SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
A new GAL4-based feed-forward loop circuit enhances β-glucuronidase (GUS) reporter gene expression in leaves and stems of stably transformed sugarcane plants.
A matrix converter system having a current control mode operation is provided. The system includes a matrix converter having a switching matrix. The matrix converter is coupled at its low-voltage side to a generator and at its output load side to a load. A controller having a pulse width modulation (PWM) control circuit is configured to control the matrix converter via its switching matrix to increase energy within the internal inductances of the generator when the switching matrix causes a short circuit. A feed forward calculator is configured to calculate a feed forward output phase angle. The feed forward output phase angle is an estimation of an angle between an output current vector and an output voltage vector that represent feedback signals of current and voltage output by the matrix converter. The angular position of the voltage output vector is adjusted as a function of the feed forward output phase angle to align angular position or phase angle of the voltage output vector that represents the voltage output with a selected angular position or phase angle.
A generalized feed-forward method for accurate tunable laser absorption spectroscopy includes generating a laser beam. The generated laser beam is directed down a reference path and a test/sample path. One or more parameters are extracted from the reference path. The one or more parameters, extracted from the reference path, are used as feed-forward, to adjust spectral analysis of the test/sample path to detect a composition and/or concentration of an analyte gas within the test/sample path. The extraction of the one or more parameters from the reference path and the spectral analysis of the test/sample path are performed substantially concurrently.
We develop new uncertainty propagation methods for feed-forward neural network architectures with leaky ReLU activation functions subject to random perturbations in the input vectors. In particular, we derive analytical expressions for the probability density function (PDF) of the neural network output and its statistical moments as a function of the input uncertainty and the parameters of the network, i.e., weights and biases. A key finding is that an appropriate linearization of the leaky ReLU activation function yields accurate statistical results even for large perturbations in the input vectors. This can be attributed to the way information propagates through the network. We also propose new analytically tractable Gaussian copula surrogate models to approximate the full joint PDF of the neural network output. To validate our theoretical results, we conduct Monte Carlo simulations and a thorough error analysis on a multi-layer neural network representing a nonlinear integro-differential operator between two polynomial function spaces. Our findings demonstrate excellent agreement between the theoretical predictions and Monte Carlo simulations.
The novel tri-port current-source medium-voltage string inverter (TCS-MVSI) is a promising candidate for large-scale PV-plus-storage (PVS) farms owing to its galvanic isolation, easy storage integration, soft-switching capability across entire load range, controlled low dv/dt and EMI, benign fault tolerance, etc. Due to its low inertia feature, traditional PI-based control cannot manage large transients effectively. Instead, a model-based predictive control (MPC) is proposed to achieve robust and stable operation. As is well known, the control performance of MPC is compromised by the sampling and computational delay during implementation significantly, if not well addressed. This paper analyzed and quantified these delays and then proposes feed-forward compensation (FFC) for the MPC method to compensate the delays and the large parameter variations due to the low-inertia nature. In addition, this method also compensates for the high dc-link ripple within each switching cycle, a unique issue for low-inertia converters. Here, the proposed method requires low computational effort, allowing it to be extended for multiple ports. The effectiveness of the proposed method has been validated in experiments. In 10 kW test, the proposed method decreases the average dc-link current by 17%, leading to ~20% conduction losses and ~0.5% increase in converter efficiency. In addition, the peak dc-link current also decreases by 15%, resulting in reduced transformer size. As a result, an increased power density can be achieved with the proposed method. Similar improvements have been observed across the power range from 2 kW to 10 kW.
