Note: Descriptions are shown in the official language in which they were submitted.
CA 02454739 2004-01-05
132250
MULTI-LEVEL RAILWAY OPERATIONS OPTIMIZATION
FIELD OF THE INVENTION
This invention relates to optimizing railway operations, and more particularly
to a
system and method of optimizing railway operations using a multi-level, system-
wide
approach.
BACKGROUND OF THE INVENTION
Railways are complex systems, with each component being interdependent on
other
components within the system. Attempts have been made in the past to optimize
the
operation of a particular component or groups of components of the railway
system.
such as for the locomotive, for a particular operating characteristic such as
fuel
consumption, which is a major component of the cost of operating a railway
system.
Some estimates indicate that fuel consumption is the second largest railway
system
operating cost, second only to labor costs.
For example. U.S. Patent No. 6,144,901 proposes optimizing the operation of a
train
for a number of operating parameters, including fuel consumption. i -iowever,
optimizing the performance of a particular train, which is only one component
of a
much larger system, including, for example, the railway network oftrack, other
trains,
crews, rail yards, departure points, and destination points, may not yield an
overall
system wide optimization. Optimizing the performance of only one component of
the
system (even though it may be an important component such as a train) may
actually
result in increased system-wide costs, because this prior art approach does
not
consider the interrelationships and impacts on other components and on the
overall
railway system efficiency. As one example, optimizing at the train ignores
potential
1
132250 CA 02454739 2004-01-05
efficiencies for a locomotive within the individual train, which efficiencies
may be
available ifthe locomotives were free to optimize their own performance.
One system and method of planning at the railway track network system is
disclosed
in U.S. Patent No. 5,794,172. Movement planners such as this are primarily
focused
on movement of the trains through the network based on business objective
functions
(BOF) defined by the railroad company, and not necessarily on the basis of
optimizing performance or a particular performance parameter such as fuel
consumption. Further, the movement planner does not extend the optimization
down
to the train (much less the consist or locomotive), nor to the railroad
service and
maintenance operations that plan for the servicing of the trains or
locomotives.
Thus, in the prior art, there has been no recognition that optimization of
operations for
a railway system requires a multi-level approach, with the gathering of key
data at
each level and communicating data with other levels in the system.
SUMMARY OF THE INVENTION
One aspect of the present invention is the provision of a multi-level system
for
management of a railway system and its operational components in which the
railway
system comprises a first level configured to optimize an operation within the
first
level that includes first level operational parameters which define
operational
characteristics and data of the first level, and a second level configured to
optimize an
operation within the second level that includes second level operational
parameters
which define the operational characteristic and data of the second level. The
First
level provides the second level with the first level operational parameters,
and the
second level provides the first level with the second level operational
parameters,
such that optimizing the operation within the first level and optimizing the
operation
within the second level are each a function of optimizing a system
optimization
parameter.
A further aspect of the present invention includes the provision of a method
for
optimizing an operation of a railway system having first and second levels
which
comprises communicating from the first level to the second level a first level
operational parameter that defines an operational characteristic of the first
level.
2
13 22) 5 0 CA 02454739 2004-01-05
communicating from the second level to the first level a second level
operational
parameter that defines an operational characteristic of the second level,
optimizing a
system operation across a combination of the first level and the second level
based on
a system optimization parameter, optimizing an operation within the first
level based
on a first level optimization parameter and based in part on the system
optimization
parameter, and optimizing an operation within the second level based on a
second
level optimization parameter and based in part on the system optimization
parameter.
Another aspect of the present invention is the provision of a method and
system for
multi-level railway operations optimization for a complex railroad system that
identities key operating constraints and data at each level, communicates
these
constraints and data to adjacent levels and optimizes performance at each
level based
on the data and constraints of adjacent levels.
Aspects of the present invention further include establishing and
communicating.
updated plans and monitoring and communicating compliance with the plans at
multiple levels of the system.
Aspects of the invention further include optimizing performance at the
railroad
infrastructure level, railway track network level, individual train level
within the
network, consist level within the train, and the individual locomotive level
within the
consist.
Aspects of the invention further include optimizing performance at the
railroad
infrastructure level to enable condition-based, rather than scheduled-based,
servicing
of locomotives, including both temporary (or short-term) servicing
requirements such
as fueling and replenishment of other consumable materials on-board the
locomotive,
and long-term servicing requirements such as replacement and repair of
critical
locomotive operating components, such as traction motors and engines.
Aspects of the invention include optimizing performance of the various levels
in light
of the railroad operating company's business objective functions, such as on-
time
deliveries, asset utilization, minimum fuel usage, reduced emissions,
optimized crew
costs, dwell time, maintenance time and costs, and reduced overall system
costs.
3
CA 02454739 2004-01-05
132250
These Aspects of the invention provide benefits such as reduced journey-to-
journey
fuel usage variability, fuel savings for each locomotive operating within the
system,
graceful recovery of the system from upsets, elimination of out-of-fuel
mission
failures, improved fuel inventory handling logistics and decreased autonomy of
crews
in driving decisions.
BRIEF DESCRIPTION OF THE DRAWINGS
Fig. 1 is a graphical depiction of the multi-level nature of railway
operations
optimization of this invention, with the railroad infrastructure, railroad
track network,
train, locomotive consist and individual locomotive levels being depicted in
their
respective relationships to each other.
Fig. 2 is a graphical depiction of the railroad infrastructure level
illustrating the inputs
and outputs to the infrastructure processor at this level.
Fig. 3 is a schematic illustrating details of optimized servicing operations
at the
infrastructure level.
Fig. 4 is a schematic illustrating details of optimized refueling operations
at the
infrastructure level.
Fig. 5 is a schematic of the railroad track network level illustrating its
relationships
with the railroad infrastructure above it and the train level below it.
Fig. 6 is a schematic illustrating details of the railroad track network
level, with inputs
to and outputs from the processor at this level.
Fig. 7 is a schematic illustrating inputs to and outputs from an existing
movement
planner at the train level.
Fig. 8 is a schematic of a revised railroad network processor having a network
fuel
manager processor for optimization of additional fuel usage parameters.
Fig. 9 is a pair of string-line diagrams, with the first diagram being an
initial
movement plan done without consideration of operational optimization and the
second diagram being a modified plan as optimized for reduced fuel
consumption.
4
CA 02454739 2004-01-05
132250
Fig. 10 is a schematic of the train level illustrating its relationship with
its related
levels.
Fig. I I is a schematic illustrating details of the inputs and outputs of the
train level
processor.
Fig. 12 is a schematic of the consist level illustrating its relationship with
its related
levels.
Fig. 13 is a schematic illustrating details of the inputs and outputs of the
consist level
processor.
Fig. 14 is a graphic illustrating fuel usage as a function of planned time for
various
modes of operation at the consist level.
Fig. 15 is a schematic of the locomotive level illustrating its relationships
with the
consist level.
Fig. 16 is a schematic illustrating details of the inputs and outputs of the
locomotive
level processor.
Fig. 17 is a graphic illustrating fuel usage as a function of planned time of
operation
for various modes of operation at the locomotive level.
Fig. 18 is a graphic illustrating locomotive level fuel efficiency as measured
in fuel
usage per unit of power as a function the amount of power generated at the
locomotive level for various modes of operation.
Fig. 19 is a graphic illustrating various electrical system losses as a
function of DC
link voltage at the locomotive level.
Fig. 20 is a graphic illustrating fuel consumption as a function of engine
speed at the
locomotive level.
Fig. 21 is a schematic of an energy management subsystem of a hybrid energy
locomotive having an on-board energy regeneration and storage capability as
configured and operated for fuel optimization.
