Search

Optimization methods are essential to improve the operation of energy conversion systems including energy storage equipment and fluctuating renewable energy. Modern systems consist of many components, operating in a wide range of conditions and governed by nonlinear balance equations. Consequently, identifying their optimal operation (e.g. minimizing operational costs) requires solving challenging optimization problems, with the global optimum often hidden behind many local ones. In this work, we propose a hybrid method that advantageously combines Bayesian optimization (BO) and Interior Point OPTimizer (IPOPT). The BO is a global approach exploiting Gaussian process regression to build a surrogate model of the cost function to be optimized, while IPOPT is a local approach using quasi-Newton updates. The proposed BO-IPOPT combination allows leveraging the parameter space exploration of the BO with the quasi-Newton convergence of IPOPT once solution candidates are in the neighborhood of an optimum. Using a challenging constrained test function, we test BO-IPOPT in accuracy and computational efficiency. Finally, we showcase the proposed method in the optimal operation of a renewable steam generation system. The results show that BO-IPOPT combines high accuracy and computational efficiency, achieving up to 50% better objective function values at the same CPU time than other state-of-the-art methods. • We propose a hybrid algorithm based on Bayesian Optimization and IPOPT. • BO-IPOPT effectively combines global exploration of BO and local efficiency of IPOPT. • The optimal operation of a nonlinear industrial system serves as real-world use case. • BO-IPOPT shows better performance than competing methods. [ABSTRACT FROM AUTHOR]

Highlights • Scaling laws for the Jet Wiping Process are derived in a one-way coupling framework. • The non-uniformities on the final film after wiping are characterized experimentally. • The wave characteristics are highly dependent on the jet to substrate standoff distance. • The results suggest a fundamental role of the jet confinement and a two-way coupling between the gas jet and the liquid. Abstract The jet wiping process is a continuous coating method that uses impinging gas jets to control the thickness of a liquid layer dragged by a moving substrate. This first part of a two-part article on the analysis of the process stability presents an experimental characterization of the achievable coating uniformity as a function of various design and operating parameters, including the nozzle opening width, stand-off distance, tilt angle and stagnation pressure, substrate speed and liquid properties. The process is reproduced on a dedicated laboratory model, instrumented with a Light Absorption (LAbs) test section for the 3D time-dependent reconstruction of the final coating film thickness. Several types of wave patterns are observed on the final coating film and characterized in terms of peak to peak wave amplitude and dominant frequency. The results are nondimensionalized following the scaling laws derived from the classical 'one-way coupling' analysis of the process. This approach, which assumes that the wiping capabilities of the jet flow are not influenced by the dynamics of the impinged liquid, is proven successful for the prediction of the average film thickness but inadequate for the scaling of the coating non-uniformity. This suggests that the dynamics of the final coat is mostly governed by the interaction of the liquid with the gas jet, investigated in the second part of the article. [ABSTRACT FROM AUTHOR]

Research in the United States evaluating ecotoxic risk to receiving waters posed by contaminants occurring in wastewater discharges typically has focused on measurements of pharmaceuticals and personal care products (PPCPs), with limited evaluations of UV filters and phenylpyrazole and neonicotinoid pesticides. In this study, concentrations of 5 representative pharmaceuticals, 11 pesticides or pesticide degradation products, and 5 ultraviolet filters were measured in 24 h composite samples of six wastewater discharges representing ∼70% of the total wastewater discharged to San Francisco Bay during the summer and fall of 2021. No significant difference was observed between concentrations measured on weekdays vs. weekends. A hydrodynamic model of San Francisco Bay was used to estimate annual average dilution factors associated with different subembayments. With and without considering dilution effects, Risk Quotients were calculated using the 90th percentile of measured concentrations in wastewater effluents and threshold concentrations associated with ecotoxicity. Risk Quotients were highest for the neonicotinoid pesticide, imidacloprid, and exceeded ecotoxicity thresholds in the lower South Bay by a factor of 2.4, even when considering dilution. Compared to commonly measured pharmaceuticals, Risk Quotients for imidacloprid were higher than those for carbamazepine, trimethoprim and diclofenac, and comparable to those for propranolol and metoprolol. Risk Quotients for the pesticide, fipronil, and the UV filter, oxybenzone, were higher than for carbamazepine. The results highlight the need to incorporate pesticides and UV filters with high Risk Quotients into studies in the United States evaluating ecotoxic risk associated with contaminants in municipal wastewater discharges. [Display omitted] • Ultraviolet filters, pesticides, pharmaceuticals measured in 6 wastewater effluents. • No difference in concentrations observed between weekday and weekend. • Imidacloprid featured the highest ecotoxic risk quotient among 21 contaminants. • Risk quotients for oxybenzone and fipronil were comparable to commonly measured PPCPs. [ABSTRACT FROM AUTHOR]

