Jakub Montewka, Floris Goerlandt, Pentti Kujala
Aalto University, Dept. of Applied Mechanics, Research Group on Maritime Risk and Safety Espoo, FINLAND
Jari Haapala, Mikko Lensu
Finnish Meteorological InstituteHelsinki, FINLAND
MODELLING A SHIP PERFORMANCE IN
DYNAMIC ICE: PART I, TRANSFORMING DATA
INTO INFORMATION
The manuscript delivered: May 2013
Abstract: For safe and efficient exploitation of ice-covered waters the knowledge about ship
performance in ice is crucial. Although ice navigation has received substantial attention over recent decades, there is still no known modelling technique to predict ship’s speed in a dynamic ice field. In order to gain an insight into this process, we need to transform the available data into information first. Only then information can be used to develop new knowledge. This paper demonstrates how to transform still data into dynamic information about operation of maritime transportation system in ice-covered waters.
For this purpose, the data from the Automatic Identification System about the performance of a selected ship is used along with a numerical ice forecast model describing the ice field in the analysed sea area.
Keywords: Winter navigation; Ship performance in ice; Bayesian Networks
1. INTRODUCTION
Ship performance in ice has been given a lot of attention in the recent years, this increased attention has led to the development of semi-empirical methods that estimate ship resistance in ice, see for example and tools that simulate ship transit in ice, see for example (Naegle 1980; LaPrairie et al. 1995; Mulherin et al. 1996; Kotovirta et al. 2009; Su et al. 2010; Lubbad & Løset 2011).
These tools are used in detailed evaluation of required power for a ship at the design stage or for predicting her performance in ice at the operational stage; however the ice conditions which are modeled are often limited to the level ice and ice channel, and in select cases, the effect of ice ridges is taken into account, see for example (Kaj Riska et al. 1997; Juva & Kaj Riska 2002). But the effect of ice compression on ship’s speed has not been researched in-depth, thus it is usually expressed in a qualitative manner, see for example (Mulherin et al. 1996). However, Kaups (2012) has allowed the quantification of compression on ship performance using a concept of added resistance, simply produced by an ice sheet in contact with a ship’s hull.
The most common modeling practice in the field is to adopt quantity-oriented models, which describe the relation among a set of ice features and ship resistance and the related ship’s speed, see for example (Naegle 1980; Lindqvist 1989; Kaj Riska et al. 1997). Although these methods have been utilized for optimizing shipping routes in ice-covered waters, see (Kotovirta et al. 2009), there are still numerous issues which need further studies, for instance: the effect of ice compression on ship performance, or the quantification of the joint effect of ice conditions (level ice, ridges, compression, the relative angle at which ice reacts on a ship) that can bring a ship to a halt. Moreover, suggestions have been made to move towards probabilistic models, see for example (Kotovirta et al. 2009).
Therefore another modeling technique can be adopted leading to an event-oriented model, which reflects the conditions (ice features) under which an event of interest occurs (ship proceeding with very low speed or a ship getting stuck in ice). This type of modeling does not provide an insight into the physics of the process of ice breaking, but simply quantifies the joint effect of various ice features on ship’s speed. Moreover, if appropriate probabilistic modeling techniques are applied, they allow for a model which is computationally fast, easy to validate and can be updated if new knowledge about the conditions/inputs is gained.
This paper describes available datasets, from which information can be extracted and turned into new knowledge about ship performance in ice. In a companion paper the methods of transmitting information into knowledge through probabilistic models are described and the obtained results are discussed.
The reminder of this paper is organized as follows: Chapter 2 presents two available sources of data which are further matched in tempo-spatial fashion: first is an accurate and detailed reanalyzed ice forecast, called hindcast, for the sea area was taken under consideration; second is detailed database containing the state vectors of an analysed ship obtained from the Automatic Identification System (AIS). Chapter 3 describes an educated use of the developed database about ship performance in dynamic ice, meaning its visualization to reproduce behavior of the analyzed system, this also allowed to capture all noteworthy events for further analysis, which is presented in a companion paper.
2. DATA
2.1. DATE AND THE AREA OF INTEREST
The date and area of interest have been selected specifically to capture challenging ice conditions, meaning high concentration of level ice, the presence of ridged ice and ice compression that changes in time. For this reason, the day of 6th of March 2011 was chosen and the sea area between two Finnish harbors, in the Bay of Bothnia (the Baltic Sea), namely Vaasa and Kokkola were selected. A single trip of the bulk carrier between two positions of boarding a pilot is considered, meaning that the stage of the open-seas navigation is addressed, where the ship is supposed to proceed with full engine power. The trajectory of the ship is overlaid on the ice chart in Figure 1. In the figure the locations where the ship was brought to a halt are marked with yellow crosses, otherwise her track is marked with the blue circles.
