Final Report Summary - ADOPT (Advanced decision support system for ship design, operation and training)
ADOPT aimed at developing a concept for a risk-based, ship specific, real-time decision support system (DSS) covering three distinct modes of use:
- Design
- Training
-Operation.
The number one priority of the ADOPT-DSS was to assist the master to identify and avoid potentially dangerous situations.
The ADOPT-DSS is an integration of the three modes design, training, and operation and was customised for a specific ship and aimed at utilising state-of-the-art know-how and technology not available widely today.
The basis was the design (or office) mode. Information generated in the design mode is passed on to the training mode. Training can be provided directly using the information generated in design mode. In addition, the master can be trained for using the operation mode implementation of the ADOPT-DSS. One of the challenges in the ADOPT project in developing such a system had been to interface and connect various data sources, hardware, and software systems, that might run on various IT platforms.
Ship accidents or near misses due to heavy seas are reported frequently, e.g. Kernchen (2008), McDaniel (2008). Apart from these safety related accidents, damages non-critical from a safety perspective occur even more frequently. These damages are of high economic concern, as their impact span from on-board repair to taking vessels out of service for docking. In the latter case, economic consequences are in the order of magnitude of EUR 100,000 IMO MSC/Circ. 707, IMO (1995), gives guidance to masters in the form of simple rules for identification of potentially harmful situations. There are also computer systems that aim at improving the situation.
All support on the bridge is subject to the warning dilemma. Core of the warning dilemma is, that safety critical systems shall issue warnings only, if action is required, but always, if action is required. Warning that are issued, if no action is required might increase the risk. Warnings that are not issued if action is required, might have two reasons - either the system is not capable in assessing a situation, or it has a malfunctions. If either of these combinations occurs during operation, the confidence of the user in the system is degraded.
On the other hand, during the design phase of a ship a massive amount of numeric simulations are carried out using direct calculation tools. Basis for these calculations are the ship in its current design stage, the operational envelope of the intended service, and environmental data. As one result of these investigations, limits of the operational envelope are determined, or the design is altered. If limits are determined, one way of presenting them are polar diagrams, giving limiting significant wave heights as a function of encounter angle and ship speed. The parameters of the polar diagrams are the wave length, the loading condition, and the exceedance of a certain threshold value that might be a particular roll angle, accelerations at a certain position of the vessel, stresses in the structure, or other.
The project was structured into 7 work packages (WP), as follows:
- WP1: Definition of the criteria
- WP2: Environment
- WP3: Numerical simulation
- WP4: Ship data
- WP5: Man machine interface
- WP6: Integration
- WP7: Validation.
ADOPT provides a process with the three major process elements:
mode 3 - design,
mode 2 - training and
mode - operation.
Decision support for mode 1 starts in mode 3. Mode 2 is required to convey the information generated in mode 3 to mode 1. Operation of a ship in mode 1 will be improved by mode 2 even without a specific computerised mode 1 DSS. A computerised mode 1 DSS will further improve the operation in mode 1 by giving real time support, which might capture scenarios beyond the experience or capabilities of the crew.
The application of complex numerical methods in a risk based framework and the high reliability requirements made for the on-board decision support revealed the following major constraints during the development of the system:
- Within a 3 hours watch it is not feasible to obtain final results from the numerical calculations, if all uncertainties are taken into account.
- The quality of the input data is not always as high as required for a reliable decision support.
- The expertise to make full use of a decision support as complex as the ADOPT-DSS is normally not available on board.
For handling the mentioned constraints, the concept of the ADOPT-DSS was extended from a pure software-hardware package for the on-board use to a procedural process consisting of three modes of use, which are the design mode (Mode 3), the training mode (Mode 2) and the operational mode (on-board use, Mode 1), each accompanied by an appropriate set of tools. By combining the possibilities of each mode of use, the limitations shown above can be diminished and overcome to a certain degree. The time consuming calculations are done in the design mode, where the time constraint of a watch length is not relevant. Insufficient quality of input data, such as the loading condition, is solved by a quality control which uses pre-calculated data, verified in the design mode to be of sufficient accuracy.
