Knowledge Fusion is a superior technology that permits NX to take advantage of engineering knowledge bases in conjunction with rules to deliver powerful applications while permitting a knowledge based extension of NX by end users. Since Knowledge Fusion is based on the principles of Knowledge Based Engineering (KBE), it would be helpful to understand some of the background of KBE and its characteristics.

Knowledge Based Engineering (KBE) is the process of capturing and structuring reusable knowledge bases to create and enhance solutions for a product during its entire life cycle. These knowledge bases can exist in many forms such as spreadsheets, handbooks, engineering formulas, drawings and documents, or in rules of thumb that are based on human judgement. KBE is able to create and reference such knowledge bases and make them readily available as an aid to the engineering process through the use of computer assistance and software tools.

A typical Knowledge Based Engineering System can evolve as follows:

1. Identify what knowledge is reusable.

2. Gather knowledge data from the various sources.

3. Format the knowledge data into rules constituting a knowledge-base.

4. Use the Rules from this knowledge base in conjunction with computer aid (knowledge Engine) to come up with new solutions and decisions.

The potential of KBE in its application throughout the entire engineering life cycle is enormous and can be beneficially employed by all disciplines from the concept of a design all the way through to its manufacture. KBE is the key to being able to answer questions of some significance that traditional CAD systems to date have not been capable of addressing such as:

1. "What was the rationale behind this design?"

2. "Have any design constraints been violated?"

3. "How much will this product cost?"

4. "Can this part be manufactured?"

5. "Will this part meet its performance goals?"

6. "Is this design optimum or are there better alternatives?"

A Knowledge Based System answers these questions and serves as the mechanism that harnesses relevant information and makes it readily available to the engineering process.

In order for KBE to become central to the engineering process, it must be capable of being applied to all relevant disciplines. The disciplines of Design, Analysis, and Manufacturing can all effectively utilize KBE. This places on the technology the demands of flexibility, openness, customization, and reusability. The technology must be deeply integrated into the system processes and not just a simple add on, or interface to an existing system; else key functionality, as described below, cannot be realized. The following are the minimum capabilities that an integrated KBE technology must possess.

• Knowledge Base Access - Knowledge bases exist in various forms, either persistently such as in a database or spreadsheet, or as the result of the execution of some external process. The KBE system must be able to take advantage of these knowledge bases and thus their accessibility and the corresponding openness of the KBE system become important factors.

• Knowledge Rules - The fundamental KBE construct that must be introduced into the CAD system is a rule, which is a textual description of an action to be performed based on some inputs. Rules must be capable of referencing various knowledge bases. In conjunction with conditional logic, the rule must be able to arrive at a result when executed. The result of a rule could be a modeling feature, a computed set of parameter values, the selection of a component that meets a specification, or a message warning that some critical constraint such as cost or weight has been violated, etc.

• Knowledge Rule Applicability - The knowledge rules must be readily available to all processes in the engineering life cycle. The flexibility of the rule form and its variable outcome should insure that it has a multi-disciplinary applicability. Thus, not only the design process should be capable of implementing a rule, a rule should be available for implementation in all downstream applications and ideally over enterprise wide disciplines.

• Knowledge Extension - KBE can be viewed as the tool to enhance the functionality of the existing Product Development Systems . It must be possible for the rules to complement and enhance the various system-provided objects and applications. System extension can only be realized if rules can interact with fundamental components of the system and its operations.

• Knowledge Rule Associativity - It is important that knowledge rules remain associated to the objects to which they are related. For example, if a rule creates a geometric feature, then without this capability of association with that particular geometric feature, the rule cannot revise or edit the same object. It only creates a new one. If the design does not maintain any relation to the rules and the rules are further not related to geometric objects, then the result is that all downstream operations based on the design model are rendered invalid when the knowledge based application is re-executed. In our example, a new design model is generated at each invocation of the rules. With such a drawback in the KBE system, concurrent engineering cannot be realized and a strictly serial life cycle process is imposed. In such a system, a design revision and a  backup in the engineering process would impose an expensive penalty due to the extensive rework that must take place.

• Knowledge Deployment - A KBE system should be deployed in a user friendly manner,  tightly integrated with the existing systems/applications, not form isolated islands of tools that are obscure to the regular users, and It should be transparent to the end users when they carry out their daily tasks. End-users should be able to carry out their routine tasks using KBE tools without incurring the need of understanding the intricacies of KBE tool and technology executions.