Reference Data Model for NEESgrid

1.  Reference Data Model

In order to facilitate collaboration within the NEES (George Brown Jr. Network for Earthquake Engineering Simulations) framework, one of the key services that NEESgrid needs to support is with respect to the data and metadata for earthquake engineering simulations. A data model is in essence a representation of the data and their interrelationship and provides a conceptual or implementation view of the data. In general, the structure of a data model can be described in different form, such as relational model and object model.  Relational modeling approach organizes data into entities and defines the relationships among the entities. The data in this model typically are saved in relational databases, such as Oracle or MySQL.  In an object data model, information is modeled as objects, which can be any sorts of (real or abstract) entities. The model and data can be represented in different data format, such as RDF (Resource Description Framework) or OWL (Web Ontology Language). 

A reference data model for shaking table and large scale structural tests has been developed in this work. The reference data model has been created in both a relational model and an "OWL-based" object model formats. Both models contain the same entities and interrelationships. However, since direct one-to-one mapping between the two models is difficult to achieve (currently there are no tools available to handle such automatic mapping), there are some minor differences in these two model formats. The following briefly describe the reference data model implemented in OWL and relational formats.

1.1.  OWL-based Model

This reference NEESgrid data model is created by using Protégé 2.1, with a OWL (Web Ontology Language) plugin.  Protégé is a open source software, which is available at http://protege.stanford.edu.  Current version of the reference NEESgrid data model is version 1.2, which is saved as two files:

• neesmeta_1.2.pprj (the project file)

• neesmeta_1.2.owl (the OWL file)

The experimental data used to test this reference data model are saved as zip files, two of them are:

• miniMOST experiment @ University of Illinois, Urbana-Champaign: miniMOST.zip

• pseudo-dynamic test @ University of Minnesota, Twin Cities: UMN_Data.zip

1.2.  Relational Model

The relational model is created with a graphical database modeling tool called Datanamic DeZign. An evaluation version of the software is available at http://www.datanamic.com/dezign/.  A graphical representation of the database schema can be viewed to see the tables and their relationships. For our current development, MySQL 4.0 is used as the backend database system. Since data ingestion tools are not available at this stage, SQL scripts are used to create database schema and to ingest data into the database. Examples of the SQL scripts for creating the database schema and data ingestion are given in the following downloadable files.

• SQL script for creating database schema: create.sql 

• SQL script for ingesting a subset of miniMOST experimental data: ingestion.sql

• SQL script for ingesting a subset of NCREE shake table test: ingestion_ncree.sql

2.  Tools for Data Browsing and Retrieval

A set of tools have been developed to support browsing and retrieving data that are saved according to the developed data model. These tools serve as demonstration on potential usage of a data repository where data are organized confirming to the defined data models. Since data model defines the structure and the relationships of data, related data in the repository can be retrieved. These related data can support further engineering needs such as generating an engineering report. The design of these data retrieval tools is explained in the Design Document. The configuration and the execution of these tools is described in the Test Plan. 

For current implementation, the tools are developed in Java. The source code of these tools is available and can also be downloaded as a zip file.  The relationship among the Java files is illustrated in this figure. As shown in the figure, metadata and data can be saved as either OWL documents or in a relational database. Software tools are developed to retrieve and present related data to the users.  DisplayNEES.java is a program that allows users to navigate through the objects saved in OWL documents.  Both OWLDataRetrieval.java and DBDataRetrieval.java are developed to allow the retrieval of related information and files of a project according to a pre-defined query set.  The query set confirms to typical engineering practice, and the current query set is based on a report structure provided by Prof. Andrei Reinhorn (SUNY-Buffalo). Once related data of a project are retrieved, the data can be reorganized to facilitate engineering report generation. 

Several external libraries were used for the development and running of these tools. Configuration and setup of these external libraries are presented in the README file. The external libraries include the following:

• Jana API to read, write and manipulate OWL files.

• MySQL Connector J/3.0 (JDBC implementation) to connect with the MySQL database.

• Java Servlet technology for dynamically generating web pages.

