|Geometric search, or searching by shape, can help with design reuse and component and supplier management.
by Dick Bourke
In Part 1, Useful Search Applications for Finding Engineering Data, I identified text-oriented solutions. Another approach is to search CAD models by shape and both shape and text. Now, I'll introduce and comment on shape-based search methods and values, identify some vendors and their offerings and suggest further actions. Even if you are reading about shape-based search for the first time, let me assure you that geometric search is far from being a bleeding edge technology with all its attendant risks. Components for geometric methods, such as feature recognition, have been around for decades. Using proven technology, several software vendors have developed shape-based searching tools as early as the mid-2000s - often with the support of university research projects.
What's Different About Geometric (Shape-Based) Searching?
Let's start with a simple definition: Geometric search, or shape-based search, uses a mathematical abstraction of a part to seek and find parts with identical or similar shapes. Shape-based searching is different from text-oriented methods that rely on keyword inquiries and accurate metadata. Effective shape-based searching depends on shape signatures for discovering and finding needed product data. A signature is a less complex version of the 3D model. Individual signatures are created and maintained in a signature file, a small file, to optimize searching for the native CAD models or neutral file formats. Most of the identified tools span both choices.
Prior to active searching, an indexing process creates the signatures. Numerous methods exist to do so, each with respective strengths and weaknesses that affect accuracy and use. The capabilities will vary for each vendor's tool. Here are a few to look for:
- Encrypting to protect intellectual property rights of the company
- Handling of file formats
- Negating need for a pre-defined classification scheme
Similarity: Searching, Discovering, Finding
With the input of a sketch or skeletal design, the signature file initiates the initial search. Similarity displays are presented to help users narrow the choices to find specific matches to their needs. These displays may be general or in detail. They show a ranking of similarity by visual and textual information, according to the vendor's algorithm - often a "secret sauce." While display styles may differ, the purpose is the same - help the user on their way to the most similar part. Additional discovering may be necessary, facilitated by secondary displays. Discovering a whole bunch of choices is a start, but not enough. What if the discovery step shows identical choices based on shape, but the materials may be different? Thus, finding a specific match may require analyzing text attributes found in PLM/PDM systems. For a complete set of searching capabilities, text searching is necessary; methods vary as described in Where Search & Discover Solutions Fit in Product Development.
What's the Value of Geometric Methods?
Design reuse is the benefit most often cited by geometric search vendors. In common with text search, geometric search also eliminates duplication, improves engineering productivity and more. Cost improvement is arguably the easiest to attain; surveys tell us that the cost could range from $6,000 to over $20,000 per avoided part assignment. If the costs have already been incurred, and inventory has resulted, part cost reductions are another matter.
When presenting their shape-based products, most tool vendors will stress the superiority of geometric searching over text-oriented solutions. Valid to a point, however, the weakness of a text-oriented solution may not be the tool itself, but the lack of valid metadata in a company's product data infrastructure hindering specific finding of design needs.
Another value is to unify and compress design activities in the supply chain that have to cope with increasing pressures, such as time-to-market and environmental regulations. A strategic initiative is based on collaborative component and supplier management (CSM) capabilities, defined as:
- Component - to support selection, design reuse, access to 3D digital component catalogs and new part introduction processes that negate assigning duplicate part numbers.
- Supplier - to support vendor selection and evaluation, enhance collaborative product development, drive part and supplier rationalization and reduce the risks in the supply chain.
As a preface to the next section, I stress that my identification of the vendors is representative, rather than comprehensive. I'm not attempting to identify all sources, but would rather show you some of the representative solutions available.
Who Are Vendors and What Do They Offer?
Vendors range from the largest, e.g., EXALEAD and Siemens PLM Software, to boutique firms focused solely on search tools. Two products, Enfinio and Similarity by Dassault Systemes, were not available for review. Other than those two, following is as complete a wrap-up as possible of the major products in shape search.
3DPartFinder technology extracts geometric descriptors for searching B-rep of native 3D formats. It supports major CAD and PLM systems. Search results in CAD are presented in an assembly ready to be used. Searching can be supplemented by textual methods in a customer's PLM system, and it follows all user rights and access. A choice of similarity algorithms (shape or size and shape) produces displays that indicate degrees of fit, such as duplicates, mirrors and similar, by color coding and ranking of the geometry found. Analytics provide dashboards and detailed reports for diagnosis of existing conditions, e.g., duplicate or quasi-duplicates analysis. PLM or non-PLM versions are available, with licenses structured for both CAD users and non-CAD users, such as purchasing.
Now at V5, 3DPartFinder is sold directly and through channels. A version for Oracle Agile PLM and AutoVue is also available.
The Product Data Intelligence (PDI) suite comprises three related applications: Data Discovery, Product Visualization and Analysis/Reporting.
