Changelog

 

​    Version 2.0.0

• New feature: subroutine meshgen1 is developed from meshgen_liu.f90 (subroutine meshgen) in MPI_v2.1.0. It is a mesh scheme that uses degenerated and size-adjusted triangles to morph the complex fault geometry with comparable element sizes. 

• New feature: Controllable parameter C_mesh is created. C_mesh == 1 indicates Duan et al. (2019) meshing scheme. C_mesh == 2 indicates the above meshing scheme.  

• New feature: Read_Input_Files.f90 is developed from the same name file from EQdyna3D_v5.1.0. It is used to read adjustable parameters from independent txt input files. This treatment makes EQdyna 2D flexible. 

• New feature: input files FE_Gloabl.txt, FE_Model_Geometry.txt, FE_Fault_Geometry.txt are required. Descriptions of quantities are in the same files.  

• New feature: subroutine interstress is developed from previous standalone program strbld.f90. It will take care of interseismic stress loading and relaxation for both the first and later earthquake cycles.     

• New feature: executable file will have version number.

• New feature: earthquake cycle numbers are imported from FE_Global.txt. They are icstart and icend. Output files will have an appendix of ‘icstart’. 

• New feature: interstress and driver are put in a do loop from icstart to icend to simulate earthquake cycles. 

• New feature: output files are redesigned. 

  • 1) Totalop.txt//{icstart} stores 5 on-fault quantities for earthquake cycles from icstart to icend (if walltime permits the finish of run_eqdyna2d_M_2.0.0). The data are formatted as 5E18.7 per row.    

  • 2) cyclelog.txt//{icstart} stores icstart and icend (or the last earthquake cycle number simulated)

  • 3) interval.txt//{icstart} stores the years used for stress to reach Coulomb failure criteria before an earthquake. 

  • 4) binaryop, which is in unformatted binary style, stores the 5 on-fault quantities for the last earthquake. It is used to continue earthquake cycles if run_eqdyna2d_2.0.0 were terminated unexpectedly. 

• New feature: there is no need to define model size. Model size is defined by the longest expected run time -- term, in seconds -- and P wave velocity. The wave cannot come back to the fault during the simulation time. 

• New feature: /script/ contains Plot_Normal_Vector.m, Plot_Single_Event_Info.m and Show_Mesh.m to visualize normal vectors of split nodes, plot details of event dynamics, and to show meshes, respectively. 

• Improvement: Import material properties by Vp, Vs and density rather than elastic modulus. 

• Improvement: On-screen output information during code run is redesigned. 

• Improvement: variable, array, and matrix transfers are extremely minimized. A great majority of them are defined in globalvar.  

• Deletion: ./mesh/ and ./bldup/ are no longer needed. 

• Bugfix: in subroutine qdct2,  ce = 16.0d0* e(m) / 15.0d0 should be used for the plain stress case. ce = 4.0d0*amu*(lambda+amu)/(lambda+2.0d0*amu) should be used for plain strain.