Search NASA⌕ Search

SEARCH · Search NASA

Results for “projection optimal parameter values”

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
All sourcesNASA NTRSDOE OSTI
NASA NTRS1996

Cascade Error Projection: A Learning Algorithm for Hardware Implementation

In this paper, we workout a detailed mathematical analysis for a new learning algorithm termed Cascade Error Projection (CEP) and a general learning frame work. This frame work can be used to obtain the cascade correlation learning algorithm by choosing a particular set of parameters. Furthermore, CEP learning algorithm is operated only on one layer, whereas the other set of weights can be calculated deterministically. In association with the dynamical stepsize change concept to convert the weight update from infinite space into a finite space, the relation between the current stepsize and the previous energy level is also given and the estimation procedure for optimal stepsize is used for validation of our proposed technique. The weight values of zero are used for starting the learning for every layer, and a single hidden unit is applied instead of using a pool of candidate hidden units similar to cascade correlation scheme. Therefore, simplicity in hardware implementation is also obtained. Furthermore, this analysis allows us to select from other methods (such as the conjugate gradient descent or the Newton's second order) one of which will be a good candidate for the learning technique. The choice of learning technique depends on the constraints of the problem (e.g., speed, performance, and hardware implementation); one technique may be more suitable than others. Moreover, for a discrete weight space, the theoretical analysis presents the capability of learning with limited weight quantization. Finally, 5- to 8-bit parity and chaotic time series prediction problems are investigated; the simulation results demonstrate that 4-bit or more weight quantization is sufficient for learning neural network using CEP. In addition, it is demonstrated that this technique is able to compensate for less bit weight resolution by incorporating additional hidden units. However, generation result may suffer somewhat with lower bit weight quantization.

Duong, Tuan A.↗
NASA NTRS1983

Space Shuttle propulsion parameter estimation using optimal estimation techniques, volume 1

The mathematical developments and their computer program implementation for the Space Shuttle propulsion parameter estimation project are summarized. The estimation approach chosen is the extended Kalman filtering with a modified Bryson-Frazier smoother. Its use here is motivated by the objective of obtaining better estimates than those available from filtering and to eliminate the lag associated with filtering. The estimation technique uses as the dynamical process the six degree equations-of-motion resulting in twelve state vector elements. In addition to these are mass and solid propellant burn depth as the ""system'' state elements. The ""parameter'' state elements can include aerodynamic coefficient, inertia, center-of-gravity, atmospheric wind, etc. deviations from referenced values. Propulsion parameter state elements have been included not as options just discussed but as the main parameter states to be estimated. The mathematical developments were completed for all these parameters. Since the systems dynamics and measurement processes are non-linear functions of the states, the mathematical developments are taken up almost entirely by the linearization of these equations as required by the estimation algorithms.

Source record↗
DOE OSTI2026

Numerical Modeling of a Two-Stage Ocean Current Turbine

The Equinox Ocean Turbines (EQOT) current energy converter has a unique design with power generation in two small-diameter turbines attached to the tips of a large-diameter passive rotor. This configuration offers some key advantages for capturing ocean currents. With no centrally placed generator, almost no reaction torque is required at the nacelle of the main large-diameter rotor, and the small-diameter tip turbine generators operate at a higher speed and lower torque. The physics that determine the performance and loads on the turbine are also unique. The interactions of the flow field between the two stages and the general architecture of the system cannot be captured with traditional mid-fidelity modeling tools. For design iterations and large sets of load cases, it is important to have mid-fidelity models that can capture the important phenomenon with enough accuracy to identify global trends. This work uses a limited set of high-fidelity computational fluid dynamics (CFD) simulations to help inform the selection of and construction of a custom mid-fidelity model. Mid-fidelity modeling approaches were verified by comparing key turbine performance quantities to those found with the CFD model. Hydrodynamic interactions of the two-stage rotor were identified through high-fidelity CFD modeling. This highlighted the impact of the main rotor tip vortex and wake on the secondary rotor apparent inflow. This results in a relative flow rotation and sharp deficit, that change the optimal secondary rotor rotation speed and adds unsteadiness to the blade loading respectively. Multiple mid-fidelity approaches were evaluated for their ability to capture these effects. A simple approximation of the combined-stage performance based on single-stage BEM provides a reasonable rough prediction, especially near the peak TSR values, with some larger discrepancy at higher TSRs. Predicting the combined-stage performance based on single-stage CFD data improves this prediction across the TSR range. Although the combined-stage modeling in OLAF was not successful in this stage of the project, it showed promise as a mid-fidelity method, assuming the parameters can be tuned to account for the significant differences in time and length scales between the main and secondary rotors. This may be addressed through code changes in future work. A significant finding from the OLAF work was the agreement between the vortex core radius values found independently via a parameter space search and via CFD. The technique of using single-stage secondary rotor BEM, with a custom inflow taken from single-stage main rotor CFD or OLAF, provides an efficient method to capture one-way coupled flow interactions. This method provided generally good predictions of the impact of the flow rotation on the secondary rotor but struggled to accurately predict the peaks of the unsteady load progression. Future work could include some superposition of a tuned main rotor trailing edge viscous wake into the custom inflow to better predict this interaction.

