TS-72
Main Engine Control for Heavy Weather Conditions: The ACME Project
Nikolaos P. KYRTATOS*, Gerasimos THEOTOKATOS* and Nikolaos I. XIROS*
ABSTRACT
Speed control of the single, large, directly coupled Diesel engine used for propulsion of commercial vessels, is nowadays achieved by use of electronic governors which implement the PID control law. Although these devices handle successfully engine running under normal conditions, there maybe cases, when the ship is sailing in heavy weather, where power/speed reduction is required for engine safety reasons. Especially, in operation of ships in heavy seas, the load demand on the propulsion engine from the propeller may change rapidly, causing significant deviations of the shaft rotational speed from the desired setpoint. Since such deviations, especially to the overspeed side, can activate limiter protection, engine speed reduction is required. If this occurs often it may result in underutilization of the powerplant
Key Words: Engine Adaptive Control, Shipboard Testing, Adverse Sea Conditions, Mathematical Modelling
1. INTRODUCTION
Answering to a practical need, the transnational ACME project (Adaptive Control of Marine Engines) was initiated, aiming to develop methods of improved control, allowing for full exploitation of the main engine power under extreme weather conditions. The methodology of the ACME project has been to develop a system for propeller load prediction a few seconds ahead in time, which could provide a driving signal to an engine control unit, that would decrease fuel temporarily in case of an imminent expected load sink, thus escaping excessive overspeed. Based on the analysis of a large number of specially conducted systematic tests of a container hull-form in rough sea conditions in a model basin, certain patterns of instantaneous torque demand have been identified. Then, after extensive simulation studies of engine operation, a first version of the load prediction system has been implemented. This version requires as input the vertical aft ship acceleration and the propeller load history. A new powerplant control unit has also been developed, which uses as additional input the signal of the load prediction system. The prototype system has been installed on-board a containership, and after a complete set of sea trials for safety assessment, has been tested in actual operating conditions for a period of several months. The data series collected from the full-scale tests further show that it may be possible to simplify the load prediction system to a version which would require only the propeller load history as input. Conclusively, the ACME project demonstrated that the use of model-based adaptive control may extend propulsion plant capabilities, especially under heavy weather conditions, enabling smoother and more reliable engine running.
2. TECHNICAL DESCRIPTION
2.1. Overview
The successful development of advanced engine control strategies and shipboard integration of prototype ECU system has been based on the potential offered by computer simulation techniques and tools.
A series of ship model basin experiments of ship operation in waves resulted in a large data set. The data series acquired included ship motion parameters, and specifically acceleration of several parts of the ship, in conjunction with shaft speed and torque [1][2][3]. A variety of mathematical simulation models have been either updated/extended or numerically configured. The models included:
a) Analytical propeller transient performance models
b) Compressor surge models
c) Marine Diesel engine models used for performance prediction and control
d) Transfer function type models for ship propulsion powerplants
The outcome of the systematic analysis and simulation work performed consisted of the following items:
1. A propeller torque demand prediction method based on statistical analysis and pattern recognition and classification that can provide engine control with a torque profile covering a few seconds ahead in time. In that, a prediction event database is created and maintained. Simulation and basin tests have shown that there exists a correlation between the current value of the independent values and the evolution of propeller shaft torque [2]. A prediction event is defined as a set of values for several variables (independent and dependent) including aft ship acceleration and instantaneous propeller shaft torque as independent variables. Prediction is effectuated by matching the current set of values of the independent variables to a prediction event from the database. If adequate similarity between the recorded and occurring situations exists then torque prediction is possible.