Additive manufacturing (AM) enables the fabrication of complex designs that are difficult to create by other means. Metal parts manufactured by laser powder bed fusion (LPBF) can incorporate intricate design features and demonstrate desirable mechanical properties. However, printing a part that is qualified for its intended application often involves reprinting and discarding many parts to eliminate defects, improve dimensional accuracy, and increase repeatability. The process of iteratively converging on the appropriate build parameters increases the time and cost of creating functional LPBF manufactured parts. This work describes a fast, scalable method for part-scale process optimization of arbitrary geometries. Additionally, the computational approach uses feature extraction to identify scan vectors in need of parameter adaptation and applies results from simulation-based feed forward control models. This method provides a framework to quickly optimize complex parts through the targeted application of models with a range of fidelity and by automating the transfer of optimization strategies to new part designs. The computational approach and algorithmic framework are described, a software package is implemented, the method is applied to parts with complex features, and parts are printed on a customized open architecture LPBF machine.
Laser plasma accelerators, typically operating at 1–10 Hz repetition rates, have the ability to produce high-quality electron beams in compact, all-optical-driven configurations, with the electron beams uniquely suited for a wide variety of accelerator-based applications. However, fluctuations and drifts in the laser delivery to the meter-scaled and below plasma target (the electron beam source) will translate into electron beam source variations that can limit their utility for demanding applications like light sources or linear colliders. Commercially available active feedback laser stabilization systems are intrinsically bandwidth limited due to their integration with multi-inch corrective mirror mounts which minimizes their effectiveness. In this manuscript, we present a Neural Network time series forecaster that can predict laser position fluctuations of the laser delivery to the final target well ahead of a future laser shot. The Root-Mean-Square-Error (RMSE) of the prediction accuracy was < 2 μ m for a 1 / e 2 beam radius of 34 μ m . Our feed-forward approach serves as a first-step in circumventing the bandwidth limitations imposed by the currently available stabilization systems since it allows for mirrors to be moved into position ahead of time to offset the predicted future position drift. This will help advance laser plasma accelerator research by providing greater robustness and stability needed for its applications.
Bio-based products derived from natural plant-derived materials offer a promising alternative to petroleum-based products, which are essential for achieving global sustainability (Gupta et al. 2022). Sugarcane provides 40% of the world’s biofuel (Brant et al. 2025). Recently, metabolic engineering of sugarcane for hyperaccumulation of biomass oil is emerging as a strategy to elevate the crop’s energy content (Cao et al. 2023; Maitra et al. 2024). Synthetic transcription factors offer a powerful tool for modulating entire metabolic pathways by enabling fine-tuned activation or repression of specific genes (Liu and Stewart 2016; Hooghvorst and Altpeter 2023). However, the application of these technologies faces challenges, including the availability of well-characterized genetic building blocks for precise manipulation of gene expression, and the inherently challenging properties of gene expression in highly polyploid crops like sugarcane (Liu and Stewart 2016).
The LAMP Low Energy Beam Transport (LEBT) transfers a continuous beam at 100 keV from the ion source to RFQ in LEBT. A LEBT buncher imposes an energy tilt to initiate velocy bunching in the chopped beam pulse about 25 ns long to form a short MPEG bunch. One possible option for the LEBT buncher based on a two-gap LC-circuit driven structure was considered, which is similar to the existing LANSCE low-frequncy buncher (LFB). Another possible option is a non-resonant element driven by a pulse-forming-network that provides a single pulse at the repetition frequency of MPEG beam, with a period of 1.8 µs. The LEBT operation is synchronized with the accelerator timing system.
Here we propose a new type of neural networks, Kronecker neural networks (KNNs), that form a general framework for neural networks with adaptive activation functions. KNNs employ the Kronecker product, which provides an efficient way of constructing a very wide network while keeping the number of parameters low. Our theoretical analysis reveals that under suitable conditions, KNNs induce a faster decay of the loss than that by the feed-forward networks. This is also empirically verified through a set of computational examples. Furthermore, under certain technical assumptions, we establish global convergence of gradient descent for KNNs. As a specific case, we propose the Rowdy activation function that is designed to get rid of any saturation region by injecting sinusoidal fluctuations, which include trainable parameters. The proposed Rowdy activation function can be employed in any neural network architecture like feed-forward neural networks, Recurrent neural networks, Convolutional neural networks etc. The effectiveness of KNNs with Rowdy activation is demonstrated through various computational experiments including function approximation using feed-forward neural networks, solution inference of partial differential equations using the physics-informed neural networks, and standard deep learning benchmark problems using convolutional and fully-connected neural networks.
