https://ntrs.nasa.gov/search.jsp?R=20110012299 2018-07-24T00:33:53+00:00Z
Smart Sensor Demonstration Payload
John Schmalzel 1 , Andrew Bracey 2 , Stephen Rawls 3 , Jon Morris 4 , Mark Turowski4,
Richard Franzl 5 , and Fernando Figueroa 6
Sensors are a critical element to any monitoring, control,
and evaluation processes such as those needed to support
ground based testing for rocket engine test. Sensor
applications involve tens to thousands of sensors; their
reliable performance is critical to achieving overall system
goals. Many figures of merit are used to describe and
evaluate sensor characteristics; for example, sensitivity and
linearity. In addition, sensor selection must satisfy many
trade-offs among system engineering (SE) requirements to
best integrate sensors into complex systems [1]. These SE
trades include the familiar constraints of power, signal
conditioning, cabling, reliability, and mass, and now
include considerations such as spectrum allocation and
interference for wireless sensors.
Our group at NASA’s John C. Stennis Space Center
(SSC) works in the broad area of integrated systems health
management (ISHM). Core ISHM technologies include
smart and intelligent sensors, anomaly detection, root
cause analysis, prognosis, and interfaces to operators and
other system elements [2]. Sensor technologies are the
base fabric that feed data and health information to higher
layers. Cost-effective operation of the complement of test
stands benefits from technologies and methodologies that
contribute to reductions in labor costs, improvements in
efficiency, reductions in turn-around times, improved
reliability, and other measures. ISHM is an active area of
development at SSC because it offers the potential to
achieve many of those operational goals [3-5].
provide the means for crew escape from on-pad to near
orbit. The nominal Ares-I LES was to be a tower-based
separation system similar to the Apollo escape system
designed by Maxime (Max) Faget and Andre Meyer [6].
The LES was deemed so important that two alternative
designs were also investigated: the alternate launch abort
system (ALAS) and the Max launch abort system (MLAS)
[7-8]. The MLAS approach is based on a fairing over the
capsule with integral rocket motors instead of the tower
mounted motors used in the Apollo LES. Fig. 1 shows the
MLAS flight test vehicle stacked on the launch stool at
Wallops Flight Facility.
MLAS: Max Launch Abort System
NASA’s Constellation Program sought to develop a new
generation of rockets as part of plans to replace the Shuttle
Transportation System (STS). The Ares-I is designed to
insert crews into orbit in the Orion capsule similar to the
architecture of the Apollo program. Ares-V is a heavy lift
vehicle to loft large payloads such as exploration vehicles
or large scientific instruments. Ares-I included a
requirement for a launch escape system (LES) that would
1
Fig. 1. MLAS flight vehicle stack.
Was with NASA-SSC (IPA); now with ECE Dept., Rowan University, Glassboro, NJ 08028
NASA, EA31, Stennis Space Center, MS 39529
3
NASA, EA22, Stennis Space Center, MS 39529
4
Jacobs Technology, M/S B8306, Stennis Space Center, MS 39529
5
Smith Research, M/S B8306, Stennis Space Center, MS 39529
6
NASA, Innovative Partnerships Office, IA30, Stennis Space Center, MS 39529
The opinions expressed are those of the authors and not those of NASA.
2
In addition to the primary flight objectives to prove the
aerodynamic stability of the LES concept, the MLAS
project also included opportunities for technology
demonstration payloads. Our group proposed a smart and
intelligent payload (SiSP) project to advance the
technology readiness level (TRL) of several sensor
technologies [9]; due to MLAS project timeline
constraints, SiSP was approved for ground operations
associated with the MLAS launch with the possibility of a
future flight opportunity.
SiSP Project Objectives and Embodiment
The SiSP project sought to demonstrate a suite of sensor
and sensor system advances including:
(1) Smart sensors. Demonstrate sensor technologies that
adhere to defined standards.
