Takashi Hasegawa

Monitoring groundwater variation by satellite and implications for in-situ gravity measurements

In order to establish a new technique for monitoring groundwater variations in urban areas, the applicability of precise in-situ gravity measurements and extremely high precision satellite gravity data via GRACE (Gravity Recovery and Climate Experiment) was tested. Using the GRACE data, regional scale water mass variations in four major river basins of the Indochina Peninsula were estimated. The estimated variations were compared with Soil-Vegetation-Atmosphere Transfer Scheme (SVATS) models with a river flow model of 1) globally uniform river velocity, 2) river velocity tuned by each river basin, 3) globally uniform river velocity considering groundwater storage, and 4) river velocity tuned by each river basin considering groundwater storage. Model 3) attained the best fit to the GRACE data, and the model 4) yielded almost the same values. This implies that the groundwater plays an important role in estimating the variation of total terrestrial storage. It also indicates that tuning river velocity, which is based on the in-situ measurements, needs further investigations in combination with the GRACE data. The relationships among GRACE data, SVATS models, and in-situ measurements were also discussed briefly.

Inverse Problem Approach Based on the Kalman Filtering and Its Applications

This paper focuses on the formulation and the numerical performances of the inverse analysis procedure which utilizes prior information in a probabilistic sense. A numerical strategy of Kalman filtering and the extended Kalman filtering technique, in conjunction with the finite element method, is adopted to characterize the material properties of a linear/nonlinear elastic medium considering noisy observations. In applying such a procedure to inverse problem, it is possible to overcome the indeterministic problems which appear in an inverse analysis even when the number of unknowns exceeds the number of measurements. Several hypothetical results of loading tests are given to characterize a distribution of the material properties.

The role of system noise in Kalman filtering for the parameter identification of ground settlement

Publisher Summary The role of system noise or process noise in the framework of Kalman filtering for the parameter identification is examined through some numerical analyses of ground settlement. System noise is considered in the estimation scheme updated with time, and has proven to be a kind of weight on measurement within its time-update algorithm. An example using Asaokas model for ground settlement is dealt with to demonstrate the above effect. It can be pointed out in the numerical example that a smaller value for the covariance of system noise introduces a more precise identification for parameters and a better accuracy for the prediction of the subsequent settlement as expected from theoretical considerations.

NUMERICAL EXPERIMENT FOR VIRTUAL PLASTER MODEL TESTS SIMULATING BLOCK SHEAR TESTS

The strength of in-situ rock masses has been estimated by in-situ rock shear tests for a long time. However, the mechanisms for the appearance of strength in such tests have not been clarified sufficiently. This paper presents the results of a numerical analysis of virtual plaster model tests used to simulate block shear tests, which are of a kind of in-situ test. In the authors’ former study, results were obtained for rock shear tests, another kind of in-situ test, along with real plaster model tests and finite element analyses. In the present study, some cases simulating block shear tests were analyzed. The appearance and propagation of cracks in the testing process were simulated with enhanced elements, which represented the displacement discontinuity in each element, as in the former analysis. The results were compared with the former results to investigate the differences between the two conditions. The shear strength in the two sets of results was found to be generally similar; however, there were some small differences. The patterns for the appearance and propagation of cracks differed from each other, while some common features also appeared. The concentration of stress in the two testing processes occurred in different parts of the materials under the two conditions, and this led to differences in both the failure mechanism and the shear strength.

Finite element analysis for the shear strength appearing in in situ rock shear tests

The mechanisms of the strength which appeared in in situ rock shear tests were examined with a finite element analysis in this paper. A set of plaster model laboratory tests simulating the in situ rock shear tests was analyzed. The plaster models were expressed initially with constant strain triangles (CSTs). Then, the displacement was imposed gradually, and each CST was replaced with a triangular element containing an embedded interface at the point when the stress in each CST reached the failure criterion of the material. The cracking patterns and the deformation obtained from the computation resembled those in the laboratory tests. The cracking pattern under each normal stress deviated from that under different conditions; and therefore, the stress path and the shear strength appeared differently under the various normal stress conditions. The relationship between the apparent shear strength and the failure criterion of the material also changed depending on the normal stress. The shear strength measured in the model tests appeared to be lower than the material strength in the lower and in the higher normal stress ranges. Such differences were thought to occur due to the influence of the stress distributions, which were not assumed, but were caused by several different cracking patterns.