Guidebook for Planning and Implementing Automated People Mover Systems at Airports (2010)

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.

209 In the current state-of-the-art APM system planning there is a range of modeling tools that are applied during the pre- design phase of the work. These tools are all facilitated by computer-based analyses intended to study the operating conditions that the APM system will serve. This appendix addresses the two most common types of computer-based tools—spreadsheets and simulation models. Effective modeling tools can be spreadsheet based, which in the current state of the art are primarily applied to assess ridership and passenger flow conditions. Spreadsheets can also be applied to approximate train performance and oper- ating conditions, although their use for this aspect is becom- ing less common. Spreadsheets generally are built around calculations from mathematical methodologies that, for the two types of models discussed herein, are simpler and more static in the conditions analyzed. Simulation models are computer-based tools that allow a more comprehensive analysis of the complex, dynamic con- ditions of APM system operations within an airport environ- ment. Such tools simulate the movement of people and APM trains (as well as other transportation elements in the case of landside models) in sequential time steps overall or a portion of the characteristic day being analyzed. These analytically oriented simulation tools are not to be confused with three-dimensional visualization tools that are intended to primarily provide a lifelike presentation of what the airport APM facilities will look like. Such visualization tools usually do not perform any specific analysis of the type described below for simulation models. That being said, 3-D presentations of simulation model analytical results are becom- ing more common in the aviation industry, and the use of 3-D visualization software to present the simulation analy- sis results can be beneficial. The following descriptions of both spreadsheet and simu- lation modeling applications provide an overview of the range of analytical results that can be obtained. The discussion of spreadsheets focuses on the modeling of APM system rid- ership demand and the associated station flows. The simula- tion modeling discussion describes the modeling of ridership demand, passenger flow through terminal/station facilities, transit user experience, train performance, and system operations. Overall, the level of detail required in the planning process will be determined by the needs of the study, and the choice of the methodology to calculate station flow will be dictated by the study scope. Generally, the use of some level of spread- sheet applications will be likely for almost every planning study. Frequently, the additional use of simulation models is often warranted as the APM system project progresses through higher levels of advanced planning and concept/ schematic design. Spreadsheet Modeling The application of spreadsheet tools has been a common practice in the airport planning field over the past twenty-five years. The analytical power of modern spreadsheets has made this a very useful and practical approach to modeling station flows and the resulting design requirements, particularly in the earlier phases of planning. This methodology commonly is used for the modeling of system ridership demands and thus the flows at each system station. Flight Schedule Processing The nature of passenger flows within an airport is funda- mentally driven by the schedule of flights and the associated enplaning and deplaning activity for the airside concourses for the planning/design day (usually an average day in the peak month). A reasonably detailed flight schedule is pre- pared in spreadsheet form that represents the number of flight arrivals and departures at a given terminal or concourse within a given period of time. Depending on the level of detail required in the modeling process, the time period increments A P P E N D I X E Modeling

