Note: Descriptions are shown in the official language in which they were submitted.
CA 02884858 2016-05-31
H8323257CA
System and Method for Monitoring Railcar Performance
Field of the Invention
[0002] This invention relates to a real-time monitoring and analysis system
for railcars, and, more
particularly, to a system that gathers data regarding various operating
parameters and conditions from
remote sensors and applies heuristics to analyze the data to detect and/or
predict operational failures.
Specific uses for monitoring temperatures on a railcar are also disclosed.
Background of the Invention
[0003] To prevent incidents and improve efficiency, railcar owners and
operators need an understanding
of how their assets are performing. With heavier cars in service, there is a
greater need to identify "bad
actors" (cars which can damage track infrastructure and lead to derailments)
as soon as their
performance becomes unacceptable. There is also a need to increase average
train speed by improving
high speed performance and reducing unplanned service interruptions through
mechanical failures. Car
owners increasingly seek to implement preventative maintenance programs to
predict and avoid
mechanical failures in the field and to efficiently schedule repairs at a
facility and time of their choice.
Finally, with more automation of rail operations and increasing regulation to
improve safety, the
railroad industry needs new ways to monitor the performance of trains, cars
and railcar trucks.
[0004] Even minor mechanical failures could lead very quickly to a
catastrophic failure, not only of a
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
single railcar, but of the entire train. Thus it is desirable to detect and
report deviations from
operational norms or predictions of impending failure to the locomotive or to
a central data handling
facility as quickly as possible, allowing for timely human intervention.
[0005] One approach in use in North America is the use of wayside defect
detectors at fixed locations
throughout the railroad network. Detectors measuring bearing temperature
(hotbox detectors) are
common, while other wayside detectors to measure wheel impacts, bearing
condition (from acoustical
signatures) and lateral forces are gradually being introduced. However, while
one detector can monitor
many freight cars as they pass, they can only provide a spot check on
performance. It is quite possible
that defects will only become apparent and escalate to a critical level
between detectors.
[0006] Another approach to railcar performance monitoring has been to use on-
board instrumentation.
One such prominent system has been developed for the Federal Railroad
Administration. In this and
other similar systems, a number of instruments on different areas of a freight
car are used to make
discrete measurements before being communicated to a central hub on the
freight car. While providing
a superior solution to that provided by wayside monitors, wiring, complexity
and costs increase the
investment required to monitor the cars and decrease efficiency and
reliability.
[0007] The current systems, however, lack the ability to apply heuristics to
act on data gathered from
more than one sensor or to detect operational deviations or trends which show
deviations from nominal
operating parameters. Furthermore, current systems are limited in that they
lack the ability to apply
such heuristics at multiple levels, for example, at the individual sensor
level, at the railcar level, and at
the train level. Lastly, current systems lack the ability for sensors to
efficiently and reliably
communicate their data to a central data gathering facility using a wireless
communications
infrastructure that has multiple redundancies and which allows communication
of data between
individual sensors.
[0008] Therefore, it would be desirable to have a system which addresses these
current deficiencies
2
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
and which improves (1) the ability to reliably collect and utilize data from
multiple sensors on each
railcar; (2) the ability to analyze collected data by the application of
heuristics to detect and predict
operational deficiencies; and (3) the ability to determine the severity of
detected conditions to determine
if immediate alarms should be raised to facilitate human intervention.
Summary of the Invention
[0009] The present invention has the objective of providing means for
monitoring the output from a
variety of sensors attached to a railcar and determining the behavior and
condition of the railcar and its
various components based on an analysis of this data. This provides regular
assurance of proper
performance and condition as well as necessary warnings of impending or actual
failure in a timely and
useful manner to the operators and owners of the train.
[0010] Some of the performance criteria that is useful to monitor, for
example, include roller bearing
temperature, temperature of the commodity being carried, position of the hand
brake, roller bearing
adapter displacement, wheel condition, truck hunting/warp/binding, brake
status and performance, load
status and load amount, whether a partial derailment has occurred and
potentially problematic track
conditions.
[0011] Given the demanding environment in which railroad trains operate, any
monitoring system
must be rugged, reliable and able to operate for long periods with little or
no maintenance. In addition,
to be cost effective, it should not add significant cost to install, maintain
or operate the system. Because
there are more than 1.5 million freight cars in North America alone, a system
of monitoring all cars in
use is highly desirable and, as such, the system needs to be able to deal with
a very large number of
potential devices.
