Intelligent Cascading Requirements Management and the Digital Thread

Summary

Product complexity manifests itself in many ways, but often results in an increase in the breakdown and distribution of product functions and responsibility across an increasingly large number of siloed product groups. Throw in the modern reality of international supply chains and you are faced with a real problem achieving the critical task of capturing, disseminating, and communicating product requirements and functionality. Miscommunication of product intent leads to many costly errors and time delays across the entire product lifecycle, including products operating in the field.

The product development process for today’s smart, interconnected, and multidisciplinary products can no longer be effectively managed using simple text-based requirements—which is the status quo. What’s needed are “intelligent” requirements that can capture various authoring methodologies (requirements come from multiple sources), content complexities (including the ability to capture parameters and expressions that define acceptable ranges), and design states (including managing product line variability). A complex requirement that is captured without “intelligence” may often lead to misinterpretations, given the varied stakeholders and influencers: business goals, regulatory compliance, technology options, interactions of systems-of- systems, design exploration choices, domain specific implementations, system integration verification, manufacturing goals and constraints, and others. Each of these interested groups has a language of their own for specifying requirements, intent and scope—often each uses a different methodology for capturing content, metadata, and traceability. This is a Tower of Babel problem, resulting in confusion and miscommunication. Therefore, requirements must be intelligently specified in a common, data centric language that eliminates ambiguity, allowing seamless integration and automation at a parameter level, into the end-to-end digital thread. This is a challenge for today’s requirements management software most of which are stand-alone and based on textual definitions.

The increasing reliance on Model-Based Systems Engineering (MBSE) and on simulation during the entire lifecycle adds to the need for intelligent requirements to effectively drive the process. MBSE and simulation simultaneously define (requirements flow-down and optimization) and consume (design test and verification) these intelligent requirements at every stage of the lifecycle.

Simulation is often used to verify the requirements during design exploration—if this is not feasible, requirements must be modified. Ultimately, simulation and testing are used to ensure that detailed designs meet the high-level system requirements and the related, cascading requirements.

More and more, semantically rich, traceable, accurate, and unambiguous intelligent requirements allow automated interpretation, monitoring, and enforcement. It is the embedding of requirements into an end-to-end digital thread that adds the required “intelligence.” The explicit numerical semantics and constraint equations for intelligent requirements allow them to be algorithmically interpreted—something that has proven impossible with text-based requirements without complicated, often incomplete and inaccurate parsing of the text. When these verification algorithms with intelligent requirements are embedded within simulation processes, specifying/enforcing inputs for simulations and automatically verifying requirements against results, they achieve more effective, automated design space exploration. This enables the automatic finding/reusing of components that meet particular requirements.

Finally, the explicitly captured semantics and rich relationships also allow the information content and intent of a requirement to be unambiguously interpreted by various stakeholders and to be visualized more easily along with associated elements of the digital thread. The digital thread becomes the common, more easily interpreted mathematical, datacentric, and graphical language that enables full requirements traceability across the lifecycle, turning simple textual requirements into “intelligent” requirements.

The goal that is achieved: requirements and simulation-driven automated design exploration, efficient reuse, and validation of complex products!

Requirements Specification

The Digital Thread to the Rescue!

Many if not most of the contextual constraints and the intent of requirements, along with the elements of the product design that satisfy these requirements, can be captured more formally using data and “rich” relationships within what is now being called a product lifecycle digital thread1,2. In addition, the digital thread must capture the evolution of the product design through the lifecycle, from conceptual design to detailed design, manufacturing, and maintenance in the field.

When simulation (and test) data are also integrated seamlessly into the digital thread, this data can be used to verify requirements, closing the Verification & Validation (V&V) loop.

When capturing relationships between requirements and design artifacts across all abstraction layers in the digital thread, the associated metadata and context must also be represented as connected data and rich relationships. There are many groups within a product team that specify, consume, and view requirements.

Some examples: divisions responsible for subsystems of the product, business teams, regulatory compliance teams, simulation and physical testing teams, manufacturing teams, and others. A well-designed digital thread will provide these stakeholders with a consistent way to capture the requirements and associated data, despite each having their own “language” for specifying their requirements. It is important to note that each company determines the content to be captured in its digital thread and how this data is connected.

The following are some of the additional aspects of “intelligent” requirements management that must be supported in the digital thread:

• Support for multiple requirements specification methodologies
• Definition of custom requirement types—different stakeholders have different needs
• Automatic enforcement of conformance to a requirement type
• Bi-Directional connectors to external requirements authoring tools with the ability to author/edit requirements on both sides
• Support a collaborative requirements-authoring environment, especially for complex sets of requirements that are generated and managed by multiple stakeholder teams
• Traceability of data across the lifecycle, supporting the generation of an up-to-date Requirements Verification Matrix

The explicitly captured semantics and rich relationships of the data in the digital thread allow the information content and intent of a requirement to be unambiguously interpreted by various stakeholders and to be visualized more easily along with associated elements of the digital thread. The digital thread becomes the common, more easily interpreted mathematical, datacentric, and graphical language that enables full requirements traceability across the lifecycle, turning simple textual requirements into “intelligent” requirements. The visualization of the digital thread also reveals what percentage of a requirement is satisfied by a particular system element, providing key guidance to cross-domain optimization algorithms.

Conclusions

Text-based requirements leave room for misinterpretation and result in delays to product release dates and cost overruns.

The semantically rich, traceable, accurate, and unambiguous nature of intelligent requirements, embedded within the context-rich environment of a digital thread, make automated interpretation, monitoring, and enforcement of these requirements feasible. It is the embedding of requirements into an end-to-end digital thread that adds the required “intelligence.”

The explicit numerical semantics and constraint equations of intelligent requirements allow them to be algorithmically interpreted — something that has proven impossible with text-based requirements, without complicated, often inaccurate parsing of the text. When these verification algorithms with intelligent requirements are embedded within simulation processes, specifying/enforcing inputs for simulations and automatically verifying requirements against results, they achieve more effective, automated design space exploration for complex products with large numbers of requirements. This also enables the automatic finding and reusing of previously designed components that meet particular requirements.

The recent, successful launch of the James Webb Telescope,4 with its complex deployment and calibration processes done remotely and automatically, is a cutting-edge example of handling large numbers of complicated and often conflicting requirements in a complex, connected, multidisciplinary system (Figure 2). Given the final observing spot of the telescope (Lagrange point 2, L2), nearly one million miles away from Earth, NASA had just one chance to get it right–it cost approximately $3.5B for the design, development, launch, and commissioning of this system, and $1B for ten years of operation. This is the most powerful space telescope ever launched, and if it works as designed, it will significantly extend our knowledge of the universe and its origins.

As organizations continue to navigate growing product complexity, strengthening their approach to requirements management becomes essential for reducing risk, improving traceability, and accelerating development cycles. If you're ready to take the next step toward more intelligent, connected engineering practices, explore how a robust digital thread solution can unify data, processes, and verification activities across the lifecycle. To go even further, discover Aras’ comprehensive digital thread and PLM solutions designed to support modern, model based systems engineering and drive more resilient, future-ready product innovation.