Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
22 C H A P T E R 3 3.1 Introduction This chapter provides a detailed description of capacity analysis methods and tools to assess the ability of a rail line segment to carry a given volume and mix of railroad traffic, while meeting the service quality goals of the operator or operators. Capacity is a function of: ⢠Physical characteristics of the line segment, such as single or double track, distance between passing sidings, signal system characteristics, permitted speeds for different train types, cur- vature and gradients. ⢠Traffic volume and characteristics, such as the numbers of trains of each type traveling over the line in a specific period of time (e.g., 24 hours), speeds, train length and weight, locomo- tive power assigned to each train, and stops for passenger stations or to drop off and pick up freight cars from industry sidings. ⢠Management practices and protocol, including dispatch procedures, safety regulations, and treatment of train movements through passenger terminal areas. As discussed earlier, capacity may be considered adequate when each user of the line segment is able to meet service quality goals for its rail services using a line segment. For a passenger ser- vice operator the service quality goal may be to achieve a given percentage of on-time arrivals and/or ensuring that individual train and aggregate delays do not exceed an agreed level. For a freight service operator, service goals will depend on service type, e.g., an intermodal train may be required to meet punctuality goals reflecting commitments made to customers by the rail- road; but for other train types the railroadâs objectives may be to minimize delays and unneces- sary stops and starts that add to fuel, employee and other costs. Given the large number of factors that must be considered in assessing a rail lineâs capacity to carry a defined traffic volume and mix, use of a structured analysis method is essential. These vary from relatively simple manual methods used on low-traffic lines or for simple scoping studies, up to complex simulation models for busy lines. This chapter provides detailed discus- sions of the following aspects of capacity analysis: ⢠The complexity of railroad operations, and why structured analyses are essential for the suc- cessful planning and implementation of rail passenger service on shared corridors. ⢠The principal factors that affect line capacity and which must be addressed in capacity analyses. ⢠The principal classes of capacity analysis tools and their application to rail service planning. This includes both preliminary or scoping analysis, and highly detailed analysis to support major infrastructure investment decisions and contractual commitments. ⢠Descriptions of individual capacity models. These include data requirements and strengths and weaknesses. Analytical Approaches to Line Capacity in Shared-Use Corridors
Analytical Approaches to Line Capacity in Shared-Use Corridors 23 3.2 Complexity of Shared Railroad Operations Section 1.3 of Chapter 1 provided a short introduction to railroad operations, describing some of the major features and defining common terminology. This section amplifies that discussion by introducing some other key factors that must be considered in capacity analysis. In particular, rail operations are confined to rails and must be actively managed to work efficiently. This is unlike highways, where individually operated vehicles are free to navigate the highway network as they wish provided they observe traffic laws and signals. Some of the key factors are as follows. ⢠Most rail lines and certainly lines shared by freight and passenger trains have to accommodate trains with very different performance characteristics. A loaded bulk commodity trains, such as a 120-car coal unit train, could weigh 18,000 tons (including four locomotives), have a maxi- mum speed of 40 mph, and be assigned locomotives providing only one horsepower per ton (hp/ton) of train. Acceleration and braking are slow. An intermodal train, carrying highway trailers or shipping containers, will be assigned 2 to 4 hp/ton and will accelerate more quickly, but could be up to two miles in length and slow to enter and exit sidings through low speed switches. In contrast, an intercity passenger train will be relatively short, be provided with up to 9 hp/ton power, be able to brake and accelerate relatively quickly and be quick to enter and exit sidings. This variability of train lengths, power and braking characteristics presents dis- patchers with a typical dilemma: stop a freight train at a passing siding, causing a substantial delay and possible capacity impacts; or stop a passenger train at a siding to allow the freight train to pass at line speed, maximizing capacity but delaying the passenger train. Please see Figures 3-1, 3-2, and 3-3 showing aforementioned train types, and Table 3-1 showing illustrative characteristics of these train types. ⢠Train performance over a line segment is a simple function of geography. Permitted speeds on curves differ between passenger and freight trains, and uphill speeds on grades are a direct function of train power-to-weight ratios. As with the other factors mentioned in the previ- ous bullet, these have to be considered by the dispatcher in managing operations over a line segment. ⢠Many freight trains and some passenger trains travel long distances between terminals. Distances can vary from several hundred miles to over 2,000 miles, for example, between the West Coast and Chicago. Operating events several hundred miles away can affect operations over a specific Figure 3-1. Coal unit train. Photo by Walt Schuchmann
24 Capacity Modeling Guidebook for Shared-Use Passenger and Freight Rail Operations line segment, causing delays that can propagate quickly through a large railroad system. This is why it is often necessary to perform capacity analysis well beyond the territory proposed for passenger service. Large railroads have concentrated most of their operations management and dispatching in network-wide centers so as to better manage operations over a wide area. ⢠Track and signal systems must receive regular inspections and maintenance to function reliably. This requires access to the track for maintenance crews, preferably in daylight, and for move- ments by automated inspection vehicles. Even with well managed inspection and maintenance programs, unplanned failures of any element in the system will occur, especially in extreme weather. Ice and snow will clog switches, high temperatures cause track to buckle laterally, Figure 3-2. Intermodal train. Photo by Justin Fox Figure 3-3. Intercity corridor passenger train. Photo by Justin Fox Train Type Length Gross Weight in Tons Typical Max Speed Horsepower per Ton Coal unit train 6,500 18,000 40 1 Intermodal 7,500 8,000 60 2 to 4 Corridor/commuter trains 600 600 80 6 to 9 Table 3-1. Illustrative characteristics of different train types.