[0012] In one embodiment of the invention, sensing units or devices, herein
referred to as "motes",
are deployed at various locations around the railcar. The motes can include a
sensor, a power source,
3
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
circuitry to read the sensor and convert the readings to a digital form, and
communication circuitry
which allows the mote to wirelessly transmit the sensor readings to an
external receiver. In some
instances, a mote may also have the capability to perform low-level analysis
of the data to determine if
an alarm needs to be raised, and the ability to communicate the alarm to an
external receiver.
[0013] Each railcar also can be equipped with a communication management unit
(CMU) which
communicates with each of the motes deployed on the railcar. The CMU is
capable of wirelessly
collecting data from each of the motes and performing higher-level analysis of
the data to detect
imminent or actual failures. During such data analysis, heuristics may be
applied to determine potential
failures based on statistical models and empirical data. The CMU is also
capable of communicating
both the data and the results of any analysis to a receiver remote from the
railcar.
[0014] The remote receiver may be located on the locomotive or other central
location on the train, or
may be off-train. The remote receiver may also be able to perform higher-level
analysis of the condition
of the train by applying heuristics and statistical models to data collected
from a plurality of CMUs,
located on different railcars in the train. The analysis of the data collected
can be carried out at any of
the different event engines distributed among the various components in the
present invention,
including the sensor units, CMU, and mobile or land base stations.
[0015] It is therefore an objective of this invention to provide a
comprehensive system which allows
the wireless collections of data and the analysis of that data to predict
operational failures and to
provide adequate warning of those failures to allow for human intervention
before a catastrophic failure
occurs.
[0016] It is another object to provide specific sensor applications, such as
temperature sensors that can
monitor the temperature of various components and items on the railcar.
[0017] The discussion which follows describes the system as in the context of
a freight car, however,
it will be understood by one of skill in the art that the same methods are
applicable to any railroad
4
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
vehicle. Furthermore, while the description which follows features a freight
car with two trucks (or
bogies), it is applicable to any configuration with more or less trucks or
axles.
Brief Description of the Drawings
[0018] Figure 1 is an exploded perspective view of a mote in accordance with
the present invention;
[0019] Figure 2 is a cross sectional view of the mote shown in Figure 1;
[0020] Figures 3 and 3A are views of the mote of Figure 1 shown mounted on a
railcar bearing
adapter;
[0021] Figure 4 is view of a mote, configured as a temperature sensor, mounted
on a railcar to obtain
ambient air temperature readings;
[0022] Figure 5 is a view of an alternative temperature sensor mote;
[0023] Figure 5A is an exploded view of the sensor mote of Figure 5 shown with
components for
mounting to the railcar;
[0024] Figure 5B is a schematic view of the sensor mote of Figure 5 shown in
use with a tank railcar;
and
[0025] Figure 6 is a schematic diagram showing the communications pathways in
accordance with the
primary embodiment of the invention.
Detailed Description of the Embodiments
[0026] In broad terms, a novel means for monitoring the performance and
operation of a railcar is
provided. This includes a system for monitoring the railcar and sensors
mounted on the railcars for use
with the system. These sensors communicate with a communication management
unit preferably
mounted on the railcar. The sensors monitor and/or collect data on particular
parameters and conditions
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
of the railcar. If a problem is detected, alarms can be forwarded for further
action. The sensors are
describe below with an exemplary sensor directed to monitoring temperature.
This is followed by a
detailed description of the monitoring system using the sensors.
[0027] In a preferred embodiment of the invention, the sensors are contained
and deployed in a self-
contained housing which generally includes the sensor, long-life batteries, a
processor board and
communications unit. As previously mentioned, these remote units are referred
to herein as "motes".
The motes can be configured for the parameter or condition to be monitored,
and can be placed on the
train in the location chosen for such monitoring.
[0028] With reference to Figures 1, 2, 3 and 3A, an exemplary mote 10 is now
described. In this
particular embodiment, the mote 10 is configured to be mountable to the
surface of bearing adapter 12
of a railcar for monitoring the temperature of the wheel bearing. The mote 10
has a housing 14
having a first or lid section 16 and a second or base section 18. Preferably,
the housing sections 16 and
18 are composed of a hard plastic resistant to environmental damage, such as a
UV rated polymer, e.g.,
a polycarbonate/ABS blend, and when fully assembled is weatherproof After the
various components
are installed within the housing 14 as described below, a potting material
(not shown) is provided
through openings in the housing 14 to maintain, encapsulate and
environmentally seal the components
within. Any suitable electrical potting material capable of protecting the
electric circuitry and
components from the harsh railroad environment can be used, where harsh
weather, UV exposure,
humidity, vibration, mechanical impact, thermal shocks and abrasion might
occur while the device is in
operation. Such materials include epoxies, polyurethanes and silicone
compounds. A flexible urethane
suitable for electrical use and through which wireless signals of the
frequencies to be used can be
transmitted is preferred.
