Estudio del comportamiento tenso-deformacional de pedraplenes inundables mediante simulaciones numéricas formuladas en diferencias finitas y calibradas con lecturas de instrumentación [thesis]

Ignacio Escuder Bueno
This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of Computer printer. The quality of th is reproduction is d ep e n d e n t upon th e quality of th e copy subm itted. Broken or indistinct prínt, colored or poor quality illustrations and photographs, prínt bleedthrough, substandard margins, and improper alignment can
more » ... ersely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorízed copyright material had to be removed, a note will indícate the deletion. La herramienta básica de análisis ha sido un programa de cálculo basado en la técnica de las diferencias finitas (FLAC, Itasca 1994) cuyos resultados han sido contrastados con los registros de presiones y asientos procedentes de la instrumentación de control. Para un completo aprovechamiento de los mismos, tomados a escala semanal, se ha simulado la secuencia constructiva real, lo cual permite contrastar los resultados del análisis con más de doscientas medidas registradas. En particular, se han llevado a cabo análisis numéricos basados en modelos de naturaleza elástica y lineal, elástica no lineal, elastoplástica, visco-elástica asi como reproducido fenómenos de reblandecimiento por humectación. Destacan por su originalidad algunas de las rutinas redactadas para su uso en los modelos elásticos no lineales, así como la metodología de simulación de las deformaciones de fluencia y humectación en el entorno de las diferencias finitas. El conjunto de los trabajos desarrollados ha permitido caracterizar y poner de manifiesto la importancia en orden de magnitud de las deformaciones de naturaleza viscosa desde el comienzo de la construcción de los grandes rellenos de escollera, así como las capacidades y deficiencias de los modelos habituales de cálculo en cuanto a la reproducción del comportamiento observado. III R eproduced with ABSTRACT This Thesis studies the stress-strain behavior o f floodable rockfills, using data obtained from quality control o f materials, control o f construction and instrumentation records. As a case o f study, a rockfill part o f the final works for a new Madrid-Valencia motorway, located at Contreras Reservoir is used. Data were collected during construction (December 1997-August 1998) and are extended to July o f 2000. After reviewing the State of art on properties of usual materials, models o f behaviour, numencal tools and experiences dealing with studies based in combined analysis and field measurements, several works have been developed. Namely the synthesis o f all available data, study o f construction procedures, implementation o f an analysis methodology and its application to the study of the stress-strain behavior during and añer construction. FLAC 2D (Itasca, 1994), an explicit fínite difference code, has been selected as numerical tool to perform the analysis, and results have been compared with measurements registered by total pressure and settlement cells. In order to improve the quality o f analysis and to make use o f all collected records to calíbrate the models (taken on a weekly basis), the real constructive sequency has been simulated. Numerical calculadon based in linear elastic, non linear elastic, elastoplastic and viscoelastic models have been performed. Newly developed routines have permited to accomplish the upgrading o f tangent parameters involved in non-linear hvperbolic formulation, calculadon of creep deformation and settlements due to reservoir filling. As a result o f the works, the stress-strain behavior o f the structure has been characterized, the importance o f creep deformation from fírst stages o f construction has been identified, and capability o f usually assumed models in reproducing observed behavior has been evaluated.
doi:10.4995/thesis/10251/57270 fatcat:lv2tsmzcynej7cx4iwpzvvfqt4