Author | Schulten, P.J.M. |

Title | The interaction between diesel engines, ship and propellers during manoeuvring |

Conference/Journal | PhD-thesis, Delft University of Technology |

Month | May |

Year | 2005 |

Abstract

This thesis describes a model (the 'Ship Mobility Model') that predicts the behaviour of the propulsion system of a manoeuvring ship. The model is set up for a twin shaft, diesel engine-propeller driven slender ship. The thesis has two abstraction levels. The basic level consists of the presentation of the general model layout and the specific details of the sub-models. Significant is the fact that the model is a 'total model' in which the complete mobility function of a ship, from the diesel engine to the movements of the ship, is modelled with considerable detail. In this way it is possible to study the interaction between all relevant sub-systems. On a higher abstraction level, this thesis deals with the question how such a total model is to be developed. Because of the extensiveness of the model, various scientific areas are involved and the system integrator has to 'smart-buy' models from other specialists. In other words, the system integrator has to be able to make a substantiated choice between the available sub-models in a limited period of time. Related to both abstraction levels of the thesis is the uncertainty analysis that has been performed on the Ship Mobility Model. This analysis shows that if a large and complex total model is constructed, the uncertainty of the model results may be high. The relevance and applicability of the Ship Propulsion Model is found in many ways. The model can be used to investigate the in-service, true operational behaviour of the mobility system which may lead to improved performance. Ultimately, this it help reducing the maintenance costs and engine downtimes. Secondly, the model could be used to design new control regimes. Two questions are worked out in detail: why is the diesel engine overloaded in a turn and what if the pitch is controlled in such a way that cavitation of the propellers is reduced.

The Ship Mobility Model consists of three main sub-models: the diesel engine model, the propeller model and the manoeuvring model. The diesel engine model is a mean value model based on fundamental physical principles. The essence of the mean value model is that for the various work and flow processes mean (time averaged) values are calculated. The engine torque is calculated by adding the net work of the various processes. The outlet receiver temperature is calculated by thermodynamically mixing the various flows (e.g. scavenge flow and exhaust flow) in the outlet receiver. The propeller model is based on open water propeller diagrams, applied to a manoeuvre dependent effective advance ratio. For each time step, the wake is built up from the straight ahead nominal wake field, superimposed with a prescribed, oblique inflow. The method by Gutsche (1975) is then applied to this composed wakefield to find the effective advance ratio. The manoeuvring model is a four degrees of freedom model (surge, sway, roll and yaw). The forces on the ship's hull are calculated using a full non-linear model. The linear parts of the forces are determined by empirical formulas based on the main particulars of the ship, while the non-linear part also includes the influence of local hull details such as the local sectional draught.

Using a sensitivity and uncertainty analysis, the uncertainty of the main interesting model results is determined. The uncertainty of the diesel engine variables is 3-7% for the engine speed and 6-9% for the outlet receiver temperature. The propeller torque has an uncertainty of 6-12%. The uncertainty of the manoeuvring variables (u, v and r) is 4-12%. These uncertainties have been determined using the total model. The uncertainties of the sub-models when regarded separately are different. The model is validated using independent measurements taken on board HNLMS De Ruyter, an air defence and command frigate of the Royal Netherlands Navy. The validation showed no significant difference between the modelled and measured diesel engine variables. The difference between the modelled and measured propeller variables as well as between the modelled and measured manoeuvring parameters is significant.

The three main sub-models (diesel engine, propeller and manoeuvring) have in common that they are all mean value models. The differences between the models consists of the details that are known to the system integrator, reflected by the way the models are described in this thesis. A substantial part of an existing diesel engine model in this case has been developed by the system integrator. Not only the concept and philosophy, but also specific details are known and presented. Apart from the primary model output also a large amount of other relevant data is calculated by the model, e.g. the speed of the turbocharger and the maximum pressure in the cylinder. The propeller model is based on an existing theory (developed by Gutsche) that has been implemented by the system integrator. The propeller model produces the desired model outputs (thrust and torque), but is unable to produce internal data (for instance local cavitation information) with sufficient accuracy. The details of the manoeuvring model are unavailable because it is a commercially protected model (the manoeuvring model 'Fresim' developed by MARIN). In this thesis only the concept and philosophy of this model is presented.

