To reduce global warming and air pollution, the maritime industry is searching for solutions to reduce fuel consumption. This thesis focuses on the reduction of generator running hours and fuel consumption of Dynamically Positioned (DP) vessels. In this ship type, redundancy of the power generation system is regularly achieved by running on one generator more than strictly necessary in every engine room. Thereby, these vessels make superfluous running hours and unnecessary fuel consumption. Previous research in this field focused on installing a hybrid system in every separate engine room to replace this redundant running generator and achieve the power generation redundancy by a battery which is temporarily used in case of a generator failure. Since DP vessels normally operate on two separate engine rooms for safety, this results in two separate hybrid systems, i.e. a double-battery hybrid system. To reduce investment costs of a hybrid DP vessel, this thesis concentrates on finding a power management strategy that allows a single hybrid system to be installed on board, which is only connected to one of the engine room grids in case of a generator failure. A single-battery hybrid system cannot be connected to both engine room grids continuously -in contrary to the double-battery hybrid system- because this would cancel the independence of the separate engine rooms. Therefore, in the single-battery hybrid system, no back up power is present in case of an increase in power demand in one of the engine rooms. In present DP vessels, this back up power is supplied by the redundant generator. When using a double-battery hybrid system, this back up power is supplied by the continuously connected hybrid system. Therefore, for a single-battery hybrid system, the starting control decision algorithm of the power management system (PMS) should decide on starting an additional generator prior to a rise of the power demand.
Two methods for this starting decision algorithm were tested. Firstly, a mathematical-statistical method using a mathematical model for DP load forecasting and a statistical model for pipe laying load forecasting was tested. This model was found not to work properly. Secondly, a multiple linear regression model that predicts the height of power peaks based on environmental parameters like wind and waves was tested which was found to work appropriately. To assess the feasibility of using either a double-battery hybrid system in combination with a simple starting decision algorithm, or a single-battery hybrid system in combination with a smart starting decision algorithm, these two hybrid system options were compared to the current situation that uses a redundant generator in every engine room. These three system options were compared in terms of investment costs, fuel consumption, generator running hours, planned maintenance costs, and reliability. For this comparison, the DP pipelaying vessel Pioneering Spirit was used as a case vessel. To find the investment costs, a hybrid system was designed for this purpose. To find the fuel consumption, generator running hours, and planned maintenance costs, full year time domain simulations of the hybrid systems in operation were performed using measured power demand data from the case vessel during pipelaying. For this analysis, a mathematical model of the power generation system was used. The reliability assessment of the different systems was performed using fault tree analysis.
The single- and double-battery hybrid systems both show a significant fuel consumption reduction of 8.4 and 9.0% respectively. When also taking into account the investment costs, generator maintenance cost reduction, and battery depreciation, the payback time of these hybrid systems is 0.9 and 1.7 years respectively. The difference in payback time is mainly caused by the difference in investment costs between the single- and the double-battery hybrid system.
For a single-battery hybrid system for DP vessels, a multiple linear regression model based starting decision algorithm works appropriately. However, although the multiple linear regression model based starting decision algorithm was tested positively in simulations using the available power data, in case of an unforeseen power peak that was not related to one of the input parameters of the regression model, the power generation system lacks back up power. Therefore, the reliability of the single-battery hybrid system is insufficient when using it for DP applications. Especially when taking into account the minor absolute difference in payback time, for the replacement of the redundant generators in DP vessels, the robust and simple double-battery hybrid system is preferred over the small and smart single-battery hybrid system.