Agile Test, Moving V&V Back Upstream Process
Systems Modeling, Design, Development, Integration, Test, Validation, Certification are the waterfall process stages. Unfortunately, it is commonplace practice to be optimistic schedule wise and correspondingly underestimate project complexity. Design and development groups push their schedule out and this also shortens the integration, test, validation and certification.
1) Zachman Framework
2) Digital Twin / MBSE / SysML
3) Brainstorming Session - Dr. Bahill
4) Periodic non-advocate tech review - Dr. Bahill
5) Risk Management - order by risk
6) DARPA Style Teaming - small cross-functional teams
7) Teaching alternate analytic tool sets
8) Preserving Context / Correlated Index / Associative Memory / Similarity Metrics
Zachman Framework
Interacting with Senior IPT Leads and their Initial Models & Designs, I would advocate and
demonstrate benefits of utilizing the Zachman Framework, a method to summarize the system
state across all necessary functional elements.
Digital Twin / MBSE / SysML
Digital Twin / MBSE / SysML are intertwined disciplines to replicate within a digital framework
the models, along with the systems engineering elements. All complex domains require
extensive documentation especially with regard to interfaces between disparate subsystems.
SysML can help conveyance between subject matter experts and software engineering
implementation practitioners easing traceability and associated change management.
Brainstorming Session - Dr. Bahill
Dr. Terry Bahill is a well-known INCOSE senior technical fellow known as an early advocate of
SysML. He also enthusiastically presents baseball flight analysis as a common thread in our
American lives that serves as examples of rigorous analysis methods. Introducing Dr. Bahill in
a “Brainstorming event” helps introduce new and practicing systems engineers to an
invigorating session that helps cross-functional knowledge transfer that may later enable the
necessary collaboration within IPT teams.
Periodic non-advocate tech review - Dr. Bahill
Internal professional etiquette often inhibits rigorous internal assessments. I would advocate
for Dr. Bahill, his colleagues and other external tech fellows participate in annual non-advocate
review (NAR) of all significant internal projects; as means to interject independent review,
suggestions and alternative paths forward. Since I reported directly to Dr. Allen Adler, I
participated in an annual review of NAR with the audience of Dr. Alder, Dr, Ganz (both SVP
Boeing), two former DARPA TTO’s and national labs directors. All participants were permitted to
meet and watch the ongoing discussions in US Person and if necessary classified
environments. Many participants received critical reviews that changed the forward trajectories.
Risk Management - order by risk
Projects often have many required “shall” requirements. Not all requirements are of equal risk.
DARPA teams order schedules and go/no-go periodic risk assessments, along with specific
team members working these high visibility requirements. Dev test, initial integration test and
V&V activities can be pushed to the front of the line with high-risk items.
Teaching alternate analytic methods:
Senior technical leaders have little or no time left to examine and learn about new or existing
methods in adjacent domains. From a companies standpoint horizontal knowledge transfer of
analytic methods that have shown benefit in other program areas are good candidates for
introduction to other domains. Internal advocates of new methods can easily convey benefits or
limitations. Existing methods have the benefit of being deployed in fielded operations, along with
mature test scenarios. These experiments can help collaborative teaming of adjacent domains;
many times adjacent work lead to new IP if the new methods have the same or better
performance as existing methods. Alternatively, the new methods may exhibit limitations.
Preserving Context / Correlated Index / Associative Memory / Similarity Metrics:
Continuous testing and validation requires consistency across program elements. Currently
extensive engineering and analysis documentation is created; however, it remains inaccessible
to many technical staff members. An alternate method that I have utilized in IVHM is called
Memex or Associative Memory Index. Conceptually the idea is that given modern computing
architectures it is now possible to memorize an entire enterprise worth of knowledge whether it
be in text or numerical form. A benefit is that numerical data profiles with many parameters can
be captured in design tests or operational tests within a time series index of indexes.
Separately, given a similarity metric space, the memex can be queried to find all similar
numerical data profiles. This enables V&V testing as it may confirm sufficient similarity between
a certified product element and a newly manufactured product element. Attached is a short
slide deck discussing this alternate method and its benefits. Several of my patents mention
associative memory; one in particular replaces FM4, a common gearbox condition indicator
statistic that reduces a complex spectral frequency profile associated with structural vibration
degradation to a single statistic. The associative memory instead captures the “full spectrum”
and uses similarity metric spaces to determine whether the complex full spectrum is “like”
normal, degraded or critically degraded gearbox. In ground station environments this leads to
above 99% true positive assessment as judged by Dr. Shanthakumaran, advanced tech lead in
Mesa, Az.