This thesis describes a model (the 'Ship Mobility Model') that predicts the behaviour of the propulsion system of a manoeuvring ship. The model is set up for a twin shaft, diesel engine-propeller driven slender ship. The thesis has two abstraction levels. The basic level consists of the presentation of the general model layout and the specific details of the sub-models. Significant is the fact that the model is a 'total model' in which the complete mobility function of a ship, from the diesel engine to the movements of the ship, is modelled with considerable detail. In this way it is possible to study the interaction between all relevant sub-systems. On a higher abstraction level, this thesis deals with the question how such a total model is to be developed. Because of the extensiveness of the model, various scientific areas are involved and the system integrator has to 'smart-buy' models from other specialists. In other words, the system integrator has to be able to make a substantiated choice between the available sub-models in a limited period of time. Related to both abstraction levels of the thesis is the uncertainty analysis that has been performed on the Ship Mobility Model. This analysis shows that if a large and complex total model is constructed, the uncertainty of the model results may be high. The relevance and applicability of the Ship Propulsion Model is found in many ways. The model can be used to investigate the in-service, true operational behaviour of the mobility system which may lead to improved performance. Ultimately, this it help reducing the maintenance costs and engine downtimes. Secondly, the model could be used to design new control regimes. Two questions are worked out in detail: why is the diesel engine overloaded in a turn and what if the pitch is controlled in such a way that cavitation of the propellers is reduced.

The Ship Mobility Model consists of three main sub-models: the diesel engine model, the propeller model and the manoeuvring model. The diesel engine model is a mean value model based on fundamental physical principles. The essence of the mean value model is that for the various work and flow processes mean (time averaged) values are calculated. The engine torque is calculated by adding the net work of the various processes. The outlet receiver temperature is calculated by thermodynamically mixing the various flows (e.g. scavenge flow and exhaust flow) in the outlet receiver. The propeller model is based on open water propeller diagrams, applied to a manoeuvre dependent effective advance ratio. For each time step, the wake is built up from the straight ahead nominal wake field, superimposed with a prescribed, oblique inflow. The method by Gutsche (1975) is then applied to this composed wakefield to find the effective advance ratio. The manoeuvring model is a four degrees of freedom model (surge, sway, roll and yaw). The forces on the ship's hull are calculated using a full non-linear model. The linear parts of the forces are determined by empirical formulas based on the main particulars of the ship, while the non-linear part also includes the influence of local hull details such as the local sectional draught.

Using a sensitivity and uncertainty analysis, the uncertainty of the main interesting model results is determined. The uncertainty of the diesel engine variables is 3-7% for the engine speed and 6-9% for the outlet receiver temperature. The propeller torque has an uncertainty of 6-12%. The uncertainty of the manoeuvring variables (u, v and r) is 4-12%. These uncertainties have been determined using the total model. The uncertainties of the sub-models when regarded separately are different. The model is validated using independent measurements taken on board HNLMS De Ruyter, an air defence and command frigate of the Royal Netherlands Navy. The validation showed no significant difference between the modelled and measured diesel engine variables. The difference between the modelled and measured propeller variables as well as between the modelled and measured manoeuvring parameters is significant.

The three main sub-models (diesel engine, propeller and manoeuvring) have in common that they are all mean value models. The differences between the models consists of the details that are known to the system integrator, reflected by the way the models are described in this thesis. A substantial part of an existing diesel engine model in this case has been developed by the system integrator. Not only the concept and philosophy, but also specific details are known and presented. Apart from the primary model output also a large amount of other relevant data is calculated by the model, e.g. the speed of the turbocharger and the maximum pressure in the cylinder. The propeller model is based on an existing theory (developed by Gutsche) that has been implemented by the system integrator. The propeller model produces the desired model outputs (thrust and torque), but is unable to produce internal data (for instance local cavitation information) with sufficient accuracy. The details of the manoeuvring model are unavailable because it is a commercially protected model (the manoeuvring model 'Fresim' developed by MARIN). In this thesis only the concept and philosophy of this model is presented.