The over-arching project objective is to fully develop and validate optimal controls frameworks that can subsequently be applied widely to different WEC devices and concepts. Optimal controls of WEC devices represent a fundamental building block for WEC designers that must be considered as an integral part of every stage of device development. Using a building-blocks approach to optimal controls development, this effort will result in the full development of a feed-forward and feed-back control approach and a wave prediction system. Phase I focused primarily on numerical offline optimization and validation using wave tank testing of three industry partners’ WEC devices, including; CalWave, Ocean Energy, and Resolute Marine Energy. These industry partnerships allowed us to identify optimal control strategies for these different WEC topologies at different maturity levels. Phase II focused on demonstrating an integrated control system on an at-sea prototype that is to be custom-built and maturing the HW and SW required to successfully run our advanced controls code frameworks on at-sea systems. A secondary focus during phase II is to adapt our systems identification, controls and wave-prediction frameworks to become more robust and comprehensive in respect to RT capability, robustness, and reliability.
The rapid integration of intermittent renewable sources into the electricity grid is driving the need for low-carbon, fossil-fuel power plant capable of rapid ramping. Therefore, a cryogenic air separation unit (ASU) as part of low-carbon, fossil-fuel power plant should be capable of rapid ramping. However, highly integrated and nonlinear processes of the ASU would significantly restrict this rapid ramping. To overcome this fundamental issue, we study the basic dynamic process of a state-of-the-art double-column ASU. A flow-driven dynamic model was established in Aspen Plus Dynamics, assuming perfect flow controls, to capture the basic dynamics of ramping ASU. We find the vapor-liquid counter-current flow structure in the low-pressure column is critical to the air separation when rapidly ramping the ASU. This flow structure is established based on a complicated heat integration process. It is simplified as an apparent counter-current heat transfer process in this work, which greatly reduces the complexity for studying the dynamics of ASU. For keeping this basic flow structure, the heat integration is maintained as its heat duty follows the ASU load, through several basic feed-forward and feed-back controllers on the critical stream flowrates. Moreover, we find a fundamental issue of mismatched dynamics in the low-pressure column, resulting in a significant loss of O2 product purity when rapidly ramping down the ASU and an obvious fluctuation of O2 product purity. Based on these explorations, we propose a basic control method for rapidly ramping the ASU without loss of O2 product purity. Results show the ASU basic dynamic process successfully ramps at a rate up to 10%/min (40-100% load) while maintaining the O2 product purity within 95.2-95.6 mol%.
A system and method are provided for a feed-forward control of an inverter to reduce, and potentially minimize, a DC link capacitor of a wireless power transfer system. The feed-forward control may be utilized to reduce the capacitance of the DC link capacitor in a single-phase series-series compensated WPT system.
Abstract In this paper, we propose a control Lyapunov‐barrier function‐based model predictive control method utilizing a feed‐forward neural network specified control barrier function (CBF) and a recurrent neural network (RNN) predictive model to stabilize nonlinear processes with input constraints, and to guarantee that safety requirements are met for all times. The nonlinear system is first modeled using RNN techniques, and a CBF is characterized by constructing a feed‐forward neural network (FNN) model with unique structures and properties. The FNN model for the CBF is trained based on data samples collected from safe and unsafe operating regions, and the resulting FNN model is verified to demonstrate that the safety properties of the CBF are satisfied. Given sufficiently small bounded modeling errors for both the FNN and the RNN models, the proposed control system is able to guarantee closed‐loop stability while preventing the closed‐loop states from entering unsafe regions in state‐space under sample‐and‐hold control action implementation. We provide the theoretical analysis for bounded unsafe sets in state‐space, and demonstrate the effectiveness of the proposed control strategy using a nonlinear chemical process example with a bounded unsafe region.
