New Software Release

ECLIPSE 2021.3 Now Available

Tuesday, 19 October 2021

Product overview

The ECLIPSE industry-reference reservoir simulator offers the most complete and robust set of numerical solutions for fast and accurate prediction of dynamic behavior for all types of reservoirs and development schemes. The ECLIPSE simulator has been the benchmark for commercial reservoir simulation for more than 35 years thanks to its extensive capabilities, robustness, speed, parallel scalability, and unmatched platform coverage. The ECLIPSE simulator covers the entire spectrum of reservoir modeling, including black oil, compositional, thermal, and streamline simulation. 

Learn more about ECLIPSE

Release highlights

The 2021.3 release of the ECLIPSE suite includes developments enhancing the performance, usability, and functionalities of the simulators.

ECLIPSE 2021.3 key developments

Improved workflows for CO2STORE option

The ECLIPSE 300 CO2STORE Thermal facility has been improved and workflows for commonly used
options have been streamlined. These enhancements improve the usability, extend the compatibility, and enhance the reporting of CO2STORE option (both isothermal and thermal).

  • In CO2STORE Thermal option, component solubility can now be defaulted (to “nonsoluble”) in KWTABt keyword family.
  • In CO2STORE thermal option, component viscosity can now be defaulted for special components, and the model will use specialist internal calculations.
  • Leverett J-function can now be used in combination with CO2STORE.
  • Solid properties are now reported in the EOS section of the PRT file for soluble solid components.
  • The component aqueous mole fraction (BAMF) summary keyword can now be used in nonthermal CO2STORE cases.
  • Placeholder component values used in the KVCR keyword will no longer make the CO2 and H2O components nonlive, which would previously cause the run to fail.


As part of our ongoing commitment to improving the ECLIPSE simulator, ECLIPSE 2021.3 contains several changes to address issues discovered in previous releases. For instance, for both ECLIPSE 100 and ECLIPSE 300 simulators:

  • An issue which could cause incorrect calculations for the final connection of a well when applying binary operations to a combination of well and connection UDQs has been fixed.

Additionally, for ECLIPSE 100 simulator:

  • The output of the GWIRNB and GGIRNB summary vectors in cases where a model also has an extended production network has been corrected.
  • Reporting for parallel runs using region-specific conditions in ACTIONR and ACTIONX keywords has been corrected. Previously, messages were sometimes generated implying that an action had been triggered.
  • The contents of the TUNING keyword are now echoed to the PRT file.

Finally, for ECLIPSE 300 simulator:

  • An option has been added to report the ROPR and RGPR summary vectors using the region separator rather than the well separator.
  • An option has been added to make the combination of the WPIMULT and CSKIN keywords compatible with ECLIPSE 100. By default, in ECLIPSE 300, when CSKIN is specified, other well values are taken from the original keyword specification rather than the current values.
  • ACTIONX reporting has been enhanced to match ECLIPSE 100. Complex conditions are now spread over two lines, which avoids truncation of conditions and values. Additionally, it is now clear when more wells match a condition than have been reported.
  • An issue where newly specified connections could fail to be assigned the correct separator for the well has been fixed.
  • Compositional input using the ZI keyword is now compatible with other keywords. 
  • NNC output to the grid files is now correct for models using the SECTOR keyword.
  • In cases with matching EOS, PVT and saturation region counts, the EOS PRT output for Kvw is now calculated using the same table number for each region set. 
  • The solution array PRT output now reports the correct values for minimum and maximum in parallel, as well as serial runs.
  • Well messaging output has been improved to avoid large numbers of duplicated messages in parallel runs.