[0029] A sensor 20 configured for monitoring the desired parameter or
condition may be mounted
within the housing 14 or may be external to the mote and be electrically
connected thereto. Figures 1
6
CA 02884858 2016-05-31
H8323257CA
and 2 show a temperature sensor 20a, which is thermally coupled to heat
transfer element 22 which
extends through an opening 24 in the housing 14, a preferred heat transfer
element 22 being a brass
plug as shown. The thermal transfer element 22 is thermally coupled to the
sensor 20 via a thermal
epoxy. This configuration is preferred for monitoring the surface temperature
of the item to which the
mote is attached since the heat transfer element 22 will contact the surface
on which the mote is
attached. A preferred temperature sensor is a silicon temperature sensor which
is ideal for electronic
circuits. In this embodiment, the mote 10 will be mounted to place the brass
plug 22 in thermal
communication with the portion of the railcar for which a temperature reading
is desired. As one of
ordinary skill would recognize, the configuration of the motes 10 with respect
to the sensor 20 is
dependent upon the type of sensor and the type of data desired. Sensor 20 can
be any type of sensor,
including for example, a temperature sensor, a pressure sensor, a load cell, a
strain gauge, a hall effect
sensor, a vibration sensor, an accelerometer, a gyroscope, a displacement
sensor, an inductive sensor,
a piezio resistive microphone or an ultrasonic sensor. In addition, the sensor
may be a type of switch,
including, for example, reed switches and limit switches. An example of
another type of mote sensor
which uses a strain gauge, e.g. a hand brake monitor sensor, is described in
U.S. patent publication
2012/0046811 (US Patent Application 12/861,713 filed August 23, 2010).
100301 Electrical circuitry 26 is provided for the operation of the mote 10.
The electrical circuitry 26
includes the components and wiring to operate and/or receive and process the
signals from the sensor 10.
This can include, but is not limited to, analog and digital circuitry, CPUs,
processors, circuit boards,
memory, firmware, controllers, and other electrical items, as required to
operate the temperature sensor
and process the information as further described below. In the illustrated
embodiment, the circuitry 26 is
in electrical communication with the temperature sensor for receiving signals
therefrom. Two circuit
boards are provided connected to one another via a header, as further
discussed below.
7
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
[0031] The circuitry 26 includes a main board 28 which includes the
communications circuitry,
antennae and microprocessor and a daughter board 30 including the circuitry to
read the data from the
sensor 10 and may perform analog to digital conversion of the data and also
may include power
conditioning circuitry. Main board 28 may also include intelligence sufficient
to perform low-level
analysis of the data, and may accept parameters from outside sources regarding
when alarms should be
raised. For example, for the mote 10 shown in Figures 1 and 2, with a
temperature sensor 20, may be
programmed to raise an alarm when the sensed temperature exceeds a certain
threshold for several
consecutive readings.
[0032] The main board 28 also includes circuitry for wireless communications.
Preferably, each mote
on a railcar is formed into an ad-hoc mesh network with other motes 10 on the
same railcar and with
a Communication Management Unit (CMU) 32, also preferably mounted on the same
railcar 38 (see
Fig. 6). In the preferred embodiment, each mote 10 would transfer its data to
the CMU 32 mounted on
the same railcar. This transfer of data may occur directly or the data may be
relayed by other motes in
the mesh network to the CMU 32. The ad-hoc mesh network is preferably formed
using the Time
Synchronized Mesh Protocol, a communications protocol for self organizing
networks of wireless
devices.
[0033] Mote 10 also includes a long-term power source 34, preferably a
military grade lithium-thionyl
chloride battery. Daughter board 30 includes power conditioning and management
circuitry and may
include a feature to conserve battery life which keeps mote 10 in a standby
state and periodically wakes
mote 10 to deliver readings from sensor 20.
