System Dynamics Simulation in the Stoca Project

Here a short description of the four system dynamics simulation studies conducted in the project is provided. For the detailed analyses see the Research Report.  

 Research_Report_223.pdf

The simulation scenarios are based on a literature review and case studies conducted in the Gulf of Finland region in different ports and railway yards. According to previous research the functionality of maritime transportation is affected by the form co-operation and information exchange between the parties involved in the system. Special risks identified for international ports include foreign containers and recreational vessels.

Based on our case studies in the Gulf of Finland region different ports and railway yards have differing risk profiles depending on the infrastructure and cargo handled. Special kind of risk in the region is connected to oil disasters at sea or spillage of railway wagons, as a large amount of Russian oil and liquid bulk is transported via Finland and Estonia. The location of each scenario is presented in the figure below. The four simulation scenarios included are: 

1) Oil Spillage at Sea near Kotka          2) Oil Spillage in Muuga            3) Wagon Spillage in Kouvola                4) Wagon Spillage in Tapa 

 

Figure 1. The location of the four simulation scenarios in the Gulf of Finland Region.  

The models were constructed using Vensim 5.9 e. The four model files are available for free download below:

  1_FinlandSeaport_Kotka.mdl  2_EstoniaSeaport_Muuga.mdl  3_FinlandRailway_Kouvola.mdl  4_EstoniaRailway_Tapa.mdl

Vensim software is required to run the models. 

 

 First Scenario: Oil Spillage at Sea near of Kotka

In this scenario we are interested in studying the impact of insufficient hinterland capacity on the performance of sea ports in crisis situations. In the hypothetical case Hamina and Kotka sea port are going to be closed due to an oil spillage in the Gulf of Finland. The container traffic from Kotka is transferred to Helsinki sea port and we analyze what happens with different amounts of hinterland capacity. Hamina’s demand is assumed to be transferred to another sea port, so it is not included in the simulation model. 

In the simulation model both of the sea ports have an estimated daily demand and capacity. As soon as the Kotka sea port malfunctions, the sea ports start shifting some of the capacity from Kotka (for instance mobile cranes) to Helsinki. There is also a limit to the amount of additional capacity which Helsinki can absorb, which will also impact the potential movable capacity. In this scenario we assume that the sea port cannot take much additional capacity from other sea ports so the potential for additional capacity is small. The shifting operation will require some time (loading at the Kotka sea port, transporting, and finally installing at Helsinki) and in the simulation model all movable capacity has been moved after 15 working days. Also, 15 days before Kotka sea port can start serving ships again, the capacity is going to be transferred back to Kotka in a similar fashion.

In Helsinki a fixed amount of containers can be stored. In the simulation model the containers stay in the sea port for two days on average (during the crisis situation the containers will only spend a very short amount of time in the sea port and this way the average time at the sea port remains low) and this is taken into account with the storage module. 

There are two constraining factors in the maximum capacity of the sea port: available flow through the sea port (calculated with the help of hinterland capacity) and the actual cargo handling equipment. If the hinterland capacity is not large enough, the available warehouses for containers start to fill up. When the practical maximum capacity is reached, the sea port cannot handle any more ships as there is not enough space to store the goods. In this situation hinterland capacity defines the maximum capacity for the sea port. Overall the simulation model contains a lot of interactions and the total model is presented in figure below.

 

 

The system dynamics model in the first scenario.

We run nine different scenarios with the simulation model in order to evaluate the impact of hinterland capacity on the functionality of sea ports in crisis situations. Hinterland capacity will differ between 1500 (a little bit over Helsinki sea ports current demand) and 3500 containers (the demand of Kotka and Helsinki combined) per day. We will study the available storage space, the maximum capacity of the sea port, and the excess demand which cannot be handled by the seaport. 

We analyze the amount of aggregated excess demand, which the sea port cannot handle during the crisis, which starts on day number 90 and its duration is 60 days. All scenarios are presented in the figure below.

 

Aggregated excess demand with different values for hinterland capacity, TEUs.

As it is possible to notice from the figure, all scenarios have the same amount of excess demand for the first 30 days of the crisis. Approximately day 120 the scenario with the lowest amount of hinterland capacity starts to differ from the rest of the scenarios. The scenarios with a hinterland capacity of at least 2500 TEU do not differ between each other. In these cases the additional amount of hinterland capacity will not make a difference as free storage space does not run out during these simulation runs.

 

Second Scenario: Oil Spillage in Muuga  

In this scenario Muuga sea port is going to be closed due to an oil spillage in the port. 20 percent of the container traffic to Muuga (105 TEU per day) is transported via Helsinki sea port. From Helsinki containers will be transported to Paldiski on platforms with ro-ro ships. The amount of 20 percent of the containers is assumed to be sufficient in respect of security of supply. 80 percent of the containers will remain in the sea ports in Central Europe. We analyze the effect of having different amounts of platforms available for the sea transport between Helsinki and Paldiski.