Analytical Approaches to Line Capacity in Shared-Use Corridors 25 and locomotives can fail for a variety of reasons. Operations must be managed taking into account regular inspection and maintenance requirements, and an expectation of typical unplanned delays. The above paragraphs are an introduction to the complexity inherent in railroad operations. In general terms, the complexity is similar to other transportations systems, but operation on fixed tracks limits flexibility and places a premium on skilled operations management. Of course, it is also operation on fixed tracks combined with centralized management and the use of auto- mated systems that enables railroads to move high volumes of freight and passenger traffic safely and without interference from other surface transportation systems. 3.3 Main Line Capacity Factors The goal of capacity analysis is to determine the maximum practical traffic volume and mix that can be accommodated on a specific line segment, while meeting service quality expectations for each traffic type. Some of these factors are fixed in the medium term, such as track and signal system characteristics, while others may vary by time of day, day of week, or seasonally. Traffic volumes are particularly subject to short term variability, as well as allowances for maintenance and typical service disruptions. In summary, these factors are: Infrastructure capacity factors ⢠Number of tracks and distance between passing sidings and crossovers. ⢠Curves and grades. ⢠Signaling and train control method, such as ABS, CTC, ATC or PTC as defined in Section 1.3. ⢠Allowable speeds for each type of train, taking into account curvature, grades, switch types, and signal and train control method used. Operational capacity factors ⢠Volume and mix of traffic: the number of trains per day for passenger trains and each type of freight train. ⢠Expected variability in traffic mix, including daily, weekly, and seasonal variations. ⢠Train characteristics: number of cars and locomotives assigned to each freight and passenger train, with aggregate train weight and locomotive horsepower. ⢠Train priorities. ⢠Availability of train crew and other operating personnel, especially at crew change locations. Train crew must be replaced when they reach the maximum hours of service prescribed by FRA regulations. Unplanned crew changes away from regular crew change points are very disruptive and must be avoided. ⢠Estimated time periods and locations when track will be unavailable for service for mainte- nance or to respond to unplanned events. Taken together, these factors govern the usable capacity of a line segment. Capacity is not a hard mathematical number. Rather, the capacity of the line segment is better expressed in terms of average delay to each type of train. A capacity limit is reached when delay statistics exceed acceptable limits for each type of traffic. Operations may still be feasible with more trains, but the delays will prevent the railroad from meeting customer expectations and ultimately damage the business and/or increase costs. When traffic increases, the negative effects of rail traffic congestion increase, requiring the railroad to become more efficient or invest to increase capacity. Physical infrastructure changes, such as adding sidings and crossovers or shortening signal block lengths, are obvious measures,
26 Capacity Modeling Guidebook for Shared-Use Passenger and Freight Rail Operations but are likely to be costly. Operations changes to increase capacity are likely to be less costly and can be implemented more quickly, but may be limited to smaller incremental improvements. Some examples are: ⢠Lengthening freight trains, including using distributed power (remote-control locomotives inserted part way along a train or at the end). However, train lengths may be limited by the length of passing sidings on the line segment. ⢠Working with freight customers to enhance rail efficiency at industry sidings. For example, a longer siding may mean that main track is not occupied while cars are dropped off and picked up at the siding. ⢠Working with all users of the line segment to adjust schedules and train sequencing to reduce conflicts. ⢠Reducing the variability in train mix, where possible. This may be an option where a railroad has alternative routes and can concentrate traffic types by route. ⢠Improving inspection and maintenance equipment and practices to reduce maintenance track time and the need for unplanned repairs. One passenger service provided funding for overnight maintenance to free up capacity for daytime passenger operations. ⢠Investing in freight yard and terminal infrastructure and efficiency, especially additional track to accommodate trains entering the yard. This will reduce the need for trains to occupy the main line while waiting to enter the yard. ⢠Provide for directional running where parallel lines exist. This will eliminate meets of opposing trains and thus enhance capacity and fluidity. It is important to note that most operations changes are to freight operations, and a freight railroad hosting a passenger service will look for win-win opportunities where the change ben- efits both parties: providing capacity for the passenger service at the same time as maintaining or improving service to freight customers. 3.4 Rail Line Capacity Analysis Methodologies and Applications 3.4.1 Introduction This section discusses the principal methodologies used in rail capacity analysis, as distinct from individual proprietary software packages available to the railroad industry. Together with the analyses, the discussion provides guidance as to when the analyses should be used. Simple analysis that can be accomplished in a few days might be suitable for initial screening of a wide range of rail service options, whereas complex simulation modeling will be needed for the detailed planning of a major infrastructure investment. Another factor is data requirements. A simple analysis might make basic generic assumptions about infrastructure and operations from basic line geography and daily train numbers, whereas the detailed analysis requires comprehensive data on train schedules and make-up, as well as relevant track and signal system information. The following paragraphs identify and describe the various modeling methodologies used in rail line capacity modeling, including analyses that are important elements in capacity modeling. 3.4.2 Scoping Models and Building Blocks This section discusses less detailed modeling processes suitable for initial scoping capacity analysis. Scoping analyses might be used to compare alternative routes, or to compare between major alternatives for a transportation corridor, such as between minor upgrades to an existing line, a major upgrade providing higher speeds and more frequent service, or constructing an
