S F MUNI s CBTC System A Brave New

S. F. MUNI’s CBTC System A Brave New World Presentation by Patricia G. De. Vlieg APTA Rail Transit Conference June 14, 2000 St. Louis, Missouri

TOPICS: u. Project Thumbnail Description u. Implementation issues u. Progress on project objectives u. Hindsight

MUNI Metro Service Territory Oakland EXISTING SUBWAY Embarcadero to West Portal: 5. 5 mi Sunset Tunnel: 0. 8 mi Embarcadero Montgomery Powell Civic Center NEW SUBWAY Embarcadero to Ferry Portal: 0. 5 mi NEW INTERLOCKINGS MMT Castro Crossover Duboce Portal Van Ness Golden Gate Park Ferry Portal MUNI Metro Turnback (MMT) Cal. Train Depot Giants’ New Ballpark Sunset Tunnel N Church Castro Crossover Castro L J Forest Hill West Portal ATCS Territory Zoo K MUNI Metro Extension (MMX) City College Surface Rail Green Rail Yard S. F. State Univ. N M BART

Why did San Francisco MUNI embark on this project?

Project Objectives PRIMARY u Obtain Optimum Throughput u Improve Safety u Increase reliability & availability, reduce maintenance SECONDARY u Provide inherently flexible operation u Provide fully automated control of new track areas u Enhance passenger information systems, and improve right-of-way intrusion security

Basic Project Deliverables u u u u u Replace old Signal System on 6. 3 mi of dual track, 3 dual portals Signal new subway extension to waterfront & “Ferry” portal Replace & enhance central control signal & scheduling systems Retrofit 59 Boeing SLRVs for ATCS Equip 77 Breda LRV 2 s for ATCS Provide enhanced system performance Interface with existing systems (SLRV, Wayside interlockings, PA, etc) Provide design, installation, test, verification & commissioning Supply training, documentation, spare parts, warranties & DTE Install & Cutover system without impacting revenue service

Phased Service Introduction u LRV 2 ä December 10, 1996 u ATCS w Jan 10, 1998 ATCS in existing subway (MMX shuttle continues) ä w Operation: all ATP functions, most ATO/ATS ATCS Shuttle Service on MMX/MMT Extension only ä w Interim Cab Signaling (existing system) 1 st train June 13, 1998 (weekends only) Integrated Operation of existing subway & MMT/MMX ä ä August 22, 1998 2 -week demo at 35 tph passed October 14, 1998 u ATCS Full Functionality (Substantial Completion) We are here ä Final Software Release: June 2000 ä ä 2 -week demo at 48 tph: July 2000 Reliability/Availability/Maintainability Demos: Summer, Fall 2000

How did the implementation go?

ATCS Shuttle Service Area SF Giants

So far so good…. Now let’s integrate ATCS with the MMX extension into the existing subway…. .

The Meltdown August 24 1998

September 2: After tense meetings in the Mayor’s office, ATCS was given one chance to perform: To eliminate unequipped trains (non-retrofit Boeings) from the system, running “pure” ATCS.

What went wrong?

Simultaneous system & service changes New Train Control System +1 st use of new subway extension & turnback +New service on “N-Line” out the new Portal +1 st day of “Proof-of-Payment” fare system +Elimination of “Reverse Riders” = Classic System Engineering Problem

Flawed Outreach to Patrons u Massive passenger confusion u Few PR personnel at problem platforms u Absence of Passenger “incentives” Train Control System Problems u Inadequate dual-mode throughput u High incidence of failed non-communicating trains u Design & Spec deficiencies: e. g. : stopping points & door-open scenarios u Passenger Information bugs

Is this an exceptionally exotic project?

Comparable Projects

Project Estimates, then & now Start date August 31, 1992 Original u u u u Completion: Aug 31, 1995 (3 yrs) Alcatel Contract (with options exercised) $48. 7 M Consultant Contracts (Safety $0. 4 M+Engr $2. 4 M=) $ 2. 8 MUNI support Budget $ 4. 96 Tax $ 2. 7 Contingency $ 5. 27 Total Project Budget Scope changes: w restructured fleet make-up w 2 nd generation Central Control w enhanced “dual-mode” signaling w multiple functional enhancements w phased implementation $68. 43 M Current est March, 2001(8. 5 yrs) $50. 8 M ($4. 37 M + $12. 5 M =) $16. 87 $ 9. 93 $ 2. 3 $ 0 $79. 9 M

Where is the project today? Is the project meeting its objectives? Is the system a “better mousetrap”?