can be as small as 1 minute or as large as 1 hour. Further, the incremental part of the airport terminal that is defined for flight arrival and departure may be as small as a single gate or as large as an entire concourse. Variable rates of enplanement and deplanement can be used to define the passenger flows in and out of the airport facilities. Often, the size and complex- ity of spreadsheet models soon reach a practical limit; thus not all spreadsheets attempt to model a high level of detail. Passenger Flow Analysis Once a flight schedule and enplaning/deplaning dataset is defined for the planning day, the spreadsheet model can be expanded to calculate the distribution of the air passenger flows through the airport facilities. Spreadsheet models can encompass airside concourses, the terminal(s), and/or the landside facilities. Depending on the placement of the APM system and its functional purposes within the airport, the simplest models would typically focus on only the part(s) of the airport to be served by the APM system. A time distribu- tion factoring technique is then applied to spread the air pas- senger activity before the time of enplanements and following the time of deplanements. The time period modeled for the flight activity and for calculating the movement of air passen- gers through the APM system should be consistent. The simplest spreadsheet models do not calculate the com- plete distribution of passenger flows for each flight in small increments of time before and after the flight arrival and departure. Rather, a factoring technique is commonly used to represent typical peaking effects within each hour for the composite of all flight activity. Such peaking factors may be derived from other aspects of the airport planning process, or a similar airport/APM application may be a suitable source of empirical data for determining such time-related flow factors. In the next level of spreadsheet models, air passenger flows between the points of origin and destination within the air- port study area are derived by factoring the flight schedule data portion of the model. This spatial flow factoring is based on information such as (1) the percentage distribution of air- port access by mode/landside facilities, (2) the percentage dis- tribution of air passenger utilization of ticketing/baggage check stations and other terminal processing functions, and (3) the distribution of air passengers between airlines and their associated airside concourse inherent to the flight schedule/ air passenger activity database. Once the basic flows of air passengers and the origin/ destination trip data are calculated, spreadsheet models then calculate the distribution of other populations that are to be served by the APM. These could be (1) flight crews (with flight schedule related distributions), (2) escort visitors (with flight schedule related distributions), (3) airline/airport employees (with work shift related distributions), and/or (4) other populations at the airport, such as office workers in adjacent buildings. Once the spreadsheet model accounts for all of the dif- ferent population movements through the portion of the airport served by the APM, the flow of APM riders passing through each station during all time periods of the day can be estimated. These flows can be calculated for any time period that the base data are suitable to derive. For calculat- ing station flow time periods that are less than the resolution of the base flight schedule data time period(s), the peaking effects can be estimated for planning purposes through the application of peaking factors, as discussed previously. Station Activity Analysis The overall APM passenger flow estimating spreadsheet can include separate worksheets to calculate the key aspects of station flows and thus physical requirements. Alternately, separate, smaller spreadsheets can be used for each station. These models are normally developed only for the peak period at each station (which could vary by station) and focus on the following station requirements: • Boarding platform occupancy sufficient to establish level-of-service indicators and related minimum plat- form sizing. The models would be designed to distribute passengers among the station platform doors, either evenly or in a ratio to the proximity of each door to the entry/exit point(s). The space requirements for queuing at each door area can be calculated using average areas for each rider, which are based on level-of-service spaces from Fruin or other sources, as discussed elsewhere in this guidebook. Given the train lengths determined in separate analysis (see Section 8.3), the spreadsheet calculations can inform the planner if the postulated station queuing area (particularly width) is adequate for the level of service desired or if adjust- ments are needed. • Vertical circulation requirements for sizing elevators, escalators, and stairways. The spreadsheet model mod- ules (or separate sub-models) can be used to estimate the requirements for these vertical transport devices if they are needed in the station design. Volumes entering and leav- ing the station and using each entry/exit are estimated from earlier modeling. Assumptions are made concern- ing the percentage using each type of device. These can be based on data from other airport APM stations or factors used in the overall airport planning effort. Typically 5–10% will use the elevators (disabled and riders with strollers or many bags), another 5–10% the stairs, and the rest the escalators. Factors are available or can be calculated within 210

the model for elevator speed (car capacity can be a variable based on general station design), escalator capacity per minute (at no less than one passenger every two steps), and stairs. The spreadsheet model can then determine whether an initial set of devices is adequate, or how many are needed. As discussed elsewhere in this guide, spreadsheet models can also be used to calculate the NFPA-130–based (or building code) requirements for emergency egress routes. This could affect the number and sizes of the ver- tical transport devices in the station. • Circulation areas between the vertical transportation elements and the platform doors. Spreadsheet models can also be used to estimate the width of areas for passenger circulation, given volumes and assumed walking speeds. The station size and vertical transport devices that result from this modeling are then input to general station design drawings and facility cost estimates. Spreadsheet models typically allow a reasonable approx- imation suitable for planning-level studies of APMs, partic- ularly for conditions where the airport or areas served by the APM are not large or the ridership population types are fairly homogeneous. The more complex the composite flows of multiple passenger types and the more refined the time increment and/or the size and complexity of the airport APM application, then the greater the importance of applying a higher level of station flow modeling. Spreadsheets are limited (or at least cumbersome) for parsing the station flow data to the demands on a minute-by-minute basis. If capacity limits appear to be important in station design, spreadsheet models may be insufficient to define the requirements for specific station elements such as platform queuing area and numbers and sizes of vertical circulation elements. Simulation Modeling Passenger flow simulation models are often quite sophis- ticated, and several different models are available from dif- ferent consulting companies. Generally, airport simulation models are created to represent the spatial areas of the terminal and landside facilities through which passengers, employees, and escort visitors move. Within these spatial models there can be comprehensive representation of the dif- ferent levels that typically comprise the airport facilities, allow- ing the complete travel paths of the different populations to be represented. Usually, APM access and station models are only a part of an overall airport terminal model, although simpler simulation models, such as with boundaries at the station entry and exit points, can also be used. Depending on the capabilities of the simulation tool, the input data can be as fundamental as the entire airport’s flight schedule and air passenger enplanement/deplanement data for a given planning/design day. Alternately, the input data may be as extensive as pre-processed passenger flow data derived from spreadsheets that have already calculated basic passenger flows for each APM station. In the former case, the passenger flow models would typically encompass a larger airport-wide scale, and in the latter, the passenger flow models may only encom- pass the specific station(s) of interest. This range of model applications therefore presents opportunities for alternative levels of combined spreadsheet and simulation model use. Flight Schedule Processing As an example, Figure E-1 shows both the originating pas- senger activity with the associated time distribution for the passengers’ early arrival in advance of flight time compared with the corresponding terminating and connecting passen- ger activity for a given flight schedule. Note how the arrival distribution of the originating passenger is smoothed by the early arrival patterns, whereas the terminating and connecting passenger activity is very spiked in profile due to their imme- diate entry into the airport facilities upon the flight arriving at the gate. The simulation model’s processing of the flight schedule database and the assignment to the passengers’ travel paths will reflect these different patterns for originating, ter- minating and connecting passengers. APM System Ridership Depending on the way that these very different activity profiles are assigned onto the APM system (a function of the APM system configuration of terminals, landside elements, and the airside concourses), the simulation model derives the ridership loading on all links of the APM system. Figure E-2 shows the actual APM system ridership for a link that is car- rying primarily terminating passengers bound for baggage claim in the terminal—a common point of peak demand con- ditions for many airport APM systems. Figure E-3 shows the peak ridership conditions for all links, but with the added precision possible (but not displayed in this figure) of the calculation of the precise time of day for the peak ridership load condition. In the example shown in the figure, each link has its peak five-minute demand condition presented as an equivalent hourly demand—a fairly extreme peaking condition for planning-level analysis purposes. The respective peak demand for each link does not occur at the same time of day, so the data as presented is not in concurrent time between the links. Of course, any other time interval ridership demand analysis can also be conducted with the simulation model, depending on the needs of the planning study and the purposes of the ridership analysis. 211