CA 02454739 2004-01-05
132250
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
Referring to Fig. 1, the multi-level nature of a railway system 50 is
depicted. As
shown, the system comprises from the highest level to the lowest level: a
railroad
infrastructure level 100, a track network level 200, a train level 300, a
consist level
400 and a locomotive level 500. As described hereinafter, each level has its
own
unique operating characteristics, constraints, key operating parameters and
optimization logic. Moreover, each level interacts in a unique manner with
related
levels, with different data being interchanged at each interface between the
levels so
that the levels can cooperate to optimize the overall railway system 50. The
method
for optimization of the railway system 50 is the same whether considered from
the
locomotive level 500 up, or the railroad infrastructure system 100 down. To
facilitate
understanding, the latter approach, a top down perspective, will be presented.
RAILWAY INFRASTRUCTURE LEVEL
Optimization of the railway system 50 at the railroad infrastructure level 100
is
depicted in Figs. 1-4. As indicated in Fig. 1, the levels of the multi-level
railway
operations system 50 and method include from the top down, the railroad
infrastructure level 100, the track network level 200, the train level 300,
the consist
level 400 and the locomotive level 500. The railroad infrastructure level 100
includes
the lower levels of track network 200, train 300, consist 400 and locomotive
level
500. In addition, the infrastructure level 100 contains other internal
features and
functions that are not shown, such as servicing facilities, service sidings,
fueling
depots, wayside equipment, rail yards, train crews operations, destinations,
loading
equipment (often referred to as pickups), unloading equipment (often referred
to as
set-outs), and access to data that impacts the infrastructure, such as:
railroad
operating rules, weather conditions, rail conditions, business objective
functions
(including costs, such as penalties for delays and damages enroute, and awards
for
timely delivery), natural disasters, and governmental regulatory requirements.
These
are features and functions that are contained at the railroad infrastructure
level 100.
Much of the railroad infrastructure level 100 is of a permanent basis (or at
least of a
longer term basis). Infrastructure components such as the location of wayside
equipment, fueling depots and service facilities are not subject to change
during the
6
CA 02454739 2004-01-05
132250
course of any given train trip. However, real-time availability of these
components
may vary depending on availability, time of day, and use by other systems.
These
features of the railroad infrastructure level 100 act as opportunities or
resources and
constraints on the operation of the railway system 50 at the other levels.
However,
other aspects of the railroad infrastructure level 100 are operable to serve
other levels
of the railway system 50 such as track networks, trains, consists or
locomotives, each
of which may he optimized as a function of a multilevel optimization criteria
such as
total fuel, refueling, emissions output, resource management, etc.
Fig. 2 provides a schematic of the optimization of the railroad infrastructure
level 100.
It illustrates the infrastructure level 100 and the infrastructure level
processor 202
interacting with track level 200 and train level 300 to receive input data
from these
levels, as well as from within the railroad infrastructure level 100 itself,
to generate
commands to and/or provide data to the track network level 200 and the train
level
300, and to optimize operation within the railroad infrastructure level 100.
As illustrated in Fig. 3, infrastructure processor 202 may be a computer,
including
memory 302, computer instructions 304 including an optimization algorithms,
etc.
The infrastructure level 100 includes, for example, the servicing of trains
and
locomotives such as at maintenance facilities and service sidings to optimize
these
servicing operations, the infrastructure level 100 receives infrastructure
data 206 such
as facility location, facility capabilities (both static characteristics such
as the number
of service bays, as well as dynamic characteristics, such as the availability
of bays,
service crews, and spare parts inventory), facility costs (such as hourly
rates,
downtime requirements), and the earlier noted data such as weather conditions,
natural disaster and business objective functions. The infrastructure level
also
receives track network level data 208, such as the current train system
schedule for
the planned arrival and departure of' railroad equipment at the service
facility, the
availability of substitute power (i.e., replacement locomotives) at the
facility and
scheduled service. In addition, the infrastructure level receives train level
data 210,
such as the current capability of trains on the systems, particularly those
with health
issues that may require additional condition-based (as opposed to scheduled-
based)
servicing, the current location, speed and heading of trains, and the
anticipated
servicing requirements when the train arrives. The infrastructure processor
202
7
CA 02454739 2004-01-05
132250
analyzes this input data and optimizes the railroad infrastructure level 100
operation
by issuing work orders or other instructions to the service facilities for the
particular
trains to be serviced, as indicated in block 226, which includes instructions
for
preparing for the work to be done such as scheduling work hays, work crews.
tools,
and ordering spare parts. The infrastructure level 100 also provides
instructions that
are used by the lower level systems. For example, track commands 228 are
issued to
provide data to revise the train movement plan in view of a service plan,
advise the
rail yard of the service plan such as reconfiguring the train, and provide
substitute
power of a replacement locomotive. Train commands 230 are issued to the train
level
300 so that particular trains that are to be serviced may have restricted
operation or to
provide on-site servicing instructions that are a function of the service
plan.
As one example of the operations of the infrastructure level 100, Fig. 4 shows
an
infrastructure level optimized refueling 400. This is a particular instance of
optimized
servicing at the infrastructure level 100. The infrastructure data 406 input
to the
infrastructure level 400 for optimizing refueling are related to fueling
parameters.
These include refueling site locations (which include the large service
facilities as
well as fuel depots, and even sidings at which fuel trucks can be dispatched)
and total
fuel costs, which includes not only the direct price per gallon of the fuel,
but also asset
and crew downtime, inventory carrying costs, taxes, overhead and environmental
requirements. Track network level input data 408 includes the cost of changing
the
train schedule on the overall movement plan to accommodate refueling or
reduced
speeds if fueling is not done, as well as the topography of the track ahead of
the trains
since it has a major impact on fuel usage. Train level input data 410 includes
current
location and speed, fuel level and fuel usage rate data (which can be used to
determine
locomotive range of travel) as well as consist configuration so that
alternative
locomotive power generation modes can he considered. Train schedule as well as
train weight, freight and length are relevant to the anticipated fuel usage
rate. Outputs
from the optimum refueling infrastructure level 400 include optimization of
the
fueling site both in terms of the fueling instructions for each particular
train but also
as anticipated over some period of time for fuel inventory purposes. Other
outputs
include command data 428 to the track network level 200 to revise the movement
plan, and train level commands 430 for fueling instructions at the facility
site,
8
CA 02454739 2004-01-05
132250
including schedules, as well as operational limitations on the train such as
the
maximum rate of fuel usage while the train is enroute to the fuel location.
Optimization of the railroad infrastructure operation is not a static process,
but rather
is a dynamic process that is subject to revision at regular scheduled
intervals (such as
every 30 minutes) or as significant events occur and are reported to the
infrastructure
level 100 (such as train brake downs and service facility problems).
Communication
within the infrastructure level 100 and with the other levels may be done on a
real-
time or near real-time basis to enable the flow of key information necessary
to keep
the service plans current and distributed to the other levels. Additionally,
information
may be stored for later analysis of trends or the identification or analysis
of particular
level characteristics, performance, interactions with other levels or the
identification
of particular equipment problems.
RAILROAD TRACK NETWORK LEVEL
Within the operational plans of the railroad infrastructure, optimization of
the railroad
track network level 200 is performed as depicted in Figs. 5 and 6. The
railroad track
network level 200 includes not only the track layout, but also plans for
movement of
the various trains over the track layout. Fig. 5 shows the interaction of the
track
network level 200 with the railroad infrastructure level 100 above it and the
individual
trains below it. As illustrated, the track network level 200 receives input
data from
the infrastructure level 100 and the train level 300, as well as data (or
feedback) from
within the railroad network level 200. As Illustrated in Fig. 6, track network
processor 502 may be a computer, including memory 602, computer instructions
604
including an optimization algorithms, etc. As shown in Fig. 6, the
infrastructure level
data 506 includes information regarding the condition of the weather, rail
yard.
substitute power, servicing facilities and plans, origins and destinations.