[0034] The individual motes 10 are mounted on the areas of interest on a
railcar 38. As an example,
Figures 3 and 3A show a temperature sensing mote 10 of the type described
above mounted to a bearing
adapter 12 of a railcar wheel bearing 39 of a railcar. The unit may be
attached using a thermally
conductive epoxy adhesive between the brass plug 22 and the adapter 36 to
ensure good heat transfer to
8
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
the temperature sensor and mechanical fasteners such as self tapping screws to
hold the mote 10 in
place. In this particular example, motes 10 may be attached to all 8 wheel
bearing adapters 12 of each
railcar wheel 40. In addition, an ambient temperature sensor mote 10 may also
be mounted on an area
of the railcar 38 receiving free airflow. Figure 4 shows a mote 10 mounted on
the body of a railcar 38
for monitoring the ambient temperature. The device electrical circuitry 26
obtains information related to
the temperature being monitored, e.g., a bearing or ambient. Since the
temperature sensor is not in
direct contact with the bearing, but the bearing adapter, the device is
calibrated so that the temperature
reading is indicative of the bearing temperature. This calibration takes into
account the temperature of
the bearing adapter at the point measured (see Fig 3a), the ambient
temperature as measured by the
second temperature sensor mote mounted elsewhere on the rail car so as to
sense ambient temperature
(see Fig 4), and information about the type of bearing adapter (different
models have different sizes and
configurations). Calibration information for use by the circuitry 26, such as
a calibration algorithm, can
be developed with suitable testing. Bearing temperature information that
constitute various alarm states
can also be provided to the circuitry 26, preferably stored thereon, allowing
the device 10 to monitor the
bearing temperature and, based on the ambient temperature, determine the
bearing temperature and take
action as desired. On a typical railcar 38, there will be eight sensors 10,
one on each bearing adapter
12 (at each wheel 40); and one sensor 10 placed to measure ambient
temperature. The ambient
temperature sensor 10 will communicate the ambient temperature to the CMU,
which provides this
information to the sensors at the bearing adapters as they call for the
information. This will allow the
sensors 10 at the bearing adapter 12 to determine the bearing temperature and
then determine if further
action is required, such as communicating an alarm of high temperature. Again,
time Synchronized
Mesh Protocol, a communications protocol for self-organizing networks of
wireless devices, is
preferred to communications between the devices 10 and the CMU 32.
[0035] An alternative temperature sensor mote 10 is illustrated with reference
to Figures 5, 5A and 5B
9
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
showing a mote 10 for monitoring the temperature of a part of the railcar
spaced from the housing 14 of
the mote 10. For example, it may be desirable to monitor the temperature of a
commodity stored in a
vessel or the temperature inside of a container by monitoring the temperature
of the vessel or container
wall, but a mote 10 as described above for attaching directly to the surface
to be monitored is not
suitable. Here the mote 10 is constructed the same as that described above but
with the temperature
sensor 20 positioned in a probe 48 spaced from the housing 14. The temperature
sensor 20 sits within a
sensor probe head 50 electrically attached to the housing 14 via wires 52. The
head 50 is preferably
cylindrical having a ring shaped head body 54 made of a thermally conductive
material such as stainless
steel. Fitting within a head body opening 56 is the temperature sensor 20
electrically attached to an
electrical connector 58a. A ring shaped magnet 60 with an opening in its
center fits around the sensor
20 within the head 50. Thermally conductive epoxy or other suitable potting
material seals the magnet
60 and temperature sensor 20 within the body 50 and ensures good heat transfer
to the sensor 20. A
corresponding electrical connector 58b connected to corresponding connector
58a connects temperature
sensor 20 via wire 52, preferably a flexible PVC coated cable 62, to the mote
housing 14. To aid in the
mounting of this housing 14, a silicon gasket 64, metal mounting plate 66, and
housing silicon
mounting gasket 68 are provided. The mote 10 continuously monitors the
commodity temperature and
alerts when specific criteria are met such as rapid temperature change,
temperature thresholds exceeded,
etc.
