amplification in Leninakan and Kirovakan are described. The fifth paper, also in a later issue of the journal, presents data and analysis of liquefaction‐induced embankment failure case histories.

This is a companion paper dealing with engineering aspects of the surface‐wave magnitude ${M_s}^{\approx }$ 6.8 1988 Armenia earthquake. The first paper, by the same writers and in the same journal issue, provides an overview of the seismological, geological, geotechnical, and structural aspects of the earthquake. The present paper offers building‐damage statistics in the cities of Leninakan and Kirovakan, Armenia. These statistics are qualitatively correlated to the soil profiles to assess the potential importance of any soil amplification effects. The results indicate that in Leninakan, located in a very wide valley, the soil had some effect but could not have been solely responsible for the tremendous damage in the city. In Kirovakan, the soil profiles vary significantly within the city. Damage was primarily within a limited region where there is a valley filled with stiff clays and having width‐to‐maximum‐depth ratio of approximately 5:1. In this region, 74% of the buildings collapsed or were heavily damaged. Two subsequent papers by the present writers, in the following month's issue of the journal, describe theoretical analyses of soil amplification in Leninakan and Kirovakan. The results of these analyses are used to determine if current state‐of‐practice procedures could explain the damage statistics and their local and geographic distribution described in the present paper.

Complete compaction curves of miscellaneous natural soils and sand‐clay mixtures were determined. The simple compression resistances or the penetration resistances were simultaneously determined. On compaction curves a particular water content appeared, the compaction sensitivity threshold (CST), linked to the clay content, the compactive force value, and the clay mineralogy. It is shown that this CST appeared also as a particular water‐content value on resistance curves. These results showed that the maximum dry density had no effect on variations in these resistances. As the water content increased from the CST, the penetration resistance or the simple resistance decreased to zero, a value at which soils might be considered saturated. It is shown that when the water content was lower than the CST aggregates remained individualized and then the hypothesis that the clay content still did not have its lubricating action was reinforced. This lubricating action appeared as the water content was higher than the CST.

The dynamic analysis of surface footings may be performed approximately, but quite accurately, by idealizing the soil as a truncated cone instead of an elastic half‐space. In this paper, cone models are extended to represent embedded foundations. To this end, a double‐cone model is introduced to represent a disk in the interior of a homogeneous full space. Symmetry considerations applied to a pair of double cones enable the simulation of a free surface midway between the two disks. By using a stack of such disk pairs, an embedded foundation is modeled. The subsequent calculation is conducted by conventional matrix methods familiar from elementary finite element analysis. The accuracy of both frequency‐domain and time‐domain formulations is just as good as for surface footings. This is confirmed by comparing the dynamic‐stiffness coefficients for cone models with rigorous solutions. In addition, cone models enable simple determination of the effective foundation input motion imposed by earthquakes upon structures with embedded foundations.

Results of miniature piezocone penetration tests (PCPT) on cohesive soil specimens in a calibration chamber system are presented. The influence of soil type, stress history, penetration boundary conditions, and filter locations on PCPT data are investigated. An increase in the strength of the soil around the cone penetrometer is observed when penetration is resumed after a dissipation test. The penetration depth required to attain a steady excess pore pressure value at the tip of the cone is influenced by the overconsolidation ratio and the lateral stress coefficient. The empirical values of the cone factor and the pore pressure factor are found to be strongly influenced by the lateral stress and the plasticity index. The tip resistance measured during the miniature quasi‐static friction cone penetration tests (QCPT) are consistently lower than the values obtained from the PCPT. Interpretation of the dissipation results to evaluate the radial coefficient of consolidation should be based on the initial dissipation values of the excess pore pressure, and not the penetration pore pressure.

The change in stresses in the rock mass generated by the internal loading of a pressurized conduit, has traditionally been estimated assuming that the rock mass is impermeable and that the internal pressure in the tunnel can be treated as a mechanical load applied to the walls of the opening. The deleterious effects of pressurized seepage, percolating through the natural discontinuities in the rock mass, have generally been ignored, resulting in tunnel alignments with inadequate lateral and/or vertical cover as well as in erroneous interpretations of the in‐situ hydrojacking test results. This paper presents an approach that takes into account the effect of seepage forces by treating the fractured rock mass as a continuous porous elastic medium. The distribution of effective stresses along a radius intersecting the springline of the tunnel was estimated from closed‐form solutions derived for the simplified case of negligible circumferential seepage forces in the medium. The validity of the simplifications was assessed using the finite‐element method. The analysis presented in this report resulted in specific recommendations to establish minimum cover criteria in pressure tunnels, and also provided the basis for reevaluating the methodology currently used to determine the minimum in‐situ stress from hydrojacking test results.