The ship covered distance of 94 NM in 14 hours, and the ice conditions hampered her significantly, making her ram the ice several times, and forced her to idle in ice for three hours, as depicted in Figure 3.
Fig. 1. Location of the area of interest (above) and ice chart (below) for the time of interest and the track of the analyzed ship from Vaasa to Kokkola, source: (SMHI 2011, maps.google.com).
Vaas Kokkola
2.2. SHIP DATA
The chosen ship is a bulk carrier having the DNV ice class of +1A1 Ice 1A super, which is equivalent to IA Super according to Finnish-Swedish Ice Class rules, which is the highest possible ice-class. This means that a structure, engine output and other properties of a ship make her capable of navigating in difficult ice conditions without the assistance of icebreakers. The design requirement for this ice class is a minimum speed of 5 knots in 1 m thick brash ice channels with a 0.1 m thick consolidated layer of ice on top, see (Transport Safety Agency 2010). For the ship particulars see Table 1, whereas ice class requirements are gathered in Table 2. In order to quantify the joint effect of ice conditions on ship’s speed, the following parameters of ship motion were retrieved from the AIS records:
x time, x ship position, x speed over ground, x course over ground, x true heading.
Then for each time step and position, the relevant ice characteristics were obtained from the ice forecast model. Once the ship parameters had been aligned with the ice forecast, one additional parameter was calculated allowing for the quantification of an effect of direction of compression with respect to a ship at a certain speed. This is called “relative direction of compression”, which is an angle between ship’s centerline and the resultant direction of the ice compression. This parameter is expressed on a scale from 0 deg (ice pressing from the bow) to 180 deg (ice pressing from the stern), where the value of 90 deg means that the ice compression acts perpendicularly to a ship.
The resolution of data describing ship motion is much finer (10 sec for most of the time) than the data obtained from the ice forecast (1 hour time interval and 1x1 NM in space), therefore a piece of code was created to extract the ice features which correspond to ship location in the given time. The difference in resolutions produces the variability in the value of parameters describing ship performance even if the modelled ice conditions remain unchanged. However, this effect is removed at the stage of model development, where all variables are discretized and divided into classes, as required by Bayesian learning algorithms that have been applied. However, we made two assumptions regarding relevant features of the analysed ship. First, refers to ship’s inertia and her ability to break the ice, meaning mass of the ship (displacement), which cannot be determined accurately. Based on our knowledge about the route of the ship and harbours visited we assumed half-load conditions during the analysed journey. Second assumption is about full engine power used to proceed through the ice, meaning that the speed fluctuations are mainly because of ice conditions and are not invoked by the ship’s crew changing the engine settings.
Table 1 Ship particulars
Type General cargo
Ice class IAS
DWT 21353 t Length 149.3 m Breadth 24.6 m Draught 9.4 m Power 9720 kW Year of construction 2006 Table 2 Ice class requirements
Ice Class Ice Class Requirements
IA Super Ships with such structure, engine output and other properties that they are normally capable of navigating in difficult ice conditions without the assistance of icebreakers.
IA Ships with such structure, engine output and other properties that they are capable of navigating in difficult ice conditions, with the assistance of icebreakers when necessary. IB Ships with such structure, engine output and other properties that they are capable of
navigating in moderate ice conditions, with the assistance of icebreakers when necessary. IC Ships with such structure, engine out and other properties that they are capable of navigating
in light ice conditions, with the assistance of icebreakers when necessary.
II Ships that have a steel hull and that are structurally fit for navigation in the open sea and that, despite not being strengthened for navigation in ice, are capable of navigating in very light ice conditions with their own propulsion machinery.
III Ships that do not belong to the ice classes referred above.
2.3. ICE DATA
The ice data was obtained from the reanalyses performed with the use of the HELMI multicategory sea-ice model, see (Haapala et al. 2005). The model resolves ice thickness and concentration for five level ice categories. These categories include rafted ice and ridged ice together with the thermodynamics of sea-ice, horizontal components of ice velocity and internal stress of the ice pack.
Ice motion is determined by the momentum balance equation, which takes into account the Coriolis force, wind and water stresses, sea surface tilt term and an internal stress. The magnitude of internal friction is used as the principal model variable to describe compression. It is to be noted that the viscous-plastic rheology does not describe elastic stresses and the internal stress arises from the interactions of moving ice. Forces arising in a static ice are included by assuming a negligibly slow viscous creep. Roughly, the internal friction term can be interpreted to describe the forces arising when ice floes are pushed and sheared against each other, or broken and heaped into ridges. Thus it is a good descriptor for the interaction between dynamical ice cover and an ice-going ship. This is manifested as ice forces against the ship hull and as the closing of channels, or other phenomena that navigators associate to compressive ice conditions. The internal friction magnitude has
typical values ranging from 0 to 10 N/m2. The magnitude, possibly scaled to semi-empirical compression numeral 0-4, where 0 means no compression and 4 stands for severe compression, acts as a proxy for ice compression, see Table 3. However, to estimate the actual local forces additional scaling arguments must be taken into account such as floe size and other ice cover geometry.