In Helsinki the handling capacity is annually 500 000 TEU. In year 2009 it handled about 350 000 TEU. Helsinki has a fixed amount container storage at the sea port. In the simulation model the containers stay in the sea port for about 1-2 days on average. Muuga cargo handling devices are not moved, they remain in the port. Tallinn and Helsinki have at least two ro-ro connections daily. As Muuga is closed the ferries from Helsinki visit Paldiski port. A standard platform is assumed to carry two TEUs. Empty platforms are transported back to Helsinki. The turnaround time for the platforms between Helsinki and Paldiski is assumed to be two days. Although the same platforms are not returned directly, the number of platforms dedicated to the transportation loop between Helsinki and Paldiski equals the number of daily containers. In different simulations, the number of dedicated platforms receives the values from 10 to 110 with an increment of 10. The duration of the malfunction is 60 days.  

As can be seen in the figure below, during Muuga malfunction Helsinki is able to receive the daily 105 containers of Muuga without any problem.

 

Free storage space in Helsinki, TEUs.

However, in Estonian perspective consequences are more dramatic. If the amount of platforms is not sufficient, receiving the containers will take months. As container handling capacity in Estonia is concentrated in the port of Muuga, the system is vulnerable to local disturbances.

 

Third Scenario: Wagon Spillage in Kouvola

In this scenario a major node of the Finnish railway network, Kouvola, malfunctions due to methanol wagon spillage. As a result, no cargo or passengers can be transported between Kouvola and the Russian border. Many bulk materials are transported using railways and it might be difficult to find specialized trucks for this kind of material in a short period of time or it is not cost-efficient to use trucks. Furthermore, we assume a normal situation, where passenger trains are still given priority over freight traffic. Thus, the capacity estimates used for the railway network in the model do not resemble the maximum that can be acquired when recovering from a crisis situation.

The most important parts in the simulation model are the major railway yards connected to the malfunctioning node, Kouvola. All of the nodes have a fixed capacity. This is seen as a limited amount of storage space for trains in railway yards and sea ports.

In the simulation model we are only studying the impact on transit. Transit through Kouvola includes traffic from Hanko, Helsinki, Kotka, and Hamina to the border crossing railway stations. As Hanko and Helsinki, and on the other hand Kotka and Hamina, use the same route to access Kouvola, these locations are aggregated into two respective pairs. As such, Kouvola is going to be connected to Lahti (Hanko and Helsinki), Kotka / Hamina (mostly the same route, only a small divergence near the cities of Kotka and Hamina) and Russian border. There are delays connected to all of these routes, which have been taken into account in the simulation model. Also, each of the yards has a fixed capacity which can be reserved for temporary holding area for wagons. These two values create constraints for the system.  

The amount of material transferred between the nodes depends on the destination nodes’ free capacity, amount of material to be transferred and the amount of capacity in the railways available between the locations. The free capacity needs to be divided between the different origins, which needs to be taken into account as well when estimating the amount of material to transfer between the nodes. This is presented in the figure below.

 

Estimating the transportation capability of a single link.

The model boundaries are in increase of the amount of exports and imports in Kotka / Hamina, Lahti, and Russian border, as well as the final import and export when materials leave Finland. In this version of the model imports and exports are assumed to remain the same all of the time and the amounts are independent. In a crisis situation the available capacity could be affected by prioritization of the trains. In this study such measures are not assumed to be taken. 

In this model we analyze the impact of the length of the malfunction. It is going to vary between 112 and 448 hours. This equals one to four weeks in length. Following figure shows the amount of export transit lost.

 

Aggregated amount of missed exports, tons.

In overall it can be stated that a malfunction in the hinterland capacity will have heavy financial implications for sea ports. Especially export transit would suffer as most of the transit is conducted using railways. It might be possible to conduct part of the transit using trucks, but this is not cost-efficient and would still have a financial impact on the sea ports.

 

Fourth Scenario: Wagon Spillage in Tapa

Estonian ports carry a large amount of Russian oil transit. The oil is transported to the ports by rail. In this scenario the connection between Tapa and Vaivara is going to malfunction. In comparison to freight traffic, passenger traffic on Estonian railways is negliable. During the malfunctioning the leave values from Tapa to Muuga and Paldiski and vice versa is going to be zero. When this happens, the whole oil transportation stops and the rail yards starts to fill up in Vaivara and Sillamäe, but oil export will continue as long as oil storages last in the ports of Muuga and Paldiski. Because of the oil reserves in the tanks consequences of disruptions are experienced at a later time and have a shorter duration than the malfunction itself.   

The structure of the model is similar to Scenario 3 above, but there exists only one way traffic as Russia does not conduct import transit through Estonia. We are assuming bad winter conditions and Sillamäe is able to handle only half of the normal capacity. As it is possible to notice from the figure below, there are large differences between the scenarios with different durations for the malfunction.

 

Missed export during malfunction in Muuga, tons. 

If the port of Sillamäe would operate in optimal conditions, it should be able to handle all of the oil. It can be stated the port of Sillamäe can substitute the port of Muuga momentarily, but longer stoppages will be hard to substitute with the limited oil tank storage and oil pumping capacity.

 

 
Agent-based Simulation
System Dynamics Simulation