University of Twente Student Theses


Simulatie railgoederenvervoer tweede Maasvlakte: een onderzoek naar de dynamiek van de spoorprocessen op eeen volledig bebouwde en benutte Maasvlakte in het jaar 2033

Stive, F.D. (2008) Simulatie railgoederenvervoer tweede Maasvlakte: een onderzoek naar de dynamiek van de spoorprocessen op eeen volledig bebouwde en benutte Maasvlakte in het jaar 2033.

[img] PDF
Abstract:The transhipment of containers in the port of Rotterdamrises annually. However it is not possible to growany faster because the port has literally reached its limits. Maasvlakte 2 should provide outcome. Stevedores settling on Maasvlakte 2 are contractually committed to transport at least 55%of the hinterland volume by rail and barge. On the basis of future developments in rail cargo, the Rotterdam Port Authority has estimated what the infrastructure needs on the Maasvlakte in the year 2033 will be. However these estimations are based on fairly basic analytical calculations and do not provide any insight into the systemdynamics. The need towards further insight forms the basis for this thesis. The calculated infrastructure needs in combination with expected cargo flows are used in a dynamic simulation model to compute the exact conditions under which 20% of all hinterland cargo fromtheMaasvlakte can be transported by rail. For the exact location of the terminals and shunting yards on the Maasvlakte, use has been made of the Maasvlakte MasterPlan version 3.3.1. Four train concepts (all container trains) are introduced for transport to the hinterland. A distinction can be made into shuttle trains (direct shuttle, hub shuttle, feeder shuttle) and block trains. The length of the trains in the model is 650 meter. An acceleration of 0.2 (m/s2) and a retardation of 0.5 (m/s2) are used in the model. The tracks on the Maasvlakte are separated into segments of 1,000 meters and for safety reasons only one train can be situated in a segment at once. The 8 rail terminals on the Maasvlakte are merged into 3 terminals in the model. The two shunting yards South and West are included in the model as well. The Maasvlakte is regarded as an integral 40 km/hour zone. Several scenarios have been examined, by varying the different model parameters. The five model parameters that have been examined are: (1) The distribution of cargo over the various train concepts. (2) The different routes of the trains over the Maasvlakte. (3) The number of trains during normal hours and during peak hours. (4) The different process times at the shunting yards and terminals. (5) The policy rules for the internal traffic. In order to quantify the model output four criteria are used: (1) Analysis of the traffic flow around nodes. (2) Analysis of the number of trains on the yards/terminals. (3) Analysis of the cycle time and waiting time of the trains. (4) Analysis of the systemthroughput. The simulation shows that the deployment of direct shuttles, feeder shuttles and block trains accounts for a significantly higher need of yard tracks (53 56) in comparison to the deployment of direct shuttles, feeder shuttles and hub shuttles (35 37). In the current MasterPlan a reservation has been made for 43 yard tracks, leaving a shortage of 10 13 tracks if block trains are to be deployed. It is therefore advised to avoid the use of block trains as much as possible. iv Summary (English) | Havenbedrijf RotterdamN.V. It is expected that the amount of needed yard tracks can be lowered by 8% to 20 % by making trains interchangeable between the two yards. The track layout however is currently not sufficient to accomplish this, further research is advised. The utilization of the two yards greatly fluctuates fromhour to hour. By reducing the variance in the visit ratios, a reduction in the amount of needed yard tracks is expected to be obtained. However it is not possible to reduce the variance by using static operation rules. It is advised to further research the possibility of variance reduction through the use of dynamic rules, in order to bring down the number of yard tracks needed. The mere deployment of feeder shuttles causes an extra need of yard tracks up to 7% in comparison to the scenario in which direct shuttles, feeder shuttles and hub shuttles are deployed. The mere deployment of direct shuttles has a positive effect of 8% on the number of yard tracks needed. However both the deployment of solely direct shuttles and the deployment of feeder shuttles have a significant lower impact on the number of yard tracks needed in comparison to deploying block trains. The use of the same train concept has a positive influence on the distribution of the cycle times and waiting times of the trains. It is recommended to work with as few as possible train concepts. The train intensities at the nodes depend on the chosen model parameters. In practice it shows that on average 80% of the trains can pass a node unhindered. An important model assumption however is that trains drive from segment to segment. It is unclear whether such a system is practicable, further research is recommended. The average cycle times of the various train concepts in the simulation scenarios are fairly constant. In a peak hour, the average cycle time per train concept increases 6 to 40 minutes over an average of 7 hours (compared to a normal hour). Once a peak hour (2.2 times more trains than in a normal hour) is simulated the standard deviation of both the cycle times and waiting times significantly increases, in some cases by a factor 9. The feeder shuttles have the longest waiting times and are most sensitive to disrupting to the other processes. Research has to be done to figure out what amount of variance in cycle times is workable/accepted by the market. It is further recommended that no more than 8 trains per hour are to be accepted on the Maasvlakte. This will avoid conflicts between trains departing and arriving fromthe shunting yardsWest and South. Finally it is stressed that the number of yard tracks mentioned in this study should be seen as a lower limit. It is expected that the number of tracks in practice will turn out to be higher. The terminals in the model (almost) never form a bottleneck, reality will show otherwise. Due to poor planning and control the amount of needed yard tracks will rise. In order for trains to depart on time, slack time will have to be scheduled to account for the deviation in the cycle times. This extra slack will partly be spend on the yard tracks which further increases the number of tracks needed. On the basis of above it is recommended to investigate how the model performs compared to an initial planning.
Item Type:Essay (Master)
Havenbedrijf Rotterdam N.V ,World Port Center (WPC)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:85 business administration, organizational science
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page