Systems Engineering Status of Industrial Use, Opportunities and

Needs

Clas A. Jacobson Chief Scientist

Systems & Controls Engineering LCCC Lund

September 19, 2012

TEAM

2

Alberto Sangiovanni Vincentelli, Alberto Ferrari, Mark Myers, John Cassidy, Richard Murray, Andrzej Banaszuk, Sean Meyn, Johan Akesson, Hubertus Tummescheit, Karl Astrom, Manfred Morari, Eelco Scholte, Rich Poisson, Satish Narayanan, Kevin Otto, John Burns, Igor Mezic, Marco Di Natale, Scott Bortoff…

AGENDA

3

System Design Systems engineering: (1) requirements, (2) architecture, (3) model based design, (4) (design/development) process Platform Based Design – design flows (orthogonalize concerns; hierarchy) Opportunities & progress System level modeling – positive on reusabality, speed… Architecture exploration – not fully exploited - but enabled Requirements – potential to move between formal languages (in progress for embeddded systems) Model based development – positive on controls - MPC (and optimization), uncertainty (and use for robust design not there yet) Process – progress on integration of tool chains; level of abstraction change (slightly) with domain (but separate into main product development cycles)

DRIVERS

4

System interactions (“emergent behavior”) Requirements & acceptance testing (verification) Safety (critical) (software intensive) systems Reusable architectures (modularity) Robustness (risk, lifing)

SYSTEMS ENGINEERING (DESIGN)

5

Process

From process to analysis (model based development) Bring forward in time the verification testing (SIL => HIL => acceptance) Orthogonalize requirements (requested behavior) and architecture (delivering services)

DESIGN PROCESS

6

Status & Opportunities

Physical Layer

Software Layer

Communications & Control Layer

Architecture Instance #1

Architecture Instance #2

Functional Requirements #1

Functional Requirements #2

Abstraction Layers

Abstraction Levels

Sizing AnalysisInstallation AnalysisVariation source IdentificationFailure Modes & Effects AnalysisModel-Driven Variation AnalysisNetwork Embedded Systems AnalysisSystem Dynamics AnalysisRobust Control DesignCritical Parameter Analysis

Analysis Viewpoints

Analysis Viewpoints

Range of analyses (views) Hierarchy (refinement) Separation of concerns (requirements, architecture, analysis)

SYSTEMS ENGINEERING (DESIGN)

7

Definition

Systems engineering is a methodology for product system level design, optimization and verification that:

1. Provides guarantees of performance and reliability against customer requirements while achieving business cost and time-to-market objectives;

2. Produces modular, extensible architectures for products incorporating mechanical components, embedded systems and application software;

3. Exploits model-based analytical tools and techniques to determine design choices and ensure robust system performance despite variations caused by product manufacturing, integration with other products and customer operation; and

4. achieves these objectives through the coordinated execution of a prescriptive, repeatable and measurable process.

REQUIREMENTS

8

Status & Opportunities

Enabler – formal language (not just equation based language) Status – strong for embedded systems; weak for continuous time (non-simulation based verification) Opportunity – robust design/uncertainty

MODEL BASED DEVELOPMENT

9

Status & Opportunities

Enabler – equations; interconnection structure Status of use of equation based language – strong for optimization (MPC; Akesson-Optimica) ; not exploited for robust design; weak for architecture exploration Opportunity – robust design/uncertainty

ROBUST DESIGN

QFD Test Planning

FMEA DFMA

Voice of the Customer KJ Clustering

Critical Parameter Management

Production Verification

Characterization Experiments

Robustness & Scalability Experiments

Capability Cpk Scorecards

Process Mapping

D & R Document

Business Opportunity

Proposal

Statistical Verification Test

Plans & Number of Prototypes

Off-Nominal Design Verification Experiment Tests

Pugh Process

Reliability Verification

Experiment Tests

135

82

4 13 9

243

8

44

4 6 1

63

8

36

2 6 1

53

0

50

100

150

200

250

300

Coils Compressors Fans Fan Motors Fan Deck Total

Qua

ntity

Drop-In DesignModular ApproachModular Approach, Obsolete A Series

Cpk Forecast and 10X Quality Plan

Design elementin new product

Corresponding design element in

existing design Failure modes By mode(ppm)

Total(ppm)

Observed defects and failures

Existing Product Performance New Design Targets

Improvement goals

Current estimate Improvement strategy Status

Marciano Chart Reliability Forecast & Plan

Modleingg

Functional Modularity

Off-Nominal Production Verification

Experiment Tests

ROBUST DESIGN & UNCERTAINTY

11

Status & Opportunities: Exploit Structure

Probability Distribution of input parameters

Exploit Weak Interaction

Find Weak Interaction

Utilize interconnection structure to tear system into strong & weak connections; propagate uncertainty (Meyn-Mathew (and others) DARPA RUM 2008)

SUMMARY

12

System Design Systems engineering : (1) requirements, (2) architecture, (3) model based design, (4) process Platform Based Design – design flows (orthogonalize concerns; hierarchy) Opportunities & progress System level modeling – positive on reusabality, speed… Architecture exploration – not fully exploited - but enabled Requirements – potential to move between formal languages (in progress for embeddded systems) Model based development – positive on controls - MPC (and optimization), uncertainty (and use for robust design not there yet) Process – progress on integration of tool chains; level of abstraction change (slightly) with domain (but separate into main product development cycles) Summary Big needs on uncertainty/robust design (much wider view of product development); Opportunity for realizing potential of tool integration (FMI) and with PLM (data management)

KEY POINTS

13

System Design Systems engineering : (1) requirements, (2) architecture, (3) model based design, (4) process Platform Based Design – design flows (orthogonalize concerns; hierarchy) Opportunities & progress System level modeling – positive on reusabality, speed… Architecture exploration – not fully exploited - but enabled Requirements – potential to move between formal languages (in progress for embeddded systems) Model based development – MPC (and optimization), uncertainty (not there yet) Process – integration of tool chains; level of abstraction change (slightly) with domain (but separate into main product development cycles) Summary Big needs on uncertainty/robust design; Opportunity for realizing potential of integration (FMI) with tool chain and PLM