Survivable Computing Environment to Support Distributed Autonomic Automation

1
Survivable Computing Environment to Support
Distributed Autonomic Automation

• Dr. Andrés Lebaudy, Mr. Brian Callahan,
• CDR Joseph B. Famme USN (ret)
• ASNE Controls Symposium
• Biloxi, MS
• December 10-11, 2007

1
2
Damage Control Requirements

• Naval studies show that ships are seldom lost to
  primary damage (direct blast effects) but the
  result of secondary damage the progressive
  spreading of fire and flooding into surrounding
  areas
• Key Challenge is to Increase Control System
  Survivability Decrease Casualty Response Time
• Past experience has demonstrated that when
  engineering casualties or damage occurs a human
  is too slow and vulnerable, and requires enormous
  logistical and medical support
• Distributed, Survivable Autonomic Processing
  Contributes to Reduced Response Time

2
3
Learning from Experience
3
4
ONR Multi-level Control Integration
Defining the Requirements for Survivable Computing
4
5
What is a Smart Valve?

• Smart Valves sense or infer valve and fluid
  parameters
• valve (actuator) position
• fluid flow rate
• upstream and downstream fluid pressure
• fluid temperature
• Embedded, programmable microprocessor-based
  controller
• controls valve actuator
• filters sensor data
• estimates flow rate
• perform valve actuator diagnostics
• can be programmed to be intelligent
• Communication interface
• interface with device- or field-level network
• send/receive information to/from other devices on
  the network
• send/receive information and commands to/from
  next highest control system tier

Courtesy of Tyco International Ltd.
6
Smart Valve Applications

• Requires only pre-hit communication
• Each valve independently determines whether it
  lies along the rupture path
• Valves initiate a closure sequence after
  pre-configured time delay
• Activates only when pressure and flow conditions
  are abnormal

• Requires full or partial communication between
  adjacent smart valves
• Neighboring smart valves calculate flow balance
• Rupture detected when flow into the zone is not
  equal to flow out of the zone
• Valves operate to isolate zone
• Allows for estimating rupture or leak size
• Number of branches and uncertainties in
  individual flow estimates determines size of
  rupture that can be reliably detected

7
DDG 1000 Fire Suppression
7
8
Live Fire Test of SmartValve Technology
Autonomic Fire Suppression System

• AFSS EDM successfully responded to all of the
  live-fire test scenarios (Shadwell 2002)
• Follow-up testing of an AFSS prototype was
  demonstrated successfully during a Weapons
  Effects Test (WET) on ex-USS Peterson (Peterson
  2003).

8
9
PAC Component Modular Design

• Multi-domain functionality-including logic,
  motion, and process control-on a single very
  flexible and highly configurable platform.
• Mil Qualified Shock, Moisture

9
10
Multi-level Mil-spec Control Modules

• Computational and storage resources that grow
  with application demands
• Resistant to component failures by distributing
  the processing load

10
11
Next Generation Control Software

• Survivable, reconfigurable third-generation
  graphical design tool
• Windows-based software package that relies on
  intuitive drag-and-drop, undo-redo, and
  cut-copy-paste functionality

11
12
Next Generation Graphical Design Environment

• Comprehensive set of field-proven function
  blocks
• state-diagramming features allow design engineers
  to define operational states

12
13
Field-proven function blocks
Examples

1. Controller Blocks (e.g., PID controller,
   lead-lag controller)
2. Signal Conditioning Functions (e.g.,
   characterizer, rate limiter, track hold)
3. Signal Comparator Blocks (e.g., high/low alarm,
   equality, thresholding)
4. Mathematical Operators (e.g., addition, natural
   log, exponent, sine)
5. Logic Functions (e.g., NAND gate, XOR gate, RS
   flip flop)
6. General Purpose Operators (e.g., timer, ramp
   profile, multiplexer, A/B switch)
7. Hardware Access (e.g., analog input, barograph
   display, pushbutton)
8. Networking Operators (e.g., broadcast, receiver,
   parameter synchronization)
9. Diagnostic Operators (e.g., data recorder,
   hardware status monitor)
10. Text Manipulation (e.g. string constants,
    concatenation, left, right, etc.)