- Centro 3D Search (Data Discovery) provides geometric search capabilities supplemented with text retrieval from PLM/PDM systems to access related documents and attribute information. On an SaaS subscription basis, it accesses native CAD models in the customer's systems to visually present similarity rankings. Mashups and text detail aids users to find specific needs for reuse or duplicate reduction analysis. ShapeSpace, Ltd. provides the core search engine under a licensing agreement.
- Spinfire Ultimate (product visualization) offers 3D/2D CAD viewing of all major formats with 3D CAD-to-CAD translation and markup capabilities, without the need for multi-CAD licenses. Operational choices include cloud-based.
- Insight (visual analysis and reporting) supports the decision-making process with features including publishing to related applications, such as ERP and dashboards to display the results of data mining.
Actify sells the products globally, directly and through channel partners, with a choice of entry-level versions and user needs, such as power or casual.
The products are based on a full range of digital technologies enabling manufacturers to create and find parts through a variety of methods, including geometric and full-text search. The company offers two primary products and services:
- eCATALOGsolutions is a service for part manufacturers and distributors of industry standard and supplier parts, enabling them to provide configurable digital models in more than 150 native and neutral CAD formats. CADENAS provides CAD models through either desktop CAD or web browser to support product design activities, such as internal development of parts catalogs and integration of company systems, e.g., PLM and ERP. The system enables engineers to search their components in addition to more than 400 catalogs of supplier parts. Searching methods (input criteria) include topology (attribute), geometry, color and pricing. Searching can be initiated with 2D sketches. A native CAD approach supports both native and neutral formats to find and use product information about parts: industry standard, supplier and company.
- Search results are presented with geometric and attribute comparisons to find specific part needs. PARTsolutions supports Strategic Parts Management (SPM), similar in scope to the CSM definition. SPM stresses "purchaneering" as a process in which the Part Catalog is collaboratively used for part identification, qualification and selection.
Dassault Systemes EXALEAD OnePart Portfolio
EXALEAD OnePart provides both geometric and full-text (semantic) searching capabilities for finding and navigating product data extracted from 2D and 3D files, and related enterprise systems. The product is browser-based for on-premise application supporting most CAD formats and sources, and many non-CAD formats, such as Microsoft Office. Individual CAD licenses are not required. Integration with related systems, such as ERP, can be done with connectors.
OnePart combines the attributes of the PLM system and the attributes and geometric features extracted from CAD files. Geometric searching uses similarity logic to produce a short list of thumbnail images that can then be explored to find exact part(s) needs. User decision-making is helped by displays of two or more parts showing a comparison of attributes, spotting relationships in assemblies, mining of 3D models and navigating related documents. The semantic (full-text) capabilities include facets (when, who, what, where, why and how) and filters.
OnePart Reduce is a companion product to manage the part library presented to end-users in OnePart. It does so by providing part classification tools, based on proprietary or imported taxonomies, and by enabling large-scale part deduplication based on shape. Similar parts are grouped into "clusters" and subject to a workflow of master/preferred part selection.
The company provides end-user products and supporting technology for the design and manufacturing phases of the product lifecycle and in support of consulting and system integration services.
Among several end-user products, DFMPro is a solution to validate designs for X of DFX (Design for eXcellence) guidelines, including manufacturability, assembly, serviceability and others, such as reuse and environment.
The supporting technology capabilities include:
- Feature Recognition (patented) to extract features from 3D models
- GeomDiff to identify geometry differences
- 3DSearchIT (patented) to enable design reuse
The 3D SearchIT technology applies to 3D native and neutral CAD files. For indexing, it generates 2D projections from the 3D models that are used to check for similarity by comparing the factors of 2D representations. Searching is supplemented with text attributes from PLM systems. Resulting displays are automatically clustered for analysis.
The company's suite spans the searching needs of product development personnel and the supply chain community. CADseek Polaris employs its patented geometric search engine to encode, classify and visualize a CAD model's full 3D shape, including internal geometry. Inquiries can be generated by selecting any model from the Navigator library to search for identical or similar shapes, by text keyword and a combination of keyword and shape. Optical scanning or photographs from a mobile device, in addition to CAD sketches, are available. A similarity algorithm presents matches in ranked sequence with a comparison capability that identifies overlapping features. Connections to CAD and PLM systems provide integrated workflow, and users can filter searches with both geometry and text entries on a number of variables, e.g., attributes or product group. Integration for leading PLM and CAD packages is available.
The CADseek Connect version is a lightweight, browser version, without the Navigator interface, for users that aren't as CAD-centric. The CADseek Pegasus module spans both elements of CSM, component and supplier, with analytical reporting for managing the processes of CSM.
As a service, the company offers analytical processing of CAD libraries to provide analysis, for instance, of duplicate parts and data cleansing (attribute creation and maintenance).