Station Activity Analysis When the planning analysis has first predetermined the criteria/goals of the APM system capacity to be provided rel- ative to the peak demand conditions, the simulation modeling tool can then assess the impacts on the passengers waiting in the station, both during the peak demand interval and through- out the day. Figure E-4 shows the results of a simulation model’s 24-hour accumulation of station waiting time— i.e., the time until the passenger was able to board a train. As shown in the figure, some stations have conditions where pas- sengers cannot board the first train to arrive in the station and they have to wait for the next train before they can board. This can be caused by either the station having extremely high 212 Figure E-1. Passenger trip processing and assignment from the flight schedule.

213 0 2,000 4,000 6,000 8,000 10,000 12,000 5 n o de sa B )hpp( pihsredi R k niL - sl a vr et nI et u ni M Terminating Passengers Connecting Passenger Employees Total Link Volume Figure E-2. Typical APM system ridership data from simulation analysis. 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,000 )hpp( e m ul oV k niL Station Figure E-3. Peak link volume (pph) equivalent hourly flow rate by station, based on the peak 5-minute interval throughout the day. Link demands can be analyzed by the simulation for precise times of day (equivalent hourly flow rates for peak 5 minute intervals not concurrent in time).

demand during peak intervals and/or the station having a location that is significantly impacted by operating condi- tions where some trains arrive completely full, not allowing all passengers to board. Based on the simulated ridership, associated with pedes- trian flows and system capacity limitations, models are used to assess the dynamic station operating conditions that result and the related level of service provided within each element of the passenger’s travel path—the passenger experience pro- vided by the APM. For purposes of APM system passenger flow modeling, these travel paths may include considerable details of the sta- tion elements, including: • Access and egress corridor locations and widths; • Station platform areas and individual boarding areas and locations for each APM vehicle position; • Numbers, widths, locations, and configurations of escalators; • Numbers, sizes, locations, and lobby configuration of eleva- tor banks; and • Numbers, widths, configurations, and locations of stairways. Figure E-5 shows animated modeling of an APM station at which all passengers alight to travel to baggage claim in the ter- minal above. As shown in the figure, the surge flow conditions are dramatic, with the full cars arriving at many times of the day. Figure E-6 shows an example of the demand conditions over several hours on a 1-minute time-step basis for the dual escalator set serving the station. The graph and legend clearly indicate the impacts of the heavy surge flows, with flow “In” and flow “Out” being accumulative over the time step, “Vol- ume” being the instantaneous occupancy at the end of the time step, and “PE Volume” being the equivalent number of pedestrians when the additional space claim of their luggage is included. It should also be noted that the flow capacity constraints to escalators is at the point where the pedestrian boards the unit, and this constraint in the simulation caused a bulk queue to build and dissipate over a few moments time whenever the surge flow rates exceeded the capacity of the escalator loading process. As long as the simulation showed that the queue dissipated in a reasonable period of time and in particular the peak-of-peak conditions did not remain until the next train arrived, the vertical circulation system was judged to be adequate for the demands imposed. The simulation model depictions of the station flow con- ditions are commonly analyzed for a complete 24-hour day in order to understand the complexities of the demand vari- ations for each flight arrival and departure complex. In large hub airports, these flight complexes often drive unique capac- ity requirements for each station such that they have peak demands occurring at times of the day that are different from the other stations. For example: • At stations serving a large component of airport and airline employees, the shift change patterns typically dictate the periods of highest demand; • At stations serving a particular airside concourse, the flight schedule characteristics of the airline(s) served on that 214 0 200 400 600 800 1000 1200 1400 1600 1800 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8 4 4.2 N um be r o f P eo pl e (D ail y) Wait Time (min) Concourse D Concourse C Concourse B Concourse A APM Headway = 1:39 Passengers Unable to Board the First APM Train to Arrive in Station Figure E-4. Station waiting time distribution (APM inbound).

215 Figure E-5. Heavy demand station with alighting surge flows to vertical circulation system. Reference Capacity PE Volume Volume Out In Escalator All Population Time (hr) 23222120191817161514131211109876543210 Co un t 5 m in 600 580 560 540 520 500 480 460 440 420 400 380 360 340 320 300 280 260 240 220 200 180 160 140 120 100 80 60 40 20 0 Figure E-6. Dual escalator demands for an APM station where most passengers alight to reach baggage claim through vertical circulation.

