Paulo Vieira, Train Control Systems Expert, MRS Logística SA
After Brazil’s railways were privatised at the end of 1996, MRS Logística found that traffic on its 1 674 km network was rising at an extraordinary rate. To cope with this growth, the company invested in improvements to the infrastructure and rolling stock, which in many cases was run-down or in poor condition. However, little was spent on signalling and train control, other than minimal investment to keep the existing CTC in working order.
By 2003 we realised that just keeping the CTC going would not suffice to handle the forecast rise in demand, so MRS was faced with a choice: either invest in enhancing the existing CTC or replace it with something different that offered a higher capacity.
The CTC suffered from low availability and poor reliability of some components, which was not helped by the difficulty of obtaining spare parts for obsolete equipment. The level of safety was no longer deemed adequate, as train drivers had no back-up to ensure that a train was brought to a halt at a signal. Assessment of these issues pushed the decision towards complete replacement. Once that was agreed, we had to decide whether to keep the CTC concept and add cab-signalling with speed codes along the line or install communications-based train control where onboard systems would communicate with servers in a control centre and enforce train stops using radio links.
After analysing the options and their likely performance and cost, the company opted for CBTC. A Request for Proposals was issued at the end of 2004, but tendering took more than a year. A contract was finally signed with Alstom, EADS, and Accenture-Atan in March 2006.
Technical choices
During the tendering process MRS had to take some important technical decisions. One of the critical issues was whether or not to retain maintenance-intensive track circuits. Initially, we expected to be able to do away with them, but finding an equivalent technology to provide the same functions of track occupation, train integrity detection and broken rail detection was easier said than done. None of the alternatives was practical or economically acceptable, and eventually we decided to incorporate the existing track circuits in the new system.
Another issue was the need to ensure that trains could not proceed beyond their limit of authority or target location representing the safe limit of operation. This meant that the location of the front of the train had to be known in a vital way — and track circuits cannot provide this information. We investigated train positioning technology such as differential GPS, as used in the North American PTC programme, or balises and tachometers as with ETCS in Europe.
The choice fell on a simpler option that used the transition from one track circuit to another as a ‘virtual positioning balise’. Along with a vital onboard tachometer, this offered a good position location system which is similar to the ETCS approach. The process works in a logical sequence:
- the train moves from one track circuit to the next;
- the track circuit occupation event is detected by the field signalling Object Controller;
- the occupation event is time-stamped using GPS, and sent to the onboard ATC computer;
- the ATC processes the event, calculating the distance the train has moved since the actual occupation until the event is received, based on the GPS time-stamp;
- the ATC resets the train position, adjusting any tachometer error.
Communications technology
As there are no regulations in Brazil governing railway control communications technology, we were able to choose any system that was technically suitable. However, this freedom was tempered by the fact that analysing the best form of communications technology was the most complex part of the whole process.
Several issues exacerbated the problem. For example, the market offered a wide range of options, some of which were already available and some which would be available imminently. There was little historic information about the application of CBTC, and there were only a few consultancies with sufficient expertise able to help with evaluation of the proposals. Communications companies lacked exposure to the rail market or did not have experience of the applications we needed.
All this meant that evaluating the proposals was especially difficult as they all used different technologies and system architectures. After examining a GPRS data transmission option (similar to the ETCS standard GSM-R), Tetra (Terrestrial Trunking Radio), APCO25, and Data Radio systems, MRS selected a Tetra system.
System architecture
Known as Siaco, from the Portuguese initials for Integrated Operations Automation & Control System, the MRS train control system is equivalent to ETCS Level 2. However it is not ETCS compliant, because of the Tetra radio and the use of a specific signalling protocol and architecture without balises. It also has special functional requirements for the onboard user interface.
Siaco has four major components (Fig 1):
- an Integrated Operational Control Centre, responsible for overall control of the system and its interfaces;
- a Signalling & Control System, responsible for the signalling modules in the field (Object Controllers) and in the control centre (Safety Logic System);
- an Onboard Control System (SCB in Portuguese), responsible for the driver interfaces (onboard computer and voice radio) and guaranteeing the safe movement of trains (ATC); the SCB also includes an event recorder, provided by Accenture-Atan, that handles real-time monitoring, acquiring and processing data from onboard sensors;
- a Train Telecommunications System, handling voice and data communication for all trains using the Tetra radio backbone.
Besides the essential CBTC characteristics, Siaco is being built with a strong emphasis on integrating the different functionalities so that all systems related to train operations are managed by the IOCC. This means that managers can also use it for corporate tactical planning and overseeing maintenance. So it incorporates field systems such as hot axlebox detectors and power supply control systems for remote installations. Information from the different systems is monitored and compared, and alarms or warnings are issued, together with suggestions for the best action to take, to users on a single interface. This significantly improves the railway’s overall performance.