Track
network data 508 includes information regarding the existing train movement
schedule, business object functions and network constraints (such as
limitations on the
operation of certain sections of the track). Train level input data 510
includes
information regarding locomotive location and speed, current capability
(health),
required servicing, operating limitations, consist configurations, trainload
and length.
9
CA 02454739 2004-01-05
132250
Fig. 6 also shows the output of the track network level 200 that includes data
526 sent
to the infrastructure level, commands 530 to the trains and optimization
instructions
528 to the track network level 200 itself. The data 526 sent to the
infrastructure level
100 includes wayside equipment requirements, rail yard demands, servicing
facility
needs, and anticipated origin and destination activities. The train commands
530
include the schedule for each train and operational limitations enroute, and
the track
network optimization 528 includes revising the train system schedule.
As with the infrastructure level 100, the railroad track network 200 schedule
(or
movement plan) is revised at periodic intervals or as material events occur.
Communication of the input and output of critical data and command may be done
on
a real-time basis to keep the respective plans current.
An example of an existing movement planner is disclosed in U.S. Patent No.
5,794,172. Such a system includes a prior art computer aided dispatch (CAD)
system
having a power dispatching system movement planner for establishing a detailed
movement plan for each locomotive and communicating to the locomotive. More
particularly, such a movement planner plans the movement of trains over a
track
network with a defined planning horizon such as 8 hours. The movement planner
attempts to optimize a railroad track network level Business Objective
Function
(BOF) that is the sum of the BOF's for individual trains in the train levels
of the
railroad track network level, The BOF for each train is related to the
termination
point for the train. It may also be tied to any point in the individual
train's trip. In the
prior art, each train had a single BOF for each planning cycle in a planning
territory.
Additionally, each track network system may have a discrete number of planning
territories. For example, a track network system may have 7 planning
territories. As
such, a train that will traverse N territories will have N BOF's at any
instance in time.
The BOF provides a means of comparing the quality of two movement plans.
In the course of computing each train's movement plan each hour, the movement
planner compares thousands of alternative plans. The track network level
problem is
highly constrained by the physical layout of track, track or train operating
restrictions,
the capabilities of trains, and conflicting requirements for the resources.
The time
required to compute a movement plan in order to support the dynamic nature of
CA 02454739 2004-01-05
132250
railroad operations is a major constraint. For this reason, train performance
data is
assumed, based on pre-computed and stored data based upon train consist, track
conditions, and train schedule. The procedure used by the movement planner
computes the minimum run time for a train's schedule by simulating the train's
unopposed movement over the track, with stops and dwells for work activities.
This
process captures the run time across each track segment and alternate track
segment in
the train's path. A planning cushion, such as a percentage of run time, is
then added
to the train's predicted run time and the cushioned time is used to generate
the
movement plan.
One such prior art movement planner is illustrated in Fig. 20, where the train
(and
thus the train level, consist level, locomotive level/engine) is at an optimum
speed S,
along the speed/fuel consumption curve 2002 resulting in reduced fuel
consumption at
the bottom 2004 of curve 2002. Typical train speeds exceed the optimum train
speed
F1, so that reducing average train speeds usually results in reduced fuel
consumption.
Figs. 7 and 8 illustrate details of an embodiment of the invention and its
benefits to
movement planning of the track network level 200. Fig. 7 illustrates an
example of a
movement planner 700 to analyze operating parameters to optimize the train
movement plan for optimizing fuel usage. The movement planner 702 receives
input
from the train level 300. The Fig. 7 embodiment of the movement planner 702
receives and analyzes messages to the movement planner 702 from external
sources
712 with respect to refueling points and the Business Objective Functions
(BOF) 710
including a planning cushion as mentioned above. A communication link 706 to
the
fuel optimizers 704 on trains in the train levels 300 is provided in order to
transmit the
latest movement plan to each of the trains on the train level 300. In the
prior art, the
movement planner attempted to minimize delays for meets and passes. In
contrast,
the system according to one embodiment of the present invention utilizes these
delays
as an opportunity for fuel optimization at the various levels.
Fig. 8 illustrates a movement planner for analyzing additional operating
parameters
beyond those illustrated in Fig. 7 for optimizing fuel optimization. The
network fuel
manager 802 provides the track network level 200 with functionality to
optimize fuel
usage within the track network level 200 based on the Business Objective
Function
11
CA 02454739 2004-01-05
132250
(BOF) 810 of each of the trains at the train level 300, the engine performance
812 of
the trains and locomotives comprising those trains, congestion data 804 and
fuel
weighting factors 808. The movement planner at the track network level
receives
input 708 from the train level optimizer 704 and from the network fuel manager
802.
For example, the train level 200 provides the movement planner 702 with engine
failure and horsepower reduction data 708. The movement planner 702 provides a
movement plan 706 to the train level 200 and congestion data 804 to the
network fuel
manager 802. The train level 200 provides engine performance data 812 to the
network fuel manager 802. The movement planner 702 at the track network level
200
utilizes the Business Objective Function (BOF) for each train, the planning
cushion
and refueling points 806 and the engine failure and horsepower reduction data
708, to
develop and modify the movement plan for a particular train at the train level
200.
As mentioned above, the Fig. 8 embodiment of the movement planner 702
incorporates a network fuel manager module 802 or fuel optimizer that monitors
the
performance data for individual trains and provides inputs to the movement
planner to
incorporate fuel optimization information into the movement plan. This module
802
determines refueling locations based upon estimated fuel usage and fuel costs
as well.
A fuel cost weighting factor represents the parametric balancing of fuel costs
(both
direct and indirect) against schedule compliance. This balance is considered
in
conjunction with the congestion anticipated in the path of the train. Slowing
a train
for train level fuel optimization can increase congestion at the track network
level by
delaying other trains especially in highly trafficked areas. The network fuel
manager
module 802 interfaces to the movement planner 702 within the track network
level
200 to set the planning cushion (amount of slack time in the plan before
appreciably
affecting other train movements) for each train and modifies the movement plan
706
to allow individual train planning cushions to be set, with longer planning
cushions
and shorter meets and passes than typical to provide for improved fuel
optimization.
A further enhancement specifics a higher planning cushion for trains that are
equipped
with a fuel optimizer 704 and whose schedules are not critical. This provides
savings
to local trains and trains running on lightly trafficked rail. This involves
an interface
to the movement planner 702 to set the planning cushion for the train and a
12
CA 02454739 2004-01-05
132250
modification to the movement plan 706 to allow the planning cushion to be set
for
individual trains.
Fig. 9 illustrates a representative set of string line graphs for the planned
movement
(movement plan 706) of two trains (i.e., trains A and B) moving in opposite
directions
on a single track, thereby requiring that the trains meet and pass at a siding
906. The
string line shows the train location as a function of travel time for the
trains, with line
A illustrating the travel of train A as it moves from its initial location 902
near the top
of the chart to its final location 904 near the bottom of the chart, and the
travel of train
B from its initial location 908 at the bottom of the chart to its final
location 910 at the
top of the chart. The "original plan" 900 as shown in the First string line of
Fig. 9 is
generated solely for the purpose of minimizing the time required to effect the
train
movements. This string line shows that train A enters a siding 906 represented
by the
horizontal line segment 906 at time ti, so as to let train B pass. Train A is
stopped and
idle at siding 906 from t1 to t2. Train B, as shown by line 908-910, maintains
a
constant speed from 908 to 910. The upper curved line 909 and curved dotted
line
extension 911 represents the fastest move that train A is capable of
performing. The
"modified plan" 950 as shown in the string line on the right of Fig. 9 was
generated
with consideration for fuel optimization. It requires that train A travel
faster (steeper
slope of line 918-912 from ti to t4) so as to reach a second and more distant
siding
912, albeit at a somewhat later time t4, e.g., t4 is later than ti. The
modified plan also
requires that train B slow its rate of travel at time t., so as to pass at the
second siding
912. The modified plan reduces the idle time of train A to is - t:, from the
previous t, -
ti and reduces the speed of train B beginning at t3 to create the opportunity
for fuel
optimization at the train level 300 as reflected by the combination of the two
particular trains, while maintaining the track network level movement plan at
or near
its earlier level of performance.
