EPANET/ReleaseNotes2_3.md

Release Notes for EPANET 2.3

This document describes the changes and updates that have been made in version 2.3 of EPANET.

• The check for at least two nodes, one tank/reservoir and no unconnected junction nodes was moved from EN_open to EN_openH and EN_openQ so that partial network data files could be opened by the toolkit.
• A EN_setcurvetype function was added to allow API clients to set a curve's type (e.g., EN_PUMP_CURVE, EN_VOLUME_CURVE, etc.).
• A EN_setvertex function was added to allow API clients to change the coordinates of a single link vertex.
• The index of a General Purpose Valve's (GPV's) head loss curve was added to the list of editable Link Properties using the symbolic constant name EN_GPV_CURVE.
• The EN_getlinkvalue and EN_setlinkvalue functions were updated to get and set the value of EN_GPV_CURVE.
• Negative pressure values for EN_SETTING are now permitted in the EN_setlinkvalue function.
• The EN_STARTTIME parameter was added into the EN_settimeparam function.
• A EN_DEMANDPATTERN parameter was added as the index of the default time pattern used by demands with no specific pattern assigned. It can be set or retrieved with the EN_setoption and EN_getoption functions, respectively, and is saved to file when the EN_saveinpfile function is called.
• The EN_getaveragepatternvalue function will now accept a pattern index of 0 which represents the constant pattern assigned to junction demands by default.
• The adjustment of a tank's minimum volume (Vmin) when its parameters are changed using EN_setnodevalue or EN_settankdata has been corrected.
• A pump whose status is set to CLOSED in the input file now also has its speed setting set to 0 which fixes having a simple pressure control activate the pump correctly.
• A failure to raise an error condition for a pipe roughness <= 0 in the input file has been fixed.
• The calculation of head loss gradient for low flow conditions was corrected.
• Improved updating and convergence tests were added to pressure dependent demand analysis.
• Improved checks to prevent outflow from empty tanks or inflow to full (non-overflow) tanks. (Still needs more work).
• The CI regression test protocol was modified by:
  • changing the absolute tolerance used to compare the closeness of test results to benchmark values from 0 to 0.0001
  • dropping the "correct decimal digits" test
  • dropping the check for identical status report content since it prevents accepting code changes that produce more accurate solutions in fewer iterations.
• A possible loss of network connectivity when evaluating a Pressure Sustaining Valve was prevented.
• Having the implied loss coefficient for an active Flow Control Valve be less than its fully opened value was prevented.