Search NASA⌕ Search

SEARCH · Search NASA

Results for “Feedforward optimization”

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.

At least 55 records · Page 3

Optimal microgravity vibration isolation - An algebraic introduction

Certain experiments contemplated for space platforms must be isolated from the accelerations of the platforms. An optimal active control is developed for microgravity vibration isolation, using constant state feedback gains (identical to those obtained from the Linear Quadratic Regulator (LQR) approach) along with constant feedforward (preview) gains. The quadratic cost function for this control algorithm effectively weights external accelerations of the platform disturbances by a factor proportional to (1/omega) (exp 4). Low frequency accelerations (less than 50 Hz) are attenuated by greater than two orders of magnitude. The control relies on the absolute position and velocity feedback of the experiment and the absolute position and velocity feedforward of the platform, and generally derives the stability robustness characteristics quaranteed by the LQR approach to optimality. The method as derived is extendable to the case in which only the relative positions and velocities and the absolute accelerations of the experiment and space platform are available.

Hampton, R. D.↗

Machine Learning-Driven Conservative-to-Primitive Conversion in Hybrid Piecewise Polytropic and Tabulated Equations of State

We present a novel machine learning (ML)-based method to accelerate conservative-to-primitive inversion, focusing on hybrid piecewise polytropic and tabulated equations of state. Traditional root-finding techniques are computationally expensive, particularly for large-scale relativistic hydrodynamics simulations. To address this, we employ feedforward neural networks (NNC2PS and NNC2PL), trained in PyTorch (2.0+) and optimized for GPU inference using NVIDIA TensorRT (8.4.1), achieving significant speedups with minimal accuracy loss. The NNC2PS model achieves 𝐿 1 and 𝐿 ∞ errors of 4.54 × 10 −7 and 3.44 × 10−6, respectively, while the NNC2PL model exhibits even lower error values. TensorRT optimization with mixed-precision deployment substantially accelerates performance compared to traditional root-finding methods. Specifically, the mixed-precision TensorRT engine for NNC2PS achieves inference speeds approximately 400 times faster than a traditional single-threaded CPU implementation for a dataset size of 1,000,000 points. Ideal parallelization across an entire compute node in the Delta supercomputer (dual AMD 64-core 2.45 GHz Milan processors and 8 NVIDIA A100 GPUs with 40 GB HBM2 RAM and NVLink) predicts a 25-fold speedup for TensorRT over an optimally parallelized numerical method when processing 8 million data points. Moreover, the ML method exhibits sub-linear scaling with increasing dataset sizes. We release the scientific software developed, enabling further validation and extension of our findings. By exploiting the underlying symmetries within the equation of state, these findings highlight the potential of ML, combined with GPU optimization and model quantization, to accelerate conservative-to-primitive inversion in relativistic hydrodynamics simulations.

conservative-to-primitive conversion↗

VLSI neuroprocessors

Electronic and optoelectronic hardware implementations of highly parallel computing architectures address several ill-defined and/or computation-intensive problems not easily solved by conventional computing techniques. The concurrent processing architectures developed are derived from a variety of advanced computing paradigms including neural network models, fuzzy logic, and cellular automata. Hardware implementation technologies range from state-of-the-art digital/analog custom-VLSI to advanced optoelectronic devices such as computer-generated holograms and e-beam fabricated Dammann gratings. JPL's concurrent processing devices group has developed a broad technology base in hardware implementable parallel algorithms, low-power and high-speed VLSI designs and building block VLSI chips, leading to application-specific high-performance embeddable processors. Application areas include high throughput map-data classification using feedforward neural networks, terrain based tactical movement planner using cellular automata, resource optimization (weapon-target assignment) using a multidimensional feedback network with lateral inhibition, and classification of rocks using an inner-product scheme on thematic mapper data. In addition to addressing specific functional needs of DOD and NASA, the JPL-developed concurrent processing device technology is also being customized for a variety of commercial applications (in collaboration with industrial partners), and is being transferred to U.S. industries. This viewgraph p resentation focuses on two application-specific processors which solve the computation intensive tasks of resource allocation (weapon-target assignment) and terrain based tactical movement planning using two extremely different topologies. Resource allocation is implemented as an asynchronous analog competitive assignment architecture inspired by the Hopfield network. Hardware realization leads to a two to four order of magnitude speed-up over conventional techniques and enables multiple assignments, (many to many), not achievable with standard statistical approaches. Tactical movement planning (finding the best path from A to B) is accomplished with a digital two-dimensional concurrent processor array. By exploiting the natural parallel decomposition of the problem in silicon, a four order of magnitude speed-up over optimized software approaches has been demonstrated.

