To demonstrate the Open Simulation Platform the OSP Joint Industry Project has developed three use cases. The purpose has been to give users a sense of how the platform works and can generate value for the industry as well as providing important feedback to the OSP development.
The use cases have developed simulation set-ups using OSP software and interface specifications for:
- Design of a hybrid ferry propulsion system (lead by Sintef Ocean)
- Virtual commissioning of a coastal service vessel (lead by Kongsberg Maritime CM)
- Operational planning for crane operation on R/Y Gunnerus (lead by DNV GL)
Design
In the first use case, we demonstrate the capability of OSP as a collaboration tool in the conceptual design phase where the uncertainty of the design is high, and one needs to explore as many options as possible to find the optimized design to start with.
The target system for this use case is a hybrid power system in a passenger ferry, combining electrical propulsion system with energy storage, see an example in figure 1. This is an industrial area where technology is rapidly developing in order to minimize emissions and fuel consumption. Having batteries combined with electric and mechanical power distribution systems allows the plant to operate beyond the strict energy balance requirement as in conventional systems and allows a designer flexibly to choose the rated power of the main power sources.
However, this added degree of freedom in operation and design comes with challenges in the complexity of design. For example, the sizing of the energy storage system depends on the actual load profile and control strategy of the system. And the number of possible configuration and a set of dimensions for the power sources for evaluation increase significantly. In this regard, static analysis for the power balance does not provide the enough design input for the hybrid system. Furthermore, the knowledge of each main component of the system still remains with the specific supplier and makes it challenging for designer to create all the necessary knowledge to perform the system assessment.
The use case demonstrates how to tackle such challenges by using simulation models as a tool for description of requirements, evaluation of designs and system integration and for assuring the transfer of knowledge through different stages of the design. Furthermore, system simulations enable the designer to efficiently explorer a large design space which is essential for designing a system with large flexibility in an optimal way.
Open Simulation Platform allows co-simulation where multiple parties provide their own component models in FMU form to a system integrator without risk of disclosing their proprietary knowledge. The system integrator for the use case has been SINTEF Ocean who has configured a system model and performs the system simulation. The system model for the simulation is presented in Figure 2 and shows the partners that have delivered models to the set-up; SINTEF Ocean, Damen Shipyards, DNV GL, Corvus Norway and Vard.
The first challenge for the design case was to choose proper models. A proper model has a certain level of fidelity to answer the question with acceptable accuracy and can be run fast enough with reasonable resources. When the system is multi-disciplinary system, this is especially important as it is necessary to establish the balance of the fidelity among different component models.
A second requirement was to have a co-simulation environment that enables to run many simulations in parallel, this was solved by using the command line interface (CLI) cosim. Using cosim, retrieving the information from the FMU component model, testing the component and running system simulations could be easily integrated in a design tool developed as a demo application for the JIP project. In this integrated tool, a user can manage the FMUs of different sources, configure a system and scenario and run many simulations in parallel, and present the simulation results. All the created parameters, configurations, FMUs and results are saved in a database.
The simulation study will be performed with two design parameters: rated power for the gensets and power upper limit for zero-emission operation of a given load profile for the passenger ferry. Combination of these two parameters are sampled using Sobol sequence sampling in order to effectively cover the design space. The simulation is run in two stages for each case. The first simulation will determine the rated capacity of the battery depending on the power and energy usage during the simulation. Then, the second simulation will be performed to find out actual fuel consumption and running hours which are key contribution to OPEX.
Video 1: Showing a summary of the OSP use case, design tool and proof of concept.
Virtual commissioning
The main objective of the virtual commissioning use case was to show that the OSP software, libcosim, could be used as a virtual commissioning tool for control system(s) software.
When performing virtual commissioning, the suppliers of equipment to a vessel will share their models and digital twin components, including exact copies of control system software embedded in the physical systems, such that system simulations can run to perform integration testing with multiple scenarios far exceeding what is possible to test during a traditional sea trial. The concept is shown in figure 3.
A key element of the Open Simulation Platform is that stakeholders can collaborate in building the digital twin vessels based on common standards. This is necessary in order to create an efficient integration process.
In order to demonstrate OSP software for virtual commissioning, the use case has covered several aspects:
- Combining multiple control systems running in closed loop with simulation models (FMUs) where FMUs and control systems are delivered from multiple suppliers.
- Simulations run in real-time and models with sufficiently high fidelity to operate control systems as if they were controlling the real asset.
- Handling control systems with different sampling rates.
- Performing efficient interfacing, both from models to models, and models to control systems.
- Running software-in-the-loop-testing in the virtual setup to replicate testing during real commissioning.
- Virtualization of selected control systems and control system HMIs
The setup is shown in figure 4 and includes two different control systems from Kongsberg Maritime CM (former Rolls-Royce Marine), and a K-Pos DP system from Kongsberg Maritime. On the simulation side, a hull model from DNV-GL, and thruster, power system and sensor models from Kongsberg Maritime CM has been included. Due to the large number of signals needed for the PMS in this case, the main part of the power simulation had to run on the control side, using the internal PMS simulator.