Inputs to the track network level movement planner 702 also includes locations
of
fuel depots, cost of fuel ($/gallon per depot and cost of time to fuel or so-
called "cost
penalty"), engine efficiency as represented by the slope of the change in the
fuel use
over the change in the horsepower (e.g., slope of fuel use/ HP), fuel
efficiency as
represented by the slope of the change in the fuel use over the change in
speed or
time, derating of power for locomotives with low or no fuel, track adhesion
factors
13
CA 02454739 2004-01-05
132250
(snow, rain, sanders, cleaners, lubricants), fuel level for locomotives in
trains, and
projected range for fuel ofthe train.
The railroad track network level functionality established by the movement
planner
702 includes determination of required consist power as a function of speed
under
current or projected operating conditions, and determination of fuel
consumption as a
function of power, locomotive type, and network track. The movement planner
702
determinations may be for locomotives, for the consist or the train which
would
include the assigned load. The determination may be a function of the
sensitivity of
the change of fuel over the change of power (A Fuel / A HP) and/or change in
horsepower over speed (AHP/ A Speed). The movement planner 702 further
determines the dynamic compensation to fuel-rate (as provided above) to
account for
thermal transients (tunnels, etc.), and adhesion limitations, such as low
speed tractive
effort or grade, that may impair movement predictions, e.g., the expected
speed. The
movement planner 702 may predict the current out-of-fuel range based on an
operating assumption such as that the power continues at the current level or
an
assumption regarding the future track. Finally, the detection of parameters
that have
changed significantly may be communicated to the movement planner 702, and as
a
result, an action such as a change in the movement plan may be required. These
actions may be automatic functions that are communicated continuously,
periodically,
or done on exception basis such as for detection of transients or predicted
out-of-fuel
conditions.
The benefits of this operation of the track network level 200 includes
allowing the
movement planner 702 to consider fuel use in optimizing the movement plan
without
regard to details at the consist level, to predict fuel-rate as a function of
power and
speed, and by integration, to determine the expected total fuel required for
the
movement plan. Additionally, the movement planner 702 may predict the rate of
schedule deterioration and make corrective adjustments to the movement plan if
needed. This may include delaying the dispatch of trains from a yard or
rerouting
trains in order to relieve congestion on the main line. The track network
level 200
also will enable the factoring of the dynamic consist fuel state into
refueling
determination at the earliest opportunity, including the consideration of
power loss,
14
CA 02454739 2004-01-05
132250
such as when one locomotive within a consist shuts down or is forced to
operate at
reduced power. The track network level 200 will also enable the determination
(at the
locomotive level or consist level) of optimum updates to the movement plan.
This
added optimization data reduces the monitoring and signal processing required
in the
movement plan or computer aided dispatch processes.
The movement plan output from the track network level 200 specifies where and
when to stop for fuel, amount of fuel to take on, lower and upper speed limits
for
train, time/speed at destination, and time allotted for fueling.
TRAIN LEVEL
Figs. 10 and 11 depict the train level operation and relationships between the
train
level 300 and the other levels. The train processor 1002 may include a memory
1 102
and computer instructions 1104 including an optimization algorithm, etc. While
the
train level 300 may comprise a long train with distributed consists, each
consist with
several locomotives and with numerous cars between the consists, the train
level 300
may be of any configuration including more complex or significantly simpler
configurations. For example, the train may be formed by a single locomotive
consist
or a single consist with multiple locomotives at the head of the train both of
which
configurations simplify the levels, interactions and amount of data
communicated
from the train level 300 to the consist level 400 and on to the locomotive
level 500.
In the simplest case, a single locomotive without any cars may constitute a
train. In
this case, the train level 300, consist level 400 and locomotive level 500 are
the same.
In such as case, the train level processor, the consist level processor and
the
locomotive level processor may be comprised of one, two or three processors.
Assuming for discussion purposes a more complex train configuration, then the
input
data at the train level 300, as shown in Fig. 10 and 11, includes
infrastructure data
1006, railway track network data 1008, train data 1010, including feedback
from the
train, and consist level data 1012. The output of the train level includes
data sent to
the infrastructure level 1026 and to the track network level 1028,
optimization within
the train level 1030 and commands to the consist level 1032. The railroad
infrastructure level input data 1006 includes weather conditions, wayside
equipment,
CA 02454739 2004-01-05
132250
servicing facilities and origin/destination information. The track network
level data
input 1008 includes train system schedule, network constraints and track
topography.
The train data input 1010 includes load, length, current capacity for braking
and
power, train health, and train operating constraints. Consist data input 1012
includes
the number and locations of the consists within the train, the number of
locomotives
in the consist and the capability for distributed power control within the
consist.
Inputs to the train level 300 from sources other than the locomotive consist
level 400
include the following: head end and end-of-train (EOT) locations, anticipate
up-
coming track topography and wayside equipment, movement plan, weather (wind.
wet, snow), and adhesion (friction) management.
The inputs to the train level 300 from the consist level 400 is typically the
aggregation
of information obtained from the locomotives and potentially from the load
cars.
These include current operating conditions, current equipment status,
equipment
capability, fuel status, consumable status, consist health, optimization
information for
the current plan, optimization information for the plan optimization.
The current operating conditions of the consist may include the present total
tractive
effort (TE), dynamic braking effort, air brake effort, total power, speed, and
fuel
consumption rate. These may obtained by consolidating all the information from
the
consists at the consist level 400, which include the locomotives at the
locomotive
level 500 within the consist, and other equipment in the consist. The current
equipment status includes the ratings of locomotives, the position of the
locomotives
and loads within the consist. The ratings of units may be obtained from each
consist
level 400 and each locomotive level 500 including derations due to
adhesioniambient
conditions. This may be obtained from the consist level 400 or directly from
the
locomotive level 500. The position of the locomotives may he determined in
part by
trainline information, GPS position sensing, and air brake pressure sensing
time delay.
The load may he determined by the tractive effort (TE), braking effort (BE),
speed
and track profile.
Equipment capability may include the ratings of the locomotives in the consist
including the maximum tractive effort (TE,,,,x), maximum braking effort
(BE,,,,.),
Horsepower (HP), dynamic brake HP, and adhesion capability. The fuel status,
such
16
CA 02454739 2004-01-05
132250
as the current and projected amount of fuel in each locomotive, is calculated
by each
locomotive based on the current fuel level and projected fuel consumption for
the
operating plan. The consist level 400 aggregates this per-locomotive
information and
sends the total range and possibly fuel levels/status at known fueling points.
It may
also send the information where the item may become critical. For example, one
locomotive within a consist may run out of fuel and yet the train may run to
the next
fueling station, if there is enough power available on the consist to get to
that point.
Similarly, the status of other consumables other than fuel like sand, friction
modifiers,
etc. are reported and aggregated at the consist level 400. These are also
calculated
based on current level and projected consumption based on weather, track
conditions,
the load and current plan. The train level aggregates this information and
sends the
total range and possibly consumable levels/status at known servicing points.
It may
also send the information where the item may become critical. For example, if
adhesion limited operation requiring sand is not expected during the
operation, it may
not be critical that sanding equipment be serviced.
