1. Model-Based Product Line Architecture and Analysis
SEDC 2014 Conference
Friday, 04 April 2014
Sean McGervey
Systems Architect

2. Acknowledgements Tamara Valinoto
Chairperson of Northrop Grumman’s MBE Community of Practice
Sounding board
Source of sustained motivation for finishing my presentation
… and Dear Friend --- Thank you!

3. Agenda Why is a new paradigm for development needed?
Three enabling elements for the new paradigm
Model-Based Systems Engineering
Multidisciplinary Analysis and Optimization
Product Line Architecture
Rationale for combining these enabling elements?
How can these enabling elements be combined?

4. So Why is a New Development Paradigm Needed?
Product development is still often based on opportunistic reuse or one-off modification of components to fit needs
Components that are modified must go through time consuming and costly verification and validation again
New and modified components have no pedigree or usage history making them riskier to use in products
Component development often involves multidisciplinary design aspects, conflicting requirements, and management of complex sets of system attributes and their parametric relationships
Product A
Payload
Radar
Structural
Frame
EO/IR
Engine
Comms
Product B
Radar
Payload
Structural
Frame
Weapon System
EO/IR
Engine
Comms

5. Enablers for a Better Development Paradigm: Model-Based Systems Engineering
Well-defined interfaces and self-consistent specifications for products and components
Robust traceability of flow down of requirements to design of products and components
Ability to assert structural, behavioral, and parametric relationships between components of products
Requirements
Behavior
Structure
Parametrics

6. Enablers for a Better Development Paradigm: Multidisciplinary Analysis and Optimization
MDAO has long been a staple of the aerospace industry
Notably, championed by NASA’s Glenn Research Center
MDAO attacks the problem of multidisciplinary system design, analysis, and optimization
Federates computational models of the system from the perspective of various design disciplines

7. Integration Mechanism
Implementing a Better Development Paradigm, Part 1: MBSE + MDAO = Integrated Analytical Design Verification
During Product Development, system attributes captured in the system model can be parametrically related to technical measures and analyzed by an integrated MDAO framework
MBSE Workflow
Integration Mechanism
MDAO Workflow

8. Enablers for a Better Development Paradigm: Product Line Architecture
Automobile
Use Cases
Product lines enable reuse at each phase of the systems engineering lifecycle
Everything including product requirements, features, and configurations can be reused during product development
Many requirements will be common across all products in a product line, while others are optional
Key to realizing “economy of scale” benefits lies in preplanned reuse of product components
Kernel Use Case
<<kernel>>
Brake Car
<<kernel>>
Steer Car
<<kernel>>
Start Car
<<optional>>
Open Roof
Optional Use Case
Automobile
Features
Common Feature
Optional Feature
Automobile
Product

9. Enablers for a Better Development Paradigm: Product Line Lessons Learned from Lego™
Lego™ products are not just popular for the predefined designs that can be built, but for the reusable blocks
Blocks have well-defined and standardized specifications which governs how they can be used
While an entire predefined design has limited reuse value, the blocks have tremendous reuse value

10. Enablers for a Better Development Paradigm: Reuse Should Be Pre-Planned, Not Opportunistic
Failing to appreciate that can result in unforeseen time and money modifying a component to fit a new product context
Reusing a component means reusing all the requirements and architectural decisions that drove its development

11. Enablers for a Better Development Paradigm: Techniques for Product Line Architecture Modeling
PLUS (Product Line UML-Based Software Engineering)
PLUS extends UML for modeling commonality and variability in software product lines
PLUS can also be used to extend SysML to model system-level product lines
System level requirements, use cases, domain features, and architecture can all be modeled as required, optional, mutually exclusive, and more

12. Implementing a Better Development Paradigm, Part 2: PLA + MBSE = Domain Specific Modeling Language
Combining PLA with MBSE allows unambiguous and self-consistent specification of systems in terms of the components that are common, optional, and variable across members of its product line
Models can be built using standardized elements that are domain-specific with rules for their use that can be validated and enforced by tools

13. Implementing a Better Development Paradigm, Part 2: PLA + MBSE = Domain Specific Modeling Language
The application of MBSE to the specification of a PLA also sets the stage for using MDAO for verifying design assertions
Parametric relationships between families of components can be specified and analyzed to guide an optimal system assembly

14. Putting the Pieces Together: MBSE + MDAO + PLA = Robust Product Development
Development starts with specifying the new product in terms of the product line’s architecture constraints, and determining which components are already provisioned and which must be developed.
MBSE
(Product Specification)
New components are developed using the domain specific modeling language for the product line, and their design is verified using the integrated MDAO framework.
Verified components are added to the collection of provisioned components for the product line, and can be used in future product development.
PLA
(Component Provisioning)
MDAO
(Product Verification)

15. Summary
Products can be made faster, cheaper, and better by adopting a product line approach to their development
Opportunistic reuse and one-off modification of components is replaced by pre-planned reuse for product development
Development of new components is managed in accordance with the constraints of the product line architecture
The architecture of each component, product, and product line is specified using a SysML-based modeling language that is specific to that product line’s domain
Development of products and new components is guided by continuous verification of their design using an integrated multidisciplinary analysis and optimization framework

17. Integrated System Model
Analysis Framework
Cost Estimation
CAD
CAE / CFD
Optimization
Simulation
Mathematical