AGS Guide to Validation and Use of Geotechnical Software
Association of Geotechnical & Geoenvironmental Specialists (AGS) Guide – Validation and Use of Geotechnical Software (Version 1.0, 1994)
The guide was written by a Working Party set up by AGS in January 1993 to establish a procedure to help specialist and non-specialist engineers in validating computer programs for geotechnical design. Applicable programs include user-developed spreadsheets and programs, in-house-developed software, bespoke and package software, created to assist in designing or analysing geotechnical structures.
Having explained the terms verification and validation, the guide uses a Software life-cycle flow chart to show the relationship between verification and validation, and illustrate where the Developer and the end-User fit into the overall software testing process. The guide is primarily concerned with validation by the end-User.
The testing of software is the heart of the guide. Flow charts are also used to illustrate the steps in Program-based testing (i.e. does the software perform as claimed?) and Problem-based testing (i.e. does the software do what is needed?).
Within the Chapter on Program-based testing under the heading of “Reviewing the Developer’s quality assurance procedures”, reference is made to the NAFEMS Quality System Supplement (QSS) to ISO 9001 relating to finite element analysis in the design and validation of engineering products. An extract from the QSS is included in the appendices that defines three categories of finite element analysis relevant to the design and validation of engineering software, i.e. Vital, Important and Advisory.
In the Chapter on Problem-based testing, the categorisation of geotechnical problems by the degree of risk involved is suggested, to help determine the quantity and breadth of testing. The NAFEMS QSS is referred to as a basis for subdividing the degree of risk into categories. The guide also covers some aspects of geotechnical modelling and the effects of simplifications and omissions.
Once the user has established that testing is required, there is another flow chart and supporting text taking them through testing planning, the establishment of worked examples as controls, implementation and reporting. There is a useful validation case study based on an actual slope stability program that follows the principles of the guide.