The health of the consist may be reported and may include failure information,
degraded performance and maintenance requirements. The optimization
information
for the current plan may be reported. For example, this may include fuel
optimization
at the consist level 400 or locomotive level 500. For fuel optimization, as
shown in
Fig. 14, data and information for consist level fuel optimization is
represented by the
slope and shape of the line between operating points 1408 and 1410.
Furthermore,
optimization information for the plan optimization may include the data and
information as depicted between operating points 1408 and 1412, as shown in
Fig. 14,
for the consist level 400.
Also as shown in Fig. 11, the output data 1026 sent by the train level 300 to
the
infrastructure level 100 includes information regarding the location, heading
and
speed of the train, the health of the train, operational derating of the train
performance
in light of the health conditions, and servicing needs, both short-term needs
such as
related to consumables and long-terns needs such as system or equipment repair
requirements. The data 1028 sent from the train level 300 to the railroad
track
network level 200 includes train location, heading and speed, fuel levels,
range and
usage and train capabilities such as power, dynamic braking, and friction
17
CA 02454739 2004-01-05
132250
management. Optimizing performance within the train level 300 includes
distributing
power to the consists within the train level, distributing dynamic braking
loads to the
consists levels within the train level and pneumatic braking to the cars
within the train
level, and wheel adhesion of the consists and railroad cars. The output
commands to
the consist level 400 includes engine speed and power generation, dynamic
braking
and wheel/rail adhesion for each consist. Output commands from the train level
300
to the consist level 400 include power for each consist, dynamic braking,
pneumatic
braking for consist overall, tractive effort (TE) overall, track adhesion
management
such as application of sand/lubricant, engine cooling plan, and hybrid engine
plan.
An example of such a hybrid engine plan is depicted in greater detail in Fig.
21.
CONSIST LEVEL
Figs. 12 and 13 illustrate the consist level relationships and exchange of
data with
other levels. The consist level processor 1202 includes a memory 1302 and
processor
instructions 1304 which includes optimization algorithms, etc. As shown in
Fig. 12,
the inputs to the consist level, as depicted in the consist level 400 with
optimization
algorithms, include data 1210 from the train level 300, data 1214 from the
locomotive
level 500 and data 1212 from the consist level 400. The outputs include data
1230 to
the train level 300, commands 1234 to the locomotive level 500, and
optimization
1232 within the consist level 400,
As an input, the train level 300 provides data 1210 associated with train
load, train
length, current train capability, operating constraints, and data from the one
or more
consists within the train level 300. Information 1210 sent from the locomotive
level
500 to the consist level 400 may include current operating conditions and
current
equipment status. Current locomotive operating conditions includes data that
is
passed to the consist level to determine the overall performance of the
consist. These
may be used for feedback to the operator or to the railroad control system.
They may
also he used for consist optimization. This data may include:
1. Tractive effort (TE) (motoring and dynamic braking) - This is calculated
based on
current/voltage, motor characteristics, gear ratio, wheel diameter, etc.
18
132250 CA 02454739 2004-01-05
Alternatively, it may be calculated from draw bar instrumentation or train
dynamics knowing the train and track information.
2. Horsepower (HP) - This is calculated based on the current/voltage
alternator
characteristics. It may also be calculated based on traction motor
current/voltage
information or from other means such as tractive effort and locomotive speed
or
engine speed and fuel flow rate.
3. Notch setting of throttle.
4. Air brake levels.
5. Friction modifier application, such as timing, type/amount/location of
friction
modifiers, e.g., sand and water.
Current locomotive equipment status may include data, in addition to one of
the above
items a to c, for consist optimization and for feedback to the train level and
back up to
the railroad track network level. This includes:
Temperature of equipment such as the engine, traction motor, inverter, dynamic
braking grid, etc.
A measure of the reserve capacity of the equipment at a particular point in
time and
may be used determine when to transfer power from one locomotive to another.
Equipment capability such as a measure of the reserve capability. This may
include
engine horsepower available (considering ambient conditions, engine and
cooling
capability), tractive effort/braking effort available (considering track/rail
conditions,
equipment operating parameters, equipment capability), and friction management
capability (both friction enhancers and friction reducers).
Fuel level/fuel flow rate - The amount of fuel left may be used to determine
when to
transfer power from one locomotive to another. The fuel tank capacity along
with the
amount of fuel left may be used by the train level and back up to the railroad
track
network level to decide the refueling strategy. This information may also be
used for
adhesion limited tractive effort (TE) management. For example, if there is a
critical
19
CA 02454739 2004-01-05
132250
adhesion limited region of operation ahead, the filling of the fuel tank may
be planned
to enable filing prior to the consist entering the region. Another
optimization is to
keep more fuel on locomotives that can convert that weight into useful
tractive effort
For example, a trailing locomotive typically has a better rail and can more
effectively
convert weight to tractive effort provided the axle/motor/power electronics
are not
limiting (from above mentioned equipment capability level). The fuel flow rate
may
be used for overall trip optimization. There are many types of fuel level
sensors
available. Fuel flow sensors are also available currently. However, it is
possible to
estimate the fuel flow rate from already known/sensed parameters onboard the
locomotive. In one example, the fuel injected per engine stroke (mmt/stroke)
may be
multiplied by the number of strokes/sec (function of rpm) and the number of
cylinders, to determine the fuel flow rate. This may be further compensated
for return
fuel rate, which is a function of engine rpnm, and ambient conditions. Another
way of
estimating the fuel flow rate is based on models using traction HP, auxiliary
HP and
losses/efficiency estimates. The fuel available and/or flow rate may be used
for
overall locomotive use balancing (with appropriate weighting if necessary). It
may
also be used to direct more use of the most fuel-efficient locomotive in
preference to
less efficient locomotives (within the constraint of fuel availability).
Fuel/Consumable range - Available fuel (or any other consumable) range is
another
piece of information. This is computed based on the current fuel status and
the
projected fuel consumption based on the plan and the fuel efficiency
information
available on board. Alternatively, this may be inferred from models for each
of the
equipment or from past performance with correction for ambient conditions or
based
on the combination of these two factors.
Friction modifier level - The information regarding the amount and capacity of
the
friction modifiers may be used for dispensing strategy optimization (transfer
from one
locomotive to another). This information may also be used by the railroad
track
network and infrastructure levels to determine the refilling strategy.
Equipment degradation/wear - The cumulative locomotive usage information may
he
used to make sure that one locomotive does not wear excessively. Examples of
these
may include the total energy produced by the engine, temperature profile of
dynamic
CA 02454739 2004-01-05
132250
braking grids, etc. This may also allow locomotive operation resulting in more
wear
to some components if they are scheduled for overhaul/replacement any way.
Locomotive position - The position and/or facing direction of the locomotive
may be
used for power distribution consideration based on factors like adhesion,
train
handling, noise, and vibration.
Locomotive health - The health of the locomotive includes the present
condition of
the locomotive and its key subsystems. This information may be used for
consist
level optimization and by the track network and infrastructure levels for
scheduling
maintenance/servicing. The health includes component failure information for
failures that do not degrade the current locomotive operation such as single
axle
components on an AC electro-motive locomotive that does not reduce the
locomotive
horse power rating, subsystem degradation information, such as hot ambient
condition, and engine water not fully warmed up, maintenance information such
as
wheel diameter mismatch information and potential rating reductions like
partially
clogged filters.
Operating parameter or condition relationship information - A relation to one
or more
operating parameters or conditions may be defined. For example, Fig. 17 is
illustrative of the type of relationship information at the locomotive level
that can be
developed which illustrates and/or defines the relationship between fuel use
and time
for a particular movement plan as shown by line 1402. This relationship
information
may be sent from the locomotive level 500 to the consist level 400. This may
include
the following:
Slope 1704 at the current operating plan time (fuel consumption reduction per
unit
time increase for example in gallons/sec). This parameter gives the amount of
fuel
reduction for every unit of travel time increase.
