Special Issue: Integrated Design and Operation of Engineering Systems With Predictive Modeling

Growing trends toward increased complexity and prolonged useful lives of engineering systems present challenges for system designers in accounting the impacts post-design activities on performance (e.g., costs, reliability, customer satisfaction, environmental impacts). Examples include manufacturing, condition monitoring, remaining life prediction, operations maintenance, service logistics, as well end-of-life options. It is difficult to develop accredited lifecycle models because these only occur after built operated. Thus, design decision-making have traditionally been addressed separately, leading suboptimal over system’s lifecycle.With significant technological advances computational modeling, simulation, sensing machine learning artificial intelligence, capability predictive modeling has grown exponentially past decade, demonstrated benefits such improved availability reduced operation maintenance costs. Predictive can bridge stages provide an optimal pathway effectively account future at stage. While potentially enables more holistic decisions, there a need acquire knowledge various aspects this emerging topic, novel methodology activities, concepts incorporating into decision-making.With 13 papers, special issue brings together fundamental scientific contributions across different areas related integrated (IDOES) with approaches, concepts, applications. Based research objective, articles are broadly grouped five themes: (i) current literature IDOES, (ii) approach (iii) new concept (iv) digital-twin design, (v) applications IDOES. In following, briefly summarized within identified themes accordingly. Current Literature IDOESIn article titled Towards Integrated Design Operation Complex Engineering Systems With Modeling: State-of-the-Art Challenges, Liu et al. conducted study where approaches strategies integrating processes categorized. Although handled from data-driven, statistical, analytical, empirical aspects, recent problems started evaluate performance, still field that require active investigation exploration. Toward end, provides summary directions closure, encouraging collaborations among communities interested design.Novel Modeling Approach IDOESMulti-fidelity calibration data fusion tasks ubiquitously arise design. paper Data Fusion Latent Map Gaussian Processes, Eweis-Labolle critical general techniques jointly fuse multiple datasets varying fidelity levels while also estimating parameters. The authors introduced converts latent space problem using latent-map processes, relations sources be automatically learned. By assimilating simultaneously proposed method, prediction multi-fidelity improved, shown reported case studies.Condition monitoring plays crucial role improving failure resilience, preventing tragic consequences brought by unexpected events avoiding consequential increases To integrate designs operations, systematic framework needed assess value stage, which would allow decisions adopting maximize benefits. Valuation Continuous Monitoring System Recurrent Maintenance Decision Scenarios, presented based information. expected cost reductions under specific modes considering effectiveness continuous predicting failures.In Reinforcement Learning-Based Sequential Batch-Sampling Bayesian Optimal Experimental Design, Ashenafi sequential batch-sampling method developing models. developed referred sampling via reinforcement experimental used optimize black-box expensive-to-compute experiments or computer codes. algorithm select batches queries entire budget hand, retains nature elements reward domain deep learning.New Concept Artificial Intelligence Ship Structures: A Variant Multiple-Input Neural Network-Based Resistance Prediction, Ao data-driven intelligence (AI)-based model assist ship hull process. Specifically, AI-based multiple-input neural network was implemented realize real-time total resistance structure inconsistent estimates types input tool process accurately hulls real time.Path tracking error control essential functionality autonomous vehicles follow planned trajectory, path errors could lead collision affect vehicle significantly. model-based currently available, bias baseline may result errors. Bias-Learning-Based Model Controller Reliable Path Tracking Autonomous Vehicles Under Environmental Uncertainty, Ren bias-learning coupled improve few experiments, so operation. their study, regression recurrent were employed learning, compared uncertainty scenarios.In Unmanned Cable Shovel Multiobjective Co-Design Optimization Structural Control Parameters, Zhang multi-stage multi-objective co-design optimization strategy unmanned cable shovels excavation loading performed stages, making it obtain global solution. point-to-point motion trajectory energy consumption working process, then synchronously structural parameters, thereby mining efficiency reducing scenarios.Novel Methodology Digital-Twin DesignIn Functional Modeling-Based Digital Twin Architecture Representation: An Instructional Example COVID-19 Breathalyzer Kiosk, Kotecha functional modeling-based representation architectures illustrated instructional example testing breathalyzer kiosk They provided review existing frameworks intended use product digital twin architecture opens venues helping twins.In Probabilistic Additive Manufacturing Process Control, Nath constructing probabilistic laser power bed (LPBF)-based additive manufacturing (AM) incorporates variability. resulting thus tailored individual part being produced AM parameter online, adjustment LPBF porosity manufactured part.In Sensing Strategies Interpretable Machine Learning, Kapteyn sensor placement dynamic scheduling twins. assimilation posed classification problem, train trees represent map observed estimated states. addition providing rapid updating capability, yield interpretable mathematical queried inform decisions.Novel Applications Mobility Prediction Off-Road Ground Using Dynamic Ensemble NARX Models, surrogate mobility off-road ground (AGVs), AGV mission planning especially early ensemble nonlinear autoregressive exogenous (NARX) time synthetic limited number high-fidelity simulations. advantages prediction.In Reliability-Based Multivehicle Planning Uncertainty Bio-Inspired Approach, bio-inspired multi-vehicle AGVs subjected reliability constraints environments. Identifying reliable uncertain environments designing operations. adaptive physics-based simulations utilized predict state operation, subsequently, called Physarum-based conjunction navigation mesh identify satisfying requirement.In Resource Allocation System-of-Systems Heuristic-Based Deep Chen two-tier framework, resource allocation acknowledging autonomy constituents (SoS) decouples SoS manager ensures not compromised interventions executed manager. applied customized Open AI Gym environment results methods show superior set key parameters.In summary, range models, methodology, As such, they representative wide variety intersection systems. upshot insight how agenda further developments generally knowledge, technology, organizational capacity our great privilege organize assemble Journal Mechanical Design. We hope you will find papers offering information inspiration advance frontier important area research. contents continue stimulate advanced techniques.