13
14
Fleet Modernization INSTALLATION EXAMPLES

• Naval Surface Warfare Center (NSWC) in
  Philadelphia to accomplish Ship Alteration 480D
  for the following ships USS Boone, USS McInerny
  (FFG 8), USS Gary (FFG 51), and USS Vandergrift
  (FFG 48).
• To regulate the cooling of the four SSDGs, as
  well as the SSDG waste heat temperature,
• the fuel temperature in two sets of oil service
  and transfer heaters, the hot water tank
  temperature, and the start-air-mixer air
  temperature.
• The PACs also control the main engine lube oil
  purifier, cooler, and service pressure loops.

14
15
Weight and Cost Savings - Table
Design Element for 20,000 Point Engineering Control System Conventional Data Acquisition Unit Design (DAU) Survivable Distributed Design Process Closest to Machinery
Enclosure Size including mounts 24x24x14 24x11x6.5 small or mini PACs
Points Density 160 max. Assume 100 36 max. Assume 25
Enclosure WT w/ mounts 140 lbs 16 lbs
No. I/O Drops 200 800
Volume per Drop 1,067 ft3 571 ft3
Weight / Drop 18,000 lbs 12,800 lbs
Cable WT 53,800 lbs 17,000 lbs
Cost Est./Drop 25,000 4,500
Total Cost 5.0 M 3.6M
Est. Weight Savings CVN-21 42,000 lbs gt 18 tons, or 1.4 times the weight of one F/A-18F
15
16
Distributed I/O Processing Saves Cable Cost
Chameleon PAC Can Interface With Any Control
System
Machinery Control System HMI Processors

• Enclosureless Mini-RTU/DAU
• Highly distributed, located in close proximity to
  machinery - Reduced Cable Cost
• Wired or secure wireless communications
• Topologies supported Ring, Bus, Star, Mesh
• Interface to smart sensors 1451.4 and 1451.5
• DDS Publish / Subscribe
• Industrial Communications
• Network Gateways
• Legacy I/O

TSCE Network
e
c
u
r
S
e
n
k

L
i
C
E
T
S
1451.4
1451.4
PWM
4-20mA
RTD
Secure Bluetooth
or 802.11 a/b/g
Pressure
1451.4
RPM
1451.5 /
LonTalk
Temperature
ZigBee
Vibration
Copper/Fiber 10/100MBps
Ethernet Ring with DDS
Communications
ProfiBUS
Ethernet/IP
0-5V
4-20mA
e
r
u
c
S
e
/
g
b
a
/

RTD
1
.
1
2
0
8
TSCE Network
17
Compare Conventional Wiring to Distributed
Process Wiring
TSCE
Conventional Compartment I/O Drop
Distributed Compartment I/O Drop
Distributed Compartment I/O Drop

• Distributed
• Savings
• Installation Costs
• Weight

Ethernet etc.
MIL-SPEC RTUs
Machinery
Machinery
18
CONCLUSIONS

• New Shp Classes will be able to employ
  Decentralized Ship System Architectures with
  Distribute Control Systems in order to Improve
  Rapid System Recovery / Ship Survivability and
  Fight Through Capability
• Survivability is Achieved through Computational
  and Process Electronics Protection Provided by
  Hardware, Hardware Architectures / Control
  Software that is Mil-Spec and Locally
  Reconfigurable
• Using Control Hardware that has been Tested to
  Highest Level of Survivability to Reduce
  Vulnerability to Damage and Ensure No Critical
  Single Points of Vital System Failure
• This solution Supports Reduced Crew Size, Lowers
  the Weight of Wire, and the Cost to Install
  Control Systems thus Improving Ship Production.
• Proposed solutions are Technical Readiness Levels
  7, 8 9.

18