IbIPP

Image-based Initialization and Post-Processing code for Topology Optimization

IbIPP

IbIPP which stands for Image-based Initialization and Post-Processing is an open-source code for initializing and post-processing free-form 2D topology optimization problems.

Syntax

• ibipp(file, nelx, volfrac) Performs topology optimization on the image provided in file using the number of elements in the x-direction nelx, the required volume fraction volfrac, and other default parameters.

• ibipp(file, nelx, volfrac,fmag,fang) Performs topology optimization using file, nelx, volfrac, vectors of force magnitudes fmag, and force angles fang, and other default parameters.

• ibipp(...,'PropertyName',VALUE,'PropertyName',VALUE,...) performs topology optimization using the following property values:

Table 1: Name-value inputs

---

Name-value	Description	Value
pressure	Pressure loads (multiple loads must be in a vector)
optimization	Type of topology optimization approach	Density, BESO, Level Set, SEMDOT
densityType	Type of density-based method	SIMP, RAMP
preserveSupport	Preserve elements within the support region in the image	0 – none
		1 – Only completely fixed elements
		2 – Only elements fixed in the x-axis
		3 – Only elements fixed in the y-axis
		4 – the union of 1 and 2
		5 – the union of 1 and 3
		6 - the union of 2 and 3
		7 – the union of 1, 2, and 3
preserveLoad	Preserve elements within the load region in the image	0 – none
		1 – elements in the force region
		2 – elements in the pressure region
		3 – the union of 1 and 2
filterRadius	Specify filter radius for density-based and BESO approaches	>1
filter	Specify filter type for density-based approaches	1 – density
		2 – sensitivity
		3 - Heaviside projection
beta	Regularization parameter for Heaviside projection	>1
penaltySIMP	Penalty value for SIMP density-based approach	>1
penaltyRAMP	Penalization factor for RAMP density-based approach	>1
ER	Evolution ratio for BESO	>0
tau	Regularization parameter for level set method (reaction diffusion)	>0
YoungsModulus	Youngs Modulus material property	>1
PoissonRatio	Poisson Ratio material property	>0
modelName	Name given to model obtained by extrusion or revolution of the optimized topology. Must end in .stl
modelType	Modeling by extrusion or revolution	Extrude or revolve
extrudeLength	Length for extrusion - a factor multiplied by min(nelx,nely) to obtain the length of the 3rd dimension	>0
symmetry	Specifies the position of the line of symmetry for symmetry-based modeling	None, Left, bottom, right, and top
distancetoaxis	Distance between the optimized topology and its revolution axis to the left for revolution-based modeling	≥0
revolutionAngle	The angle of revolution for revolution-based modeling	≥0

Table 2: Color codes (RGB) for creating an input image file

Features	Color code rules	Recommended values
Designable domain	R∩G∩B<200	0 0 0
Non-designable domain	R∩G∩B>200	255 255 255
Point load	200≤R≤255, G=0, B=0	200 0 0
Pressure	200≤R≤230, 100≤G≤150, B=0	200 100 0
Preserved region	R=0, 200≤G≤255, B=0	0 200 0
Fixed region	R=0, G=0, 200≤B≤255	0 0 200
Region fixed along x	100≤R≤150, G=0, 200≤B≤255	100 0 200
Region fixed along y	R=0, 200≤G≤255, 200≤B≤255	0 200 200

• Note: for multiple forces or pressure, R should be in multiples of 5 starting from 200 for forces (200, 205, 210,...etc) and G should be in multiples of 5 starting from 100 for pressures (100, 105, 110,..etc)

Examples:

The following examples show how ibipp can be used. Images of some sample design problems and their corresponding input images for ibipp have been prepared and are in this repository. When developing an input image for a design problem, Table 2 should be used while Table 1 outlines the possible name-value inputs. Any image tool can be used to prepare the input file but .png is the preferred file extension.

• Example 1: Optimize a half-MBB using default values of optional inputs

  • ibipp('Example_pics\half_mbb.png',150,0.4)

• Example 2: Optimize a Hammerhead pier using the BESO approach

  • ibipp('Example_pics\hammerhead.png',200,0.5,[2 1 1 2],[180 180 180 180],'optimization','BESO','filterRadius',3)

• Example 3: Optimize a 2-point loading mechanical part while preserving load and support elements

  • ibipp('Example_pics\2point.png',300,0.45,[1 1],[0 0],'optimization','levelset','tau',5e-5,'preserveLoad',1,'preserveSupport',1)

• Example 4: Optimize a half-spanner design and generate an extruded full spanner STL model

  • ibipp('Example_pics\half_spanner.png',250,0.5,'pressure',[1,1],'preserveload',2,'preservesupport',1,'modelname','spanner.stl','symmetry','right','modeltype','extrude','extrudelength',0.2)

Supporting Open-Source Codes

IBIPP utilizes other open-source codes such as top88.m by Andreassen et.al, esoL.m by Xia et. al., levelset88.m by Otomori et.al, revolve2D.m by Treeby and Cox, and stlwrite.m by Sven.

To Cite

If you find this code helpful in your work, please cite this open access paper