Analytical Approaches to Line Capacity in Shared-Use Corridors 27 all-new right-of-way. This section also describes common modeling building blocks that are not complete models in themselves, but are often elements in a capacity model. The methods described are: ⢠Train Performance Calculator ⢠String Line Analysis ⢠Grid Time Analysis ⢠Other Preliminary Planning and Scoping Approaches 3.4.2.1 Train Performance Calculator (TPC) A TPC (sometimes called a Train Performance Simulator or TPS) is used to calculate uncon- strained journey time for a single train over a rail line segment. The train related inputs are train weight, locomotive power characteristics such as a speed versus tractive effort curve, train resistance from rolling friction and aerodynamic drag, and brake performance characteristics. Infrastructure and operations data include gradients, curvature, speed limits and location, and dwell time at station stops. A TPC does not include any consideration of other trains operating on the corridor, but it is common to add a percentage to the minimum journey time to estimate a practical scheduled time for planning purposes. A TPC can also be used to estimate journey time changes resulting from increasing locomotive power, raising speed limits, adding or removing station stops, and similar changes. As an element in a capacity model, TPCs are used to calculate travel times between points where a train must slow or stop for a meet or pass, or to use a crossover or siding entry switch. A TPC is also an essential component of all rail operations simulation models. Simulation models calculate the movements of all trains on a specific line segment in parallel over time, sav- ing a snapshot of the operation at the end of each time interval and re-starting the calculation for a new time interval. Time intervals boundaries are set after either a fixed length of time or when an event occurs, such as a dispatcher decision to route a train into a siding. The TPC is used to cal- culate the movements of each train for each time interval given initial speed, terrain, train weight, locomotive power, and operating instructions applicable to each train at that time and location. 3.4.2.2 String Line Analysis A string line chart is a representation of rail operations over a line segment on a time-distance plot. Figure 3-4 is a typical string line chart. A string line chart is a time-distance plot showing all trains operating on a line segment over a given time period, most often 24 hours. Usually distance in miles is shown on the Y or vertical axis, which will also show station, passing siding or crossover locations. The X or horizontal axis shows time in hours and minutes. Train movements are shown as forward or backward sloping lines depending on the direction of travel. Steeper slopes indicate a faster train. A stationary train, usually at a passing siding or at a station stop is shown as a horizontal line. String line charts are used in almost all capacity analyses. They illustrate present capacity problems by displaying what actually happens, especially delays at different points along the line segment. They are provided as one of the outputs from complex simulation analyses, along with delay statistics, journey time data and animations. They can also be produced manually. Furthermore, they are also a capacity analysis tool in themselves. String line charts display the results of âwhat ifâ exercises, such as adding additional trains, adding passing sidings or cross- overs, adding double track sections, or changing schedules. String line analyses, typically of a representative 30-day period, provide realistic estimates of journey time and operating delays for each train operating on the line segment for each scenario and highlight problem locations for further study.
28 Capacity Modeling Guidebook for Shared-Use Passenger and Freight Rail Operations 3.4.2.3 Grid Time Analysis Another relatively simple scoping method is known as a grid time analysis, which is used to test the upper limit for the number of daily trains a corridor can handle, without consideration of individual train service commitments. The theoretical capacity to handle traffic on single track is dictated by the time it takes a train to travel the distance between two sidings and clear the way for an opposing train. The time a train takes to traverse the single track section and be in the clear for the opposing movement is called the one-way grid time. The grid time varies as a function of the spacing between sidings and the average of train speeds in each direction. The single track segment on which the trains take the highest amount of time dictates how many trains can traverse a line in a day. This segment defines the capacity over the line as the number of trains that can be handled daily. Figure 3-5 is an illustration of a representative grid time analysis calculation. Source: North Sound Rail Operations Simulation, Whatcom Council of Governments, 2011. Figure 3-4. Typical string line chart showing conflicts of opposing trains.
Analytical Approaches to Line Capacity in Shared-Use Corridors 29 In the upper illustration, an eastbound train is progressing past one siding to another. The time it takes to run from being stopped at point A to point C is 30 minutes. C is 1.5 miles past point B, being a typical length of a longer freight train today. It is assumed that the eastbound train will come to a complete stop before it reaches the end of the eastern siding, until the track ahead is clear of any opposing trains. A westbound train is waiting in the eastern siding for the eastbound train to clear point B. In the lower illustration, a stopped westbound train pulls out of the eastern siding just after the eastbound train has passed point B. It takes another 30 minutes for the westbound train to run past point D, 1.5 miles west of point A. It is assumed that the westbound train will come to a complete stop by the time it reaches the end of the western siding. In all it takes the two opposing trains 60 minutes or one hour of grid time to cover the length of single track between points C and D. Thus the maximum theoretical capacity of this segment is two trains per hour or 48 trains per day. The theoretical capacity calculation assumes the availability of an unlimited supply of trains at both ends of a line segment throughout the day. This is not realistic for actual operations, so a downward scaling factor is applied to derive the practical capacity of the line. The scaling factor reduces the theoretical number to reflect a typically uneven sequence of trains entering the track segment, as well as allowing for typical delays due to track maintenance work, slow orders, and unplanned service disruptions. As a general rule, the practical capacity of a line with one or more single track segments falls in the range of 50% to 75% of the theoretical capacity. A lower percentage could reflect assumptions about particular conditions due to the time of year, such as during the spring snow melt when ground may be soft over extended periods of time, requiring slower operations. The range can also reflect the bias of the analyst: a more conservative analyst, knowing fewer details about operations of a line, may assume a lower percentage to guard against potentially overstating practical capacity. Over segments of double and triple tracks, calculating the practical capacity of a route is not as straightforward as on single track. Theoretically, the capacity is extremely high because trains can fleet behind one another, unimpeded by opposing traffic and limited only by the train speed and by the spacing between trains provided by the signal system. In practice, however, numerous other factors combine to reduce the effective capacity of multiple track segments. The most important ones are traffic mix (trains with different speeds, characteristics and customer Figure 3-5. Representative grid time analysis calculation.