16 TIDAL AND WAVE POWER↗
NASA NTRS2018

An Active Subspace Method for Accelerating Convergence in Delaunay-Based Optimization via Dimension Reduction

Delaunay-based derivative-free optimization, ∆DOGS, is an efficient and provably-convergent global optimization method for the problems which has computationally expensive objection function and the analytical expression for the objective function is not available. ∆-DOGS is a novel optimization scheme in the family of response surface methods (RSMs); however, it suffers from the curse of dimensionality since the computational cost increases dramatically as the number of design parameters increases. As a result, the number of design parameters in ∆-DOGS algorithm is relatively low (n.10). To avoid such problems, this paper proposes a combination of derivative-free optimization, seeking the global minimizer of an expensive and nonconvex objective function f(x) and active subspace method, detecting the directions of the most variability using evaluations of the gradient. The contribution of other directions to the objective function is bounded by a sufficiently small constant. This new algorithm iteratively applied Delaunay-based derivative-free optimization to seek the minimizer on the d-dimensional active subspace that has most function variation. Inverse mapping is needed to project data from active subspace to full-model for evaluating function values. This task is overcome by solving an inequality constrained problem that curves the response surface of the objective function. The test results show that this strategy is effective on a handful of optimization problems.

Bewley, Thomas R.↗
DOE OSTI2022

A Bayesian Approach for In-Situ Stress Prediction and Uncertainty Quantification for Subsurface Engineering

Many subsurface engineering applications require accurate knowledge of the in-situ state of stress for their safe design and operation. Existing methods to meet this need primarily include field measurements for estimating one or more of the principal stresses from a borehole, or optimization methods for constructing a 3D geomechanical model in terms of geophysical measurements. These methods, however, often contain considerable uncertainty in estimating the state of stress. Here, in this paper, we build on a Bayesian approach to quantify uncertainty in stress estimations for subsurface engineering applications. This approach can provide an estimate of the 3D distribution of stress throughout the volume of interest and provide an estimate of the uncertainty arising from the stress measurement, the rheology parameters, and a paucity of measurements. The value of this approach is demonstrated using stress measurements from the In Salah carbon storage site, which was one of the world’s first industrial carbon capture and storage projects. This demonstration shows the application of this Bayesian approach for estimating the initial state of stress for In Salah and quantifying the uncertainty in the estimated stress. Also, an assessment of a maximum injection pressure to prevent geomechanical risks from CO 2 injection pressures is provided in terms of the probability distribution of the minimum principal stress quantified by the approach. With the In Salah case study, this paper demonstrates that using the Bayesian approach can provide additional insights for site explorations and/or project operations to make informed-site decisions for subsurface engineering applications.