[0036] As an example of such a device and a method of installing the device,
shown in Fig 5B
representing a liquid holding tank 70 of a rail tank car 38, the temperature
mote 10 with the spaced
probe 48 can be used to monitor the temperature of a liquid 72 stored within
the tank 72 on a rail tank
car 38 that is covered with a fairly thick insulation jacket 74 having an
outside cover 80. Here, a small
opening 76, about 2" diameter, can be cut into the jacket 74 allowing the
probe head 50 to be attached
to the cleaned outside surface 78 of the vessel 70 with thermally conductive
epoxy. The magnet 60
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
within the head 50 helps hold the probe 48 in place while the epoxy cures. The
housing 14 of the mote
is mounted on the outside of the jacket cover 80, connected to the head 50
vial the cable 62 which
can be coiled if necessary to fit within the opening 76. To attach the housing
14 to the jacket cover 80,
the silicone gasket 64 is placed against the jacket cover 80 over the opening
76, followed by the metal
mounting plate 66, followed by silicone gasket 68, the gaskets 64 and 68
ensuring a good weather seal
of the housing 14 to the jacket cover 80. The metal mounting plate 60 is
screwed or bolted to the jacket
cover 80, and the mote housing 14 is screwed to the mounting plate 66 with
screws 82. Calibration of
the mote circuitry ensures an accurate temperature reading of the liquid 72
within the vessel 70, and
ambient temperature, monitored from another sensor 10, can be used to
calculate accurate temperature
readings.
[0037] To communicate data collected by the motes 10, each mote is in two-way
communication with
a CMU 32 mounted on the railcar 38, which collects data from each mote and can
also send instructions
to the motes, as shown in Fig. 6. As previously stated, CMU 32 and each of the
motes 10 connected to
the same railcar 38 form a local area ad-hoc mesh network, to facilitate
communications therebetween.
The term "Communication Management Unit" or "CMU" as used herein means any
device capable of
receiving data and/or alarms from one or more motes 10 and capable of
transmitting data and alarm
information to a remote receiver. CMU 32 is preferably a single unit that
would serve as a
communications link to other locations, such as a mobile base station 42
(e.g., the locomotive 46), a
land-based base station 44, etc, and have its own circuitry, CPUs, processors,
memory, power source,
etc. to process the data received. In a preferred embodiment, it can also
communicate with, control and
monitor motes 10 on the railcar 38.
[0038] CMU 32 is capable of performing advanced data analysis, using data
collected from multiple
motes 10, and may apply heuristics to draw conclusions based on the analysis.
The chart below contains
examples of the types of mote sensors 10 and high level descriptions of the
heuristics applied to analyse
11
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
the data.
[0039]
Parameter Sensor Type Output Heuristic
Sensed
Adapter Temp. Temperature Sensor Bearing Temp. Adapter temperature is
correlated to bearing cup
temperature using empirical
data.
Hatch Position Reed Switch Hatch open/close Determine open/closed
state dependent upon state
of switch.
Pressure Pressure Transducer Brake pressure The pressure
transducer is
fitted directly to the
trainline for measuring
pressure.
Handbrake Link Strain Gauge Handbrake On/Off Handbrake link strain is
Strain correlated to the ON/OFF
status of the handbrake.
Bolster Hall Effect Sensor Car Load Bolster/side frame
Displacement displacement is measured
and spring stiffness data is
used to convert
displacement to load.
Bolster position Reed Switch Car The relative
position of
Empty/Full bolster/sideframe is
measured. The LOADED
position is determined
using empirical data or
spring stiffness.
Inner Jacket Temp. External Temperature Tank Car Inner jacket
surface
Sensor Commodity temperature on a tank car
Temp. is determined and
commodity temperature
can be estimated using
theoretical
12
CA 02884858 2014-12-15
WO 2015/025188
PCT/IB2013/003267
conduction/convection
laws.
Bolster Position Limit Switch Car Empty/Full A limit switch is
mounted
to the sideframe and
activated when the
bolster/sideframe position
is in the loaded state.
Sill Acc. Accelerometer Coupler Force Impact data is
collected.
Using empirical data, a
modal influence matrix
can be computed for
different coupler types that
relates the impact data to
the output. Using an FFT
on the sampled data, and
multiplying by the inverse
of the modal matrix yields
the input in the frequency
domain. This input can be
converted to the time
domain to yield the
coupler force.
Adapter Acc. Accelerometer Bearing Fault An adapter mounted
Indicator accelerometer can be
used
to sample dynamic bearing
data. An FFT can be used
on data sets and plotted
over time to isolate
dominant modes and any
shifting or relative
amplification.
Amplification at rolling
frequency indicates a
likely fault.
Radial Axle Acc. Accelerometer Vehicle Speed An axle mounted
accelerometer can be used
to measure radial
acceleration. The radial
acceleration can be
converted to vehicle speed
using simple dynamics
using the wheel and axel
diameters.
