Benchmarks for Finite Element Pre-Processors

Hose, D R, and Rutherford, I A

First Published - 1993
Softback Report - 44 Pages

This report was produced under NAFEMS contract number NEL/F/520 at the request of the General Technology Working Group. The objective of the contract was to define principles and procedures for the evaluation of finite element pre-processor software with respect to accuracy, robustness, breadth of application and ease of use. The work was divided into two phases. The first was essentially a definition of functionality and a survey of commercially available pre-processor software and was reported in Reference 1. In the second phase, criteria were identified by which to assess such software and a set of benchmark problems was developed.

This document deals with mainly the second phase. Accordingly Section 3 discusses the attributes required of pre-processor software and hence the criteria to be considered in evaluating its suitability. Section 4 presents a number of benchmark problems which are designed for software assessment.

Section 3 and 4 are intended, as far as possible, to be self-sufficient. However, the discussion of pre-processor attributes in Section 3 necessarily assumes some familiarity with the abilities and operations of current commercial pre-processors. Therefore, for completeness, a summary description of the functionality of such software is provided in Section 2. This has been derived from the survey performed in the first phase of the project which is reported in more detail in Reference 1.

The work as a whole is the responsibility of the authors, but it has been guided and assisted by members of the General Technology Working Group under the Chairmanship of Prof. G.A.O. Davies. In addition the opinions and assistance of the finite element user community have been sought. To this end, two questionnaires were devised. The first questionnaire covered in some detail the functionality anticipated of existing pre-processor software, and was distributed to the authors and/or vendors of fifteen commercially available programs. Comments were invited from these sources to ensure that all major facilities were recognised and fairly represented. The second, relatively short, questionnaire was distributed to all U.K. addresses on the NAFEMS mailing list. This questionnaire gave all interested (and sometimes frustrated) users the opportunity to highlight any features or problems which they believed should receive particular attention, either by direct inclusion in a functionality checklist or inclusion implicitly in a benchmark problem.

CONTENTS

FUNCTIONALITY OF EXISITING SOFTWARE

• • Hardware
  • Operating Environment
  • Geometry Definition
    • Hierachical Approach
    • Solid Modelling
  • Mesh Generation
  • Element Properties
  • Material Properties
  • Loading/Boundary Conditions & External Constraints
  • Model Viewing/Verification
  • Model Editing
  • General Capabilities
  • Analysis Capabilities
  • Translators
  • Technical Support

PROCEDURES FOR SOFTWARE EVALUATION

• • General Considerations

The Role of Benchmarks

• • Accuracy
    • Geometrical Accuracy
    • Appropriateness of Finite Element Mesh
  • Speed
    • CPU Time
    • Number of Key Strokes
  • Robustness
  • Ease of Use
    

PROPOSED BENCHMARKS

• • Benchmark Classification

TABLES FOR USER RESPONSE

REFERENCES