58 GEOSCIENCES↗
DOE OSTI2026

Evaluating the Limits of QAOA Parameter Transfer at High-Rounds on Sparse Ising Models With Geometrically Local Cubic Terms

The emergent practical applicability of the Quantum Approximate Optimization Algorithm (QAOA) for approximate combinatorial optimization is a subject of considerable interest. One of the primary limitations of QAOA is the task of finding a set of good parameters, which is usually done using a variational optimization loop. Parameter transfer, or parameter concentration, is a phenomenon where QAOA angles trained on problem instances that are self-similar tend to perform well for other problem instances from that similar class. This suggests a potentially highly efficient and scalable non-variational learning method for QAOA angle finding. In this work, we systematically study QAOA parameter transferability from small problem sizes (16 and 27 decision variables) onto large problem instances (up to 156 qubits) for heavy-hex graph Ising models with geometrically local higher order terms using the Julia based QAOA simulation tool \texttt{JuliQAOA} to perform classical angle finding for up to $49$ QAOA layers ($p$). Parameter transfer of the fixed angles is validated using a combination of full statevector, Projected Entangled Pair States (PEPS), Matrix Product State (MPS), and LOWESA numerical simulations. We find that the QAOA parameter transfer from single instances applied to other (unseen) problem instances does not in general provide monotonically improving performance as a function of $p$ - there are many cases where the performance temporarily decreases as a function of $p$ - but despite this the transferred angles have a general trend of improved expectation value as the QAOA depth increases, in many cases converging close to the true ground-state energy of the $100+$ qubit instances. We also sample the hardware-compatible Ising models using the ensemble of transfer-learned QAOA parameters on several superconducting qubit IBM Quantum processors with 127, 133, and 156 qubits. We find continuous solution quality improvement of the hardware-compatible QAOA circuits run on the IBM NISQ processors up to $p=5$ on \texttt{ibm\_fez}, up to $p=9$ on \texttt{ibm\_torino}, and up to $p=10$ on \texttt{ibm\_pittsburgh}.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
NASA NTRS2018

Determination of Ballistic Limit for IM7/8552 Using LS-DYNA MAT 261

The goal of the NASA ACC High Energy Dynamic Impact Project is to determine the state of the art of dynamic fracture simulations for high velocity impact for composite fuselage shielding applications. Using a building block approach, several computational models considered under NASA ACC are being validated against test data, starting at unconfigured panels and progressing to configured panels under combined out-of-plane and in-plane loading due to ballistic impact. The computational models being evaluated in this project include MAT 162, MAT 213, MAT 261, SPG, and Peridynamics. In this paper, the simulation results using LS-Dyna Material MAT 261 are presented. In particular, a series of blind predictions for unconfigured panels were performed to determine the ballistic limit or V50 velocity. MAT 261 employs failure approach that is generally physically-based using fracture toughness criteria. The overall material model relies on typical ply-level stiffness properties, similar to MAT 162 and other composite continuum damage material models. The fracture toughness values are based on standard tests, and thus are not subject to extensive calibration. This approach is more efficient than performing extensive optimization studies for calibration of parameters. Also, this approach of relying on physical properties reduces the uncertainty of results, as questions concerning the quality and extent of the calibration studies is no longer relevant. However, it was found that carefully controlled coupon-level tests are needed to accurately obtain the required fracture toughness values. Additionally, it should be noted that there is one significant parameter in MAT 261 that does appear to require calibration, and that is the overall failure strain. This is the strain at which the element is deleted, and is not the same as the strain at which damage begins to accumulate. This failure strain is termed EFS (Effective Failure Strain), and is the maximum effective stain for element failure. Simulations have shown that this value will significantly affect impact response and failure. The paper presents the effect of this element failure strain parameter, along with possible uncertainties in fracture toughness values. With an adjusted appropriate value for EFS, it is seen that simulation results compare well with impact test data for predicted penetration velocity.

