Link to Report: Coming Soon
Background :
Bridge engineers manage large inventories of bridges, with many bridges located in regions of high seismicity. The preponderance of these bridges are supported on reinforced concrete columns, so the post-earthquake functionality of the traffic network strongly depends on the performance of reinforced concrete columns subjected to axial and cyclic lateral loading. To support the development of accurate and computationally efficient seismic performance models for individual columns, the proposed research will develop tools to enable easy and rapid access to experimental data sets characterizing the response and performance of reinforced concrete columns subjected to cyclic loading in the laboratory. Research will be accomplished by expanding, modernizing, and improving the capabilities of an existing structural performance database hosted by the Pacific Earthquake Engineering Research Center. The data repository and the tools that facilitate selection and compilation of individual bridge column data sets stored in this repository will be shared with the public and state departments of transportation to ensure broad impact.
Scope :
Task 1 – Data Processing.
Since 2004 when the SPD database was originally developed, the University of Washington team has collected additional column test data. Specifically, data have now been collected for:
- 439 tests of rectangular-reinforced concrete columns, of which 355 have cyclic histories, 61 have force-displacement envelopes and 23 have only maxima.
- 293 tests of spiral-reinforced columns, of which 227 have cyclic force-displacement histories, 2 have digitized envelopes, and 94 have only maxima.The first step in the proposed work will be to organize the bridge column data, so that it is compatible with the existing PEER SPD database.
Task 2 – Subset Identification Based on Test Properties
A new feature will be deployed to allow a bridge engineer to find tests that are most similar to a column that exists in the bridge inventory. The matching will be conducted using a relational database, such as SQL, combined with an optimization function to be developed as part of the proposed work. The optimization function will be used to rank similarity between field inventory columns and database tests involving a series of properties, such as span-to-depth ratio, axial-load ratio, longitudinal reinforcement ratio, transverse reinforcement ratio, concrete compressive strength and reinforcement tensile strength. The exact form of the optimization function will be tested by repeated implementation of the algorithm and consultations with potential users.
Task 3 – Subset Identification Based on Code Compliance
A series of Jupyter notebooks will be created that facilitate extraction, from the database of column tests, data that meet the specific requirements of the AASHTO, Caltrans, and WSDOT bridge specifications. These specifications are publicly available and well known to the research team, so they are easily incorporated into Jupyter notebooks using a series of Python functions. For a given specification, the notebooks will compare each specification requirement (e.g., limits on longitudinal reinforcement, minimum amounts of transverse reinforcement for confinement and shear) with the properties for each test specimen. The results of these checks will be incorporated into database searches, so a bridge engineer can quickly identify tests that meet specifications and are similar to bridge columns in the field. Additionally, the notebooks will include functionality to identify the specific ASSHTO, Caltrans and WSDOT specification requirement that an individual column, in the database or in the field, does not meet.
Task 4 – Response Parameters
For each column test, the researchers will identify key performance parameters (e.g., column stiffness, strength, failure mechanism) using Python code. These precalculated values can be used as targets against which to evaluate the accuracy of various analytical models for various subsets of column, or for individual columns.
Task 5 – Analytical Models
For any laboratory test column and for any bridge inventory column, the researchers will develop the capability of automatically generating analytical models. This capability will be developed using a combination of Python Code, which can be accessed via Jupyter notebooks, or alternatively, installed directly on DOT servers.
Task 6 – Dissemination
The research products, including the column database, Jupyter notebooks, and Python scripts, will be made available online, with the option of downloading them to local servers. The software tools will be accompanied with a detailed report that describes the assumptions that are built into the tools, as well as how to use the tool. This report will be available online as a user’s manual.
Using the models and test data generated by this research, it will be possible to quantify the likelihood of an individual column reaching a specific damage state based on the properties of the test and field columns. For example, at a particular level of column deformation it will be possible to identify the likelihood that a particular field column would experience cover spalling, bar buckling and/or shear failure, based on: (1) identification of similar columns in the database and their failure modes as that level of deformation, and (2) the results of standard evaluation methodologies, calibrated by the test data. The combination of the test data and methodology results would allow the engineer to quickly identify columns that are susceptible to damage, as opposed to those that are less susceptible to damage. The susceptibility will be expressed not as the meeting or failure to meet a code requirement, but instead, the susceptibility will be expressed in terms of a likelihood of reaching a particular damage state.
Research Team :
Principal Investigator: Marc Eberhard
Co-Principal Investigator: Laura Lowes