(2) Advanced commercial off the shelf (COTS) data
acquisition system. Demonstrate the use of flexible
COTS data acquisition architectures to make
redundant (and other) ground based sensor (GBS)
measurements.
(3) RFID. Integrate radio frequency identification (RFID)
technology for tracking system elements and to link to
transducer electronic data sheets (TEDS).
(4) PoE. Use power over Ethernet (PoE) to provide
subsystem power.
(5) Wireless sensors. Employ wireless sensor
technologies.
(6) Intelligent sensors. Incorporate intelligent sensors with
embedded health assessment.
(7) Other sensor-related opportunities as time and budget
permitted.
Smart Sensors
Objective: Incorporate smart sensors adhering to defined
standards to reduce SE and maintenance costs.
Rationale: Methods are needed that offer “plug-and-play”
to support automatic configuration of a data acquisition
system to recognize the collection of sensors present as
well as to allow simple reconfiguration when sensors are
replaced. IEEE 1451.x standards define various “Smart
Transducers and Actuators” and associated logical
function and physical interface definitions that can
contribute to achieving the goal of “plug-and-play” sensor
technologies [10-11].
One of the most immediately useful elements defined
in the standards is the Transducer Electronic Data Sheet
(TEDS). IEEE 1451.4 defines TEDS data structures for
storing and sharing basic sensor information including
manufacturer, model, calibration status, calibration
coefficients, etc. Systems with TEDS-enabled sensors can
be rapidly configured and easily updated as sensors are
exchanged for maintenance actions. The use of TEDS
helps minimize configuration errors made during repetitive
sensor database updates. IEEE 1451.4 provides detailed
TEDS templates for commonly encountered sensors
including the bridge (pressure), accelerometer, and
thermocouples used by SiSP, as well as the thermistors
used as part of the MLAS GBS measurements.
IEEE 1451.0 and 1451.1 address network centric
sensor capabilities. The network capable application
processor (NCAP) supports many plug-and-play functions
including sensor discovery and data exchange with a
publish-subscribe approach. One advantage of a shared
data acquisition system typical of legacy systems is shared
timing and voltage references. Disadvantages include the
reliability risks of a single system and the increased
cabling costs due to all sensors having to be routed to a
single point. Advantages of a distributed smart sensor are
reduced risk of failure due to distributed functionality and
minimization of cabling if a bus topology is used.
Disadvantages include timing uncertainty and the need for
redundant elements such as voltage references for each
sensor.
Embodiment: A tri-axial accelerometer (PCB Piezotronics,
TLD 356A16) containing integral IEEE 1451.4 TEDS was
incorporated. Fig. 2 shows the 3-axis accelerometer. Signal
conditioning/acquisition functions and TEDS reader
capability were provided by the NI 9234 IEPE module (see
below). IEEE 1451.1 NCAP functionality was
implemented using an ARM-based microcontroller to
communicate with internal and external elements.
Fig. 2. 3-axis accelerometer with TEDS.
COTS Data Acquisition System
Objective: Employ a small reconfigurable commercial off
the shelf (COTS) data acquisition system to make
redundant GBS and other measurements.
Rationale: The CompactRIO (C-RIO) chassis is a field
programmable gate array (FPGA) based system that offers
a low power (< 25W fully populated) chassis that supports
up to eight signal conditioning modules [12]. Software is
developed using the LabVIEW graphical programming
environment; resulting code is downloaded to the FPGAs.
The result is a rapid development environment. Our
group’s experience with the C-RIO hardware and software
suggested it as the way to acquire redundant IOP
measurements and to make additional thermal, acoustic
and accelerometer measurements.
Embodiment: An 8-slot C-RIO system was used; Table 1
summarizes the mapping between conventional sensor and
signal conditioning module.
Table 1. SiSP CompactRIO data acquisition functions.
Module
Chan.