The knowledge about the ship specific behaviour with respect to the individual hazards gathered during the design process is transferred to the crew on board via training sessions in Mode 2. Within this framework the master and the crew are also trained and familiarised with the concepts and the handling of the ADOPT-DSS. For this purpose the Mode 1 tools can be used in a simulated environment. This is an important prerequisite for the appropriate and responsible use of the ADOPT-DSS in operating conditions.
A further conclusion of the ADOPT-project is the necessity of a clear distinction between economic and safety risk as both modes are associated with fundamentally different risk acceptance criteria.
Further, two parallel branches of ship motion calculation and probability assessment have been developed, where a frequency domain provides fast and efficient decision support with respect to all limit states which show a mainly linear response behaviour, whilst a time domain implementation is provided for strongly non-linear responses such as parametric rolling.
The feasibility of the ADOPT concept has been proven by a demonstration prototype which was tested in a simulated environment. The consortium concludes that the ADOPT concept has clearly demonstrated its feasibility.
However, the ADOPT-DSS demonstrator has to be improved and further developed, especially with respect to response time, before practical on-board use is possible. Amongst others, the numerical processes of the ADOPT-DSS demonstrator have not been optimised so far. Additionally it is anticipated that the constantly increasing available computer power will assist in solving this problem in a few years.
The inclusion of relevant uncertainties from all sources (high level requirement 5) cannot be fulfilled due to time constraints. A sensitivity and uncertainty analysis to identify relevant and unimportant uncertainties may reduce the number of relevant uncertainties and thereby increase response time and/or reliability.
In the case of safety related hazards such as capsize the high level requirements regarding the display of reliability of the decision support (high level requirement 6 and 29) is only satisfied in the way of non-display, when the input data is unreliable. Improvements of the quality control and adequate displayed of the decision support reliability may be subject to further research.
Further some identified, generic hazards, such as broaching to, cannot be assessed by the numerical ship motion tools applied in the ADOPT-DSS today. Development progress on the applied numerical methods may in the future extend the possibility to include all generic hazards. The modular layout of the ADOPT-DSS also gives opportunity to include additional numerical assessment of limit states parallel to the two branches of ship motion simulation.
- Design
- Training
-Operation.
The number one priority of the ADOPT-DSS was to assist the master to identify and avoid potentially dangerous situations.
The ADOPT-DSS is an integration of the three modes design, training, and operation and was customised for a specific ship and aimed at utilising state-of-the-art know-how and technology not available widely today.
The basis was the design (or office) mode. Information generated in the design mode is passed on to the training mode. Training can be provided directly using the information generated in design mode. In addition, the master can be trained for using the operation mode implementation of the ADOPT-DSS. One of the challenges in the ADOPT project in developing such a system had been to interface and connect various data sources, hardware, and software systems, that might run on various IT platforms.
Ship accidents or near misses due to heavy seas are reported frequently, e.g. Kernchen (2008), McDaniel (2008). Apart from these safety related accidents, damages non-critical from a safety perspective occur even more frequently. These damages are of high economic concern, as their impact span from on-board repair to taking vessels out of service for docking. In the latter case, economic consequences are in the order of magnitude of EUR 100,000 IMO MSC/Circ. 707, IMO (1995), gives guidance to masters in the form of simple rules for identification of potentially harmful situations. There are also computer systems that aim at improving the situation.
All support on the bridge is subject to the warning dilemma. Core of the warning dilemma is, that safety critical systems shall issue warnings only, if action is required, but always, if action is required. Warning that are issued, if no action is required might increase the risk. Warnings that are not issued if action is required, might have two reasons - either the system is not capable in assessing a situation, or it has a malfunctions. If either of these combinations occurs during operation, the confidence of the user in the system is degraded.
On the other hand, during the design phase of a ship a massive amount of numeric simulations are carried out using direct calculation tools. Basis for these calculations are the ship in its current design stage, the operational envelope of the intended service, and environmental data. As one result of these investigations, limits of the operational envelope are determined, or the design is altered. If limits are determined, one way of presenting them are polar diagrams, giving limiting significant wave heights as a function of encounter angle and ship speed. The parameters of the polar diagrams are the wave length, the loading condition, and the exceedance of a certain threshold value that might be a particular roll angle, accelerations at a certain position of the vessel, stresses in the structure, or other.