There have been no or little issues with the performance of libcosim for this use case. Several control systems using different sample times are running in closed loop. There have been no stability issues in this use case, and signals can easily be manipulated using the built-in functionality of the cosim demo app.
In addition, a set of network FMUs has been developed to take care of communication between the control systems and simulator. Network FMUs are network protocols wrapped as FMUs. These enable suppliers of control systems to implement their preferred protocol, wrap it as an FMU, and deliver it to the system integrator along with the control system and any model FMUs. This means that the simulator core does not have to support any specific interfaces. The communication between network FMUs and model FMUs is set up by the system integrator in the same way as model FMU to model FMU communication is set up. Also, since libcosim and the cosim demo app is open source, any supplier can use these to verify that models and Network FMUs are running according to expectation before supplying them to the system integrator.
To keep the communication modular and preserve system topology, it was decided to use several network FMUs for each control system, depending on the type of interface. As an example, the DP system requires sensor and position reference feedback from the simulator in order to perform station keeping of the simulated vessel. This information is normally communicated using one or more standard NMEA 0183 messages per device (sensor/position reference system). Since the number of devices varies from vessel to vessel, it was decided to make one network FMU for each device, enabling reuse and a one to one relationship between sensor models and network FMUs.
The main challenges when integrating systems from several suppliers will be the large number of signals for some systems and lack of standardization of interfaces and models in the maritime industry. The work done on connections in the OSP project has helped improve this, but further work on standardization and good tools for connecting control systems and models will clearly be very important in the future.
The use case is demonstrated in the following two videos.
Video 2 demonstrates the interaction between the K-Pos DP system, Mcon thruster control system and simulation running in the cosim demo app.
Video 2: K-Pos DP, Mcon thruster control and simulation
Video 3 shows ACON PMS and K-Pos DP connected to the same simulation running in the cosim demo app. The Mcon thruster control system is also connected but is not shown in the video for simplicity.
Video 3: ACON PMS, K-Pos DP, Mcon thruster control and simulation
Operational Planning
A strong motivation for the establishment of a system digital twin is easy access to simulation results during vessel operation. New tools and services for operational planning based on co-simulation is one application that can add significant value to vessel owners and operators assuring efficient and safe operations. A use case for operational planning was therefore established to investigate how the OSP software can:
- simulate an operation to support choice of operational strategies and system settings
- efficiently simulate changing environment (wind, waves and current) to investigate limiting conditions for specific operations
- handle the challenge of simulating tightly coupled systems (for example crane and vessel interaction)
- provide output for visualization tools considered essential for training and operational planning
The rationale behind using Open Simulation Platform for operational planning is that:
- Model protection enabled by use of FMI standard means less need to share system specific information (IPR) between different stakeholders
- That simulation tools can be easily updated by suppliers to reflect the current status of the vessel, through replacing or updating component models and software
- Reuse of models from the design and commissioning phase is facilitated by complying with common interface specifications, increasing the value of investing in validated and “flexible” models for a newbuild project
The NTNU research vessel R/V Gunnerus performing a crane lift operation for installation of subsea equipment was chosen as basis for this case study. Component models (FMUs) representing Gunnerus hull and main propulsion system were already available developed by Sintef Ocean and NTNU in previous research projects such as ViProMa. The vessel is also case study for the ongoing project TwinShip. In 2019 Gunnerus hull was lengthened midships and a new knuckle boom crane was installed, the set of FMUs are therefore updated with a modified hull model and a new crane model.
Re-implementing the use-case in OSP software environment demonstrated that libcosim and the cosim demo app was very efficient and stable for running simulations. The crane model runs at 1000Hz while the remaining models run at 20Hz. Challenges when establishing the use case emerged mainly from lack of documentation of the FMUs and their interfaces. In addition, the Sintef software VeSim generates FMUs that must be run in a distributed environment making debugging of system set-up slightly more challenging. Figure 5 shows the components included in the simulation set-up, where green frames marks components run by distributed co-simulation using the FMU-proxy framework.
An updated interface configuration for the models and configuration according to OSP-IS was implemented using the configuration tool kopl. The motivation was to allow for more efficient and verifiable reuse of FMUs and system configurations in the future.
A crane welded to a ship is tightly coupled through causality and pose challenges for time-domain simulations. The optimal numerical solution would be to combine crane and vessel model in one FMU. To allow separation between suppliers, the use of separate FMUs was demonstrated in this use case. Coupling of crane and vessel was solved by having vessel motion as input to the crane and the reaction forces from the crane are input to vessel model. The vessel model represents hydrodynamic properties of hull calculated for a specific center of gravity. When changing position of crane and load on a vessel these properties will change as center of gravity changes, thus the current set-up must be used with care or be further developed for detailed planning of heavy lift operations.
A scenario file is implemented and used to specify crane operation, environment and DP control parameters. A video visualizing one of these scenarios was produced by OSC and can be seen below in video 3.
Video 3: Visualization of scenario
Figure 6 show the cosim demo app plotting the crane tip position (NED frame, relative to vessel center of gravity) for the same scenario.
A short video will soon be available and shared here showing a summary of the use case demonstrating the use of the OSP software.