Participation in following tasking list:
Model Testing, Dev Test, Integration Test / V&V Projects
1) Materials Analysis - Corrosion, Fatigue, Crack Method References
2) IVHM - AH-64 Structures, F18 FADEC, C-130J, CH47-all ATA, SOSA, HUMS/CPHE
3) Terabeam EO - Free Space Optics, Point/Track, Embedded ELM, BIT/Analytics, Service
4) Process Control - Oil,Gas,Ethylene, Pharma, Pulp/Paper, ManufacturingESM, Emitters
5) Telco EO - Laser, Coupler, WDM, Fiber Endpoints, Fiber, Photodiodes/APDStandard Practice
References:
1) 3D Metal Prints - Materials Analysis
a) https://pennstate.pure.elsevier.com/en/publications/reliability-centered-additive-
manufacturing-computational-design- ( Harris, Reutzel, Earthman, Laskowski )
b) https://www.researchgate.net/profile/EW_Reutzel
c) https://engineering.uci.edu/users/james-earthman
d) https://www.spiedigitallibrary.org/profile/notfound?author=Bernard_Laskowski
2) Integrated Vehicle Health Management
a) USAF C!30J OEM SHM
https://www.researchgate.net/publication/
321289112_Verification_of_the_RAF_C-
130J_Structural_Health_Monitoring_System_Through_Operational_Loads_Meas
urement
b) UK / USAF / US Army OEM SHM - SUMS Structural Usage Monitoring System
http://web.mit.edu/hchin/Public/HAI/AHS_Papers/Usage%20Based%20Fatigue
%20Damage%20Calculation%20for%20AH-64%20Apache%20Dynamic
%20Components.pdf
c) US Army Rotorcraft Condition Indicator Dev, Test, Field -
https://core.ac.uk/display/10554227
d) CH-47 Safety Escapes
http://ww.chinook-helicopter.com/maintenance/CH-47_Maintenance_Messages/
ASAM_Message_Summaries.html
3) Free Space Optical Network / Distributed Ethernet System
a) Point / Track - FPGA C Implementation
https://patentimages.storage.googleapis.com/c6/0d/9c/4a9e30d7193cd1/
US6381055.pdf
Circular / polar coordinate tracking optimizing cross-link performance. Cross-link
optical signal carries performance data enabling observability at edge devices.
b) Cross link optical / ethernet performance
https://patentimages.storage.googleapis.com/d4/ea/97/54aadb85e0d8a5/
US7012897.pdf
Cross link monitor and control via user parametric input on optical link APD
power detection and ethernet switch interface bit errors, when parameters below
set thresholds failover to alternate links via software agents acting like humans,
logging into switch, adding new link VLAN/VPN routes in parallel at either end of
link with automatic recovery; and reverse sequencing to restore original link.c) Cross link optical / ethernet performance
https://patentimages.storage.googleapis.com/d4/ea/97/54aadb85e0d8a5/
US7012897.pdf
Cross link monitor and control via user parametric input on optical link APD
power detection and ethernet switch interface bit errors, when parameters below
set thresholds failover to alternate links via software agents acting like humans,
logging into switch, adding new link VLAN/VPN routes in parallel at either end of
link with automatic recovery; and reverse sequencing to restore original link.
d) A network interface controller enables connection from the edge device to the
FPGA device and the upstream network management control system.
e) A distributed network management system that secures. manages, controls and
monitors all devices in the network. Aggregator services for where, when, what
devices are associated with customers.
f)
Customer and device analytics performance assessment services layers will
need to interact with the edge devices, authenticating with NMS systems.
Creating performance maps of devices, cross-link performance, failures and
failovers to other routes, congestion control.
g) Customer services on top of everything else with Service Level Agreements.
4) Process Control - AspenTech InfoPlus.21 + PID control, Flow, Threshold Alerts
Lead Development Engineer reporting to Founders, my team created core, control
systems and applications that are still in use today.
a) https://www.aspentech.com/en/products/msc/aspen-infoplus21
b) https://www.aspentech.com/en/products/pages/aspen-watch-performance-monitor
c) https://www.aspentech.com/en/industries/pharmaceuticals
5) Telco GR-253 - SONET Physical Layer Characterization
a) https://telecom-info.njdepot.ericsson.net/site-cgi/ido/docs.cgi?
ID=SEARCH&DOCUMENT=GR-253&
b) GR253 4.2.2 - Transmitter Spectral Characteristics
c) GR253 4.2.3 - Receiver Sensitivity, Overload & Reflectance
d) GR253 4.2.4 - Optical Path Reflectance, Attenuation, Dispersion
Comments
Post a Comment