Fuel increase between the fastest plan 1710 and the present plan 1706. This
value
corresponds to the difference in fuel consumption between points F3 and F1, as
shown
on Fig. 17.
21
CA 02454739 2004-01-05
132250
Fuel reduction between the optimum plan 1712 and the present plan 1706. This
value
corresponds to the difference in fuel consumption between points F i and F4 of
Fig. 17
Fuel reduction between the allocated plan and current plan. This value
corresponds to
the difference in fuel consumption between points F, and F, of Fig. 17.
The complete fuel as a function of time profile (including range).
Any other consumable information.
For optimizations at the consist level 400, multiple closed loop estimations
may be
done by the consist level and each of the locomotives or the locomotive level.
Among
the consist level inputs from within the consist level are operator inputs,
anticipated
demand inputs, and locomotive optimization and feedback information.
The information flow and sources of information within the consist level
include:
6. Operator inputs,
7. Movement plan inputs,
8. Track information,
9. Sensor/model inputs,
10. Inputs from the locomotives/load cars,
11. Consist optimization,
12. Commands and information to each of the locomotives in the consist,
13. Information flow for train and movement optimization, and
14. General status/health and other info about the consist and the locomotives
in the
consist. The consist level 400 uses the information from/about each of the
locomotives in the consist to optimize the consist level operations, to
provide
feedback to the train level 300, and to provide instructions to the locomotive
level
500. This includes the current operating conditions, potential fuel efficiency
CA 02454739 2004-01-05
132250
improvements possible for the current point of operation, potential
operational
changes based on the profile, and health status of the locomotive.
There are three categories of functions performed by the consist level 400 and
the
associated consist level processor 1202 to optimize consist performance.
Internal
consist optimization, consist movement optimization, and consist monitoring
and
control.
Internal optimization functions/algorithms optimize the consist fuel
consumption by
controlling operations of various equipments internal to the consist like
locomotive
throttle commands, brake commands, friction modifier commands, anticipatory
commands. This may be done based on current demand and by taking into account
future demand. The optimization of the performance of the consist level
include
power and dynamic braking distribution among the locomotives within the
consist, as
well as the application of friction enhancement and reducers at points along
the
consist for friction management. Consist movement optimization functions and
algorithms help in optimizing the operation of the train and/or the operation
of the
movement plan. Consist control/monitoring functions help the railroad
controllers
with data regarding the current operation and status of the consist and the
locomotives/loads in the consist, the status of the consumables, and other
information
to help the railroad with consist/locomotive/track maintenance.
The consist level 400 optimization provides for optimization of current
consist
operations. For consist optimization, in addition to the above listed
information other
information can also be sent from the locomotive. For example, to optimize
fuel, the
relationship between fuel/HP (measure of fuel efficiency) and horsepower (HP)
as
shown in Fig. 18 by line 1802 may be passed from each locomotive to the
consist
level controller 1202. One example of this relationship is shown in Fig. 18.
Referring
to Fig. 18, the data may also include one or more of the following items:
Slope 1804 of Fuel/HP as a function of HP at the present operating horsepower.
This
parameter provides a measure of fuel rate increase per horsepower increase.
Maximum horsepower 1808 and the fuel rate increase corresponding to this
horsepower.
23
CA 02454739 2004-01-05
132250
Most efficient operating point 1812 information. This includes the horsepower
and
the fuel rate change to operate at this point.
Complete fuel flow rate as a function of horsepower.
The update time and the amount of information may be determined based on the
type
and complexity of the optimization. For example, the update may be done based
on
significant changes. These include notch change, large speed change or
equipment
status changes including failures or operating mode changes or significant
fuel/HP
changes, for example, a variation of 5 percent. The ways of optimizing include
sending only the slope (item a above) at the current operating point and may
be done
at a slow data rate, for example, at once per second. Another way is to send
items a, b
and c once and then to send the updates only when there is a change. Another
option
is to send only item d once and only update points that change periodically
such as
once per second.
Optimization within the consist considers factors such as fuel efficiency,
consumable
availability and equipment/subsystem status. For example, if the current
demand is
for 50 ,%o horsepower for the whole consist (prior art consists have all of
the
locomotives at the same power, here at 50% horsepower for each), it may be
more
efficient to operate some locomotives at less than a 50% horsepower rating and
other
locomotives at more than a 50%horsepower rating so that the total power
generated
by the consist equals the operator demand. In this case, higher efficiency
locomotives
will be operating at a higher horsepower than the lower efficiency
locomotives. This
horsepower distribution may be obtained by various optimizing techniques based
on
the horsepower as a function of fuel rate information obtained from each
locomotive.
For example, for small horsepower distribution changes, the slope of the
function of
the horsepower as a function of the fuel rate may be used. This horsepower
distribution may be modified for achieving other objective functions or to
consider
other constraints, such as train handling/drawbar forces based on other
feedback from
the locomotives. For example, if one of the locomotives is low on fuel, it may
be
necessary to reduce its load so as to conserve fuel if this locomotive is
required to
produce a large amount of energy (horsepower/hour) before refueling, even if
this
locomotive is the most efficient one.
24
132250 CA 02454739 2004-01-05
Other input information from each locomotive at the locomotive level 500 may
be
provided to the consist level 400. This other information from the locomotive
level
includes:
Maintenance cost. This includes the routine/scheduled maintenance cost due to
wear
and tear that depends on horsepower (ex. S/kwhr) or tractive effort increase.
Transient capability. This may be expressed in terms of the continuous
operating
capability of the locomotive, maximum capability of the locomotive and the
transient
time constant and gain.
Fuel efficiency at each point of operation.
Slope at every point of operation. This parameter gives the amount of fuel
rate
increase per horsepower increase.
Maximum horsepower at every point of operation and the fuel rate increase
corresponding to this horsepower.
Most efficient operating point information at every point of operation. This
includes
the horsepower and the fuel rate change to operate at this point.
Complete fuel flow rate vs. horsepower curve at every point of operation.
Fuel (and other consumable) range, based on current fuel level and the plan
and the
projected fuel consumption rate.
It' the complete profile information is known, the overall consist
optimization
considers the total fuel and consumables spent. Other weighting factors that
may he
considered include cost of locomotive maintenance, transient capability and
issues
like train handling, and adhesion limited operation. Additionally, if the
shape of the
consist level fuel use as a function of time as depicted by Fig. 14 changes
significantly
due to its transient nature (for example, the temperature of the electrical
equipments
such as traction motors, alternators or storage elements), then this curve
needs to be
regenerated for various potential power distributions for the current plan.
Similar to
CA 02454739 2004-01-05
132250
the previous section, the data may be sent periodically or once at the
beginning and
updates sent only when there is a significant change.
As input to the movement plans, optimization information may be developed at
the
consist level 400. Information may be sent from the locomotive level 500 to he
combined by the consist level with other information or aggregated with other
locomotive level data for use by the railroad network level 200. For example,
to
optimize fuel, fuel consumption information as a function of plan time, e.g.,
the time
to reach the destination or an intermediate point like meet or pass, may be
passed
from each locomotive to the consist controller 1202.
To illustrate one embodiment of the operation of optimization at the consist
level 400,
Fig. 14 illustrates the consist level as a function of fuel use versus time. A
line
denoted as 1402 represents fuel use vs. time at the consist level for a
consist
scheduled to go from point A to point B (not illustrated). Fig. 14 shows the
fuel
consumption as a function of time as derived by the train. The slope of line
1404 is
the fuel consumption vs. time at the present plan. Point 1406 corresponds to
the
current operation, 1408 to the maximum time allocated, 1410 corresponds to the
best
time it may make and 1412 corresponds to the most fuel efficient operation.