Count
Samples/s
(sps), Size
Measurement
Function
Bridge
NI 9237
4
1.6k –
51.2k
Digital IO
NI 9203
32
0 – 100k
IEPE Analog
Input
NI 9234
4
1.6k –
51.2k
Accelerometer
Microphone
TC, Type K
NI 9211
4
0 – 14
Internal
Temperature
Universal
Analog Input
NI 9219
4
2 – 100
Radiometers,
Battery
Voltage
SD Memory
Card
NI 9802
2
2 GB per
slot
Data storage
Ignition Over
Pressure
T-600, T0
Triggers
IRIG-B DC
RFID Tags
Objective: Integrate radio frequency identification (RFID)
technology for tracking system elements and to link to
transducer electronic data sheets (TEDS).
Rationale: The benefits of TEDS functionality are out of
reach of systems composed of conventional sensors unless
some means of retrofitting TEDS is available. The
standard makes provision for a virtual TEDS (VTEDS) by
allowing a keyed lookup into a database where the TEDS
information is stored. Radio frequency identification
(RFID) tags offer one means of retrofitting TEDS to
legacy sensors. Simple passive RFID tags contain a 128-bit
memory with a unique code value. A reader excites the
RFID tag with RF energy sufficient to power the device in
order to transmit its code to the reader. The unique RFID
tag is then used as the key for a database search to find the
associated VTEDS. IEEE 1451.7 is a recently approved
standard that incorporates RFID into the smart sensor
framework. Another application for RFID in a system is
for subsystem identification. This offers a means to
validate system builds to confirm that all the elements have
been qualified for flight.
Embodiment: Medium frequency, 13.56 MHz, RFID tags
(Texas Instruments, RI-I03-112A-03) were used for
VTEDS and inventory control. These tags have an
additional 2048-bit user memory to allow future upgrades
to local TEDS storage in the RFID tag. Fig. 3 shows a
typical passive RFID tag.
Fig. 3. Passive RFID tag.
Shared Signal and Power Cabling
Objective: Provide power using network cabling.
Rationale: Cabling and interconnections for wired
applications represents high costs as measured by mass,
volume, electromagnetic interference (EMI) susceptibility,
etc. Ways to reduce cabling and connector mass could
prove beneficial in many aerospace applications. IEEE
802.3af defines Power over Ethernet (PoE). Up to 12.95 W
of dc power at approximately 48 Vdc is available to a
powered device. Power sourcing equipment can source up
to 15.4 W. Table 2 summarizes the power levels available
at a device.
Table 2. IEEE 802.3af power levels.
Class
0
1
2
3
4
Maximum Power at Device, Watts
0.44 – 12.95
0.44 – 3.84
3.84 – 6.49
6.49 – 12.95
(Reserved)
A recently approved extension, IEEE 802.3at-2009
extends the available power up to 25W for a powered
device.
Embodiment: PoE was used to power the C-RIO data
acquisition system and two external sensors: (1) an
intelligent sensor located at the base of the launch stool,
and (2) a wireless sensor node. PoE injectors were
contained within the SiSP package.
Wireless Sensors
Objective: Employ wireless sensors.
Rationale: Wireless communication offers the means to
reduce cable mass. There are many wireless
communication standards in widespread use such as IEEE
802.11 (WiFi), IEEE 802.15.1 (Bluetooth®), and IEEE
802.15.4 (ZigBee®). Some of the considerations affecting
the possible use and choice of wireless sensors include
distance, interference with other Rf devices, signaling rate,
and power consumption. The importance of each of these
factors is highly dependent on application; for example,
applications such as a quick turn monitoring problem may
not be concerned about power consumption.
Embodiment. The original plan was to monitor one
channel of ignition over pressure using a Zigbee® wireless
smart sensor MOBEE (Mobitrum Corp.). Because the 1
sps sampling rate was too low for dynamic pressure
measurements, the wireless sensor was applied to the much
lower bandwidth requirements of tracking ambient
temperature. A frequency utilization request was
submitted—and permission obtained—for this low-power
RF application. Redundant launch stool temperature
measurements were made in the days prior to launch. Fig.
4 shows the 4-channel wireless sensor.
