Search

Effort is geared towards development of large-scale nonlinear ground-structure seismic response simulations. Mechanisms to allow for modeling of transmitting boundaries are incorporated, mainly relying on the Domain Reduction Method (DRM) approach. Parallel computing is employed to permit the execution of these large-scale simulations. A range of geometric configurations are addressed in order to explore various aspects of the involved seismic response characteristics. The OpenSees computational platform is employed throughout. To accommodate nonlinear response and soil/structure element stiffness considerations, an implicit time integration scheme is adopted. This scheme poses severe limitations on the number of parallel computing processors that can be used with reasonable efficiency (due to the required taxing communications between the different processors). Within the available constraints on time and computing resources, and the necessary additional OpenSees parallel- implementation machine-specific adaptions, the conducted DRM investigations mostly employed a soil domain 3D 8-node brick element of a 20 m side length (with about 150,000 such elements in the mesh). Consequently, severe limitations are imposed on the frequency content of the propagated seismic waves and the resulting system response. Future extensions in this direction of research can build solidly on the developments in this report and provide more accurate higher frequency system response. Significant attention is given to the simulation of a large-scale highway interchange system under seismic loading. A three-dimensional (3D) Finite Element model of an existing bridge interchange at the intersection of Interstates 10 and 215 (San Bernardino, CA) is developed. This interchange consists of three connectors at different bridge superstructure elevations. Initial focus is placed on modeling the three bridges, evaluation of vibration properties, and validation of one of the bridge models (North-West connector, NW) based on available earlier recorded earthquake response. A strategy to incorporate the above bridge structural models into a bridge- foundation-ground system (BFGS) is implemented based on the Domain Reduction Method (DRM) as developed by Bielak and his co-workers. A numerical implementation of this DRM by Petropolous and Fenves is employed and adapted as the soil domain. In this implementation, seismic waves are propagated from a realistic fault rupture scenario in southern California. As such, the BFGS can include the three-bridge interchange subjected to a 3D seismic excitation scenario. Within this numerical analysis framework, the effect of foundation soils of different stiffness and strength are investigated. The results are compared to the more conventional bridge model response under uniform as well as multi-support base excitation. In addition to this DRM-based implementation, a nonlinear ground-bridge model based on the actual local soil conditions at the interchange is investigated (with the NW only as the super-structure). Efforts include implementation and validation of a classical transmitting boundary at the base of the soil domain. Using this formulation, the BFGS response is compared and validated with earthquake recorded response at the bridge and local site. Under a potential site specific strong ground motion, computed force demands from the employed linear column models are compared to the strength as defined by a representative nonlinear column formulation. Finally, the seismic response of a large rigid structure with different embedment depths is assessed. Dynamic interaction between the structure and the surrounding soil is studied based on changes in soil elastic properties, depth of embedment, and characteristics of input excitation

Additional file 1: Supplementary Figure 1. HPLC chromatograms of (a) water extract and (b) bio-converted MF extract without LLE. Supplementary Figure 2. Effect of acidic modifier on HPLC chromatograms of MF at 320 nm; (a) non acidic modifier in modifier and (b) 0.1% formic acid in mobile phase. Supplementary Figure 3. Change of HPLC chromatograms of MF at 320 nm by controlling gradient conditions; M0: B% 5-20-100 (115 min), M1: B% 10-20-61 (70 min), M2: B% 15-20-45 (70 min) and M3: B% 15-20-45 (60 min). Supplementary Figure 4. Effect of solvent on MF extraction at 320 nm; (a) 100% methanol, (b) 50% methanol and (c) water. Supplementary Figure 5. Specificity of five major peaks in sample by comparing three points of UV spectrums at same peak: (a), (b), (c), (d) and (e). Supplementary Table 1. Optimization of sample preparation by RSM. Supplementary Table 2. Comparison of predicted and experimental values based on peak area (n = 6). Supplementary Table 3. Linearity, LOQs and LODs of the analytes. Supplementary Table 4. Intra-day and inter-day accuracy and precision of major compounds (n = 5). Supplementary Table 5. Repeatability of major compounds (n = 6). Supplementary Table 6. Recovery of the developed method (n = 6).