30 Capacity Modeling Guidebook for Shared-Use Passenger and Freight Rail Operations requirements);, track outages for repairs and maintenance; spacing between block signals and interlockings; and queuing at entrances of terminals and junctions. When dealing with multiple tracks, there is no cut and dried method of calculating this capac- ity and one can only address cases on an individual basis, through simulations or other methods, each with the casesâ specific operating conditions. However, this does not preclude identifying trends and capacity ranges that these individual analyses have provided over time. 3.4.2.4 Other Scoping and Planning Approaches Not all rail service capacity assessments are targeted to development of specific rail projects. Broad scale planning exercises may rely on the tools described below as a more cost-effective proxy for the detailed and time consuming modeling techniques which support specific con- tract and capital upgrade proposals for specific line segments. Some examples of these broader approaches include: Parametric capacity estimate. Capacity is calculated from a âcapacity formulaâ where the inputs are number of tracks, signal block lengths, train speeds, siding and crossover spacing, and mix of train speeds. Alternatively, tables or graphs derived from the formula may be used. The formula is derived from the capacity of a representative sample of line segments. Generic linear programming and cost-benefit algorithms. These algorithms are attempts to optimize capacity by identifying the lowest-cost combination of actions to reach a specified capacity goal. This approach is promoted by many independent consultants and academic insti- tutions as an attractive way of resolving capacity problems without using time consuming and costly simulation analysis. However, the approach treats service output as a dependent variable that is specified by the capacity-maximizing algorithm, and for this reason it is wholly inappro- priate for situations with strict service requirements such as passenger rail. National Rail Freight Infrastructure Capacity and Infrastructure Study method. This 2007 study used a simplified variant of the parametric relationship method. The study sponsor, the Association of American Railroads (AAR), in collaboration with American Association of State Highway and Transportation Officials (AASHTO), sought to depict upcoming rail network con- gestion nodes on a national scale. Capacity estimates were derived from analysis of national rail traffic and infrastructure data by route segment using only three key variables: number of tracks, signal system type, and train type. The specific variables were: ⢠Number of tracks: Between one and six. ⢠Signal systems: â No signals and Track Warrant Control. Track warrants are structured voice radio messages giving a train permission to occupy a defined track segment, usually between sidings. â Automatic Block System (ABS). â Centralized Traffic Control (CTC) or Traffic Control System (TCS). ⢠Train types: â Heavy bulk commodity and general merchandise freight. â Fast freight: intermodal and multi-level auto carrier trains. â Passenger service. The analysis resulted in a table relating capacity to the three key variables as shown in Table 3-2. These estimates are broad averages and do not include many factors known to affect capac- ity, such as siding spacing and length, curves and grades, and the power assigned to trains. These capacity definitions were useful in developing a national ârail congestionâ map in that the required engineering data was consistent with that required of rail industry players in their annual regulatory filings.
Analytical Approaches to Line Capacity in Shared-Use Corridors 31 3.4.3 Operations Simulation Analysis Methods Simulation analysis has become the principal method by which line capacity issues are resolved, both in North America and overseas. These models provide a step-by-step simulation of all trains operating on a specific line segment to create a complete and accurate picture of operations. The models incorporate a routine to simulate, or look up from schedule data, the arrival of trains at both ends of the segment, a TPC to simulate train movement over the line between signals, sidings, and crossovers along the segment, and a dispatching algorithm that mimics the behav- ior of a typical dispatcher making meet-and-pass and similar decisions. Model outputs include string line charts, journey time and delay statistics, and animations. The models are also capable of introducing random disruptions into the simulation, such as from unplanned maintenance of track and of equipment failures, and to test the robustness of the operation to recover from such delays. Four simulation models have been identified. ⢠NCFRP Web-based Freight-Passenger Rail Corridor Project Screening Tool. This is a model developed under a National Cooperative Freight Research Program (NCFRP) contract. The research is described in NCFRP Report 27: Web-Based Screening Tool for Shared-Use Rail Cor- ridors. It is a web-based tool designed for initial planning analysis and to be easily accessible to interested parties. Information on the web-based screening tool can be accessed at http:// www.trb.org/main/blurbs/171116.aspx. ⢠Berkeley Simulation Softwareâs Rail Traffic Controller (RTC). RTC is used by the Class I railroads, government agencies, commuter and passenger operators, and consultant groups throughout the railroad industry. The RTC system is particularly notable for its ability to simulate actual dispatch behavior on a North American freight railroad as it copes with high variability in the timing and volume of train movements. ⢠SYSTRAâs RAILSIM Program. RAILSIM is primarily used by commuter and passenger agen- cies, Class I railroads, and consultant groups. Strengths of RAILSIM are its ability to simulate complex schedules of passenger operations and its associated features for planning equipment and staff resources needed for an intensive passenger service. ⢠CANACâs RAILS2000 Program. This tool has experienced a decrease in use over time and now has relatively limited exposure within the industry. It is used primarily within CANACâs consulting services. Because of their complexity and importance, simulation models are discussed in detail in Section 3.5. Number of tracks Type of Train Control Prac cal Maximum Trains per Day Mul ple Train Types Single Train Type 1 N/S or TWC 16 20 1 ABS 18 25 2 N/S or TWC 28 35 1 CTC or TCS 30 48 2 ABS 53 80 2 CTC or TCS 75 100 3 CTC or TCS 133 163 4 CTC or TCS 173 230 Note: Es mates for 5 and 6 tracks omi ed. Table 3-2. Estimated average capacities of typical freight railroad corridors.