2.1.  Data Saved in OWL

For the OWL-based model where metadata are saved in OWL files and data are saved in files, two tools are available.  One is a Project Viewer (DisplayNEES.java) to retrieve the saved data and to view the data on a web browser according to the data model. The other is a Data Retrieval (OWLDataRetrieval.java) that retrieves related information and files about a particular project from the repository.  Example usage of these tools is shown in these slides.

2.2.  Data Saved in MySQL

For the relational data model where metadata and data are saved in MySQL database, we can also retrieves related information and files about a particular project from the repository by using a Data Retrieval (DBDataRetrieval.java). Example usage of this tool is shown in these slides.

2.3.  Integrated Data Retrieval for Report Generation

For engineering data/metadata that are saved in different format and different locations, a uniformed access interface is developed to retrieve these data. The retrieved data can be reorganized to facilitate engineering Report Generation (ReportGen.java).

3.  Links and Publications

3.1.  Related Links

The NEESgrid data model was created after a series of review of the following data models:

• [OrSt Model] Oregon State University and Network Alliance for Computational Science and Engineering. NEES Database and Metadata Structure, Version 1.3, white paper, Network for Earthquake Engineering Simulation, 2003. http://nees.orst.edu/IT/data.model/.

• [Ontology of Science] http://protege.stanford.edu/ontologies/ontologyOfScience/ontology_of_science.htm. 

• [Berkeley CUREE/Kajima] http://www.ce.berkeley.edu/~boza/research/Curee-Kajima/.

• [SensorML] M. Botts (ed.) Sensor Model Language (SensorML) for In-situ and Remote Sensors, OpenGIS Interoperability Program Report, OGC 02-026, Version 0.7 Open GIS Consortium Inc, 2002. http://vast.uah.edu/SensorML/.

• [COSMOS/PEER] Geotechnical Data Model, part of 2L02 the Data Archiving and Dissemination Project. http://geoinfo.usc.edu/gvdc.

• [NEES Site Specification Database] http://www.nacse.org/neesSiteSpecs/do/siteSelection. 

3.2.  Reports and Publications

The following are several reports and papers related to the data modeling efforts for NEESgrid:

• K. H. Law, J. Peng and P. Demian, "An Assessment of Data Curation Issues for NEES," Technical Report, September, 2005.

• J. Peng and K. H. Law. "Reference Data Models for Supporting the Network for Earthquake Engineering Simulation (NEES)," Proceedings of ASCE International Conference on Computing in Civil Engineering, Cancun, Mexico, July 12-15, 2005.

• J. Peng and K. H. Law. Data Retrieval Tools: Software Design Specification, NEESit Technical Report, 2005.

• J. Peng and K. H. Law. Data Retrieval Tools: Test Plan, NEESit Technical Report, 2005.

• J. Peng and K. H. Law. A Brief Review of Data Models for NEESgrid, Technical Report NEESgrid-2004-01, 2004.

• J. Peng, K. H. Law, and G. Pekcan. Reference NEESgrid Data Model for Shake Table Experiment, NEESgrid Technical Report, 2004.

• J. Peng and K. H. Law. Validity and Usability of the NEESgrid Reference Data Model, Technical Report NEESgrid-2004-44, 2004.

• K. H. Law. Summary Report on NEESgrid's Data Curation Summit, Technical Report NEESgrid-2004-43, 2004.

• J. Peng and K. H. Law. Reference NEESgrid Data Model, Technical Report NEESgrid-2004-40, 2004.

• J. Peng and K. H. Law. "A Reference Data Model for NEESgrid Shake Table Experiments," Proceedings of International Symposium on Earthquake Engineering in the Past and Future Fifty Years, Harbin, China, 2004.

This work was supported by the George E. Brown, Jr. Network for Earthquake Engineering Simulation (NEES) Program of the National Science Foundation, Award CMS-0117853. Any opinions, findings, and conclusions or recommendations expressed in this material are, however, those of the authors and do not necessarily reflect the views of others and the National Science Foundation.