Kemeny, Sabrina E.↗

Investigation, development, and application of optimal output feedback theory. Volume 3: The relationship between dynamic compensators and observers and Kalman filters

Relationships between observers, Kalman Filters and dynamic compensators using feedforward control theory are investigated. In particular, the relationship, if any, between the dynamic compensator state and linear functions of a discrete plane state are investigated. It is shown that, in steady state, a dynamic compensator driven by the plant output can be expressed as the sum of two terms. The first term is a linear combination of the plant state. The second term depends on plant and measurement noise, and the plant control. Thus, the state of the dynamic compensator can be expressed as an estimator of the first term with additive error given by the second term. Conditions under which a dynamic compensator is a Kalman filter are presented, and reduced-order optimal estimaters are investigated.

Broussard, John R.↗

Preliminary Work for Examining the Scalability of Reinforcement Learning

Researchers began studying automated agents that learn to perform multiple-step tasks early in the history of artificial intelligence (Samuel, 1963; Samuel, 1967; Waterman, 1970; Fikes, Hart & Nilsonn, 1972). Multiple-step tasks are tasks that can only be solved via a sequence of decisions, such as control problems, robotics problems, classic problem-solving, and game-playing. The objective of agents attempting to learn such tasks is to use the resources they have available in order to become more proficient at the tasks. In particular, each agent attempts to develop a good policy, a mapping from states to actions, that allows it to select actions that optimize a measure of its performance on the task; for example, reducing the number of steps necessary to complete the task successfully. Our study focuses on reinforcement learning, a set of learning techniques where the learner performs trial-and-error experiments in the task and adapts its policy based on the outcome of those experiments. Much of the work in reinforcement learning has focused on a particular, simple representation, where every problem state is represented explicitly in a table, and associated with each state are the actions that can be chosen in that state. A major advantage of this table lookup representation is that one can prove that certain reinforcement learning techniques will develop an optimal policy for the current task. The drawback is that the representation limits the application of reinforcement learning to multiple-step tasks with relatively small state-spaces. There has been a little theoretical work that proves that convergence to optimal solutions can be obtained when using generalization structures, but the structures are quite simple. The theory says little about complex structures, such as multi-layer, feedforward artificial neural networks (Rumelhart & McClelland, 1986), but empirical results indicate that the use of reinforcement learning with such structures is promising. These empirical results make no theoretical claims, nor compare the policies produced to optimal policies. A goal of our work is to be able to make the comparison between an optimal policy and one stored in an artificial neural network. A difficulty of performing such a study is finding a multiple-step task that is small enough that one can find an optimal policy using table lookup, yet large enough that, for practical purposes, an artificial neural network is really required. We have identified a limited form of the game OTHELLO as satisfying these requirements. The work we report here is in the very preliminary stages of research, but this paper provides background for the problem being studied and a description of our initial approach to examining the problem. In the remainder of this paper, we first describe reinforcement learning in more detail. Next, we present the game OTHELLO. Finally we argue that a restricted form of the game meets the requirements of our study, and describe our preliminary approach to finding an optimal solution to the problem.