32 Capacity Modeling Guidebook for Shared-Use Passenger and Freight Rail Operations 3.4.4 Modeling Objectives and Model Data Requirements A key issue in capacity analysis, which affects how the models are used and the usefulness of results, is the level of detail in infrastructure and operations data needed by the models. All capacity models need these data at some level of detail to achieve their objectives. The accuracy of model results for a specific line segment depends directly on how closely the data represent actual infrastructure and operating conditions, including details of individual trains operating over the segment. An analysis relying on a generic parametric model using coefficients derived from a variety of actual operations cannot provide the level of detail and credibility that can be achieved by using a detailed simulation model. It follows that the appropriate use of the simpler models is to study broad transportation alternatives, such as between alternative routes, upgrade levels or between rail and non-rail alternatives. Once a broad alternative has been selected, then detailed modeling must be used to determine what capacity investments are required to meet planned rail service objectives. In most cases, such modeling is the only way to provide proper support for major infrastructure investments and to finalize contract agreements between users of a line segment. Detailed analysis requires detailed data. If the rail line segment is operated by a public agency, such as a commuter rail authority or Amtrak, then detailed operation, track, and signal system data will normally be available to any responsible party interested in capacity analysis. However, if the line segment is operated by a private freight railroad, then much of the detail concerning rail operations may be proprietary and market sensitive. Data will only be released under strict confidentiality conditions, and the railroad will be sensitive to the interpretation of any results obtained. The most common situation is where a public agencyâfor example, a state or regional pas- senger rail authorityâis seeking to implement passenger rail service over one or more freight railroad line segments. There is tension inherent in this situation: the freight railroad needs assur- ance that a proposed passenger service will not interfere with freight operations, and the passenger authority needs to assure that state and federal funding is being spent responsibility and passenger rail service objectives will be achieved. Recent practice has been for one party (usually the freight railroad) to manage the analysis, and for the other party (the passenger rail authority) to have access to capacity model methods, inputs and results. Trust and cooperation between the host freight railroad and the passenger authority are essential, and may involve considerable effort to overcome initial suspicions on both sides and to maintain trust over time. A fundamental step forward is for both parties to agree on the modeling input assumptions that drive the analysis. Section 3.5 describes the mechanics of detailed modeling, including data requirements and the pros and cons of alternative simulation models. 3.5 Detailed Simulation Models 3.5.1 Technical Modeling Process and Data Needed The detailed simulation models are used to determine whether a specific line segment has suf- ficient capacity for each rail traffic type moving over the segment to meet its service quality goals. Simulation analysis provides the most reliable and accurate way of making this determination. Five simulation modeling approaches have been identified, each with different strengths and weaknesses, as described in this section.
Analytical Approaches to Line Capacity in Shared-Use Corridors 33 3.5.1.1 Input Data and Model Operation While different models may vary in level of detail and emphasis given to different aspects of a rail operation, the primary data input categories will be as follows: ⢠Details of the trains operating over the line segment, including weight, locomotive power and braking characteristics. The level of detail can vary between a generic freight or passenger train and the make-up of each individual train. These factors most clearly distinguish pas- senger from freight trains, and between different types of freight trains. As well as differences to power-to-weight ratios mentioned earlier, braking capabilities differ greatly between pas- senger and freight trains. Short passenger trains, typically less than 1,000 feet with automatic wheel slide protection systems (similar to ABS on an automobile), have higher braking rates and shorter stopping distances. In addition, because the train is short, the time taken to release the brakes by restoring air pressure in the brake system is short. With long freight trains (up to 10,000 feet), brakes must be applied slowly to avoid excessive longitudinal forces in the train, and the time taken to restore air pressure and release the brake is much longer than with pas- senger trains. ⢠Line segment infrastructure details, including passing siding locations and lengths, signal locations and signal block lengths, crossover locations, switch types, curvature and grade, and speed limits applicable to each traffic type. Signal system characteristics are critical to capac- ity. Signal block lengths and siding spacing govern the distance between trains, and they are a fundamental limit on capacity. Although most line segments being analyzed will be equipped with CTC, with remotely controlled switches and signals, some lower-traffic locations will only have ABS with manually operated switches at passing sidings. In this case, the time taken by train crews to operate switches must be factored into the simulation. PTC, when imple- mented, may impose conservative braking characteristics on trains to ensure they can always stop before a stop signal even under adverse braking conditions. ⢠Traffic and operations data. For passenger trains the data may include the planned schedule and statistics for average schedule deviations due to factors other than interference from other rail traffic, such as over-staying time at a station stop and operating delays outside the line segment being analyzed. For freight trains, many of which are likely to be unscheduled, arrival time at the line segment must be represented by a probability function. The modeling process can start once the train, infrastructure, and operations data have been entered into a modelâs database. The process involves first using passenger and freight opera- tions data to initiate trains entering the line segment in chronological order. The model may use Monte Carlo randomization methods to represent the uncertainty in freight train operations. Then, train movements are simulated by the TPC function of the model applied to each train, and the dispatcher simulator function of the model is used to resolve operating conflicts as they arise. The model can also introduce operations disruptions, such as slow orders and delays due to external factors, using data representative of typical operations. The result is a detailed description of the movement of each train through the line segment. It is usually necessary to run the simulation for several days to fully capture random effects, typically a week to a month (7 to 30 days). A highly detained simulation involving hundreds of train movements per day over several days or weeks can be run in a matter of a few hours, or less. A range of operating statistics may be derived from each model run; the statistics include trip times and trip time variability over the line segment and average delays to each type of train. Other outputs include string line charts, animations, locations where delays occur, and train speed graphs. A train speed graph is shown in Figure 3-6.