Under
the current plan, it will consume a certain amount of fuel and will get there
after a
certain elapsed time ti. It is also assumed that between points A and B. the
train at the
consist level assumes to operate without regard to other trains on the system
as long
as it can reach its destination within the time currently allocated to it,
e.g., ti-.
Optimization is run autonomously on the train to reach point B.
As noted above, the outputs of the consist level 400 include data to the train
level 300,
commands and controls to the locomotive level 500 as well as the internal
consist
level 400 optimization. The consist level output 1230 to the train level
includes data
associated with the health of the consist, service requirements of the
consist, the
power of the consist, the consist braking effort, the fuel level, and fuel
usage of the
consist. In one embodiment, the consist level sends the following types of
additional
information for use in the train level 300 for train level optimization. To
optimize on
fuel only, fuel consumption information as a function of plan time (time to
reach the
destination or an intermediate point like meet or pass) can be passed from
each of the
26
CA 02454739 2004-01-05
132250
consists to the train/railroad controller. Fig. 14 discloses one embodiment of
the
invention for fuel optimization and identifies the type of information and
relationship
between the fuel use and the time that can be sent by the consist level to the
train
level. Referring to Fig. 14, this includes one or more of the items listed
below.
Slope 1404 at the current operating plan time (fuel consumption reduction per
unit
time increase: gallons/sec). This parameter gives the amount of fuel reduction
for
every unit of time increase.
Fuel increase between the fastest plan and the current plan. This value
corresponds to
the difference in fuel consumption between points 1410 and 1406.
Fuel reduction between the best and current plan. This value corresponds to
the
difference in fuel consumption between points 1406 and 1412, of Fig. 14.
Fuel reduction between the allocated plan and current plan. This value
corresponds to
the difference in fuel consumption between points 1406 and 1408 of Fig. 14.
The complete fuel as a function of time profile as depicted in Fig. 14 by the
line 1402.
As noted in Fig. 13., the consist level 400 provides output commands to the
locomotive level 500 about current engine speed and power generation and
anticipated demands. Dynamic braking and horsepower requirements are also
provided to the locomotive level. The signals/commands from the consist level
to the
locomotive level or the locomotive within the consist level include operating
commands, adhesion modification commands, and anticipatory controls.
Operating commands may include notch settings for each of the locomotives,
tractive
effort/dynamic braking effort to be generated for each of the locomotives,
train air
brake levels (which may be expanded to individual car air brake in the event
electronic air brakes are used and when individual cars/group of cars are
selected),
and independent air brake levels on each of the locomotives. Adhesion
modification
commands are sent to the locomotive level or cars (for example, at the rear of
the
locomotive) to dispense friction-enhancing material (sand, water, or snow
blaster) to
improve adhesion of that locomotive or the trailing locomotives or for use by
another
consist using the same track. Similarly, friction lowering material dispensing
27
CA 02454739 2004-01-05
132250
commands are also sent. The commands include, the type and amount of material
to
be dispensed along with the location and duration of material dispensing.
Anticipatory controls include actions to be taken by the individual
locomotives within
the locomotive level to optimize the overall trip. This includes pre-cooling
of the
engine and/or electrical equipment to get better short-term rating or get
through high
ambient conditions ahead. Even pre-heating may be performed (for example,
water/oil may need to be at a certain temperature to fully load the engine).
Similar
commands may be sent to the locomotive level and/or storage tenders of a
hybrid
locomotive, as is depicted in Fig. 21, to adjust the amount of energy storage
in
anticipation of a demand cycle ahead.
The timing of updates sent to and from the consist level and the amount of
information can be determined based on the type and complexity of the
optimization.
For example, the update may occur at a predetermined point in time, at
regularly
scheduled times or when significant changes occur. These later ones may
include:
significant equipment status changes (for example the failure of a locomotive)
or
operating mode changes such as the degraded operation due to adhesion limits,
or
significant fuel, horsepower, or schedule changes such as a change in the
horsepower
by 5 percent. There are many ways of' optimizing based on these parameters and
functions. For example, only the slope (item a above) of the fuel use as a
function of
the time at the current operating point may he sent and this may be done at a
slow
rate, such as once every 5 minutes. Another way is to send items a, b and c
once and
only send updates when there is a change. Yet another option is to send only
item d
once and only update points that change periodically, such as once every 5
minutes.
As indicated in the earlier discussion, with simplified versions of train
configurations,
such as single locomotive consists and/or single locomotive trains, the
relationship
and extent of communication between the train level 300, consist level 400 and
locomotive level 500 becomes less complex, and in some embodiments, collapses
into
a less than three separately functioning levels or processors, with possibly
all three
levels operating within a single functioning level or processor.
28
CA 02454739 2004-01-05
132250
LOCOMOTIVE LEVEL
Figs. 15 and 16 illustrate the locomotive level 500 relationship with the
consist level
400 and optimization of the locomotive internal operation via commands to the
various locomotive subsystems. The locomotive level includes a processor 1502
with
optimization algorithms, which may be in the form of a memory 1602 and
processing
instructions 1604, etc. The input data to the locomotive level includes
consist level
data 1 512 and data 1514 from the locomotive level (including locomotive
feedback).
The output from the locomotive level includes data 1532 to the consist level
and
optimization of performance data 1534 at the locomotive level. As shown in
Fig. 16,
the input data 1512 from the consist level includes tractive effort command,
locomotive engine speed and horsepower generation, dynamic braking, friction
management parameters, and anticipated demands on the engine and propulsion
system. The input data 1514 from the locomotive level include locomotive
health,
measured horsepower, fuel level, fuel usage, measured tractive effort and
stored
electric energy. The later is applicable to embodiments utilizing hybrid
vehicle
technology as shown and described hereinafter in connection with the hybrid
vehicle
of Fig. 21. The data output 1532 to the consist level include locomotive
health,
friction management, notch setting, and fuel usage, level and range. The
locomotive
optimization commands 1534 to the locomotive subsystems include engine speed
to
the engine, engine cooling for the cooling system for the engine, DC link
voltage to
the inverters, torque commands to the traction motors, and electric power
charging
and usage from the electric power storage system of hybrid locomotives. Two
other
types of inputs include operator inputs and anticipated demand inputs.
The information flow and sources of information at the locomotive level 500
include:
a. Operator inputs,
b. Movement plan inputs,
c. Track information,
d. Sensor/model inputs,
c. Onboard optimization,
29
CA 02454739 2004-01-05
132250
C. Information flow for consist and movement optimization, and
General status/health and other information for consist consolidation and for
railroad optimization/scheduling.
Three categories of functions performed by the locomotive level include
internal
optimization functions/algorithms, locomotive movement optimization
functions/algorithms, and locomotive control/monitoring. Internal optimization
functions/algorithms optimize the locomotive fuel consumption by controlling
operations of various equipments internal to the locomotive, e.g., engine,
alternator,
and traction motor. This may he done based on current demand and by taking
into
account future demand. Locomotive movement optimization functions and/or
/algorithms help in optimizing the operation of the consist and/or the
operation of the
movement plan. Locomotive control/monitoring functions help the consist and
railroad controllers with data regarding the current operation and status of
the
locomotive, the status of the consumables and other information to help the
railroad
with locomotive and track maintenance.
Based on the constraints imposed at the locomotive level, operation parameters
that
may be optimized include engine speed, DC link voltage, torque distribution
and
source of power.
For a given horsepower command, there is a specific engine speed which
produces the
optimum fuel efficiency. There is a minimum speed below which the diesel
engine
cannot support the power demand. At this engine speed the fuel combustion does
not
happen in the most efficient manner. As the engine speed increases the fuel
efficiency improves. However, losses like friction and windage increase and
therefore an optimum speed can be obtained where the total engine losses are
the
minimum. This fuel consumption vs. engine speed is illustrated in figure 20
where
the curve 2002 is the total performance range of the locomotive and point 2004
is the
optimum performance for fuel usage vs. speed.