The ice-forecasting model was discretized in a c-crid and in curvi-linear co-ordinates. The grid has 415 nodes from west to east and 556 nodes from south to north. The SW lower corner coordinates are 56.74° N 16.72° E, NE corner coordinates 65.99° N 30.48° E and the increment is 1/30 degrees eastwards and 1/60 degrees northwards. This is approximately 1 NM in both directions at 60°N. The model was forced by the HIRLAM regional weather model forecasts or reanalyses of winds and air temperature. The length of forecast is 54 hours and interval of 3 hours, while reanalyses are stored at 1 hour intervals. The present set-up of the ice prediction system does not include any dynamical ocean component, thus ocean currents are neglected. Sea surface temperature is prescribed and updated once a day and thus an ice edge is very much controlled by this procedure. Ice forecasts have been validated against the observed ice situations; see for example (Lehtiranta et al. 2012).
Table 3 Ice compression conversion
Ice pressure obtained from the HELMI
model[10 N] Practical scale 0 - 500 0 - no significant compression 500 - 1000 1- mild pressure 1000 - 2000 2- moderate pressure 2000 - 3000 3- severe pressure
3000 -> 4- extreme severe pressure
3. EXTRACTING INFORMATION FORM THE DATA
Two data sets described in previous chapter contain very detailed and the most accurate data available. If the data is properly handled it can deliver relevant information, which in turn can provide a new knowledge about the analyzed system, see for example (Aven, 2013). In this chapter we describe a way in which we transform data into information.
The speed profile of the voyage of the analyzed ship is shown in Figure 2, along with annotations regarding the navigational status and noteworthy events. The data range used for the model development is shown as well. The detailed information of the vessel’s movement is obtained through the Automatic Information System (AIS). This AIS data was linked to available vessel characteristics regarding vessel type, ice class and tonnage as obtained through the PortNet system and to the local ice conditions as obtained from the HELMI ice dynamics model, see (Haapala et al. 2005).
While the speed profile of Figure 2 and the ship trajectory of Figure 1 are insightful, the specific navigational conditions and traffic conditions are better understood with animations of the tempo-spatial data. Thus, we made videos, which enable a detailed understanding and analysis of the voyage and allow us transforming still data into dynamic and context dependent information. Moreover, the crew of a Finnish icebreaker was consulted to assure the proper identification of all manoeuvers performed by the analyzed ship and the assisting icebreakers as evident from the animation.
Figure 3 and Figure 4 show still images of animations of the analyzed vessel. The left pane shows the vessel in the center, along with local vessels within a 2NM radius around her. Situational parameters regarding speed, heading and course over ground are shown, as well the time and location, air and sea temperature, wind speed and direction, ice drift speed and direction and ice compression magnitude and direction. The upper right pane shows the complete trajectory of the analyzed vessel and the traffic situation in the entire area. The mean ice thickness, a weighted average of the level, ridge and rafted ice thicknesses is shown in lower right pane, with indication of the current vessel’s position. On the right, local ice parameters at the location of the analyzed vessel and parameters of the closest vessels to the analyzed ship are shown.
As seen in Figure 2, the analyzed vessel left the waiting area in front of the harbor of Vaasa at 0545 UTC, two icebreakers assisted her. Figure 3 shows the vessel under tow in the presence of mild to moderate compression, in line with the taxonomy of Table 2. After decoupling the tow, one of the icebreakers leaves the vessel, while the other icebreaker escorts the vessel to open water conditions under decreasing ice pressure and ice concentration. The towing operation occurs at a variable speed between 8 and 12 kn, while the escort takes place under variable speed between 12 and 14 kn. From 0730 until 1040, the vessel proceeds alone through open water conditions at an open water speed of 15.5 kn. At 1040 the ship enters the ice field and her speed drops and falls in the range between 10 and 12kn. At 1105 the ship encounters another ship, which is stopped in ice, approaches her and cuts her off; the speed of analyzed vessel is still high about 9 kn. At 1130 she passes two other immobile ships at the distance of 1.5 NM, her speed begins to decrease and at 1210 she is brought to a halt for the first time. She made an attempt of backing and ramming in order to release but did not succeed. At 1236 an outbound ice breaker towing a ship approaches her on reciprocal course and cuts her off, so she can resume her voyage and she follows the ice channel made by these two ships. Her speed falls in the range 8-10 kn for about half an hour, untill 1300, when it decreases gradually and at 1422 she is stopped for the second time, some 15 nm from the harbour. Figure 4 shows that the other ships in the area (marked with green) also are barely able to proceed due to the severe ice pressure, and that the prevailing ice conditions are rather demanding. After few attempts of backing and ramming she and another nearby vessel are cut off by an outbound ice breaker at 1440, and they continue approach, passing several stopped ships along their way. At 1540 both ships are brought to a halt again, the analyzed vessel makes one or two attempts to release by herself, but she gives up, and both ships need to wait. They are cut off at 1800 by an inbound icebreaker, which overtakes them and proceeds to the harbour. Since that point the analyzed ship follows the ice channel made by the icebreaker, with the speed of 10kn, at 1900 she slows down to 2kn at the pilot boarding position, then speeds up and continues to the harbour, which is finally reached at 1930 UTC.