34 Capacity Modeling Guidebook for Shared-Use Passenger and Freight Rail Operations Normally the initial and final periods of the simulation are discarded as unrepresentative. These warm-up and cool-down periods can range from a few hours to a day each. These data are compared with the requirements for each type of rail service to determine whether or not capacity is adequate. 3.5.1.2 Analyzing a Shared Passenger/Freight Rail Corridor The analysis of a real passenger freight corridor normally involves defining a series of rail traffic scenarios and determining capacity adequacy for each. Observed service deficiencies are then iso- lated for each such scenario and additional model runs performed to test alternative investment and operations approaches. The outcome of this iterative process is to define the best combination of inputs for increasing capacity and enabling each user of the facility to meet service speed and integrity goals, and to determine a fair distribution of investment costs between the users. Figure 3-6. A typical train speed graph. Source: North Sound Rail Operations Simulation, Whatcom Council of Governments, 2011.
Analytical Approaches to Line Capacity in Shared-Use Corridors 35 It is also necessary to consider planned or likely changes in passenger and freight traffic over a term of up to 20 years. This is because track access agreements between freight railroad and a passenger rail operator should have a long life to avoid frequent renegotiations. FRA also requires 20-year service and infrastructure assessments as a condition of federal funding participation in such projects. Only by first taking a longer view can private and public sector stakeholders be assured that early phase commitments will not be wasted as a more mature service structure evolves. While each corridor will need an analysis plan tailored to local circumstances, a typical sequence of capacity model runs for new or additional passenger rail service would be as follows: 1. Base Case. The base case simulation is of present passenger and freight rail traffic and other conditions on the corridor under study. Model results are compared with actual corridor per- formance over a minimum 7-day period, and are used to calibrate the model for this specific corridor. The calibrated model is then available for further capacity analyses of the corridor. 2. Freight Traffic Growth Case, without Capacity Improvements. This case includes estimates of freight traffic growth, usually for the 5-, 10- and 20-year time horizons, to indicate when addi- tional capacity would be needed to maintain freight service quality. Generic freight volume growth estimates (such as the Freight Analysis Framework from FHWA) are used to develop longer term projections in the absence of specific market and lane data. This case can actually be a series of cases, depending on freight growth assumptions and the timing of same. 3. Freight Traffic Growth Case, with Capacity Improvements. This case will be guided by the results of Case 2 showing the locations and types of capacity problems, if any, that need to be corrected to maintain service quality. Like Case 2, it can be a series of cases, depending on the freight growth assumptions and timing of same. 4. Initial New Passenger Service Case, without Capacity Improvements. Additional passenger train trips are added to the base case and results reviewed for both passenger and freight traffic. If both still meet service requirements, then no immediate capacity improvements are required. If service requirements are not met, then analysis of capacity improvements is required, as in Case 6. 5. Initial New Passenger Service Case with Capacity Improvements. This case explores candidate capacity improvement options, guided by the results from Case 4 as to where and when capacity problems (e.g., delays) occur. The results will allow the analyst to select the most cost-beneficial improvements for passenger service. 6. Ongoing Analysis Cases. These cases combine new or additional passenger service with expected growth in freight traffic to determine what capacity improvements are needed to achieve desired service quality goals for all traffic types using the corridor. The cases will consist of various scenarios with differing assumptions of passenger and freight traffic on the corridor. While capacity analysis is essential for planning improvements on a busy rail corridor, it does not provide a complete answer. Capacity analysis does not take into account all factors that must be taken into account in decision making. Most models do include impacts for entering and exit- ing freight terminals, where a lack of capacity can affect adjacent main line segments. Furthermore, the incidence of service disruption due to track maintenance or slow orders, as well as unplanned events, can be underestimated. That noted, it is possible to simulate a number of conditions that may affect operations over a specific corridor, such as a track outage due to weather conditions, signal failures, train coupler failure, etc. The goal is to test the recoverability of the operation given a fixed rail infrastructure. Also, capacity analysis in part is an art, where experience of past analyses and actual perfor- mance outcomes will influence the interpretation of results. Openness and good communica- tions among all stakeholders is essential to acceptance of results and buy-in by all.
