Quality Assurance Policy

From wiki.olisystems.com
Jump to: navigation, search


OLI Systems, Inc. (Cedar Knolls, New Jersey, United States) employs a multi-tiered quality assurance policy for the development and deployment of the its software packages.

This policy covers the OLI Simulation Software including:

  • OLI Studio: Stream, Corrosion and ScaleChem
  • OLI Engine: Developer Edition (sometimes called the OLI API or the callable Engine)
  • OLI Flowsheet: ESP (planned for 2015)
  • OLI Pro


The OLI Original Software including:

  • OLI ESP Original
  • OLI ScaleChem Standard


And the OLI Alliance Software including:

  • OLI Engine in Aspen HYSYS
  • OLI Engine in Aspen Plus
  • OLI Engine in gPROMS
  • OLI Engine in IDEAS
  • OLI Engine in PRO/II
  • OLI Engine in UniSim Design


Each software package contains three basic units, each which are tested individually and then as a full package. These units are

  • Databanks
  • Numerical Solver (also known as the Engine)
  • Interface


Databanks and data consistency

The OLI Databanks are developed from published literature data. This data is reduced via a data regression program and thermodynamic parameters developed. These parameters and then used by the numerical engine to predict the data used from the literature. This is called data validation. All new data validation results are stored and are available to end users upon request for specific data systems. A listing of the source material is also stored and is also available upon request. The actual literature data is stored in hardcopy form and is not available to end users from OLI Systems as this could potentially be a copyright violation. Prior to each release of the databanks a difference program is run to compare the contents of the new databank against the contents of the previous databank. The differences that are recorded are:

  • Additions
  • Deletions
  • Changes

These differences are recorded and published in the release notes for the databank. The release notes are available from the OLI Systems, Inc. support website http://support.olisystems.com Numerical Solver Each business day the code base for the numerical solver is compiled. This is called the nightly build. This nightly build is then run against a standardized set of test cases. These test cases are for the OLI ESP software. Any numerical failures of non-converged cases or crashes are then investigated for cause. Prior to the release of the software a final build is created and tested. If there are no errors the build is then incorporated into the interface and an installation file created. A sub-set of the ESP software cases used to test the numerical engine is run against the installation file. These results are available on the OLI support website. Interface


Prior the release of each software build the interface is tested. The testing of the interface varies depending on the program being tested. The following programs are not tested at OLI Systems:

  • OLI Engine in gPROMS
  • OLI Engine in IDEAS

These programs undergo acceptance testing at the testing facilities of their respective Alliance Partners. The testing of the following programs is tested manually according to testing scripts developed by OLI. These scripts test user input and functionality:

  • OLI ESP Original
  • OLI ScaleChem Standard
  • OLIPro
  • OLI Engine in Aspen HYSYS
  • OLI Engine in Aspen Plus
  • OLI Engine in PRO/II
  • OLI Engine in UniSim Design

The testing of the OLI Studio program consists of a manual test according to testing scripts and automated testing. The automation testing uses a program that can be purchased separately and uses a database of tests. The automated tests can test the following calculation types in the OLI Studio program:

  • Isothermal
  • Isenthalpic
  • Bubble point
  • Vapor Fraction
  • Precipitation Point
  • pH set point
  • Temperature surveys
  • Pressure Surveys
  • Composition Surveys
  • pH surveys
  • Water Analysis
  • Equilibrium calculations
  • pH/Alkalinity reconciliation
  • pH reconciliation
  • Alkalinity calculation
  • Corrosion Rate calculations

The automatic tester cannot test OLI Studio ScaleChem objects. The results of each test are compared to the stored version. Any software variable with an absolute difference of 0.0001 (in the native unit of the variable) is flagged for cause. Any relative difference of 0.05 (in the unit of the variable) is also flagged for cause. Possible sources of differences are:

  • Databank changes
  • Numerical solver changes
  • Interface changes

Each flagged difference is investigated for cause. In the case of databank changes, the databank may have added new chemistry which means the difference is acceptable. In some databank changes, there was an error in the databank which is then further investigated. The numerical solver may have added a new feature or improved the calculation of a parameter. If this is the case then the differences are deemed acceptable. Interface changes may introduce errors which are investigated. If this is the case, the error is corrected. Trademarks

Aspen Plus and Aspen HYSYS are trademarks of Aspen Technology, Inc. Cambridge, Massachusetts

IDEAS is the trademark of Andritz AG, Graz, Austria

gPROMS is the trademark of Process Systems Enterprise, LTD., London, United Kingdom

PRO/II is the trademark of Schneider Electric, Rueil-Malmaison, Ile-de-France

UniSim Design is the trademark of Honeywell, Morristown, New Jersey