Transnet Freight Rail
TFR’s requirement was for a Resource Scheduling System for TFR Coal Traffic. A hybrid dynamic scheduling tool was developed with a high level of integration with TFR’s existing resource demand and tracking systems.
Following a soft go-live in March 2011 and full go-live in May 2011, TFR has achieved an increase of 9% in their coal volumes, a decrease in total loaded to offloaded duration of 3.07%, improved visibility, and improved identification of data inconsistencies in their base data as of the 1st April 2012.
Transnet Freight Rail (TFR) is a heavy-haul freight company dealing in general freight goods, iron ore and coal exports. TFR is the largest division of Transnet, and is South Africa’s sole national rail freight provider.
TFR was responsible for transporting the coal exports from the mines in the Vryheid area to Richard’s Bay Coal Terminal (RBCT) via Ermelo, and had been experiencing year-on-year annualised transported volume decreases from 2005, despite extensive investments and several other programmes.
Following investigations, it was determined that the existing scheduling system was unable to model TFR’s operations fully, and did not always adhere to operational restraints. The time to create a schedule via the system took roughly two days. The result was a plan which was generally unfeasible at the outset, and which became impossible to implement as the schedule commenced, resulting in manual intervention.
TFR wanted a routing and scheduling solution that would allow them to achieve the following primary objectives:
- Implement a modern scheduling system to meet TFR’s increasingly complex requirements;
- Reduce the amount of time to generate a feasible schedule;
- Effectively handle occupations and disruptions on the coal line, and be able to adjust the schedule to deal with these events;
- Improve the visibility and accountability of the scheduling system;
- Increase the coal volumes being delivered on a weekly basis in order to meet the proposed annualised targets.
After assessing TFR’s operations, a hybrid dynamic scheduling tool – PLATO.Rail – was developed which utilises simulation to generate feasible schedules from given load demands presented by the mines. It then ‘breeds’ solutions through 200 generations via a genetic algorithm to reach an optimized solution.
Main features of the project include:
Integrated Occupations Management System
Next Week’s Business – customer loads
Sprint – tracking and monitoring
Following an extensive change management programme within Transnet, and a soft go-live in March 2011 – in which PLATO.Rail ran in parallel with TFR’s existing system - the project officially went live in May 2011.
After launch, TFR began to achieve record volumes month-on-month, ultimately achieving an increase of 9% on the previous year’s annual volume from 62.2 million tons to 67.7 million tons.
As roughly 8% of new resources (wagons and locomotives) had been added to the network, there was an expectation that additional congestion would result in increased wagon cycle times - the amount of time from being loaded to being offloaded. A study was performed to compare a single month of operation from 2010 and 2011, with special consideration taken to ensure the closest possible alignment in both scenarios.
It was determined that, due to improved scheduling, PLATO.Rail was able to mitigate the cost of these additional resources and decrease overall wagon cycle time by 3.07%
The yearly amount of coal throughput has increase by 8.8% in the last financial year, contributing greatly to the 32% total increase in revenue to TFR.
Planning time was reduced from 2 days to 30 minutes, with a similar amount of time required to re-run the scheduler and amend existing schedules in the event of occupations or unexpected delays.
PLATO.Rail is expected to be rolled out to the rest of the coal network during 2012 and 2013.