36 Capacity Modeling Guidebook for Shared-Use Passenger and Freight Rail Operations 3.5.2 Descriptions of Individual Capacity Models This section describes various capacity analysis tools that rely on the simulation of rail opera- tions over a selected rail corridor. 3.5.2.1 NCFRP Web-based Freight-Passenger Rail Corridor Project Screening Tool The National Cooperative Freight Research Programâs (NCFRP) Web-based Freight-Passenger Rail Corridor Project Screening Tool is a preliminary screening tool to evaluate the effects of adding new passenger rail service to existing freight or shared-use rail lines. Also known as the Shared-use (SU) Tool, it is designed to meet the need of public agencies with limited resources in identifying rail corridors that merit further investigation as candidates for shared-use service. Representatives from Class I railroads, the Association of American Railroads, commuter rail operators, state Departments of Transportation and the Federal Railroad Administration have overseen the development of this tool. The SU Tool uses a refined TPC, incorporating graduated tractive effort and dynamic braking curves to reflect train-handling practices. Resistive forces on the train are calculated on a car-by- car basis every 500 feet or less, accounting for changes in terrain and track curvature. The tool implements a deadlock-free dispatching algorithm to closely mimic actual train operations on complex, shared-use territories. Slow orders and track outages for maintenance and inspection may also be incorporated into the simulation. The tool complies with the âRailroad Operations Analysisâ described in the FRAâs 2005 guid- ance manual for rail corridor planning (Reference: Railroad Corridor Transportation Plans: A Guidance Manual, Office of Railroad Development, RDV-10, Federal Railroad Administration, July 8, 2005, pp. 14-15). Operations over corridors are simulated with specific topographic detail, speed zones, and train schedules and their corresponding variances. As trains are simulated over alternative paths on multi-track corridors, operational string lines over a 24-hour period are developed for analysis purposes to test alternative design strategies. Results are identified with minute-by-minute train metrics on string line and block authority visualization. Train perfor- mance is measured in minutes of delay by individual train and by train type: passenger, freight, intermodal, and other types as defined by the user. The tool evaluates the capacity effects of track infrastructure improvements and scheduled track outages. The SU Tool has a meet-pass logic that allows it to automatically resolve conflicts of opposing trains on the basis of priority, as well as allowing high priority trains to pass (overtake) slower trains. RTC, described herein, has a similar capability. As with all simulation models, the reliability of output from the web-based model is a product of accurate track infrastructure data and train consist and train schedule information. Track survey data can be imported directly into the tool or entered using a graphic track visualization screen. Default equipment libraries help users develop train consists, and other input screens track train schedules and time table routes. 3.5.2.2 Berkeley Simulation Softwareâs Rail Traffic Controller A widely used model for shared-use passenger and freight operations is the Rail Traffic Con- troller program developed by Berkeley Simulation Software. It is a computer simulation model that mimics human dispatching decisions that would be made to send trains through a rail subdivision and/or network. In addition to evaluating train movements, the model has the abil- ity to estimate the impact of changes to rail infrastructure and train movements. Both cost and performance are continuously recomputed for a given track configuration to minimize cost of
Analytical Approaches to Line Capacity in Shared-Use Corridors 37 delay for trains involved. RTC is commonly used to develop operating plans, diagnose bottle- necks, recommend schedule changes, evaluate various improvements to the rail infrastructure, and assess the impact of adding new trains to the corridor. The model has been used for providing potential locations for main track improvements, sid- ings, turnout speeds, turnout locations, train control system improvements, and train operation changes. It is the most widely used capacity planning/simulation tool used in the railway industry in the United States. Its use of corridor animation has allowed technical and non-technical users to understand and comprehend the corridor operations and the needed infrastructure improve- ments. All seven Class I railroads in the U.S. plus Ferromex in Mexico and Amtrak have this soft- ware in-house and require their consultants to use it when dealing with their operations. Having the standardized model already in place, with the basic corridor information within the RTC data- base, reduces the effort and time needed in conducting detailed evaluations of a corridorâs capacity. RTC, as other sophisticated models, incorporates a Train Performance Calculator that deter- mines the minimum run time of a train between two points, taking into account the alignment, gradient, allowable speed of the track, the horsepower and tractive effort of the locomotive, the tonnage, length and make-up of the train, and the specific physics of energy and resistances in having the train move across the corridor. It is the key component of the simulation system that replicates the actual movement of a specific train over a specific corridor. The resultant perfor- mance calculations are then incorporated by the Train Dispatch Simulator (TDS) that replicates and simulates the movement of all trains over a corridor. Whereas the TPC evaluates the perfor- mance of a specific train, TDS projects dispatch management decisions for all train movements and interactions over a corridor. The simulation process to test the effect of adding trains over a corridor utilizes standard railroad capacity planning procedures. That is, RTC is run first with the base case or existing train pattern operating over the corridor. The base case includes all current freight traffic, mixed freight and passenger trains, and local freights. Subsequent simulation runs test the effects to the overall operations under increased traffic patterns (viz., planned freight traffic growth and intro- duction of passenger traffic). This testing allows for the determination of operational effects that occur to existing freight traffic, while prioritizing passenger trains. It also allows for quantifying average train velocities and delay statistics as line capacity under existing and proposed railway track conditions is consumed. RTCâs chief attractiveness for freight railroads is its meet-pass logic, reflecting priority-based opposing train conflict resolutions. This is indeed the way railroad dispatchers resolve opposing train conflicts, so railroad executives have faith that RTC can accurately simulate their opera- tions. Also, its graphical display of results, where a viewer can watch a trainâs progress along a rail line, including delays, are useful in communicating the analysis and its implications to decision makers. Another advantage is that most Class I railroads have created an RTC database for most of their main lines. This reduces the cost and duration of capacity analyses on these lines: only data on proposed infrastructure and operations changes needs to be entered before running a simulation analysis. RTC base case assumptions and output may also be calibrated against existing operations as a means of improving the credibility of future case service/infrastructure scenarios. All this capability comes at a cost. RTC analysis requires extensive data gathering and labor hours for the analysis. RTC licenses for using the program are also expensive to acquire. 3.5.2.3 SYSTRA Inc.âs RAILSIM Another widely used simulation program available for use is SYSTRAâs RAILSIM Simula- tion Software Suite which is used to model and analyze operations on the most complex rail networks, including transit (light rail and heavy rail rapid transit), freight rail, commuter rail, and mixed main line railroad traffic. Though not as popular as RTC within the North American