Siemens PLM Software, Geolus Search
The Geolus Search product delivers CAD-agnostic 3D shape search integrated with Solid Edge, NX, CATIA v5, Pro/ENGINEER, Creo and with any browser supporting HTML5. Geolus reads the neutral file formats JT, VRML and STL. It functions either in a standalone mode or through Teamcenter. The Teamcenter clients that support shape search include Active Workspace, the Teamcenter rich client and NX Manager. Users can launch searches with finished or early incomplete designs and scanned objects, supplemented with text entries. Searches launched from within Teamcenter clients can combine shapes with Teamcenter attributes, providing searching with filtering. Visual displays show individual parts with 3D thumbnails and similar parts in families. Attributes are highlighted to show differences. Both shape and Teamcenter attributes can be combined as search criteria to increase specificity. Similarity reports support initial classification creation and subsequent maintenance.
In NX, knowledge reuse capabilities support two primary applications:
- NX Reuse Library - for dragging and dropping reusable digital assets (standard parts, sketches, materials, symbols, etc.) into the NX design window. Included with every seat of NX, the library can be organized in a hierarchy of folders or integrated with Teamcenter and Geolus, optionally including Teamcenter Classification.
- NX Product Template Studio - provides code-less dialogs to capture, display and incorporate design parameters and related documentation and analyses during the design process, offered as a standalone product.
BCT Technology, a Partner of Siemens PLM Software Since 1994
For reuse/classification/standardization activities, BCT offers two products:
- BCT aClass with Shape Search to support Teamcenter with a classification/attributes system for managing structured product data with Geolus providing shape search functionality.
- BCT Engineering Package for management of standard and catalog parts by creating libraries of intelligent part and feature combinations.
The firm markets a suite of technologies for searching, classifying and analyzing product data for all major 3D CAD formats. Its Product Data Query Engine (PDQE) provides the core capabilities of partial and full shape matching, attribute display, bill of material analysis and collaborative product data analytics. The query language supports complex inquiries based on similarity, use in products and bounding by time.
ShapeSpace licenses its shape search API for integration into third-party software to a number of partners, including Aras and Actify, Inc.. The firm works collaboratively with these partners to build shape search solutions tailored to a specific customer's search requirements. PDQE is sold directly and through resellers, AESSIS and Minerva.
The ShapeSpace analytics team works on a project basis with customers to resolve issues of product complexity, rationalization, modularization and standardization using its suite of product data analytics technology.
Sketch & Search
CADFind is a suite of products for the retrieval and grouping of 2D drawings and 3D CAD models. Its database can contain both 2D and 3D models and be searched in a single pass using basic or fully detailed originals. The system displays thumbnails of search results in similarity order and a model viewer to aid further examination of the choices presented.
The primary products are for SOLIDWORKS and AutoCAD users. CADFind is integrated with both systems to allow searching, using an existing model from within the CAD, then loading back any match into CAD for modification or analysis.
CADFind's Group Technology tools allow the user to identify interactively families of similar parts that form the basis of cellular manufacturing systems.
The company was founded as a spinoff from a UK university in 2004. Its global marketing and product development strategy is driven by the popularity of the two CAD systems and the amount of legacy 2D data in many companies. Database creation and analysis is offered on a service bureau basis.
Identifying Product Alternatives Is Just One Step, Prepare for More
Identifying a choice or two is only one step to an appropriate decision for a company's circumstances. User evaluation teams need to resolve a number of factors before selecting a tool. A few I previously identified bear repeating - with additional input:
- Developing a selling strategy through a creeping commitment approach, starting with general demos and proceeding through trials and a formal Proof of Concept.
- Determining the extent of cleansing the classification/attribute data and other metadata needed to assure valid discovery and finding.
- Developing a complete ROI for C-level approval, including the total cost of ownership; don't prematurely focus on the bare cost of the tool.
To sharpen your perspective - and justification - regarding the potential of shape-based tools, consider a selective analysis of the current product data. Most of the vendors identified offer such a service, as do two vendors previously identified: Convergence Data and Noetic-Data for text-oriented analysis.
As an example of a successful application of shape-based search from the realm of component and supplier management, one manufacturing company attempted to analyze its duplication status for a purchased part. They tried to do so with a team of experienced purchasing personnel with their extensive tribal knowledge and relying on existing records. They found six duplicates identified within the part category. Then, with a trial using shape-based software, a total of forty-one parts were identified in various stages of similarity.
Should Design Reuse Be the Only Element to Move Forward?
A fresh vision may be required to excite C-level executives. A prime example of a fresh vision of thinking broadly, deeply and strategically is a recent article, Toyota Rolls Out Shared-Parts Strategy to Chop Costs. The president stated, "We're at a crossroads where we must now build a new business model." Toyota's strategy is aimed at cutting development costs by 20 percent.
In that spirit, I highly recommend a new book as a template to define your action plan. The author describes a comprehensive strategy that she calls Variety Effectiveness Process, or VEP. VEP is a broader way to look at what most of us would call design reuse. Smart Simple Design/Reloaded/: Variety Effectiveness and the Cost of Complexity, by Gwendolyn Galsworth, 2015, is available through Amazon.
Useful Search Applications for Finding Engineering Data, Part 2 is reprinted with permission from Engineering.com
For more information about SDS, contact Dick Bourke at firstname.lastname@example.org.