Project Objectives PRIMARY u Obtain Optimum Throughput u Improve Safety u Increase reliability & availability, reduce maintenance SECONDARY u Provide inherently flexible operation u Provide fully automated control of new track areas u Enhance passenger information systems, and improve right-of-way intrusion security

pre-ATCS

Remove 20 unequipped cars

Project Objectives PRIMARY u Obtain Optimum Throughput u Improve Safety u Increase reliability & availability, reduce maintenance SECONDARY u Provide inherently flexible operation u Provide fully automated control of new track areas u Enhance passenger information systems, and improve right-of-way intrusion security

SAFETY KUDOS u u u Automatically enforced safe train separation Automatically enforced speed control Centralized Control train speed control w “go-slow” zones enforcement w Station & train holds w etc w u Detailed performance logs ISSUES u u u Emergency Brake Applications Non-Communicating Trains Lingering ATCS-Vehicle interface problems

Project Objectives PRIMARY u Obtain Optimum Throughput u Improve Safety u Increase reliability & availability, reduce maintenance SECONDARY u Provide inherently flexible operation u Provide fully automated control of new track areas u Enhance passenger information systems, and improve right-of-way intrusion security

May 23, 00 Jan 31, 00 July 2, 99 May 3, 99 Dec 31, 98 Sep 3, 98

Project Objectives PRIMARY u Obtain Optimum Throughput u Improve Safety u Increase reliability & availability, reduce maintenance SECONDARY u Provide inherently flexible operation u Provide fully automated control of new track areas u Enhance passenger information systems, and improve right-of-way intrusion security

CENTRAL CONTROL Vehicle Control Center (VCC) (Vital system controller) Central Control Operator Workstation: SMC + VCC

Julio Says: It’s Great! Much better than the old system !

Project Objectives PRIMARY u Obtain Optimum Throughput u Improve Safety u Increase reliability & availability, reduce maintenance SECONDARY u Provide inherently flexible operation u Provide fully automated control of new track areas u Enhance passenger information systems, and improve right-of-way intrusion security

EMBARCADERO MMT 1063’ to 1 st MMT switch Ferry Portal MMX

Project Objectives PRIMARY u Obtain Optimum Throughput u Improve Safety u Increase reliability & availability, reduce maintenance SECONDARY u Provide inherently flexible operation u Provide fully automated control of new track areas u Enhance passenger information systems, and improve right-of-way intrusion security

Train Arrival Message: At Door Open Next 6 trains in subway displayed & Next 3 announced on PA

Where are we headed? In the right direction…. .

Why was it so hard?

Anticipated Issues u SYSTEM BOUNDARY PROBLEMS w Open loop (mixed street & subway operations) ä ä w Interface Control ä ä w u SLRV retrofit, LRV 2 interface design Public Address system, etc. Dual Mode (simultaneous old & new signal systems) ä u Failed entries & Non-Communicating Trains Schedule adherence ATCS control of existing block system & signals STAFFING w Central Control skill set & training w Maintenance skill set & training PROPRIETARY TECHNOLOGY

Issues Less Well Anticipated u u ORGANIZATIONAL w “Cultural” shift & resistance to change w Ambivalent Management support w Heavy dependence on Consultants w Intensive Training requirements SPECIFICATION & PROJECT INITIATION LIMITATIONS w Software requirements & development cycle w Testing Requirements; extended night subway shutdown w Inadequate budget for support staff & contingency w Unrealistic project schedule

Maintenance issues u Major shift in technology u Difficulties diagnosing trouble “to the interface” u Configuration Management burden greatly increased u Increased importance of PMs u Training, training

Next Steps w w w w Acceptance of Final Software Release 48 -TPH Demonstration Reliability, Availability, Maintainability Demos Final Deliverables (Manuals, Source Code, etc) Claims resolution Contract Closeout: target Spring 2001 Final Cutover

Is CBTC/ATCS a better mousetrap?

Judge for yourself…. . . …. . at the APTA Annual meeting San Francisco, September 24 -28, 2000 This presentation will be posted at TSD. ORG