Clouse, Jeff↗

Combined magnetic and kinetic control of advanced tokamak steady state scenarios based on semi-empirical modelling

Our paper shows that semi-empirical data-driven models based on a two-time-scale approximation for the magnetic and kinetic control of advanced tokamak (AT) scenarios can be advantageously identified from simulated rather than real data, and used for control design. The method is applied to the combined control of the safety factor profile, q(x), and normalized pressure parameter, beta(N), using DIII-D parameters and actuators (on-axis co-current neutral beam injection (NBI) power, off-axis co-current NBI power, electron cyclotron current drive power, and ohmic coil). The approximate plasma response model was identified from simulated open-loop data obtained using a rapidly converging plasma transport code, METIS, which includes an MHD equilibrium and current diffusion solver, and combines plasma transport nonlinearity with 0D scaling laws and 1.5D ordinary differential equations. The paper discusses the results of closed-loop METIS simulations, using the near-optimal ARTAEMIS control algorithm (MoreauDet al 2013 Nucl. Fusion 53 063020) for steady stateAT operation. With feedforward plus feedback control, the steady state target q-profile and beta(N) are satisfactorily tracked with a time scale of about 10 s, despite large disturbances applied to the feedforward powers and plasma parameters. We also show the robustness of the control algorithm with respect to disturbances of the H&CD actuators and of plasma parameters such as the H-factor, plasma density and effective charge.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Multiscale topology optimization using neural network surrogate models

In this study, we are concerned with optimization of macroscale elastic structures that are designed utilizing spatially varying microscale metamaterials. The macroscale optimization is accomplished using gradient-based nonlinear topological optimization. But instead of using density as the optimization decision variable, the decision variables are the multiple parameters that define the local microscale metamaterial. This is accomplished using single layer feedforward Gaussian basis function networks as a surrogate models of the elastic response of the microscale metamaterial. The surrogate models are trained using highly resolved continuum finite element simulations of the microscale metamaterials and hence are significantly more accurate than analytical models e.g. classical beam theory. Because the derivative of the surrogate model is important for sensitivity analysis of the macroscale topology optimization, a neural network training procedure based on the Sobolev norm is described. Since the SIMP method is not appropriate for spatially varying lattices, an alternative method is developed to enable creation of void regions. Lastly, the efficacy of this approach is demonstrated via several examples in which the optimal graded metamaterial outperforms a traditional solid structure.

42 ENGINEERING↗

Optimal command generation for tracking a class of discontinuous trajectories

Commands are found to drive a linear system to optimally track a class of prescribed trajectories, each of which contains a point of discontinuity. The paper focuses on the guidance problem of command generation, to be implemented in a feedforward fashion; the necessary additional control or feedback regulation structure is not studied in detail, but assumed to be provided in applications as a conventional error-feedback regulator. Solutions are found for the infinite-time problem which are optimal with respect to a quadratic performance criterion; suboptimal controls which satisfy a continuity condition are also found. The controls have applications to certain problems in aircraft guidance where command trajectories are piecewise continuous. Several examples are worked out in detail, with comparisons with conventional, nonfeedforward solutions to the problem, and a brief discussion of a simpler, suboptimal solution.

Weissenberger, S.↗

Wave forecast and its application to the optimal control of offshore floating wind turbine for load mitigation

Control algorithms play an important role in energy capture and load mitigation for offshore floating wind turbines (OFWTs). One of the advanced and effective control techniques is the feedforward or model predictive control approach, which requires the forecast of incoming environment conditions. For OFWTs, wave loading is one of the dominant sources to excite structural responses. This study is thus motivated to develop forecasting algorithms for wave elevations and wave excitation forces with the purpose of applying feedforward controllers on OFWTs. Two forecasting algorithms, the approximate Prony Method based on ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques) and SVM (Support Vector Machine) regression, are developed and validated using wave records from tank tests. Utilizing the forecasted wave elevations and wave excitation forces, a feedforward LQR controller is designed to mitigate structural loads of an OFWT system.

