Speed loss analysis and rough wave avoidance algorithms for optimal ship routing simulation of 28,000-DWT bulk carrier

Highlights

• The analytical model with time-domain speed loss evaluation is constructed.

• Optimal route is simulated for 28,000-DWT bulk carrier with measured data.

• Differences of simulations are clarified in each condition with actual voyage.

• Speed loss is evaluated by introducing precise seakeeping and propulsion theories.

Abstract

Restrictions on greenhouse gas emissions and the existence of developing countries has diversified ship routes and ship types, making the ship operation more complex. Many types of simulation methods have come up in the last decade due to the enforcement of the energy efficiency design index. However, few studies that validated their simulations with actual data. This study performs the optimal ship routing of a 28,000-DWT-class bulk carrier to evaluate the optimal ship route with the speed loss analysis. The simulation results are compared with the measurement data of the bulk carrier in conditions including rough sea voyage in the Pacific Ocean. For comparison, a modified version of the speed loss analysis algorithm is developed. These algorithms are combined with the isochrone method. The differences in the simulation results with the American and European weather databases without higher wave avoidance and deliberate speed reduction are obvious. The simulated routes are closer to the measured ship trajectory if higher wave avoidance is considered. There are a couple of remarkable speed drops in the measured result, and the deliberate speed reduction is considered. Finally, some factors to control the reproducibility of ship routing including in rough sea voyages are summarized.

Introduction

The global society depends on the logistics of natural resources, energy resources, foods, manufactured goods, etc. Efficiency and safety have been key topics from various points of view in the field of naval architecture and ocean engineering. Maritime transportation sustains more than 95% of global logistics. The relationship between maritime transportation and industrialization was analyzed by Qinetiq et al. (2012). The stable transportation of natural resources or energy resources is key to development in several Asian countries with poor resource accessibility and large populations. The main regions exporting resources to Asia are Oceania, Middle East, and Latin America. This implies that transportation is required across open seas such as the Pacific Ocean, Atlantic Ocean, Indian Ocean, Tasman Sea, etc. Major voyage routes are set to diversify to all over the world, including the Southern Hemisphere.

A ship's operation is decided either by the ship master or by the shipping company. They refer weather forecasting to make a decision through intuition and experience. With the advancement of computer technology, the spatiotemporal range of weather forecast has become more precise. The concept of optimal ship routing was proposed in the 1950s with a combination of weather forecasting and route optimization techniques. In the field of meteorology, the amount and resolution of the weather database has improved. However, there are sparse data on oceans and high latitudes because of the difficulty in installing measurement systems in such regions. The World Meteorological Organization (WMO) has collected observational weather information from ships worldwide to compensate the insufficient number of observation points. Currently, satellite monitoring is being used with the development of remote sensing technology. These data are blended with the numerical simulations of the air and the ocean, and an objectively analyzed value is fixed as the quasi-measured value. The accuracy of weather forecast depends on this value and is expected to increase with the volume of information. A few meteorological organizations have published these weather data on their websites. Greener and more efficient transport is nowadays a requirement in maritime transportation, hence the number of studies on optimal ship routing has greatly increased. Some private services that specialize in ship-routing optimization are managing commercial fleets (Japan Marine Science, 2017; WNI, 2020). Data from such private services, albeit valuable for research purposes as well as practical use onboard, are not publicly available, which makes it difficult to validate models and estimate their accuracy. In this study, we conduct optimal routing analysis for observations from a rough sea voyage across the Pacific Ocean.

Weather forecasting accuracy is validated by meteorological observations mainly over land; the reliability of weather forecasting over oceans is reduced due to fewer meteorological observations available for validation. Validation of weather forecasting over ocean occurs by comparing the sparse onboard observations with numerical simulations (Lu et al., 2017; Chen et al., 2020). In this sense, there is added uncertainty to the weather-forecasting validation by the limited, variable, or unknown accuracy of meteorological observations over ocean, including in the Southern Hemisphere or higher latitudes. The optimal ship routing is validated for the North Pacific Ocean in this study, as the first step. Although these data have been used in optimal ship routing studies, their reliability has not been sufficiently validated. Tanemoto et al. (2015) discussed the accuracy of forecasted winds by comparing them with the objective wind parameters in the Pacific Ocean and the Indian Ocean. Studies on the performance of a ship at sea is being conducted since the early 2000s. Kobe University (2012), for example, has conducted the research project Three Principles of Maritime Transportation to clarify the variation of ship performance for a newly built 28,000-DWT bulk carrier at sea. The accuracy of wind waves was discussed by comparing them with the objective data and ship motions in multiple rough sea voyages, including the North Pacific Ocean, the Indian Ocean, the South Atlantic Ocean, and the Tasman Sea (Sasa et al., 2015; Lu et al., 2017). Despite their uncertainty (Prpić-Oršić and Faltinsen, 2012; Prpić-Oršić et al., 2020), the parameters of the weather can be quantitatively evaluated with onboard measurement data and the objective weather data. At present, the weather data covers the entire Earth to a resolution of 0.75°–1.0° in NCEP FNL (Kalnay et al., 1996), American meteorological database, and ERA-interim (Dee et al., 2011). The state-of-the-art European meteorological database ERA5 has higher spatial and temporal resolutions of 0.25° and 3 h (since 2018), respectively. Naito et al. (1990) also showed that weather conditions can remain constant for a maximum of 30 min. The spatial resolution corresponding to this time should be less than 7 nautical miles (0.1°) in the case of bulk carrier, because the voyage speed is approximately 14 knots. Thus, it is necessary to estimate the wind and wave fields using numerical models and objective data.

