About PhiPsi

PhiPsi is a computational solid mechanics program, which involves the extended finite element method (XFEM), as well as the finite element method (FEM). The program is written in Fortran and compiled using the GNU Fortran compiler (gfortran). The program is named PhiPsi because φ and ψ are the basic variables for the level set method which is widely used in the extended finite element method. On the other hand, φ and ψ denote the internal friction angle and the dilation angle in the plastic theory. See the Applications page to learn what PhiPsi can do.

The data generated by PhiPsi is post processed by a Post-Processor program written in Matlab. You can download the latest version of executable PhiPsi for the windows platform, the source codes of an earlier version of PhiPsi, the source codes of the Matlab Post-Processor, and some other tools from the Downloads page.

PhiPsi GUI is the graphical user interface for PhiPsi written in Visual Basic.NET. It is designed for pre- and post-processing for PhiPsi which is written in Fortran, but It is still on the beta channel. The beta version of PhiPsi GUI is available for downloading from the Downloads page.

Please cite the papers listed on this page if the PhiPsi is helpful to you.

Features of PhiPsi

○ Supported analysis type: 2D and 3D static analysis, 2D and 3D hydraulic fracturing analysis, 2D and 3D dynamic analysis and 2D field problems analysis.

○ Support as many as 1000 fractures, voids and inclusions.

○ Randomly generating initial fractures, voids and inclusions.

○ Intersection of 2D and 3D fractures, intersection of fracture and voids or inclusion.

○ Penalty function method to determine the contact status.

○ Optimized Newton-Raphson scheme to solve the nonlinear system.

○ Sparse matrix to store the global stiffness matrix K.

○ Support DOFs coupling.

○ Fast linear system solvers including LAPACK, UMFPACK, Lis, SuperLU, and PCG-EBE.

○ Both formatted and binary files are supported for the generated data.

○ OpenMP support for some computational processes and linear solver.

○ Keywords file support with parameter definition and four operations (+, -, *, and /).

How does PhiPsi work ?

There are two options to use PhiPsi.

(1) Option 1 - Use ANSYS (or Abaqus, etc.) and Matlab for pre-processing and post-processing, respectively

The input files of PhiPsi include a keywords file and other data files contain the node coordinates, element-node information, boundary conditions, and external forces. The keywords file defines information such as work directory, analysis type, coordinates of initial fractures. Details on the PhiPsi keywords file can be seen from this page PhiPsi keywrods manual. The data files can be generated automatically by run the macro file called Ansys2PhiPsi_2D.mac or Ansys2PhiPsi_3D.mac after the model is created in ANSYS. Certainly, you can define the data files by other software such as Abaqus (Click here to download the Abaqus2PhiPsi_Matlab tool) or just generate it manually. A description of PhiPsi file types and data structures can be found in Section 4 of this page PhiPsi Instruction Manual.

Once PhiPsi starts, it will check all the input files firstly. After the analysis, the output files will be saved to the work directory. A Matlab-based program (all the source codes are available for downloading from the Downloads page ) is offered for post-processing. By default, vtk files which can be post-processed using Paraview will also be saved.

A tutorial is available in the Downloads page.

(2) Option 2 - Use PhiPsi GUI

Alternatively, you can also perform an analysis using PhiPsi GUI (PhiPsi Graphical User Interface). Within PhiPsi GUI, you can directly create the geometric model, apply boundary conditions and external force, perform meshing, edit keywords file, call PhiPsi Fortran kernel, plot deformed mesh, plot contours, etc.

It should be noted that the current version of PhiPsi GUI is still in the beta channel, some functions are not available yet. Therefore, pre-processing using ANSYS or Abaqus and post-processing using Matlab (Option 1) are suggested if the problem to be solved is relatively complicated.

Author of PhiPsi

Name: Fang Shi

Workplace: Faculty of Mechanical & Material Engineering, Huaiyin Institute of Technology, China

CopyRight © 2016-2024 All Rights Reserved.