On this page, you can find links to our internal MSDL report publications, as well as abstracts and bibliographic information for publications in peer-reviewed journal articles and conferences. Older publications by Dr. Collette before he founded the MSDL are also included. Where allowed by copyright, full text downloaded or pre-prints of articles are also provided. Papers where source code and data is provided for provenance will have a link below the abstract to access the code or data. Zotero metadata is also provided for the publications.
Department of Naval Architecture
and Marine Engineering
210 Naval Architecture and Marine Engineering
2600 Draper Dr., Ann Arbor MI 48109
Mark Groden; Chair: Matthew D. Collette. "A Probabilistic Graphical Framework Fusing Data for Model Updating and Decision Support". P.h.D University of Michigan-Ann Arbor. 2016
There is significant uncertainty in assessing the structural health and capabilities of a marine structure during both service life and after sustaining damage. Design-stage marine structural engineering models offer limited information on the as-built structure's health during service life. Despite copious amounts of data provided by structural monitoring techniques, synthesizing these different data types to update the design-stage models remains challenging. A novel decision support graph was created by extending a parametrically encoded Bayesian network data fusion framework to influence diagrams for data to decision. The data to decision framework combines observational and sensor through-life data to update the design-stage models. Once updated, these models provide predictions of future structural health and safety, decision support for inspection timing and extent, and decision support to emergency response teams for survival and mission objective satisfaction strategies. To demonstrate the effectiveness of the Bayesian network parametrically encoded data fusion, a lognormal probabilistic fatigue initiation model was developed for a series of large stiffened metallic grillages; grillages consist of identical fatigue-critical details typical of vessel and platform structures. Monte-Carlo simulations were used to compare the Bayesian network's prognosis with the synthetic data. Evidence for inference includes data acquired from visual inspection, operating conditions, and an innovative stand-alone mechanical strain sensor, the Strain Amplification Sensor, developed as a part of this work. Results demonstrated that the Bayesian network produces better estimates for fatigue crack initiation through addition of various pieces of evidence. Successful prognosis led to the adaptation of the network to provide inspection guidance, and to aid in decision-making given a damaged marine structure.
Yan Liu; Chair: Matthew D. Collette. "Surrogate-Assisted Unified Optimization Framework for Investigating Marine Structural Design Under Information Uncertainty". P.h.D University of Michigan-Ann Arbor. 2016
Structural decisions made in the early stages of marine systems design can have a large impact on future acquisition, maintenance and life-cycle costs. However, owing to the unique nature of early stage marine system design, these critical structure decisions are often made on the basis of incomplete information or knowledge about the design. When coupled with design optimization analysis, the complex, uncertain early stage design environment makes it very difficult to deliver a quantified trade-off analysis for decision making. This work presents a novel decision support method that integrates design optimization, high-fidelity analysis, and modeling of information uncertainty for early stage design and analysis. To support this method this dissertation improves the design optimization methods for marine structures by proposing several novel surrogate modeling techniques and strategies. The proposed work treats the uncertainties that are sourced from limited information in a non-statistical interval uncertainty form. This interval uncertainty is treated as an objective function in an optimization framework in order to explore the impact of information uncertainty on structural design performance. In this examination, the potential structural weight penalty regarding information uncertainty can be quickly identified in early stage, avoiding costly redesign later in the design. This dissertation then continues to explore a balanced computational structure between fidelity and efficiency. A proposed novel variable fidelity approach can be applied to wisely allocate expensive high-fidelity computational simulations. In achieving the proposed capabilities for design optimization, several surrogate modeling methods are developed concerning worst-case estimation, clustered multiple meta-modeling, and mixed variable modeling techniques. These surrogate methods have been demonstrated to significantly improve the efficiency of optimizer in dealing with the challenges of early stage marine structure design.
Thomas Devine; Chair: Matthew D. Collette. "Using and Interpreting the Bayesian Optimization Algorithm to Improve Early Stage Design of Marine Structures". P.h.D University of Michigan-Ann Arbor. 2016
Early stage naval structural design continues to advance as designers seek to improve the quality and speed of the design process. The early stages of design produce preliminary dimensions or scantlings which control the cost and structural performance of a vessel. Increased complexity in the evaluation of structural response has led to a need for efficient algorithms well suited to solving structural design specific optimization problems. As problem sizes increase, existing optimizers can become slow or inaccurate. The Bayesian Optimization Algorithm (BOA) is presented as one solution to efficiently solve problems in the structural design optimization process. The Bayesian optimization algorithm is an Estimation of Distribution Algorithm (EDA) that uses a statistical sample of potential design solutions to create and train a Bayesian network (BN). The application of BNs is well suited for nearly decomposable problem composition which closely matches rules based structural design evaluation. This makes the BOA well suited to solve complex early stage structural optimization problems. Additionally, the learning processes used to create and train the BNs can be analyzed and interpreted to capture design knowledge. This return of knowledge to the designer helps to improve designer intuition and model synthesis in the face of more complex and intricate models. The BNs are thus analyzed to augment design problem understanding and explore trade-offs within the design space. The result matches a paradigm shift in early stage optimization of naval structures. Designers gain better understanding of critical design variables and their interactions as compared to the previous focus on the single most optimal solution. This leads to efficient simulations which rapidly explore design spaces, document critical design variable relationships and enable the designer to create better early stage design solutions.