17 WIND ENERGY↗

Integrated Control Using the SOFFT Control Structure

The need for integrated/constrained control systems has become clearer as advanced aircraft introduced new coupled subsystems such as new propulsion subsystems with thrust vectoring and new aerodynamic designs. In this study, we develop an integrated control design methodology which accomodates constraints among subsystem variables while using the Stochastic Optimal Feedforward/Feedback Control Technique (SOFFT) thus maintaining all the advantages of the SOFFT approach. The Integrated SOFFT Control methodology uses a centralized feedforward control and a constrained feedback control law. The control thus takes advantage of the known coupling among the subsystems while maintaining the identity of subsystems for validation purposes and the simplicity of the feedback law to understand the system response in complicated nonlinear scenarios. The Variable-Gain Output Feedback Control methodology (including constant gain output feedback) is extended to accommodate equality constraints. A gain computation algorithm is developed. The designer can set the cross-gains between two variables or subsystems to zero or another value and optimize the remaining gains subject to the constraint. An integrated control law is designed for a modified F-15 SMTD aircraft model with coupled airframe and propulsion subsystems using the Integrated SOFFT Control methodology to produce a set of desired flying qualities.

Halyo, Nesim↗

Neural Generalized Predictive Control: A Newton-Raphson Implementation

An efficient implementation of Generalized Predictive Control using a multi-layer feedforward neural network as the plant's nonlinear model is presented. In using Newton-Raphson as the optimization algorithm, the number of iterations needed for convergence is significantly reduced from other techniques. The main cost of the Newton-Raphson algorithm is in the calculation of the Hessian, but even with this overhead the low iteration numbers make Newton-Raphson faster than other techniques and a viable algorithm for real-time control. This paper presents a detailed derivation of the Neural Generalized Predictive Control algorithm with Newton-Raphson as the minimization algorithm. Simulation results show convergence to a good solution within two iterations and timing data show that real-time control is possible. Comments about the algorithm's implementation are also included.

Soloway, Donald↗

Toward a More Robust Pruning Procedure for MLP Networks

Choosing a proper neural network architecture is a problem of great practical importance. Smaller models mean not only simpler designs but also lower variance for parameter estimation and network prediction. The widespread utilization of neural networks in modeling highlights an issue in human factors. The procedure of building neural models should find an appropriate level of model complexity in a more or less automatic fashion to make it less prone to human subjectivity. In this paper we present a Singular Value Decomposition based node elimination technique and enhanced implementation of the Optimal Brain Surgeon algorithm. Combining both methods creates a powerful pruning engine that can be used for tuning feedforward connectionist models. The performance of the proposed method is demonstrated by adjusting the structure of a multi-input multi-output model used to calibrate a six-component wind tunnel strain gage.

Stepniewski, Slawomir W.↗

Measurement-induced deterministic and probabilistic entanglement with strong and weak interactions

A scheme is proposed to transform spatial coherence of a single particle into entanglement. Two quantum systems can be entangled by having them interact in parallel with an ancillary particle in a Mach–Zehnder interferometer, then making a suitable post-selection of the particle followed by a conditional feedforward on one of the systems to be entangled. For a strong interaction between each system and the ancilla, the process works deterministically. For a weaker interaction only the probability of success is reduced, but the output continues to be a maximally entangled state. It is demonstrated that the process is optimal when the two interactions are symmetric, systems with continuous variables are considered, and the effects of the environment are taken into account. - Highlights: • Presents a scheme for deterministic entanglement based on feedforward, if a strong interaction is available. • Presents a scheme for probabilistic entanglement, if only a weak interaction is available. • Specifies the importance of an equal interaction strength in both arms. • Includes the effects of the environment and of the internal degrees of freedom of the ancilla.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Machine Learning-Based Model Predictive Control of Two-Time-Scale Systems