13
CA 02884858 2014-12-15
WO 2015/025188
PCT/IB2013/003267
Adapter Ace. Accelerometer Bearing Fault An adapter mounted
accelerometer can be used
to sample dynamic bearing
data. Kurtosis can be
computed as an indicator
of bearing damage.
Kurtosis is measured in
the time domain and
requires computation of a
probability density
function.
Adapter Acoustics Piezio-electric sensor, Bearing Fault
Sampled acoustic data can
microphone, and be used for either an
accelerometer acoustic noise response
or
Acoustic Emission which
is ring-down counts and
amplitude. Empirical data
from defective bearings is
needed.
Temp. Temperature sensor
Commodity/Fluid A temperature sensor can
Pressure be used to measure
surface
temperature of a pressure
vessel (Tubing, tank, etc.).
Heat conduction equations
can be used to convert the
surface temperature to
fluid temperature. Using
published data for the
working fluid, the
temperature can be
converted to pressure.
Displ. Displacement Sensor Coupler Force Coupler displacement
is
measured and correlated to
force using force-closure
curves.
Axle RPM Inductive Type Sensor Vehicle Speed An inductive
proximity
sensor facing the axle can
generate a signal in
response to an exciter ring
on the axle, and converted
to vehicle speed using
wheel and axle diameters.
14
CA 02884858 2014-12-15
WO 2015/025188
PCT/IB2013/003267
Adapter Ace. Accelerometer Track Damage Sensor is mounted on an
Detection adapter or other truck
component to sample
dynamic data. A Probability
Density Function and
Kurtosis can be computed
from the data. High
Kurtosis, or impulsivity, will
indicate track defects. A
transfer function relating the
wheel input to the adapter is
needed, and can be
determined empirically or
by creating a theoretical
model.
Adapter Ace. Accelerometer Truck Hunting Sensor can be mounted
on
Detection an adapter or other
truck
component to sample
dynamic data. A simple
algorithm could use an
FFT to isolate known
hunting frequencies. More
sophisticated algorithms
could detect flange
impacts using time-series
data.
Wheel Temp. IR Temperature Sensor Wheel Tread Temp Wheel temperature is
correlated to tread
temperature using
empirical data.
Proximity Ultrasonic Sensor Empty/Full status An ultrasonic sensor
could
be used to detect the
presence of lading in tank-
cars, box-cars, covered
hoppers, etc.
Strain Load Cell Car Load Load cell on multiple
places of the truck.
Displ. Reed Switch Handbrake On/Off Position of a handbrake
chain is determined and
correlated to On/Off
Status.
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
Bolster Ace. Accelerometer Truck tilt angles Using a 3-axis
accelerometer fixed to a
bolster, the gravitational
field can be used to
measure the respective
roll, pitch, and yaw angles
with respect to fixed-earth
coordinates.
Hatch Ace. Accelerometer Hatch Tilt Accelerometer measures
the relative tilt of hatch
with fixed-earth
coordinates.
[0040] As shown in Figure 6, CMU 32 also communicates data and alarms to a
remote location, either
a mobile base station 42 located on the train, usually in the locomotive, or
to a stationary, land-based
base station 44, or both. Data and/or alarms may be relayed from land-based
base station 44 to the
mobile base station 42. CMU 32 may be in constant wireless or wired
communication with the mobile
base station 42, which may communicate with the land-based base station 44 via
the cellular network,
via satellite, or via any number of other well known means.
[0041] Data collected from motes 10 may be sent to base station 44 for
analysis and further action.
The heuristics shown in the chart above may be performed by either mobile base
station 42 or land-
based base station 44. In addition, either station 42, 44 may utilize
train¨wide heuristics to predict train-
wide failures, or to spot train-wide trends, which a single CMU 32 may be
unable to do with data from
only a single railcar 38.
[0042] When an alert is detected, it is preferably sent to a display unit in
the locomotive 46 or at the
land-based base station 44. Any typical display unit of a type that would be
mounted in a mobile base
station 42, such as in a locomotive, may be used. Communications devices as
known in the art
communicate with base station 44 via satellite, and display units display the
alert to the locomotive
engineers. Incoming alerts may appear on the display and are accompanied by an
audible alarm which
16
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
must be acknowledged and cancelled by the driver. Each type of alert is
accompanied by a
recommended practice that the locomotive driver should take when an alert
appears, based on the needs
of the particular rail network. The action required to be taken by the
locomotive driver varies based on
the severity of the alert. Alerts may also be sent via email or posted to a
web site.