The thermal management of cryogenic storage tanks requires advanced control strategies to minimize the boil-off losses produced by heat leakages and sloshing-enhanced heat and mass transfer. This work presents a data-assimilation approach to calibrate a 0D thermodynamic model for cryogenic fuel tanks from data collected in real time from multiple tanks. The model combines energy and mass balance between three control volumes (the ullage vapor, the liquid, and the solid tank) with an Artificial Neural Network (ANN) for predicting the heat transfer coefficients from the current tank state. The proposed approach combines ideas from traditional data assimilation and multi-environment reinforcement learning, where an agent's training (model assimilation) is carried out simultaneously on multiple environments (systems). The real-time assimilation uses a mini-batch version of the Limited-memory Broyden–Fletcher–Goldfarb–Shanno with bounds (L-BFGS-B) and adjoint-based gradient computation for solving the underlying optimization problem. The approach is tested on synthetic datasets simulating multiple tanks undergoing different operation phases (pressurization, hold, long-term storage, and sloshing). The results show that the assimilation is robust against measurement noise and uses it to explore the parameter space further. Moreover, the work shows that simultaneously sampling from multiple environments accelerates the assimilation. • The framework identifies closure relations from real-time data in cryogenic tanks. • A learning agent adjusts the closure relations in a 0D model to match observed data. • The closure relations are approximated via a feed-forward Artificial Neural Network. • The framework benchmarking varied data noise, observation times, and batching method. • The out-of-sample error reduces as the agent concurrently learns from multiple tanks. [ABSTRACT FROM AUTHOR]

Space missions to Near Rectilinear Halo Orbits (NRHOs) in the Earth–Moon system are upcoming. A rendezvous technique in cislunar space is proposed in this investigation, one that leverages coupled orbit and attitude dynamics in the Circular Restricted Three-body Problem (CR3BP). An autonomous Guidance, Navigation and Control (GNC) technique is demonstrated in which a chaser spacecraft approaches a target spacecraft in a sample southern 9:2 synodic-resonant L2 NRHO, one that currently serves as the baseline for NASA's Gateway. A two-layer guidance and control approach is contemplated. First, a nonlinear optimal controller identifies an appropriate baseline rendezvous path for guidance, both in position and orientation. As the spacecraft progresses along the pre-computed baseline path, navigation is performed through optical sensors that measure the relative pose of the chaser relative to the target. A Kalman filter processes these observations and offers state estimates. A linear controller compensates for any deviations identified from the predetermined rendezvous path. The efficacy of the GNC technique is tested by considering a complex scenario in which the rendezvous operation is conducted with an uncontrolled tumbling target. Hardware-in-the-loop laboratory experiments are conducted as a proof-of-concept to validate the guidance algorithm, with observations supplemented by optical navigation techniques. • Coupled orbit and attitude dynamics for rendezvous in cislunar space. • Two-layer guidance and control algorithm; ArUco markers for vision-based navigation. • Proof-of-concept hardware-in-the-loop validation of rendezvous in halo orbits. • Robot Operating System (ROS) to integrate controller, robot, and pose estimator. [ABSTRACT FROM AUTHOR]

Ejectors are passive mixing devices that might operate in time-varying conditions, e.g., at start-up or during transients. Nevertheless, modeling efforts in literature have mostly been focused on steady conditions. In this work, we propose a novel 1D unsteady model for supersonic single-phase ejectors based on a pipeline analogy with two inlets and one outlet. Conservation of mass, momentum and energy is internally enforced in each domain through 1D gas dynamic equations, and across the domains by a junction model. The characteristics-based formulation of the junction admits a solution in the critical and the sub-critical regime, including backflow, as long as the primary nozzle is choked. Furthermore, the model is not restricted to sonic or supersonic ejectors. To the authors' knowledge, this formulation is the first to enable 1D ejector simulations with time-varying boundary conditions across operating regimes. The model was calibrated and validated on experimental data in steady state and benchmarked against a 2D URANS simulation featuring a sudden increase in back pressure, triggering traveling pressure waves and a burst of backflow. The results show that the model can correctly predict the ejector performance and the stream-wise evolution of relevant integral quantities in both steady and transient conditions. • An unsteady 1D ejector model based on a pipeline analogy is presented. • The model is validated in steady state against published experimental data. • The model is verified in transient conditions against a 2D URANS simulation. • Transitions between on- and off-design points, some producing backflow, are analyzed. [ABSTRACT FROM AUTHOR]

• A data-driven algebraic model for the turbulent heat flux is developed. • The ANN is trained with DNS data of near unity and low Prandtl number flows. • Invariance and realizability properties are embedded into the model structure. • The model adapts satisfactorily to different Prandtl numbers and flow conditions. • The model is highly sensitive to the accuracy of the momentum modelling. Liquid metals play a central role in new generation liquid metal cooled nuclear reactors, for which numerical investigations require the use of appropriate thermal turbulence models for low Prandtl number fluids. Given the limitations of traditional modelling approaches and the increasing availability of high-fidelity data for this class of fluids, we propose a Machine Learning strategy for the modelling of the turbulent heat flux. A comprehensive algebraic mathematical structure is derived and physical constraints are imposed to ensure attractive properties promoting applicability, robustness and stability. The closure coefficients of the model are predicted by an Artificial Neural Network (ANN) which is trained with DNS data at different Prandtl numbers. The validity of the approach was verified through a priori and a posteriori validation for two and three-dimensional liquid metal flows. The model provides a complete vectorial representation of the turbulent heat flux and the predictions fit the DNS data in a wide range of Prandtl numbers (Pr = 0.01 − 0.71). The comparison with other existing thermal models shows that the methodology is very promising. [ABSTRACT FROM AUTHOR]