concourse will determine the unique demand patterns by time of day; and • At stations serving specific landside facilities such as rental car facilities, the patterns of mode split with respect to car rentals and returns by time of day, thereby changing the distribution of boarding and alighting demands. Operational Impacts Simulation models have other application aspects that can be very beneficial to analyzing the station facilities, vertical circulation elements, and short-duration ridership demand loads. Operational anomalies are common in the real world, and airports in particular must be continually adjusting to variations in the flight schedules as weather, air traffic, and equipment malfunctions change the patterns of demand on the APM system. The nature of simulation methodology allows random effects to be studied in detail, as opposed to the deterministic, static nature of most spreadsheet models. This stochastic analysis capability can be important when station capacity limitations are a concern. The processing power of simulation models allows randomization of, for example, the flight schedule data- base in terms of aircraft arrival times, which can substantially change the peaking patterns at APM stations. Having a ran- domized series of model runs with a suitable sampling of the distribution of peak demand conditions allows a more real- world assessment of the actual design capacity that should be defined for the station elements. Simulation models also allow the study of operational impacts resulting from service disruptions of the APM system. Even in the planning phase of the project, it can be important to assess the effect of failure-induced passenger accumulations within the APM stations. For example, if train operations are stopped for an extended period during a critical time of the day (e.g., 30 minutes without service during the peak hour of the day), then a station’s boarding platform may completely fill with people. Once system service is restored and trains with a completely full passenger load begin to arrive, initial operation must address the effect of arriving trains discharg- ing unusually large numbers of alighting passengers onto the station platforms that are already full of waiting passengers. Simulation models can assist with studies of such failure recovery operating conditions. A third aspect of airport operations that is driven by policy decisions is that of allowing luggage carts into APM stations, and in some airports allowing passengers to take them onto the APM system. Increasingly, planning studies are being tasked to evaluate the implications of luggage carts being taken into the station area, on the trains, and through the entire APM system. Such studies can greatly benefit from full sim- ulation modeling of the passenger flows at stations as well as on trains when the models include the greater space that lug- gage carts require for a percentage of passengers. Luggage cart aspects can also benefit from simulation model tests of differ- ent random mixes of such large space-claim conditions. In summary, simulation models can be used to not only test the operational impacts of random effects (e.g., flight schedule/ air passenger patterns) and variations of luggage cart space claim under normal operating conditions of the APM system, but also under equipment failure-mode operating conditions. Train Performance Analysis The performance analysis of the APM system automated trains/vehicles is an area of study to which simulation mod- els are commonly applied, even in a planning-level study. Although a full range of APM technologies is not necessar- ily studied in early planning work, a generic baseline technol- ogy is frequently defined for analysis purposes. The following aspects of the APM system are usually the focus of the per- formance studies: • Acceleration/jerk and operating speed—These elements of train performance are included in most simulation models of train performance. In a planning-level analysis, the most important of these is the operating speeds along the planned alignment of APM system. Some approxima- tion of the guideway alignment, configuration, and station locations, and in particular the curve radii that are possible given the alignment, are typically the controlling factors in the maximum operating speed at which the train can progress along its route. Figure 7 shows a graph of the train performance results for a given link (defined as the guide- way between sequential station stops). In the particular simulation mode shown, the allowable operating speed is shown in accord with the guideway geometry constraints and other operating considerations. The train’s acceleration/ deceleration response to the allowable speed compensates for the trailing end of the train clearing the reduced speed zone before acceleration occurs to a higher speed. • Power and energy—The APM vehicle propulsion system and the related train resistance parameters are usually part of the input data for the simulation model. When a generic baseline APM system is defined for planning study pur- poses, the simulation of train performance provides power and energy consumption estimates that are very useful to establish capital cost and O&M cost estimates. Figure E-7 also shows power consumption for train progress along the selected link. • Round trip time—The most important end product of the performance simulation is the determination of the round trip time for the specific alignment, guideway configura- tion, and station locations. The round trip time is used in 216