[0043] Setting locomotive alarm thresholds at values that are sub-critical
will likely lead to an excess
of stoppages and delays. As such, alert messages are selected such that only
actionable messages are
sent to the locomotive 46, meaning that only those alarm levels that require
the crew to take action are
typically transmitted to the locomotive crew. In addition, rather than
requiring the train to be stopped
on the mainline, some alerts could be addressed by putting operating
restrictions in place. For example,
speed restrictions can be placed on the operation of the train at tiered alarm
levels so that the train
would be allowed to proceed to a siding or other appropriate stoppage point,
allowing other traffic to
continue on the mainline without inordinate delays or costs. Low level (Level
1 / Stage 1) alerts,
however, can still be monitored at base station 44 to make determinations
about repeat temperature
offenders and/or trending events that would signify an impending problem,
although not imminent.
[0044] As an example of operation, consider the monitoring of wheel bearings.
The goal is to monitor
bearings in motion, therefore, data that is collected while the railcars 38
are motionless does not
contribute to determining the condition of the bearing. To save power and
limit the uninformative
temperature information, data may be suppressed when the railcar was not
moving. As such, data is
only stored in CMU 32 and transmitted to the base station 42, 44 when useful
data is found. Three
conditions that might define "interesting" or "useful" data include:
1. 'Differential Condition' events;
2. 'Above Ambient Condition' events; and
3. 'Node Temp Anomalies'.
17
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
[0045] A Differential Condition event exists when the difference across any
axle is greater than or
equal to a specified variable.
[0046] An Above Ambient Condition event occurs when any bearing temperature
exceeds the value
reported from the Ambient Temperature Node by a specified constant.
[0047] A Node Temp Anomaly occurs when any data channel, bearing or ambient,
does not report
valid temperatures even though other channels are collecting data well past
the corresponding time
period. The delay allows the system a chance to recover from possible
communication errors. CMU 32
will continue to gather and save temperatures from the other bearings, even if
a full data set should have
been gathered and one or more channels are missing data.
[0048] The data suppression is confirmed by seeing all temperature data
converge to ambient (train
has stopped) before logging stops. Divergent temperatures show the bearings
are generating heat again
and the train has started moving.
[0049] Levelled Alarms and Responses These are examples of various levels of
alarms, based on
severity, and the appropriate response:
Stage 1: "Bearing Temp Alarm"
-Alarm to base station only
- Used for trending recommendations and repeat offender identifications
Stage 2: "Axle Differential Alarm": is the condition any bearing that is a
predetermined amount
hotter than its axle mate
-Action: Stop train, check journal of alarming bearing. Look for any sign that
the
bearing is "walking off" the axle, grease is being purged, or the bearing has
been
damaged.
-Proceed at predetermined reduced maximum speed until a stage 2 "alarm clear"
message is received at the locomotive terminal
-If stage 2 "alarm clear" message is received at the locomotive terminal,
proceed as
normal
Stage 3: "Above Ambient Alarm": is when any bearing is at a temperature a
predetermined
amount above ambient
-Action: Stop train, check journal of alarming bearing. Look for any signs
that the
bearing is "walking off" the axle, grease is being purged, or the bearing has
been
damaged.
-Proceed at predetermined reduced maximum speed until stage 3 "alarm clear"
message
is received at the locomotive terminal, at which point speed can be increased
18
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
-If the alarm message does not clear, a choice can be made to remove and
change the
bearing at an appropriate stoppage point; otherwise, reduced speeds are
required to
reduce the chance of a screwed journal or catastrophic bearing failure
-If Stage 2 and Stage 3 "alarm clear" messages are both received at the
locomotive
terminal, proceed as normal
Stage 4: "Critical Alarm": this is the absolute alarm set at a predetermined
bearing temperature
-Action: Stop train, remove bearing
[0050] Advanced Algorithms - Improvements to the alarms can be made based on
statistical models
of bearing temperature behaviour. The following section details examples of
advancements to the
existing data analysis as it pertains to identifying bearings that are on the
"watch list" for
degrading/trending condition.
[0051] Level 2 Algorithms The Level 2 algorithms use temperature data that had
been collected
every minute while the railcar had been moving during a period of days
directly preceding this analysis.
When at least four of the following five criteria are flagged for the same
bearing, an alert may be sent to
the customer to schedule maintenance for that bearing.
