MUNI METRO OPERATIONS HISTORY – CAPACITY AND DEMAND
Angelo Figone, Transportation Consultant; Former Chief Transportation Officer, Rail
The concept of a Market Street subway dates to Muni’s formative decades with a “ Report on Rapid Transit for the City of San Francisco With Special Consideration to a Subway Under Market Street” issued in 1931 by City Chief Engineer M.M. O’Shaughnessy. Subsequent studies throughout the 1940s and 1950s culminated with the establishment of the BARTD system and the collaborative effort to build a two-level subway with BART and MUNI rail operations. Begun in 1967, the subway construction was completed in 1977; however, initial Muni Metro (MM) revenue operations did not begin until February, 1980. Thereafter in phases, the existing surface streetcar lines N, K,L,M and J would transition from surface operation to surface-subway operation utilizing then-new Boeing-Vertol light rail vehicles (LRV).
Initial Capacity Design:
In 1967 – the commencement of subway construction – the five surface streetcar lines were scheduled for peak-period headways of 2-8 minutes with a peak-hour trip count of 72 producing a maximum capacity of 7920 volume (200 % load factor of 55 average seats/car). 100 cars were scheduled for the peak period. Initial planning prior to SLRV design and purchase called for 78 cars with an option for 14, but the final purchase totaled 100. It was assumed that replicating the 7920 peak-hour capacity would be accomplished with a 50% reduction in running time in the Market Street segment, a 25% increase in car capacity and a projected spare ratio of 20% yielding 80 maximum scheduled cars.
1980 Startup Capacity and Revisions
The phasing of PCC replacement with SLRV commenced with two subway lines: the N initially in February, 1980, followed by the K (Shuttle) in June, 1980. Surface operations continued on J, K/L and M. SLRV maximum scheduling of 80 peak cars allowed 36 on the K (St. Francis Circle-Embarcadero) and 30 cars on the full N line. The K (Shuttle) operation consisted of 4-car trains on a 6-minute headway and the N of 1-car trains on a 3-minute headway, resulting with 30 trains/hour or 120 second headway in the subway. The capacity of both trunks was 8160 ( J remained on the surface.)
Note: The scheduling of multiple-unit trains (2 car on N and 4 car on K) was a goal from inception. Maximizing train lengths and therefore maximizing subway headways mitigated running time delay inherent in the Embarcadero station stub-end turnaround.In 1981, all five (5) streetcar lines were converted to SLRV operation and initially closely replicated the former PCC headways. The limitations of the “Cabs Signal System” and the Embarcadero stub-end turnaround demonstrated that single-car trains resulted in 47 trains/hour and unreliable running times. In 1982, remaining single-car trains on the K,L,M were replaced with 2-car trains thereby reducing the hourly throughput to 38 trains/hour.
Enroute Portal Coupling/Uncoupling
Unreliability required further train reduction, and subsequently “enroute coupling” was initiated at both the Duboce junction for the J & N and the West Portal junction for K&L and M&L trains. Resulting train length at the Duboce Portal was 3-car and at West Portal was 4-car, both trunks on a 6-minute headway. Because of SLRV fleet availability – 101 car maximum peak requirement out of 131 car fleet -some coupling train lengths were less than 3 and 4 cars. Peak-hour train counts were reduced to 20 trains/hour with 180 second subway headway. In order to maximize train meets and resulting successful train couplings, a 2-minute dwell was scheduled at both portals. Surface train delays did occur as well as occasional coupling malfunctions preventing some scheduled couplings, but the resulting subway train counts did not routinely exceed 24 trains/hour or 150 seconds. The increase in the SLRV fleet from 100 to 131 allowed the 20 trains/hour to produce 69 car trips with 9384 capacity/hour (200% load factor of 68 seats/car). The delay incurred at both surface portals was perceived more positively than the running delay in the subway because of Embarcadero stub-end turnarounds.
ATCS and MMT
Conceived as a solution to subway delays and unreliable running times, two projects were planned during 1985-1990: A new multi-track turnaround east of Embarcadero (Muni Metro Turnaround/MMT) and a new Automated Train Control System (ATCS). The MMT would allow thru trains on the MMX (Muni Metro Extension to Embarcadero surface) to diverge from trains terminating at Embarcadero. It also allowed Embarcadero Station deboarding/boarding to occur without switching delays previously encountered and creating train stacking delays.
ATCS implementation would afford “moving block spacing” negating previous “Cabs system” train spacing requirements and operator train control. The ATCS goal was to increase train throughput and negate portal coupling. Note: SLRV1 Boeing cars were not all equipped for ATCS operation nor was ATCS capable of recognizing trains coupling at portals within ATCS limits. Additionally, Breda SLRV car length precluded 4 car trains at both West Portal and Foresthill platforms/stations. As a result of the 1997-1998 Breda fleet introduction, coupling could not occur as heretofore. With full ATCS and complete Breda fleet implementation, the resulting throughput was 40 trains/hour.
Subway Ongoing Delays
The startup of the T-Third rail line in 2007 resulted in an operations planning decision to utilize the subway Castro crossover and station as a terminal. The ATCS system allowed multiple-destination routing and train protection. Previously, a Castro Shuttle (S) service had been implemented during peak periods. Replacing the S peak trips, the T line did not increase the total trains beyond 40 trains/hour. The T-Third all-day Castro terminal requirement did often require supervised interventions at the Control Center and often resulted in delays to outlying trains. Consequently, by 2008 T-Third trains were thru-routed with K line trains and the previous peak-only S Shuttle was reinstated. Recognizing recurring subway delays, MTA initiated an ATCS software upgrade (THALYS) that would allow multiple train occupancy at stations. While not reducing station dwell, the running delay in awaiting station occupancy would be reduced.
SLRV Fleet and Capacity Increase
During each of the periods described here (1980-2017) MTA/Muni operations planning has been constrained by SLRV fleet availability and maximum scheduled peak car demand. The initial 80/100 (maximum schedule/total fleet) and later 103/131 and currently 120/151 has precluded service additions that measurably increase load capacity. Additional restrictions on short-line trips outside the subway has further decreased throughput. Concurrently, demand for Muni Metro service capacity has increased due to population and demographic changes.
The arrival, testing and phased implementation of a new Siemens SLRV4 fleet in 2018 affords an opportunity to increase load capacity without degrading subway reliability. In fact, the new Siemens fleet may provide improved reliability if new operating schemes are considered including portal coupling of trains. 3-car Breda consists have been run in revenue service in the MM subway, and 3-car Siemens consists (including enroute coupling of trains) should be tested within the ATCS/THALYS train control system as well. It is essential that this capability be achieved in order to increase MM capacity.