Fig. 2. Annotated time history of the analyzed ship journey
Fig. 4. Snapshot of the situation at which the analyzed ship is stopped in ice.
4. CONCLUSIONS
In this paper we described available datasets and present their usability for extracting new knowledge about ship behavior in ice-covered waters. We analyzed socio-technical system, where we anticipate some variables, which are not recorded however observable and which can affect the system behavior. One of such variables not recorded in the data or observable, is behavior of ship’s crew with respect to speed adjustment done on board ship during transit. An assumption about the use of full power throughout the whole voyage is made, which is then to the large extent proven by analyzing the visualized data. Another variable is an effect of ice channel broken by an icebreaker, which can only be found by post processing the data.
Moreover we demonstrated a way in which static data can be transformed into dynamic information, which is further used to develop knowledge. By analyzing the developed database, we eliminated to large extent the effect of some existing but not recorded variables lading to the development of correct information.
ACKNOWLEDGMENTS
The work presented here has been financially supported by FP7 project SAFEWIN on “Safety of winter navigation in dynamic ice” (www.safewin.org).
The Merenkulun säätiö – the Maritime Foundation - from Helsinki is thanked for the travel grant.
References
1. Aven, T., 2013. A conceptual framework for linking risk and the elements of the data–information– knowledge–wisdom (DIKW) hierarchy. Reliability Engineering & System Safety, 111(0), pp.30–36. 2. Haapala, J., Lönnroth, N. & Stössel, A., 2005. A numerical study of open water formation in sea ice.
Journal of Geophysical Research: Oceans, 110(C9), p.n/a–n/a.
3. Juva, M. & Riska, Kaj, 2002. On the power requirement in the Finnish-Swedish ice class rules, Espoo, Finland: Helsinki University of Technology.
4. Kotovirta, V. et al., 2009. A system for route optimization in ice-covered waters. Cold Regions Science
and Technology, 55(1), pp.52–62.
5. Lehtiranta, J., Lensu, M. & Haapala, Jari, 2012. Ice model validation on local scale., Helsinki: Finnish Meteorological Institute.
6. Lindqvist, G., 1989. A straightforward method for calculation of ice resistance of ships. In The 10th
conference on POAC. POAC. Luleå University of Technology, pp. 722–735.
7. Naegle, J.N., 1980. Ice-resistance prediction and motion simulation for ships operating in the continuous
model of icebreaking. Doctoral thesis. Michigan, USA: The University of Michigan.
8. Riska, Kaj et al., 1997. Performance of merchant vessels in ice in the Baltic.
9. SMHI, 2011. Ice chart for the Baltic Sea for 6th of March 2011. Available at: http://www.smhi.se/oceanografi/iceservice/is_prod_en.php.
10. Transport Safety Agency, 2010. Maritime safety regulation. Ice class regulations and the application thereof.
MODELOWANIE OSIGÓW STATKU NAWIGUJCEGO W DYNAMICZNYM POLU LODOWYM: CZ I, PRZEKSZTACENIE DANYCH W INFORMACJE
Streszczenie: W celu bezpiecznej oraz wydajnej eksploatacji akwenów pokrytych lodem, wiedza o
zachowaniu statku w tych warunkach jest niezbdna. Pomimo, i egluga w lodach pozostaje tematem wielu opracowa naukowych, tematyka modelowania zachowania statku w dynamicznym polu lodowym, zwaszcza w obecnoci zjawiska kompresji pokrywy lodowej, pozostaje wci kwesti otwart.
W artykule omówiono dostpne róda danych, które po odpowiednim przetworzeniu dostarcz informacji, umoliwiajcej lepsze zrozumienie procesu nawigacji statku w dynamicznym polu lodowym co pozwoli na modelowanie tego procesu.
W artykule podkrelono zasadno przeprowadzenia procesu przeksztacenia danych w informacj które nastpnie mona wykorzysta w celu uzyskania nowej wiedzy. Na przykadzie pokazano, i nawet najdokadniejsze dane, nie dostarcz penej informacji, jeeli nie zostan odpowiednio przetworzone oraz zinterpretowane, co moe prowadzi do bdnych lub niepenych wniosków.
