reactingMultiphaseEulerFoam crashes when one field is initially patched with 0 phase fraction

Summary

ReactingMultiPhaseEulerFoam crashes when one of the fields is initially patched with 0 phase fraction.

Steps to reproduce

Please unpack and simply run the attached scripts. Allrun1 replicates the first behavior and Allrun2, the second. Please see below for a description of the behaviors.

Example case

Attached is a simple two-phase test case (air and water) of a gas-stirred cylindrical tank initially half filled with water and air is injected through a small nozzle on the centre of the bottom patch.

What is the current bug behaviour?

(1) When one of the fields is initially patched with 0 alpha, reactingMultiphaseEulerFoam crashes after the first pimple iteration while going to turbulence object. This behavior is replicated with Allrun1 script.

(2) If we put a small positive value (0.001) instead of 0, the solver runs stable for a few iterations and crashes eventually. This behavior is replicated with Allrun2 script.

What is the expected correct behavior?

The solver should run stably in both cases.

Relevant logs and/or images

Environment information

• OpenFOAM version : v-1906
• Operating system : Ubuntu 16.04 LTSgasStirredTank.tar.gz
• Hardware info :
• Compiler : gcc

Possible fixes