Distributed multisensor networks record multiple data streams that can be used as inputs to machine learning models designed to classify operations relevant to proliferation at nuclear reactors. The goal of this work is to demonstrate methods to assess the importance of each node (a single multisensor) and region (a group of proximate multisensors) to machine learning model performance in a reactor monitoring scenario. This, in turn, provides insight into model behavior, a critical requirement of data-driven applications in nuclear security. Using data collected at the High Flux Isotope Reactor at Oak Ridge National Laboratory via a network of Merlyn multisensors, two different models were trained to classify the reactor’s operational state: a hidden Markov model (HMM), which is simpler and more transparent, and a feed-forward neural network, which is less inherently interpretable. Traditional wrapper methods for feature importance were extended to identify nodes and regions in the multisensor network with strong positive and negative impacts on the classification problem. These spatial-importance algorithms were evaluated on the two different classifiers. The classification accuracy was then improved relative to baseline models via feature selection from 0.583 to 0.839 and from 0.811 ± 0.005 to 0.884 ± 0.004 for the HMM and feed-forward neural network, respectively. While some differences in node and region importance were observed when using different classifiers and wrapper methods, the nodes near the facility’s cooling tower were consistently identified as important—a conclusion further supported by studies on feature importance in decision trees. Node and region importance methods are model-agnostic, inform feature selection for improved model performance, and can provide insight into opaque classification models in the nuclear security domain.
We developed a new class of physics-informed generative adversarial networks (PI-GANs) to solve forward, inverse, and mixed stochastic problems in a unified manner based on a limited number of scattered measurements. Unlike standard GANs relying solely on data for training, here we encoded into the architecture of GANs the governing physical laws in the form of stochastic differential equations (SDEs) using automatic differentiation. In particular, we applied Wasserstein GANs with gradient penalty (WGAN-GP) for its enhanced stability compared to vanilla GANs. We first tested WGAN-GP in approximating Gaussian processes of different correlation lengths based on data realizations collected from simultaneous reads at sparsely placed sensors. We obtained good approximation of the generated stochastic processes to the target ones even if there is a mismatch between the input noise dimensionality and the effective dimensionality of the target stochastic processes. We also studied the overfitting issue for both the discriminator and the generator, and we found that overfitting occurs also in the generator in addition to the discriminator as previously reported. Subsequently, we considered the solution of elliptic SDEs requiring approximations of three stochastic processes, namely the solution, the forcing, and the diffusion coefficient. Here again, we assumed data collected from simultaneous reads at a limited number of sensors for the multiple stochastic processes. Furthermore, three generators were used for the PI-GANs: two of them were feed forward deep neural networks (DNNs), while the other one was the neural network induced by the SDE. For the case where we have one group of data, we employed one feed forward DNN as the discriminator, while for the case of multiple groups of data we employed multiple discriminators in PI-GANs. We solved forward, inverse, and mixed problems without changing the framework of PI-GANs, obtaining both the means and the standard deviations of the stochastic solution and the diffusion coefficient in good agreement with benchmarks. In this work, we have demonstrated the effectiveness of PI-GANs in solving SDEs for about 120 dimensions. In principle, PI-GANs could tackle very high dimensional problems given more sensor data with low-polynomial growth in computational cost.
We introduce a new style of multileaf collimator which employs rotors with angularly dependent radius to control the masking aperture: a rotor-based multileaf collimator (RMLC). Using a padlock-inspired mechanism, a single motor can set dozens of rotors, i.e. leaves, independently. This is especially important for an ultra-high vacuum (UHV) compatible MLC, since this reduces the number of actuators and vacuum feedthroughs required by more than an order of magnitude. This new RMLC will complement previous work employing a UHV compatible MLC with an emittance exchange beamline to create arbitrarily shaped beams on demand. A feed-forward control system which abstracts away the complexity of the RMLC operation, and is adaptable to real beamline conditions, is discussed and demonstrated in simulation.