Regulatory agencies often specify a minimum thickness of compacted soil liners that will ensure that the liner performs adequately. No consensus has been formed, however, on an appropriate minimum thickness. In some states soil liners may be as thin as 60 cm (2 ft), whereas in other states soil liners are required to be at least 360 cm (12 ft) thick. Difficulties encountered in the analysis of soil liners, such as uncertainties in construction, flow in macropores, and spatial variability of hydraulic properties, and variability in past experiences of regulatory agencies are the most likely reasons for the lack of consensus on minimum thickness. In the present paper, the first in a two‐part series, two models of fluid flow in compacted soil liners are described. These models incorporate flow in macropores, spatial variability, and uncertainty via probability theory, but only consider advective transport in saturated soil. In the second paper, the models are used to identify an appropriate minimum thickness of soil liners.

The stochastic models described in a companion paper were used to analyze compacted soil liners with a variable number of 15 cm (6 in.) thick lifts. No optimum number of lifts could be defined based on first‐passage time; first‐passage time increased as the thickness increased. Both models showed that the flux through the liner and the equivalent hydraulic conductivity decreased and the mean hydraulic conductivity (modeled as a lognormally distributed random variable) of lifts decreased. There was little benefit to increasing the number of lifts beyond four to six lifts. An analysis of case histories of in situ hydraulic conductivity also showed little reduction in hydraulic conductivity when the number of lifts was increased beyond four to six. Based upon hydraulic conductivity considerations, the recommended minimum thickness of compacted soil liners is four to six lifts, or 60 cm to 90 cm (2 ft to 3 ft).

This paper presents a generalized effective stress model for describing the rate‐independent behavior of normally to moderately overconsolidated clays (with OCRs less than 8) that exhibit normalized behavior. The model formulation comprises three components: (1) An elastoplastic model for normally consolidated clay including anisotropic and strain‐softening behavior; (2) equations to describe the small strain nonlinearity and hysteretic response in unloading and reloading; and (3) bounding surface plasticity for irrecoverable, anisotropic, and path‐dependent behavior of overconsolidated clays. Model complexity is controlled through the use of input parameters that can be obtained from a small number of standard soil tests. A complete procedure is described to select the input parameters for a given clay. A detailed assessment of the predictive capabilities and limitations of the proposed model is presented in a companion paper.

This paper investigates predictive capabilities of the effective stress soil model MIT‐E3 (described in a companion paper) through detailed comparisons with laboratory test data for Boston blue clay (BBC). The main focus of these evaluations is predictions of anisotropic stress‐strain behavior using data from two special laboratory shear devices: the directional shear cell (DSC), and the multi‐directional direct simple shear apparatus (MDSS). Input parameters are first estimated from standard types of laboratory data, including one‐dimensional ($K_0$) compression and undrained triaxial shear tests. The DSC is a plane strain apparatus with the additional capability of controlling principal stress directions during shear. Measurements in this device include undrained shear tests with principal stress rotations at $\mathrm{OCR}=1$, and with fixed stress direction but different principal stress orientation at $\mathrm{OCR}=4$. The MIT‐E3 model describes accurately the observed variations in undrained shear strength and stress‐strain behavior in tests at $\mathrm{OCR}=1$, but is notably less reliable at $\mathrm{OCR}=4$. The MDSS has the same dimensions and boundary conditions as the Geonor simple shear apparatus, but can apply horizontal shear stresses in two independent directions. Measurements in the MDSS apparatus demonstrate the anisotropic properties of BBC at $\mathrm{OCR}=1$ induced by consolidation under an applied shear stress. MIT‐E3 predictions are in excellent agreement with measured variations in peak shear resistance and prepeak stress‐strain response.

The effect of repetitive lateral loads on deflections of two drilled piers in Tampa Bay were significantly greater than predicted by a p‐y procedure commonly used in practice. Reasons for the discrepancy between predicted and measured deflections are discussed. Two methods for predicting the effect of repetitive lateral loads are developed using results of 34 cyclic lateral load tests to quantify model parameters important to the behavior of piles subjected to repetitive lateral loading. The two methods model cyclic lateral load behavior of a pile by degrading soil resistance as a function of number of cycles of load, method of pile installation, soil density, and character of cyclic load. The two methods differ in the computational effort required to make the prediction. The first method is most suitable for hand calculation and rule‐of‐thumb estimation and is based upon a beam‐on‐an‐elastic foundation model with a soil reaction modulus, $K_h$, increasing proportionally with depth. The second method modifies nonlinear static p‐y curves to derive a cyclic p‐y curve. The two methods provide a simple means for estimating effects of cyclic lateral load.

This paper presents a rigorous lower‐bound solution for the bearing capacity of soft clays under embankments. The analysis is carried out by using a numerical technique that is based on the lower‐bound theorem and finite elements. It is found that the lower‐bound solution obtained in this paper agrees well with the upper‐bound solutions obtained from Hill and/or Prandtl failure mechanisms. The lower‐bound solution for the vertical stresses that a soft foundation can support in combination with a surface shear stress (or shear force) can be used to obtain a safe design for the critical heights of both unreinforced and reinforced embankments by following an iterative procedure similar to that described by Michalowski in 1992. The usual assumption of a uniform shear stress mobilization underneath the embankment is shown to be conservative as it leads to a slightly lower value of the vertical bearing capacity.