In this study, we present a general form of nonlinear two-time-scale systems, where singular perturbation analysis is used to separate the dynamics of the slow and fast subsystems. Machine learning techniques are utilized to approximate the dynamics of both subsystems. Specifically, a recurrent neural network (RNN) and a feedforward neural network (FNN) are used to predict the slow and fast state vectors, respectively. Moreover, we investigate the generalization error bounds for these machine learning models approximating the dynamics of two-time-scale systems. Next, under the assumption that the fast states are asymptotically stable, our focus shifts toward designing a Lyapunov-based model predictive control (LMPC) scheme that exclusively employs the RNN to predict the dynamics of the slow states. Additionally, we derive sufficient conditions to guarantee the closed-loop stability of the system under the sample-and-hold implementation of the controller. A nonlinear chemical process example is used to demonstrate the theory. In particular, two RNN models are constructed: one to model the full two-time-scale system and the other to predict solely the slow state vector. Both models are integrated within the LMPC scheme, and we compare their closed-loop performance while assessing the computational time required to execute the LMPC optimization problem.

97 MATHEMATICS AND COMPUTING↗

Probing Quantum Telecloning on Superconducting Quantum Processors

Quantum information cannot be perfectly cloned, but approximate copies of quantum information can be generated. Quantum telecloning combines approximate quantum cloning, more typically referred to as quantum cloning, and quantum teleportation. Quantum telecloning allows approximate copies of quantum information to be constructed by separate parties, using the classical results of a Bell measurement made on a prepared quantum telecloning state. Quantum telecloning can be implemented as a circuit on quantum computers using a classical coprocessor to compute classical feedforward instructions using if statements based on the results of a midcircuit Bell measurement in real time. We present universal symmetric optimal 1 → M telecloning circuits and experimentally demonstrate these quantum telecloning circuits for M = 2 up to M = 10, natively executed with real-time classical control systems on IBM Quantum superconducting processors, known as dynamic circuits. We perform the cloning procedure on many different message states across the Bloch sphere, on seven IBM Quantum processors, optionally using the error suppression technique X–X sequence digital dynamical decoupling. Two circuit optimizations are utilized: one that removes ancilla qubits for M = 2,3 , and one that reduces the total number of gates in the circuit but still uses ancilla qubits. Parallel single-qubit tomography with maximum likelihood estimation density matrix reconstruction is used in order to compute the mixed-state density matrices of the clone qubits, and clone quality is measured using quantum fidelity. These results present one of the largest and most comprehensive noisy intermediate-scale quantum computer experimental analyses on (single qubit) quantum telecloning to date. The clone fidelity sharply decreases to 0.5 for M > 5 , but for M = 2 , we are able to achieve a mean clone fidelity of up to 0.79 using dynamical decoupling.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Advanced Energy Harvesting Control Schemes for Marine Renewable Energy Devices

