In the early 1980s, the PLOT3D data format gained acceptance as a de facto standard to enable the storage and exchange of CFD data within analysis processes, and among collaborating organizations. This initial CFD data standard today continues to be the most common storage and exchange standard for CFD data, based on structured grids.
However, by the early 1990s several limitations in the PLOT3D standard had become apparent. Individual organizations were overcoming these limitations by defining extensions to the PLOT3D standard to meet their needs. These extensions were not coordinated among different organizations, and therefore data stored in these extended formats generally could not be utilized outside the originating organization. Further, the PLOT3D standard had not anticipated several key trends in CFD technology, such as unstructured grids, turbulence models based on solutions of partial differential equations, and the need to include chemical species concentrations as part of a CFD solution. Also, the PLOT3D format, which was originally developed simply to expedite post-processing (visualization) did not include self-documenting features. Therefore, it was necessary to rely on file-naming conventions or external notes to maintain awareness of the flow conditions and analyzed geometry of each PLOT3D data file.
The CGNS Data Standard was initially conceived in 1994 by NASA, Boeing , and then-McDonnell Douglas teams working under the Integrated Wing Design element of NASA's Advanced Subsonic Technology Program. The objective of this work was to greatly reduce the time required to design a transport wing. Implicit in this goal was increased extensive use of Computational Fluid Dynamics (CFD) and the possibility of collaborative analyses by many organizations.
To achieve this vision, it was necessary to establish a common data format suitable to meet the needs of production CFD tools in the mid- to late-1990s. This format would be used to enable interchange of data among different CFD-related tools and different computing platforms, and to provide a mechanism for archive and retrieval of CFD data. The chief tools that were taken into consideration for this goal were two structured-grid multi-block codes, OVERFLOW and TLNS3D. The available data standard, the PLOT3D format, was increasingly proving to be inadequate for this purpose. Some of these shortcomings included:
Several database options were considered by the NASA / Boeing / McDonnell Douglas team during the period December 1994 to March 1995. In March 1995, a decision was taken to build a new data standard called CGNS (Complex Geometry Navier Stokes). This standard was a "clean sheet" development, but it was heavily influenced by the McDonnell Douglas Common File Format (CFF) standard, which had been established and deployed in 1989 and significantly revised in 1992.
It should be noted that the CGNS data standard consists of two major
|Data Content and Format||The definition of the intellectual content of the data to be represented in this standard, and the format of the representation in the standard-conforming data file.|
|Implementing Software||Software packages developed to ease the process of establishing CGNS-compliant database references within an applications code.|
In accepted standards contexts such as ISO / STEP, the "standard" consists only of the first item, a definition of the data content and format. In this regard, the CGNS development team went beyond the traditional role in setting standards, by also developing software to easily implement the standard in a code. The implementing software, in turn, was developed in two layers:
The data standards are controlled by two documents, which are available at the CGNS Documentation Home Page. These key control documents are:
The ADF library was developed during 1995, and the first large-scale deployment was made by (then) McDonnell Douglas - St. Louis in November 1995, as part of an upgrade to the Common File Format system. During 1995-97, the NASA - Boeing - McDonnell Douglas team focused on adding content to the control documents, and laying out the requirements of the mid-level library.
At a review in June 1997, the CGNS team (NASA, Boeing, and McDonnell Douglas) determined that additional professional support would be required to produce an adequate mid-level library. Subcontracts were issued to the ICEM CFD Engineering Company , in Berkeley, CA, following this decision. ICEM CFD Engineering in effect became the lead organization for the development of the mid-level library. At this time, the acronym "CGNS" was re-defined to mean "CFD General Notation System", which was more in keeping with the evolved goals of this project.
An initial mid-level software library (version 1.0), which met the original goals of structured multi-block analysis codes, was released in May 1998. At this time, a decision also was taken whereby NASA and Boeing (McDonnell Douglas by this time had been absorbed by Boeing) would relinquish all rights to ICEM CFD Engineering. Concurrently, NASA and the informal CGNS committee determined that there was no need for export authority, so the CGNS standard, the ADF and mid-level library, and all supporting documentation could be distributed worldwide as freeware. Appropriate legal reviews and approvals were obtained at both NASA and Boeing to validate this decision.
At meetings in March, May, and October 1998, the mid-level library was extended to support a wide range of unstructured grid types. The SIDS document defining the standard was modified, and extended versions of the mid-level library were released at intervals in late 1998 and early 1999. By May 1999, the extension to unstructured grids was released.
Up to this time, all activities related to the development of the standard, the implementing software, and the related documentation had been coordinated and largely funded by NASA under the Advanced Subsonic Technology Program. In 1998, NASA took a decision that the Advanced Subsonic Technology program would end on September 30, 1999, which was approximately one year earlier than their original plan. Further, NASA indicated that they would not be able to manage the development of a standard or a software system such as CGNS, once it ceased to be the focus of an ongoing NASA program.
At this time, a number of U.S. and international organizations had established plans to use the CGNS standard and the ADF and mid-level library, and in several cases they had begun implementation. These organizations had a clear interest in the existence of an organization to coordinate future use and extensions of the CGNS standard and its supporting software and documentation. Also during this same period (1998-99), The Boeing Company launched an initiative to establish an ISO standard for aerodynamic data, to be based on the CGNS standard. However, in a best-case scenario CGNS will not become an ISO standard until roughly 2005-2006, and acceptance of CGNS as an ISO standard is not a certainty. It became clear that CGNS needed to find an organizational home, to coordinate its extension and utilization.
The organizations interested in the CGNS standard met in Hampton, VA, on May 20, 1999 to discuss options for a CGNS management organization. Out of this meeting, the CGNS Steering Committee was established. This Steering Committee is a voluntary organization to coordinate the further development and dissemination of the CGNS standard and its supporting software and documentation. In January 2000, the CGNS Steering Committee became an official subcommittee under the purview of the American Institute of Aeronautics and Astronautics (AIAA) Committee on Standards. The AIAA also distributes the CGNS SIDS document as an AIAA Recommended Practice. However, this AIAA affiliation does not preclude the CGNS committee from public dissemination of the SIDS and other CGNS documentation.
The following sections of this document present the vision of how the CGNS Steering Committee will operate.