Criteria 1 - Peak Analysis: Count the percentage of bearing temperature values
> a
predetermined value
= For each bearing, count the number of temperature readings that occur
above a
predetermined value
= Flag any bearing with a certain percentage of temperature values > a
predetermined
value
Criteria 2 ¨ Above Ambient Analysis: Count the percentage of bearing
temperature values > a
defined value above ambient
= For each bearing, count the number of temperature readings that occur
over a defined
value above ambient
= Flag any bearing with a certain percentage of temperature values > the
defined threshold
Criteria 3 ¨ Deviation from Wagon Average Analysis: Calculate the average
bearing
temperature and standard deviation of each bearing compared to the average of
bearing
temperatures for the rest of the wagon
= Calculate the average temperature over the time span for each bearing and
the wagon
average
= Calculate the standard deviation on each bearing temperature from the
wagon average
= Flag any bearing with a certain standard deviation
Criteria 4 ¨ Deviation from Fleet Average Analysis: Calculate the average
bearing
temperature and standard deviation of each bearing compared to the average of
bearing
19
CA 02884858 2014-12-15
WO 2015/025188 PCT/IB2013/003267
temperatures for the rest of the fleet
= Calculate the average temperature over the time span for each bearing and
the fleet
average
= Calculate the standard deviation on each bearing temperature from the
fleet average
= Flag any bearing with a certain standard deviation
Criteria 5 ¨ Heating Rate of Change Analysis: Calculate the percent of
operating time that a
bearing is heating quickly
= Calculate a linear fit to a moving window of temperature data
= Count the number of instances where the slope of the linear fit is above
a certain
threshold
= Flag any bearing with > a certain amount of the operating time having a
slope of the
linear fit > the threshold
[0052] Level 3 Algorithms The Level 2 algorithms use temperature data that has
been collected every
minute while the railcar is moving for the previous 30 days directly preceding
this analysis. When a
bearing is ranked in the top five percent for at least four of the five
criteria, an alert is sent to the
customer to schedule maintenance for that bearing.
Criteria 1 ¨ Peak Analysis: Count the percentage of bearing temperature values
> a
predetermined value
= For each bearing, count the number of temperature readings that occur
above a
predetermined value
= Rank the bearing in a league table with the rest of the fleet
= Flag the top percent of bearings in the fleet
Criteria 2 ¨ Above Ambient Analysis: Count the percentage of bearing
temperature values > a
defined value above ambient
= For each bearing, count the number of temperature readings that occur
over a defined value
above ambient
= Rank the bearing in a league table with the rest of the fleet
= Flag the top percent of bearings in the fleet
Criteria 3 ¨ Deviation from Wagon Average Analysis: Calculate the average
bearing
temperature and standard deviation of each bearing compared to the average of
bearing
temperatures for the rest of the wagon
= Calculate the average temperature over the time span for each bearing and
the wagon
average
= Calculate the standard deviation of each bearing temperature from the
wagon average
= Rank the bearing in a league table with the rest of the fleet
= Flag the top percent of bearings in the fleet
CA 02884858 2016-05-31
H8323257CA
Criteria 4¨ Deviation from Fleet Average Analysis: Calculate the average
bearing temperature and
standard deviation of each bearing compared to the average of bearing
temperatures for the
rest of the fleet
= Calculate the average temperature over the time span for each bearing and
the fleet average
= Calculate the standard deviation on each bearing temperature from the
fleet average
= Rank the bearing in a league table with the rest of the fleet
= Flag the top percent of bearings in the fleet
Criteria 5 ¨ Heating Rate of Change Analysis: Calculate the percentage of
operating time that a
bearing is heating quickly
= Calculate a linear fit to a moving window of temperature data
= Count the number of instances where the slope of the linear fit is > a
certain threshold
= Rank the bearing in a league table with the rest of the fleet
= Flag the top percent of bearings in the fleet
100531 In another alternate embodiment of the invention, one or more motes may
be housed in
alternative housings or built in to the railcar itself In one such embodiment,
motes can be built into the
form of an adapter pad similar to the type shown in U.S. Patents 7,698,962 and
7,688,218, which could
be adapted for use with the present invention.
100541 Various embodiments of the invention have been described in the context
of various examples,
however, the invention is not meant to be limited in any way. As one of skill
in the art recognizes there
may be many implementations that are within the scope of the invention, as is
described in the following
claims.
21