Another feature that once deployed we hope will lead to better performance is real-time locomotive telemetry. Together with real-time data transmission via Tetra radio, this should prevent most delays caused by locomotive failures — because data captured on board can be analysed remotely by locomotive experts at the IOCC.
Project deployment
After the contract was signed, the suppliers moved rapidly, and factory testing began in 2007 before a start was made on field deployment and testing. The system started operating in May 2008, but still in CTC mode as some of the SCB functionalities and components were not completely ready. Operation in CBTC mode began in August 2008 with some field training activities and more comprehensive field integration tests.
By December, Siaco had reached the ‘system stabilisation’ stage where commercial trains were running under CBTC on a 35 km pilot section on the Pombal – Guaíba branch, about 50 km from Rio de Janeiro, together with four shunting yards. At this stage the drivers handling the trains were assisted by technical staff from Alstom and MRS. The drivers were already trained, and theoretically could have operated the trains unaided. However, we agreed with Alstom-EADS that one more stage was needed to verify the stability of operations and to observe the reaction of the drivers and the dispatchers in the IOCC, especially when dealing with out-of-course events requiring the use of new procedures.
In early January Siaco was in use on six locomotives hauling service trains during normal working hours; this allowed for two to four trains a day to run on the pilot section in CBTC mode.
The project team quickly identified the need for adjustments that required further development, and the system stabilisation period was extended until the end of February. This allowed for the testing of a new system version deployed at the beginning of that month.
While operations are gradually being stepped up on the pilot section, deployment is proceeding in parallel with Object Controllers and TTS being installed across the network and SCB fitted to the locomotives. At the same time adjustments are being made by the technical team and checked by operators in the pilot section. Other technical staff are working in the factory and in the field with the aim of having Siaco deployed across the entire MRS Logística network by mid-2010.
Full commercial operations are due to start in the pilot section this month, and our strategy envisages that Siaco will be operational on 200 route-km of the busiest part of the network by mid-2009.
Deployment strategy
Deploying the equipment and checking it for safe operation is a very challenging task for the project team as it has to be done without interfering with operation of the railway. Trains run 24 h a day, 365 days a year, and traffic density is very high. Around 500 locomotives have to be fitted, together with any other powered vehicles that operate on the network. At the same time all drivers and IOCC staff have to be trained, and while all this is going on further development of hardware and software is taking place.
It quickly became obvious that the system could not be deployed in one go. The process had to be gradual, with tight co-ordination of training with software development. Deployment had to be sequential to avoid fragmentation, and we decided to equip the busiest areas first to test the system thoroughly and to gain experience as quickly as possible. Using these guidelines, five main installation phases were developed, with each phase divided into sections (map).
Fitting the onboard equipment is a long process, as the locomotives cannot be equipped simultaneously. Even with an aggressive programme, locomotives cannot be taken out of service simply to fit the onboard equipment as the process takes seven working days. Onboard installation has been carefully synchronised with maintenance intervals so that the equipment can be fitted when the locomotives are undergoing scheduled maintenance.
When operations began on the pilot line just a few locomotives had been equipped to run in CBTC mode. Many locomotives using the pilot section were not fitted, so the system had to be designed to work with CBTC and the original CTC functioning in parallel. An overlay arrangement has been adopted whereby the conventional signalling equipment remains live so that non-equipped trains can run through the CBTC area in CTC mode (Fig 2).
Lessons learned
Any project of this complexity demands detailed planning, strong teamwork, risk management and an intensive integration effort among the companies involved. In addition, several specific issues had to be addressed.
It is important to understand that existing systems will need to be incorporated in the new arrangements so that they can interact either during the deployment process or even on a permanent basis as part of the final operation. Detailed studies are essential if delays and bad choices are to be avoided. With the Siaco project, the problem was to incorporate the existing CTC system, and this led to delays and reworking of some parts of the scheme.
Similarly, operational processes need to be reviewed. Deployment of a new system is an opportunity to review existing problems and eliminate the limitations imposed by the old methods. We chose to implement Siaco as part of a change management project which prepared people for new working methods. In this way MRS was able to avoid delays, and the new processes actually contributed to an immediate improvement in operational performance.
The radio coverage plan and data traffic studies needed to be developed very carefully, with all communications requirements properly assigned. These must address both the needs of the signalling system and the railway alignment characteristics — particularly tunnels and curves. The studies should be conservative, otherwise additional resources may become necessary as the project develops. With Siaco, we found that additional base stations and carriers were required to achieve full coverage, which led to delays and additional negotiations with the suppliers.
Finally, we found it was important to validate the CBTC principles in the early stages. Validation is essential in any technical project, and we tried to address this, although some communications-related issues such as integration with the new signalling took much longer to resolve than envisaged.
Part of the process is the need to monitor developments systematically. Validation in the field requires monitoring capability to be powerful with specific tools and detailed logs as well as the presence of a knowledgeable technical team. These monitoring capabilities were not properly prepared when the Siaco field testing began, which caused long delays