The Advanced Energy Harvesting Control Schemes for Marine Renewable Energy Devices (Project) investigated, analyzed and modeled advanced turbine control schemes with the objective of increasing the energy harvested by hydrokinetic turbines in turbulent flow. Ocean Renewable Power Company (ORPC) implemented and validated a feedforward controller to increase power capture; and applied and tested the controls on ORPC’s RivGen® Power Systems in Igiugig, Alaska. Assessments of performance improvements were made for the RivGen® in the Igiugig environment and for ORPC’s TidGen® Power System in a reference tidal environment. Annualized Energy Production (AEP) and Levelized Cost of Energy (LCOE) improvements associated with implementation of the recommended control methodology were made for the TidGen® Power System in the DOE reference tidal environment. System Performance Advancement (SPA) goals were selected for the project. SPA targets were to improve Power to Weight Ratio (PWR) and system Availability, with the intention of reducing Levelized Cost of Electricity (LCOE). This project focused primarily reducing in PWR. Reductions in PWR of 25.5% were achieved. Reductions of 20.3% in LCOE were achieved. This project evaluated four types of controllers which were tested in simulation, emulation, a laboratory flume, and the field. The adaptive Kω2 controller performs similarly to the non-adaptive version of the same controller and may be useful in tidal channels where the mean velocity is continually evolving. Trends in simulation were largely verified through experiments, which also provided the opportunity to test assumptions about turbine responsiveness and control resilience to varying scales of turbulence. Laboratory experiments provided an essential stepping stone between simulation and implementation on a field-scale turbine. Experiments also demonstrated that using “energy loss” as a metric to differentiate between well-designed controllers operating at an optimal tip-speed ratio set-point is difficult, which anticipated the outcome from field experiments. The clear message is that the feedforward Kω2 controller out-performs the feedback controllers in almost all aspects and modes of evaluation. The controllers proved a substantial improvement over the baseline performance of the TidGen® turbine, in terms of energy capture. The effects of noise-contaminated angular velocity signals were investigated and validated by simulation as an explanation for the performance limitations observed for TidGen® turbine operations in Eastport, Maine. Measurements of loads performed as part of the laboratory testing indicate that there are limited differences in average axial thrust force between control architectures. This suggests that none of the control strategies are likely to substantially affect loads on the turbine support structure. Velocity measurements during the ORPC RivGen® turbine deployment at Igiugig, Alaska, in 2014 were used to assess the variability of the river flow. Results suggest that the river flow is approximately steady, in the mean sense, at any particular location in the river, with random turbulent fluctuations that are around 10% of the mean flow. The mean flow in the center channel of the river is 2.5 m/s, with reductions near the riverbanks and in the shallows. Spectral analysis and lagged correlation results indicate that temporal fluctuations at a given point are dominated by large scale fluctuations, such that measurements at the turbine location are just as useful for inflow control implementation as upstream measurements. At this site, and likely at many other river sites, flow is generally steady at a given location, but flow varies dramatically between locations, particularly laterally across the river. The primary result is that a lateral change in position of a few meters results in changes to flow speed that far exceed the turbulence fluctuations at any given location. The turbulence is dominated by long time scales. Following final system tests, the RivGen® device was submerged and each evaluated controller was tested across a range of gain/set point values and filter configurations for a minimum of 5 minutes, with longer runs attempted for well-performing cases. In addition to testing controllers during the 2015 deployment season, LGL Alaska Research Associates Inc. (LGL) performed a fish monitoring study in compliance with Alaska Department of Fish and Game fisheries habitat permit for the project. During the season, LGL reviewed 10% of 179 one-hour blocks of footage (6 minutes on the hour) and documented a total of over 1200 fish in the vicinity of the RivGen® device, including over 800 salmon smolt and over 350 adult salmon. No evidence of adverse effects including passage delay by upstream migrating salmon was noted. This is an important result for future deployments and has a direct impact on commercial system designs.

16 TIDAL AND WAVE POWER↗

Accelerated Training for Large Feedforward Neural Networks

In this paper we introduce a new training algorithm, the scaled variable metric (SVM) method. Our approach attempts to increase the convergence rate of the modified variable metric method. It is also combined with the RBackprop algorithm, which computes the product of the matrix of second derivatives (Hessian) with an arbitrary vector. The RBackprop method allows us to avoid computationally expensive, direct line searches. In addition, it can be utilized in the new, 'predictive' updating technique of the inverse Hessian approximation. We have used directional slope testing to adjust the step size and found that this strategy works exceptionally well in conjunction with the Rbackprop algorithm. Some supplementary, but nevertheless important enhancements to the basic training scheme such as improved setting of a scaling factor for the variable metric update and computationally more efficient procedure for updating the inverse Hessian approximation are presented as well. We summarize by comparing the SVM method with four first- and second- order optimization algorithms including a very effective implementation of the Levenberg-Marquardt method. Our tests indicate promising computational speed gains of the new training technique, particularly for large feedforward networks, i.e., for problems where the training process may be the most laborious.

Stepniewski, Slawomir W.↗