Chen et al. (2020) investigated a method to improve the estimation of wind characteristics by optimizing the parameters of the numerical conditions in the Weather Research and Forecasting (WRF) model. Besides forecasting the weather, the accurate estimation of a ship's performance is also necessary. To this end, the speed loss of a 28,000-DWT bulk carrier during rough sea voyage was validated by comparing the measured and theoretical (simulation) estimates (Sasa et al., 2017). They found that the simulated results reproduced the speed loss during automatic engine control accurately. Meanwhile, the speed loss at sea comprises deliberate speed reduction, which is a human factor, besides the natural speed loss in many related studies in the 1960s–1980s. Several seafarers have provided the criteria for estimating the necessary deliberate speed reduction, which are summarized in Prpić-Oršić and Faltinsen (2012) and Vettor and Guedes Soares (2016b). There are limited speed loss data for cargo vessels at the actual sea, and the proposed criteria should be validated by measured data. These studies have summarized the patterns of speed loss of a bulk carrier for three cases of rough sea in the Southern Hemisphere. The results show that the actual limit values may be different from those arrived at using these criteria, except for the vertical acceleration determined by NordForsk.

Studies involving an optimized ship routing solution as a scientific tool started in the second half of the 20th century with the arrival of computational weather forecasting. The development of weather forecasting, ship technology, and computation analysis has led to optimal ship routing becoming a field of navigation and ocean engineering (Vettor and Guedes Soares, 2016b). Various types of numerical methods have been proposed, such as the modified isochrone method (Szlapczynska et al., 2007; Lin et al., 2013; Roh, 2013), Dynamic Programming Method (Takashima et al., 2004; Shao and Zhou, 2011; Shao et al., 2012; Varelas et al., 2013), Dijkstra's method (Fagerholt et al., 2000), and A* method (Bentin et al., 2016; Grifoll et al., 2018), genetic algorithm (GA) (Maki et al., 2011; Veneti et al., 2015). There are also simplified methods like Kwon's method (Kwon, 1981; Lu et al., 2015). In the first stage, the problem is only solved to minimize the voyage time. With the fluctuation of oil price and GHG emissions regulations, the necessity to minimize fuel consumption increased, thereby increasing the need of a multi-objective solution (Szlapczynski et al., 2019; Zyczkowski et al., 2018). The GA was applied in optimal ship routing to account for this requirement. However, there are not many studies comparing simulated results with single-objective methods. In recent years, the algorithm has been further developed regarding the evaluation of speed loss or fuel consumption by combining propulsion theory. Zaccone et al. (2018) constructed their model based on the dynamic programming method with wave added resistance and ship motions in irregular waves as sea spectrum. Nielsen et al. (2019) introduced an algorithm including a dynamic engine-system model and a propeller–wave interaction model. These studies improved the evaluation of resistance and propulsion than earlier models in the optimal ship routing. Conversely, Sasa et al. (2019a) followed a different approach for evaluating speed loss, which included deliberate speed reduction with the vertical acceleration, the deck wetness and the slamming. The limiting wave height up to which high waves can be avoid was analyzed using the automatic identification system (AIS) data for rough sea voyages across the Pacific Ocean and the Atlantic Ocean (Fujii et al., 2019). In this study, these points are newly considered for the accurate evaluation in rough sea voyages. As mentioned, numerical simulation methods have been improved to meet the demand of ship operation. However, there is no guarantee these improvements will provide accurate solutions. It is also not simple to validate the accuracy of optimal ship routing, because the definition of ‘the correct solution’ is not necessarily precise, such as the seakeeping analysis. Furthermore, there are few studies comparing the measurements taken at the actual sea and computational results. It is important to discuss factors controlling the accuracy of solutions with measured ship data at the actual sea. In this study, the optimal ship route was evaluated for voyages by the 28,000-DWT bulk carrier, which was taking onboard measurements from 2010 to 2016, using computational methods. The numerical simulation was based on the modified isochrone method, whose procedure can be summarized as follows.

a. Two types of weather databases were used to set the wave, wind and current conditions in the simulation, and the results were compared as the optimal ship routing.

b. A physical speed loss algorithm (Sasa et al., 2017) was newly programmed, which estimates the ship speed by solving the equation of motion between engine thrust and resistance in each step. This step was performed because there are few studies evaluating speed loss on rough seas, though the route optimization algorithm has been modeled in detail. The deliberate speed reduction was also included in the algorithm by estimating the probability of deck wetness and slamming (Sasa et al., 2019a).

c. The route was simulated under various conditions for a voyage across the north Pacific Ocean from China to Mexico from September to October 2010. The results were compared with the measured data about the route, voyage time, ship speed, etc., to clarify the parameters controlling the accuracy of optimal ship routing.

In section 2, we outline the theoretical description of the algorithm. The algorithm consists of the speed loss analysis, including natural speed loss and deliberate speed reduction in combination with the modified isochrone method. We apply a physical speed loss algorithm (Sasa et al., 2017), which estimates ship speed by solving in each step the equation of motion between engine thrust and resistance. Deliberate speed reduction was also included in the algorithm by estimating the probability of deck wetness and slamming. In section 3, simulation parameters such as weather and ship conditions, and measured results of a rough sea voyage are described. Two types of weather databases were used to prescribe wave, wind, and current in the simulation, and were compared each other. In section 4, the optimal ship routing is simulated for 28,000-DWT class bulk carrier on a voyage across the Pacific Ocean in each simulation condition. The route was simulated under various conditions for a voyage across the north Pacific Ocean from China to Mexico from September to October 2010. The results were compared with the measured data, thereby evaluating the route, voyage time, ship speed, etc., in order to identify the parameters controlling the accuracy of optimal ship routing. In section 5, we discuss the difference between measured and simulated results by comparing the variation of relative water surface, etc. We conclude by elaborating on the current model's validation and future topics for further developing the optimal ship routing simulations.