The project was structured into 7 work packages (WP), as follows:
- WP1: Definition of the criteria
- WP2: Environment
- WP3: Numerical simulation
- WP4: Ship data
- WP5: Man machine interface
- WP6: Integration
- WP7: Validation.
ADOPT provides a process with the three major process elements:
mode 3 - design,
mode 2 - training and
mode - operation.
Decision support for mode 1 starts in mode 3. Mode 2 is required to convey the information generated in mode 3 to mode 1. Operation of a ship in mode 1 will be improved by mode 2 even without a specific computerised mode 1 DSS. A computerised mode 1 DSS will further improve the operation in mode 1 by giving real time support, which might capture scenarios beyond the experience or capabilities of the crew.
The application of complex numerical methods in a risk based framework and the high reliability requirements made for the on-board decision support revealed the following major constraints during the development of the system:
- Within a 3 hours watch it is not feasible to obtain final results from the numerical calculations, if all uncertainties are taken into account.
- The quality of the input data is not always as high as required for a reliable decision support.
- The expertise to make full use of a decision support as complex as the ADOPT-DSS is normally not available on board.
For handling the mentioned constraints, the concept of the ADOPT-DSS was extended from a pure software-hardware package for the on-board use to a procedural process consisting of three modes of use, which are the design mode (Mode 3), the training mode (Mode 2) and the operational mode (on-board use, Mode 1), each accompanied by an appropriate set of tools. By combining the possibilities of each mode of use, the limitations shown above can be diminished and overcome to a certain degree. The time consuming calculations are done in the design mode, where the time constraint of a watch length is not relevant. Insufficient quality of input data, such as the loading condition, is solved by a quality control which uses pre-calculated data, verified in the design mode to be of sufficient accuracy.
The knowledge about the ship specific behaviour with respect to the individual hazards gathered during the design process is transferred to the crew on board via training sessions in Mode 2. Within this framework the master and the crew are also trained and familiarised with the concepts and the handling of the ADOPT-DSS. For this purpose the Mode 1 tools can be used in a simulated environment. This is an important prerequisite for the appropriate and responsible use of the ADOPT-DSS in operating conditions.
A further conclusion of the ADOPT-project is the necessity of a clear distinction between economic and safety risk as both modes are associated with fundamentally different risk acceptance criteria.
Further, two parallel branches of ship motion calculation and probability assessment have been developed, where a frequency domain provides fast and efficient decision support with respect to all limit states which show a mainly linear response behaviour, whilst a time domain implementation is provided for strongly non-linear responses such as parametric rolling.
The feasibility of the ADOPT concept has been proven by a demonstration prototype which was tested in a simulated environment. The consortium concludes that the ADOPT concept has clearly demonstrated its feasibility.
However, the ADOPT-DSS demonstrator has to be improved and further developed, especially with respect to response time, before practical on-board use is possible. Amongst others, the numerical processes of the ADOPT-DSS demonstrator have not been optimised so far. Additionally it is anticipated that the constantly increasing available computer power will assist in solving this problem in a few years.
The inclusion of relevant uncertainties from all sources (high level requirement 5) cannot be fulfilled due to time constraints. A sensitivity and uncertainty analysis to identify relevant and unimportant uncertainties may reduce the number of relevant uncertainties and thereby increase response time and/or reliability.
In the case of safety related hazards such as capsize the high level requirements regarding the display of reliability of the decision support (high level requirement 6 and 29) is only satisfied in the way of non-display, when the input data is unreliable. Improvements of the quality control and adequate displayed of the decision support reliability may be subject to further research.
Further some identified, generic hazards, such as broaching to, cannot be assessed by the numerical ship motion tools applied in the ADOPT-DSS today. Development progress on the applied numerical methods may in the future extend the possibility to include all generic hazards. The modular layout of the ADOPT-DSS also gives opportunity to include additional numerical assessment of limit states parallel to the two branches of ship motion simulation.