The DC link voltage on an AC locomotive determines the DC link current for a
given
power level. The voltage typically determines the magnetic losses in the
alternator
and the traction motors. Some of these losses are illustrated in figure 19.
The voltage
CA 02454739 2004-01-05
132250
also determines the switching losses in the power electronics devices and
snubbers. It
also determines the losses in the devices used to produce the alternator field
excitation. On the other hand, current determines the 12r losses in the
alternator.
traction motors and the power cables. Current also determines the conduction
losses
in the power semiconductor devices. The DC link voltage can be varied such
that the
suns of all the losses is a minimum. As shown in Fig. 19, for example, the
alternator
current losses vs. DC link voltage are plotted as line 1902 the alternator
magnetic core
losses vs. DC link voltage are plotted as line 1906 and the motor current
losses vs. DC
link voltage are plotted as line 1904 which are substantially optimized at
line 1908 at
DC link voltage V1,
For a specific horsepower demand, the distribution of power (torque
distribution) to
the six traction axles of one embodiment of a locomotive may be optimized for
fuel
efficiency. The losses in each traction motor, even if it is producing the
same torque
or same horsepower, can be different due to wheel slip, wheel diameter
differences.
the operating temperature differences and the motor characteristics
differences.
Therefore, the distribution of the power between each axles can be used to
minimize
the losses. Some of the axles may even he turned off to eliminate the
electrical losses
in those traction motors and the associated power electronic devices.
In locomotives with additional power sources, for example, hybrid locomotives
such
as shown in Fig. 21, the optimum power source selection and the appropriate
amount
of energy drawn from each of the sources (so that the sum of the power
delivered is
what the operator is demanding), determines the fuel efficiency. Hence
locomotive
operation may be controlled to obtain the best fuel-efficient point of
operation at any
time.
For consists or locomotives equipped with friction management systems, the
amount
of friction seen by the load cars (especially at higher speeds) may be reduced
by
applying friction reducing material on to the rail behind the locomotive. This
reduces
the fuel consumption since the tractive effort required to pull the load has
been
reduced. This amount and timing of dispensing may be further optimized based
on
the knowledge of the rail and load characteristics.
31
CA 02454739 2004-01-05
132250
A combination of two or more of the above variables (engine speed. DC link
voltage
and torque distribution) along with auxiliaries like engine and equipment
cooling may
he optimized. For example, the maximum DC link voltage available is determined
by
the engine speed and hence it is possible to increase the engine speed beyond
the
optimum (based on engine only consideration) to obtain a higher voltage
resulting in
an optimum operating point.
There are other considerations for optimization once the overall operating
profile is
known. For example, parameters and operations such as locomotive cooling,
energy
storage for hybrid vehicles, and friction management materials may be
utilized. The
amount of cooling required can be adjusted based on anticipated demand. For
example, if there is big demand for tractive effort ahead due to high grade,
the traction
motors may be cooled ahead of time to increase its short term (thermal) rating
which
will he required to produce high tractive effort. Similarly if there is a
tunnel ahead if
the engine and other components may be pre-cooled to enable operation through
the
tunnel to be improved. Conversely, if there is a low demand ahead, then the
cooling
may be shut down (or reduced) to take advantage of the thermal mass present in
the
engine cooling and in the electric equipment such as alternators, traction
motors,
power electronic components.
In a hybrid vehicle, the amount of power in a Hybrid Vehicle that should be
transferred in and out of the energy storage system may be optimized based on
the
demand that will be required in the future. For example, if there is a large
period of
dynamic brake region ahead, then all the energy in the storage system can he
consumed now (instead of from the engine) so as to have no stored energy at
the
beginning of dynamic brake region (so that the maximum energy may be
recaptured
during the dynamic brake region of operation). Similarly if there is a heavy
power
demand expected in the future, the stored energy may be increased for use
ahead.
The amount and duration of dispensing of friction increasing material (like
sand) can
be reduced if the equipment rating is not needed ahead. The trailing axle
power/tractive effort rating may be increased to get the maximum available
adhesion
without expending these friction-enhancing resources.
32
CA 02454739 2010-01-19
132250
There are other considerations for optimization other than fuel. For example,
emissions may be another consideration especially in cities or highly
regulated
regions. In those regions it is possible to reduce emissions (smoke, Nitrogen
Oxide,
etc.) and trade off other parameters like fuel efficiency. Audible noise may
be another
consideration. Consumable conservation under certain constraints is another
consideration. For example, dispensing of sand or other friction modifiers in
certain
locations may be discouraged. These location specific optimization
considerations
may be based on the current location information (obtained from operator
inputs,
track inputs, GPS/track information along with geo-fence information). All
these
factors are considered for both the current demand and for optimizations for
the
overall operating plan.
HYBRID LOCOMOTIVE
Referring to Fig. 21, a hybrid locomotive level 2100 is shown having an energy
capture subsystem 2116. The energy management subsystem 2112 controls the
energy capture subsystem 2116 and the various locomotive components, such as
diesel engine 2102, alternator 2104, rectifier 2106, mechanically driven
auxiliary
loads 2108, and electrical auxiliary loads 2110 that generate and/or use
electrical
power. This management subsystem 2112 operates to direct available electric
power
such as that generated by the traction motors during dynamic braking or excess
power
from the engine and alternator, to the energy capture subsystem 2116, and to
release
this stored electrical power within the consist to aid in the propulsion of
the
locomotive during monitoring operations.
To do so, the energy management subsystem 2112 communicates with the diesel
engine 2102, alternator 2104, inverters and controllers 2120 and 2140 for the
traction
motors 2122 and 2142 and the energy storage subsystem interface 2126.
As described above, a hybrid locomotive provides additional capabilities for
optimizing locomotive level 500 (and thus consist and train level)
performance. In
some respects, it allows current engine performance to be decoupled from the
current
locomotive power demands for motoring, so as to allow the operation of the
engine to
be optimized not only for the present operating conditions, but also in
anticipation of
33
132250 CA 02454739 2004-01-05
the upcoming topography and operational requirements. As shown in Fig. 21,
locomotive data 2114, such as anticipated demand, anticipated energy storage
opportunities, speed and location, are input into the energy management
subsystem
2112 of the locomotive layer. The energy management sub-system 2112 receives
data from and provides instructions to the diesel engine controls and system
2102, and
the alternator and rectifier control and systems 2104 and 2106, respectively.
The
energy management sub-system 21 12 provides control to the energy storage
system
2128, the inverters and controllers of the traction motors 2120 and 2140, and
the
braking grid resistors 2124.
When introducing elements of the present invention or the embodiment(s)
thereof, the
articles "a," "an," "the," and "said" are intended to mean that there are one
or more of
the elements. The terms "comprising," "including," and "having" are intended
to be
inclusive and mean that there may be additional elements other than the listed
elements.
Those skilled in the art will note that the order of execution or performance
of the
methods illustrated and described herein is not essential, unless otherwise
specified.
That is, it is contemplated that aspects or steps of the methods may he
performed in
any order, Unless otherwise specified, and that the methods may include more
or less
aspects or steps than those disclosed herein.
While various embodiments of the present invention have been illustrated and
described, it will be appreciated to those skilled in the art that many
changes and
modifications may be made thereunto without departing from the spirit and
scope of
the invention. As various changes could be made in the above constructions
without
departing from the scope of the invention, it is intended that all matter
contained in
the above description or shown in the accompanying drawings shall be
interpreted as
illustrative and not in a limiting sense
34