38 Capacity Modeling Guidebook for Shared-Use Passenger and Freight Rail Operations rail industry, it is used by commuter and passenger agencies due to its ability to simulate train movements within controlled terminal areas. The RAILSIM package includes simulation and design capabilities for many types of train control systems, including fixed block and moving block systems. RAILSIM support for communications-based train control modeling and analysis includes flexible menu-driven inputs for buffer distances, communication times and system latency, under speed settings for typical operation, and guaranteed brake rate settings. The RAILSIM database also supports site specific definition of re-localization beacons, the wayside-to-train communications devices that serve to reset any accrued error in the determination of the current position of each train on the track. As with other complex models, RAILSIM uses a TPC as the basis for train operations over a defined corridor. Following the TPC development and network setup, complete simulations over the corridor can then be run testing the operations and quantifying the capacity under new signal and/or infrastructure design under various âwhat ifâ analyses. While popular for transit planning and frequency commuter line analysis, its lack of meet- pass logic for resolving train conflicts based on priority limits its attractiveness for analyzing unscheduled freight operations. 3.5.2.4 CANAC Inc.âs RAILS2000 Another program available for use is The Railway Analysis and Interactive Line Simulator (RAILS2000) model from CANAC Inc. The program replicates the operations of a corridor and tests the effects of changes (infrastructure changes and/or traffic changes). The software is owned by CANAC Inc., currently a wholly owned subsidiary of SAVAGE Companies. The model, originally developed by Corporate Strategies Inc. (CSI) of Washington, DC, was acquired by CANAC in 1999. RAILS2000 is an event-based simulation model and, as with other models, it contains a TPC that drives the movements of individual trains within the simulation. The software also contains a TDS which simulates the dispatching and control of trains over a defined route or network of lines. The TPC evaluates the performance of a single train over a given track, whereas the TDS is used for multiple trains over a network of tracks. The software is logic-based, with optimizing capabilities deliberately restricted to emulate real-world limitations of train dispatching. The model is capable of handling multi-track main line corridors and signals for both freight and transit operations. The program provides a consistent, reproducible, and inexpensive procedure for evaluating alternative railroad line and terminal configurations and train operations. It is a powerful tool for quickly establishing train schedules (timetables), analyzing service reliability, examining capacity issues, evaluating impacts of construction and maintenance delays (including work on road cross- ings and bridges), identifying conflicts, and evaluating alternatives. Computerization of opera- tions analysis allows rapid, economical evaluation of a large number of complex alternatives. At one time, RAILS2000 was used extensively by CSX Transportation and Canadian National Railway, but was gradually replaced by RTC because as the Class I freight industry migrated to a common modeling platform. RTC also has superior simulation graphics that more easily enable translation of technical model outputs to lay audiences. Although now much less popular than the RTC or RAILSIM, RAILS2000 has been a proven tool used on select projects throughout North America. 3.5.2.5 Other Models Other proprietary models from consultant and academic groups in the marketplace exist, though detailed information about how they work may not be readily available. Some less sophisticated models incorporate various linear programming techniques to allow varying levels of corridor analysis following field calibration. Such models require basic train running times
Analytical Approaches to Line Capacity in Shared-Use Corridors 39 between nodes to be directly input and used as the fundamental operational criteria for cur- rent and proposed operations. Their value is that a relatively quick, high level analysis can be performed but a more detailed, data intensive approach will still be required in advance of any specific project designs. 3.5.2.6 Summary of Simulation Models Within the railroad industry, the main two simulation tools currently utilized are the RTC and RAILSIM packages. Both have proven their robustness and have incorporated physical principles of equipment and specific territory. At the same time they are comparatively easy to employ for technical and non-technical users. They have developed an extensive equipment database and sophisticated and proven train dispatch algorithms and train control systems. Animation has allowed the results to be explained and represented to non-technical users and thus enabled a bet- ter buy-in by stakeholders examining specific corridors. Whereas RTC has been embraced by the freight industry, RAILSIM is used primarily by the passenger/commuter rail industry. RAILS2000 model is also an excellent model with ease of use and accurate results, but it has been eclipsed by RTC as the model of choice by the freight rail industry due to its lack of operational animation. A model with limited use in complex situations, the NCFRP web-based SU Tool is a screening tool capable of evaluating rail capacity on shared-use rail corridors. It offers some of the same basic fundamentals that the RTC or RAILSIM offer and with a quick turnaround response, but it does not obviate the need for analysis with more robust modeling tools. It can be used in the early stages of a projectâs development in support of a basic, exploratory âfeasibility study.â The use of such tools as described above may be complex and time consuming. One advan- tage, however, is that use of a model already employed by one of the host carriers may save substantial time and resources where the corridor in question has already been encoded into the software platform. Much of the time and energy consumed in employing such tools relates to the initial setup and the high level of detail required in the physical plant description. Use of a standardized modeling tool allows the analytical process to focus on evaluating the operations alternatives and developing sound recommendations. Once the fixed plant database is in place (and depending on the complexity of operations) varying scenarios can be quickly and easily tested for a specific corridor. Animation and its internal logic are perhaps the greatest attributes that RTC brought before the rail industry. Its ability to animate train operations for viewing by management and non- technical stakeholders (rather than with string lines and mathematical reports) was key to its acceptance as the tool of choice by the freight railroads. Results presented with simple animation allow management and stakeholders to quickly understand the existing issues and recommenda- tions quickly. RAILSIMâs ability to simulate high frequency train operations in a controlled environment has been key in its acceptance by several transit agencies. In summary, a simulation model is simply a tool that allows the stakeholders to model a spe- cific corridor and to develop mutually agreed upon infrastructure improvements for the chosen operation. The acceptance by the rail industry of the RTC and RAILSIM models has permitted the current developers to continually reinvest in development opportunities to maintain tech- nological relevance when dealing with current and future train operations. Whereas the use of other models is continually evolving, no other models as yet bring to the table the same level of technical sophistication and stakeholder acceptance. The Class I railroads and many transit agencies have embraced RTC and RAILSIM as essential tools for planning and extracting maxi- mum productivity from their expensive and increasingly crowded track networks.