the planning process to determine the operating fleet size and the throughput capacity for a given number of trains/ vehicles in service. System Operations Analysis The final area for which simulation models are often used in the planning phase of study concerns the analysis of the whole system’s operations. Although some aspects of system operations are not necessary to address in planning levels such as for master planning studies, higher levels of advanced planning do require this analysis when APM projects are reaching a program definition stage with budgetary cost esti- mates, programmatic sizing of facilities, and protection of right-of-way. Given below is a brief overview of system operational fea- tures that advanced simulation models are capable of apply- ing within the analysis process: • Testing of supervisory control functions—There can be some APM operational conditions that require certain con- straints and management functions to be imposed by the automatic train supervision system. When this aspect of automated operations is recognized as important during the planning of the system, then there can be substantial bene- fit to a simulation-based assessment of these ATS functions. Some examples are headway management routines that continually work to even out headway perturbations follow- ing operational interruptions, and “station-ahead-clear” controls that prevent a following train from leaving a station until the train in front has cleared the station ahead. • Emulation of moving-block train control systems—The operational benefits of moving-block control systems has resulted in a whole new set of train control products to be offered in the APM market place. The ability to operate trains closer together, compressing headways to absolute mini- mums, can be a matter of interest, even during the planning phase of the project. Figure E-8 illustrates the way that the simulation model continuously calculates the minimum safe stopping distance of each train as a function of its operating speed at each point in time. In the figure, the colored target point in front of the train represents the moving-block pro- tection, which when encroaching on a train ahead would cause the following train to slow down. • Demand responsive dispatch—The advancement of PRT technology to the first implementation at an airport brings the consideration of this operating mode to the forefront of planning studies. The simulation of the very dynamic oper- ating conditions inherently imposed by PRT application requires the emulation of a demand-responsive dispatching of vehicles—meaning that vehicles are not sent into service until there is a passenger demand imposed from a specific 217 Power Velocity Grade Track Speed Airtrain: P4 IB-TCI Distance (ft) 1,4001,3001,2001,1001,0009008007006005004003002001000 Sp ee d ( mp h) 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 Power (kW) 350 340 330 320 310 300 290 280 270 260 250 240 230 220 210 200 190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 Figure E-7. Train performance simulation station-to-station link results.

218 origin station to a specific destination station. A second aspect of complexity for the simulation model to analyze is the management of empty vehicles, including their place- ment in strategic storage locations in or near the stations where higher future demand is anticipated. When complete APM system operations are simulated, the model allows a rigorous testing of changes to system capacity through variations in train performance, system alignment/ configuration, and train size. For simulation models that also combine the system operations with the modeling of ridership and station passenger flow, these operational changes can also be used to evaluate the resulting impacts on level of service and localized overload conditions within the APM station facilities. Most importantly, the system operational simulations over a 24-hour day provide important data such as vehicle operating miles and operating hours and power consumption. Simulation models are also very useful for conducting optimization studies of the operating fleet size. While it is true that a full opera- tional analysis is not required for the planning of many APM systems, for the more complex dynamic conditions—such as with the demand responsive dispatching of vehicles—the simulation modeling of complete APM system operations is very important. Figure E-8. Simulation of a moving-block train control function.