Fig. 4. Wireless sensor.
Intelligent Sensors
Objective: Use intelligent sensors with embedded health
assessment.
Rationale: Further system benefits can be achieved if smart
sensors can perform additional assessment tasks. In
addition to converting raw data into final engineering
units, such “intelligent sensors” can also perform data
validity checks and other functions that reduce processing
overhead at a central monitoring station. The availability
of smart sensors with sufficient computing capability to
implement NCAP functions makes it possible to embed
those additional algorithms to make the smart sensor
intelligent.
Fig. 5. Intelligent Sensor (SNE).
Other Opportunities
Objective: Identify and incorporate other related sensor
technology elements as time and budget permit.
Rationale: Ways are needed to address the new challenges
posed by distributed sensors including timing and data
acquisition accuracy. The timing problem arises from
asynchronous sampling; techniques are needed to ensure
that all samples can be accurately time aligned. IEEE 1588
defines methods for achieving time synchronization across
a network that supports multicast messaging. Using the
IEEE 1588 protocol, timing jitter between network nodes
can be kept below one microsecond. This technique
provides orders of magnitude improvement over legacy
IRIG-B timing.
Embodiment: We investigated the use of IEEE 1588 in the
lab, but decided to postpone 1588 integration due to
resource constraints. IRIG-B was available at the launch
pad and was used since all other launch systems also were
based on IRIG-B.
SiSP Implementation
The SiSP development approach adopted the MLAS
Resident Engineer model; two early-career SSC engineers
were tapped as the electrical (A. Bracey) and mechanical
leads (S. Rawls). They were supported by senior NASA
personnel and a contractor team. The project started in 2Q
2008 and was to be completed by 3Q 2008 in keeping with
the original MLAS flight schedule. Rapid prototyping
approaches were used where possible including use of 3-d
ABS printing for custom housing fabrication (Fig. 6.);
panels and chassis elements were fabricated using an
abrasive water jet (AWJ).
Embodiment: We had previously collaborated with a group
at Kennedy Space Center (KSC) who had developed a
smart networked element (SNE) [13]. In addition to
supporting the IEEE 1451.1 protocol, the SNE
implemented a number of health detection routines. A
KSC SNE was modified to support ignition over pressure
measurements and was installed at the base of the launch
stool. Fig. 5 shows an unpackaged SNE.
Fig. 6. 3-d solids model of sensor housing for fabrication
using ABS printer.
Software development costs were manageable due to the
availability and reuse of significant portions of code for
data acquisition and smart/intelligent sensor support from
prior projects.
Fig. 7 depicts the block diagram of the SiSP. Ethernet
(802.3af) provides the communication core of the system.
Onboard power is available from a primary cell; however,
an external power supply was used to power the system
due to launch pad safety considerations that required
remote control of all power sources during motor arming
procedures.
COTS Data
Acquisition
(C-RIO)
3-axis
Accelerometer
w/ IEEE 1451.4
TEDS
Ethernet Switch
& PoE Injectors
(Power from
onboard battery
or external
supply)
IEEE 1451.1
NCAP
(Optional)
Legacy Sensors:
External:
2-Radiometers
1-Microphone
3-Pressure
PC-104
Controller
IEEE 1451.1
NCAP
Intelligent Sensor:
Smart Networked
Element (SNE):
Pressure
Internal:
4-Thermocouples
External Triggers:
T-600 s
T0
IEEE 1451.1
NCAP
Wireless Sensor
Control Room
Operator Interface
Wireless Portal
IEEE
802.4.15
Zigbee
IEEE 1451.1
NCAP
IEEE 802.3af
Fig. 7. SiSP block diagram showing major elements.
Fig. 8. The completed SiSP.
Controller
SiSP required an onboard controller to perform a number
of functions including power sequencing, data storage, and
communication with a remote control room computer. Fig.
