Part and Assembly Search Based on Similarity and 3D Shape Attributes
Main Participants: Satyandra
Antonio Cardone, Mukul Karnik, and Abhijit Deshmukh
Sponsors: This project is being sponsored by the Naval Surface
Warfare Center at Indian Head,
Maryland and the Center for Energetic Concepts Development at the
Keywords: Shape Similarity, Shape Search, Part Retrieval, and
Intuitively, if two products are similar, it is possible to reuse
information about one product to derive corresponding information about
the other one. There are many possible applications where reuse of
information can be of significant value. Representative examples
include part-family formation, redesign suggestion generation, supplier
selection, cost estimation, tooling design, machine selection, stock
selection, and design reuse. Representative examples are illustrated
Most users currently browse CAD models manually to locate similar
parts. This is a very time consuming step and slows down the decision
making process. Most product development engineers spend a lot of time
searching for information. A search tool that can help in locating
parts and assemblies having similar application-specific features will
help in cutting down that time significantly and allow engineers to
spend more time on creative aspect of their jobs. By improving decision
making in design and planning functions, the shape similarity based
search tool is expected to play an important role in reducing costs and
reducing the time-to-market.
- Cost Estimation for Machined Parts. Nowadays, many job
shops allow designers to submit a 3D model of the part to be machined
over the Internet and provide a cost estimate based on the 3D part
model. For some manufacturing domains such as rapid prototyping,
reasonably accurate estimates of cost can be achieved by estimating
volume or weight of the part. However, for some manufacturing domains
such as machining, cost estimate depends on the geometric details of
the object and automated procedures are not available for doing
accurate cost estimation. Currently in such cases, humans perform cost
estimation. In the Internet era, where designers solicit many quotes to
make a decision, manual cost estimation is not economical. Cost of
manufacturing a new part can
be quickly estimated by finding previously manufactured parts that are
similar in shape to the new part. If a sufficiently similar part can be
found in the database of previously manufactured objects, then the cost
of the new part can be estimated by suitably modifying the actual cost
of the previously manufactured similar part.
- Part Family Formation. In many manufacturing domains
such as sheet metal bending, machine tools can be setup to produce more
than one type of part without requiring a setup or tool change.
However, parts need to be shape compatible in order for them to share
common tools and setups. Therefore, in order to find common tools and
setups, geometrically similar and therefore compatible parts need to be
grouped into families. Shared tools and setups can be used to
manufacture objects in the same family and therefore result in
significant cost savings.
- Reduction in Part Proliferation By Reusing Previously
Designed Parts. Reusing design/manufacturing information stored
would result in a faster and more efficient design process. While
designing a new part the designer can refer to existing designs and
utilize the components used previously. Let us consider the design of
the shaft of a turbine engine. Usually the designer has two options.
The first option is to design the shaft from scratch and go through the
process and manufacturing planning. The second option is to refer to
the database of existing designs, and select an existing shaft and
either use it as it is or make minor modifications to it (e.g., drill a
few holes or cut a few slots).
The search tool locates existing parts similar to the new part based on
some geometric attributes. It creates signatures for each of the parts
in the database and stores the signatures along with the solid model of
the part. A signature is a list of geometric attributes that describe
the part and depends on the application. These pre-computed signatures
reduce the time required for comparison and, thus, improve the speed of
comparison. The search tool then uses the signatures to compare the
signature of the query part with each of the signatures of the database
parts to determine if the parts are similar.
We have developed two different techniques for performing search.
Advantages of our method include:
- Feature-Based Techniques:This
method uses feature
information obtained from feature-based modeling software such as Pro/E
SolidWorks to assess similarity between parts. We can use either design
features or manufacturing features. We
utilize feature vectors consisting of feature position vectors, feature
orientation vectors, feature types, and feature parameters as a basis
shape similarity. We have defined a distance function between two sets
feature vectors. The distance between the feature vectors is used as a
of similarity between the two parts. The features
provide a very convenient way of including critical details and
irrelevant details in search for similar parts.
- Gross-Shape Based Techniques:
The gross-shape based technique uses several signatures to identify
existing parts similar to the query part. The signatures are applied
sequentially to improve the efficiency and accuracy of the comparison.
The following types of signatures are used to compare the query part
with database part to assess similarity: part volume and surface area,
basic shape statistics such as the types of surfaces and their
corresponding areas, gross shape complexity, and detailed shape
complexity that includes the surface area and curvature
- Customizability: Our
method allows the user to customize the
search criteria by letting the user select the important feature
characteristics that s/he wants to consider such as feature type,
volume etc. Also the distance function used for comparison can also be
customized so that some feature characteristics are given more
- Manufacturing Information: Manufacturing
information such as
tolerance, surface finish etc. that is associated with a feature can
used. Thus two parts having the same shape but largely different
requirements will not be considered as similar.
- Applicable to Assemblies:
Our method utilizes the feature
information obtained from a feature tree. It can be easily extended to
assembly feature information obtained from an assembly tree.
The following papers provide more details on the above-described
Some of these papers are available at the publications
section of the website.
- A. Cardone, S.K. Gupta, and M. Karnik. A survey of shape
similarity assessment algorithms for product design and manufacturing
applications. Journal of Computing and Information Science in
Engineering, 3(2):109--118, 2003.
- M. Karnik, S. K. Gupta, and E. B. Magrab. Geometric algorithms
for containment analysis of rotational parts. Computer Aided Design,
37(2):213--230, February 2005.
- A. Cardone, S.K. Gupta, and M. Karnik. Identifying similar parts
for cost estimation. In ASME Design for Manufacturing Conference,
Salt Lake City Utah, September 2004.
- M. Karnik, D. K. Anand, E. Eick, S. K. Gupta, and R. Kavetsky.
Integrated visual and geometric search tools for locating desired parts
in a part database. CAD Conference,
Bangkok, Thailand, 2005.
- A. Deshmukh, S.K. Gupta, M. V. Karnik, and R. Sriram. A system
for performing content-based searches on a database of mechanical
assemblies. ASME International
Mechanical Engineering Congress & Exposition, Orlando, FL,
- Cardone and S.K. Gupta. Shape Similarity Assessment Based on
Face Alignment using Attributed Applied Vectors. CAD Conference, Phuket Island,
Thailand, June 2006.
- S.K. Gupta, A. Cardone, and A. Deshmukh. Content-Based Search
Techniques for Searching CAD Databases. CAD Conference, Phuket Island,
Thailand, June 2006.
- A. Cardone; S.K. Gupta, A. Deshmukh, and M. Karnik. Machining
feature-based similarity assessment algorithms for prismatic machined
parts. Computer Aided Design, 8(9):954--972, 2006.
For additional information and to obtain copies of the above papers
Dr. Satyandra K. Gupta
Department of Mechanical Engineering and Institute for Systems Research
2135 Martin Hall
University of Maryland
College Park, Md-20742