Temple, D. and M. Collette. 2016. “Understanding Lifecycle Cost Trade-Offs for Naval Vessels: Minimizing Production, Maintenance, and Resistance.” Ships and Offshore Structures (Published Oct 2016).
Current optimisers struggle to explore the multi-disciplinary trade-space that defines vessel lifecycle cost. This paper tests an enhanced multi-objective collaborative optimiser on a three-objective hull and structural lifecycle costing problem. The optimiser extends multi-disciplinary collaborative optimisers by including goal-programming and a novel genetic algorithm at the discipline level. The optimiser finds the trade-space between vessel resistance, production cost, and structural maintenance cost. Simultaneous changes to the hullform geometry and structural scantlings are used to explore this design space. The approach is demonstrated on a naval optimisation problem. With fixed maintenance schedules, the trade-space is shown to be steeply walled. This topology indicates that single-discipline optimisation for lifecycle cost may lead to large increases in lifecycle cost for the disciplines not considered. In conclusion, multi-disciplinary optimisation is shown to be a useful tool for addressing lifecycle costing during design.
Liu, Y., H. Jeong and M. Collette. 2016. “Efficient optimization of reliability-constrained structural design problems including interval uncertainty.” Computers and Structures, 177, p. 1-11 11 p.
A novel interval uncertainty formulation for exploring the impact of epistemic uncertainty on reliability-constrained design performance is proposed. An adaptive surrogate modeling framework is developed to locate the lowest reliability value within a multi-dimensional interval. This framework is combined with a multi-objective optimizer, where the interval width is considered as an objective. The resulting Pareto front examines how uncertainty reduces performance while maintaining a specified reliability threshold. Two case studies are presented: a cantilever tube under multiple loads and a composite stiffened panel. The proposed framework demonstrates its ability to resolve the Pareto front in an efficient manner.
Zhu, J., and M. Collette 2016. “A Bayesian approach for shipboard lifetime wave load spectrum updating.” Structure and Infrastructure Engineering. 1-15.
Updating design stage loading predictions with in-service measurements is an essential step for predicting the fatigue life of complex structures. This is especially true when the loading process is not statistically stationary. Ships and marine structures are an example of such structures. They are fatigue prone from being subject to alternating wave forces, and can also change their loading environment owing to their mobility. This manuscript proposes a two-level offline lifetime load updating scheme to address updating of non-stationary load processes. First, corrections to hydrodynamic predictions are established in short-term statistically stationary operating cells based upon spectral parameters. A hierarchical Bayesian model interpolates corrections between cells and infers correction factors for cells that are not yet observed. Finally, projected operational profiles of the vessel are used to combine cell predictions into an updated lifetime fatigue load profile. The effectiveness of this process is demonstrated on an example vessel in two case studies.
Temple, D., and M. Collette. 2016. “A Goal-Programming Enhanced Collaborative Optimization Approach to Reducing Lifecycle Costs for Naval Vessels.” Structural and Multidisciplinary Optimization, 1–15.
Understanding the trade-offs involved in assessing lifetime cost for engineering systems requires understanding trends in various engineering disciplines that require significantly different analysis methods to efficiently explore. The corresponding design spaces can be flat, defined by weak minima, and thus difficult to understand using traditional optimization methods. This paper presents a new multi-disciplinary framework that uses a goal-programming enhanced multi-objective collaborative optimization (eMOCO) approach to facilitate the development of the spaces. In order to further increase its efficiency in discrete or flat spaces well-suited to evolutionary optimization a unique discipline level genetic algorithm is proposed. Naval vessels are an example of an engineering system that has a difficult design space with respect to lifetime cost, however, one where it is critical to understand. As these costs are increasing, they are becoming limiting factors in a vessel’s operational life. Though they are so important, the interaction between different cost categories such as production and operation has not been explored in depth and is not always clear. Understanding the trade-offs between different aspects of a vessel’s total ownership costs early in the design stage can aid in the production of new ships where they are minimized. The proposed framework is verified on mathematical problems, and then used to develop trade-spaces between resistance and production for a nominal naval combatant vessel. These trade-spaces show both the knowledge gained by designers in understanding these trade-offs and the ability of the proposed eMOCO framework to develop them effectively.