9 depicts the state diagram for the controller. Every state
provides a path to the advance to shutdown (ATSD) state
to inert the system, which requires manual reset to
override. The power up sequence was initiated by a T-600
s trigger received from the launch pad ground support
equipment (GSE). The system could be placed into a hold
state to accommodate anticipated launch delays. The GSE
T0 trigger initiating MLAS motor ignition was used to
start the data acquisition process, which continued for 10 s
before entering the ATSD state.
Prior to shipment to WFF, a flight readiness review
(FRR) was held at SSC. The only issues outstanding were
final integration into the WFF GSE; the project was given
the green light to proceed to launch test.
Fig. 9. Controller state diagram.
Launch System Integration
Fig. 10 shows how the SiSP was integrated into the
Wallops Island GSE. A Connex shipping container located
near the launch pad provided remotely controlled power
switching, GSE triggers, network access, and IRIG-B
timing. A strict network access policy was enforced to
ensure security; the control room computer was provided
to our project team, which meant that some last-minute
software modifications were needed to work around
network and operating system features that were different
from the baseline development platform.
The SiSP package was placed between the Connex
and the launch pad behind a blast barrier. Wiring extended
through conduit to the center of the launch stool and then
was distributed to the collection of sensors. Four pressure
transducers were mounted on the launch stool to make IOP
measurements, two radiometers were attached to the blast
wall, and a microphone was collocated with one of
Marshall Space Flight Center’s acoustic measurement
positions.
Fig. 10. Integration of SiSP with Wallops Island ground support equipment.
Results
Launch
MLAS was launched on July 8, 2009 early in the morning.
See [14] for a video of the launch. Power flight lasted
approximately seven seconds followed by a series of
spectacular parachute deployments. For SiSP, the first
seconds were all that were recorded.
Post-Launch Review
SiSP performed nominally. Fig. 11 shows one example of
the data taken from the launch event, which is a plot of
thermal data obtained from the blast walls. The signal is
lost at the 1.5 s mark due to blast impingement on the
sensors and interconnect wiring. Loss of wiring was
significant as shown in Fig. 12.
Fig. 11. Radiometer measurements from blast wall.
5.
Intelligent sensors show promise. A smart sensor
combined with health detection algorithms offers
future systems data and health measures to help
identify failing sensors. Distributed sensors can also
contribute to decentralizing monitoring and control
functions.
Future Work
The lessons learned from SiSP suggest a number of
follow-on actions. Several of the most important are:
Fig. 12. Wiring damage, post launch.
Review of the data that resulted from the SiSP experiment
shows that it achieved the majority of the important
objectives.
1.
2.
3.
4.
Sensor standards (e.g., IEEE STD 1451.1/.4) provide
the ability to simplify sensor installation and
maintenance moving toward plug-and-play capability.
In particular, TEDS provides key benefits that reduce
labor and risks associated with configuration. SiSP
showed that RFID technology is a cost-effective
method of retrofitting virtual TEDS capability into
existing systems with conventional sensors. RFID also
offers a simple means for configuration management.
The advantage of a flexible COTS data acquisition
system has been demonstrated by the ease with which
core SiSP data acquisition functions were
implemented including the ability to directly connect
IEEE STD 1451.4 sensors. The utility of this approach
has been demonstrated in other follow-on applications
that used instances of the technologies demonstrated
by SiSP.
Power over Ethernet (PoE) can be used to simplify
interconnects. The range of power supported by IEEE
STD 802.3af/at covers useful subsystem functions.
Development efforts should also be directed at
supporting even more flexibility such as adding lower
power classes of operation (< 0.44 W) to
accommodate evolution in low power smart and
intelligent sensor technology.
Wireless sensor elements show promise as a means for
reducing interconnect. However, wireless sensor node
power budgets must support the application
requirements as measured by total lifetime of batteries
traded against the sampling rates achievable by the
sensor. The spectrum used for wireless is used by
many other applications so will continue to become
more crowded with attendant interference and security
issues.
