Knowledge-based engineering is about reuse of experience, expertise and other information relevant to each phase of the engineering life cycle of an end-user product. It can be either in a structured format like a spreadsheet or even in a subjective format like human judgement. Aiyappan Ramamurthi illustrates how knowledge-based engineering can be deployed to enhance the productivity of the design team

Reducing costs and bringing out products faster to market than ever before, while addressing increasing customer requirements, are two of the major challenges facing manufacturing companies today. Manufacturers have standard product lines and are continuously engaged in the process of engineering similar products, which are altered to suit changing customer needs.

Capturing knowledge, adopting best practices and providing automation of design and manufacturing processes would be the key technology enabling manufacturers to achieve efficiency and profitability. Combining high-end CAD systems with knowledge-based systems would allow manufacturers to automate their engineering and manufacturing processes capturing the design rules, experience and expertise residing in the organisations and leveraging it during new product development.

Challenges
With customer requirements as input, manufacturing companies are engaged in the process of design, development and manufacturing of products. The customer requirements are translated to performance criteria and cost. This process of design to manufacturing is a complex process, which is time-consuming, requires high skills and is also expensive. Manufacturers continue to invest in technology, consulting and services to improve the process, but engineers continue to face the following questions:

• What was the rationale behind this design?
• Have any design constraints been violated?
• How much will this product cost?
• Can parts be manufactured?
• Will the parts meet performance goals?
• Is this design optimum or are there better alternatives?

All these questions are potentially answerable with precision since the information required or the means to obtain it exists, though distributed in various forms throughout multiple organisations.

Customer needs are changing!
To overcome these challenges, manufacturers rely on solution providers for ‘intelligent’ software that can make rule based decisions based on knowledge and experience and can harness disparate information and make it readily available to automate the engineering process.

Knowledge-based engineering
Knowledge-based engineering (KBE) is fundamentally about reuse and ability to take advantage of any experience, expertise and other information relevant to each phase of the engineering life cycle of an end-user product.
These knowledge bases can exist in many forms such as:

• Spread sheets
• Hand books
• Engineering formulas
• Propriety software
• Human judgement, such as rule of thumb.

Being able to create and reference a knowledge base and make it readily available, as an aid to the engineering process constitutes knowledge-based engineering. Looking at the application areas and the value KBE offers to the industry, a leading industry analyst from D H Brown & Associates said, “By progressively automating lower level repeated tasks as well as higher level product development processes, well-structured knowledge-based engineering systems allow organisations to capture and reuse product development experience at many levels. Companies that have effectively implemented KBE systems report compressing the time and cost of design tasks by as much as 90 percent.”

While such significant benefits can be achieved using KBE, there are some major barriers in implementing a KBE system.

Barriers in implementing KBE
There are two ‘knowledge approaches’ to product development. One is the design approach and the other is the engineering approach. In the design approach, geometry is created and then constraints on the geometry are added when needed. The model is then evaluated and repeatedly modified until the desired result is obtained. This process is typical of every MCAD tool.

In the engineering approach, an ‘engineering script’ is created, which captures the engineering intent. This script is then executed to create the desired geometry. Changes can be made only to the script and it must be rerun each time to create the geometry. This process is typical of every KBE tool. In addition to the differences in the approach of design and engineering, implementing KBE has a few other barriers:

• KBE applications are specialised and disjointed from MCAD systems.
• Traditionally programmatic, KBE systems require high level of expertise.
• Knowledge capturing is a complex process.
• Although the knowledge may be documented, it is not always referred to.
• Investing in one KBE system precludes reusing knowledge in other systems.

Overcoming the barriers and confirming to key feature requirements is essential for a knowledge-based system to be widely adopted by the industry.

Key requirements
In order for KBE to become central to the engineering process, it has to be capable of being applied to all relevant disciplines. The disciplines of design, analysis and manufacturing can all effectively utilise KBE. This places on the technology the demands of flexibility, openness, customisation and, of course, reusability. The technology must be deeply integrated into the system processes and not just as a simple add on, or interface to an existing system; else key functionality, as described below, cannot be realised.

Knowledge driven automation focuses on processes that will automate many industry standard practices and provide tools for the following:

• Sales automation
• Engineering automation
• Manufacturing automation

Knowledge driven automation value

• Reduction of time for RFQ by ability to locate and modify product configuration to meet customer needs.
• Dramatic reduction of engineering on configurable product design by creating knowledge-based design rules and using rules to automatically create product design.
• Cut down the cycle time from product design to manufacturing by concurrently sharing and validation of product design and tool design.
• Gain huge productivity improvements from design inception through manufacturing by managing design variation and options.
• Improve first time quality while reducing manufacturing pilots by capturing the manufacturing intent during design and application of expert knowledge during design and manufacturing with digital validation.

The author is technical manager at EDS PLM Solutions India. He can be reached at ramamurthi.aiyappan@eds.com