1. Flight opportunity. MLAS was unique in that there were
few mass constraints—the capsule simulator was
approximately 25,000 kg, so the SiSP at 20 kg was
inconsequential. However, that is not the norm: a high
premium is placed on low mass, small volume, and low
power. Advancing the complement of SiSP sensor
technologies to TRL8 and beyond requires significant
redesign to achieve a flight-ready payload. Flight
opportunities should be sought to continue this
advancement.
2. Sensor standards. IEEE 1451.4 is well-defined and
simple enough to have found commercial adoption;
however, that is not the case for IEEE 1451.1 as one
example. Efforts to update 1451.1 are underway [15]; this
should allow simpler implementations of networked
sensors that better meet the goal of plug-and-play sensor
architectures.
3. Sensor networks. Optimization of sensor networks will
require changes in power management—e.g., network
standards that support low-power nodes, and ways to
handle distributed timing.
Finally, this project showed that a focused team could be
assembled to accomplish a significant project in a
relatively short time. The vertical experience model was
also shown to be an important way to not only accomplish
the project goals but serve as an effective means to convey
experience and values to the engineers who will be the
vanguard of the next exploration frontiers.
Acknowledgements
A year ago (as of this writing), the SiSP field team was
swatting mosquitoes on Wallops Island during the final
week of installation and checkout. The project could never
have gotten to that point without the leadership of the
NESC, the support of the NASA-SSC Innovative
Partnerships Program, the NASA-HQ Technical
Excellence Initiative Program, and the many contributing
NASA centers. L. Langford and W. Mitchell provided
continuing systems engineering and technical support. P.
Mease (Rowan University) fabricated our AWJ parts. All
are gratefully acknowledged.
For Further Reading
[1]
K.R. Fowler and J.L. Schmalzel, “Sensors: The first stage in the
measurement chain,” IEEE I&M Mag., 7:3, September 2004.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
F. Figueroa, J. Schmalzel, M. Walker, M. Venkatesh, R. Kapadia, J.
Morris, et al., “Integrated systems health management: Foundational
concepts, approach, and implementation,” 2009 AIAA
Infotech@Aerospace, Seattle, Washington.
J.L. Schmalzel, M. Turowski, R. Franzl, and F. Figueroa, “Anomaly
detection toolkit for Integrated Systems Health Management
(ISHM),” Infotech@Aerospace 2010, Atlanta, GA, 20-22 April
2010, AIAA-2010-3498.
G.D. Lecakes, Jr., J.A. Morris, J.L. Schmalzel and S. Mandayam,
“Virtual reality environments for integrated systems health
management of rocket engine tests,” IEEE Trans. on
Instrumentation and Measurement, 58:9, September 2009, pp.
3050-7.
F. Figueroa, J. Schmalzel, J. Morris, M. Turowski, and R. Franzl,
“Integrated System Health Management: Pilot operational
implementation in a rocket engine test stand,” Infotech@Aerospace
2010, Atlanta, GA, 20-22 April 2010, AIAA-2010-3454.
U.S. Patent 3,001,739, M. Faget and A. Meyer, “Aerial capsule
emergency separation device.”
J. Paulson, “Preliminary analysis of Ares I alternate launch abort
system (ALAS) configurations tested in the Boeing polysonic wind
tunnel,” NASA/CR-2007-214877.
http://www.nasa.gov/offices/nesc/home/index.html.
J. Schmalzel, F. Figueroa, and J. Lansaw, NASA TechNeeds:
“Smart and intelligent sensors,” NASA Tech Briefs, 33:12,
December 2009.
April 2008 issue of the I&M Magazine.
http://ieee1451.nist.gov/
National Instruments, Austin, TX. http://www.ni.com.
R.L. Oostdyke, C.T. Mata, and J.M. Perotti, “A Kennedy Space
Center implementation of IEEE 1451 networked smart sensors and
lessons learned,” 2006 IEEE Aerospace Conf., Big Sky, MT.
http://www.nasa.gov/exploration/features/mlas.html
TC-9 working group for P1451.1.