Byar, Alan↗
DOE OSTI2024

An Efficient Annual-Performance Model of a Geothermal Network for Improved System Design, Operation, and Control: Preprint

Geothermal district energy systems (DES), and specifically geothermal networks, provide a viable solution for decarbonizing residential and commercial heating and cooling loads. District energy systems of all kinds enable a thermal resource with a relatively high capital cost (such as a geothermal borehole field) to be shared among a large number of users. While district heating and cooling has been studied for many years, geothermal networks, fifth generation DES that utilize water-source heat pumps and an ambient temperature loop to meet heating and cooling loads, have not been implemented extensively and thus require additional technical and economic optimization to obtain maximum benefits. This paper presents a newly developed reduced-order model that captures the flow of energy within the network, including the commercial and residential users' electrical usage, at an hourly rate over a year. The model includes building loads, heat pumps, borehole fields, and auxiliary heat/cool input, all connected with an ambient-temperature thermal loop model. In the model, operational control is possible for the borehole fields, circulation pump, and auxiliary system. For a given system, the model can output the complete state parameters for each component, the thermal loop, and the collective system, such as flow rate over time, average thermal loop temperature over time, and total electricity usage. The model can also be used to optimize the system control for maximizing system efficiency or minimizing system operational cost. For example, one initial assessment of the borehole controller for an example system showed that a controller with on/off operation of the borehole field reduces annual electrical usage by 33%, compared with continuous operation mode. Hence, the model can assist in optimizing a given system's operation to get the most value out of a geothermal network installation. Future work will consider the model's application to a demonstration project, including the model validation against operational data and system operation optimization.

ambient-temperature loop↗
NASA NTRS1994

Modeling tropical Pacific sea surface temperature with satellite-derived solar radiative forcing

Two independent datasets for the solar radiation at the surface derived from satellites are compared. The data derived from the Earth Radiation Budget Experiment (ERBE) is for the net solar radiation at the surface whereas the International Satellite Cloud Climatology Project (ISCCP) data is for the downward flux only and was corrected with a space- and time-varying albedo. The ISCCP net flux is at all times higher than the ERBE flux. The difference can be divided into an offset that decreases with latitude and another component that correlates with high tropical cloud cover. With this latter exception the two datasets provide spatial patterns of solar flux that are very similar. A tropical Pacific Ocean model is forced with these two datasets and observed climatological winds. The upward heat flux is parameterized taking into account separately the longwave radiative, latent, and sensible heat fluxes. Best fit values for the uncertain parameters are found using an optimization procedure that seeks to minimize the difference between model and observed SST by varying the parameters within a reasonable range of uncertainty. The SST field the model produces with the best fit parameters is the best the model can do. If the differences between the model and data are larger than can be accounted for by remaining uncertainties in the heat flux parameterization and forcing data then the ocean model must be held to be at fault. Using this method of analysis, a fundamental model fault is identified. Inadequate treatment of mixed layer/entrainment processes in upwelling regions of the eastern tropical Pacific leads to a large and seasonally varying error in the model SST. Elsewhere the model SST is insufficiently different from observed to be able to identify model errors.

Seager, Richard↗
DOE OSTI2023

An empirical method for geometric calibration of a photon counting detector-based cone beam CT system

BACKGROUND: Geometric calibration is essential in developing a reliable computed tomography (CT) system. It involves estimating the geometry under which the angular projections are acquired. Geometric calibration of cone beam CTs employing small area detectors, such as currently available photon counting detectors (PCDs), is challenging when using traditional-based methods due to detectors’ limited areas. OBJECTIVE: This study presented an empirical method for the geometric calibration of small area PCD-based cone beam CT systems. METHODS: Unlike the traditional methods, we developed an iterative optimization procedure to determine geometric parameters using the reconstructed images of small metal ball bearings (BBs) embedded in a custom-built phantom. An objective function incorporating the sphericities and symmetries of the embedded BBs was defined to assess performance of the reconstruction algorithm with the given initial estimated set of geometric parameters. The optimal parameter values were those which minimized the objective function. The TIGRE toolbox was employed for fast tomographic reconstruction. To evaluate the proposed method, computer simulations were carried out using various numbers of spheres placed in various locations. Furthermore, efficacy of the method was experimentally assessed using a custom-made benchtop PCD-based cone beam CT. RESULTS: Computer simulations validated the accuracy and reproducibility of the proposed method. The precise estimation of the geometric parameters of the benchtop revealed high-quality imaging in CT reconstruction of a breast phantom. Within the phantom, the cylindrical holes, fibers, and speck groups were imaged in high fidelity. The CNR analysis further revealed the quantitative improvements of the reconstruction performed with the estimated parameters using the proposed method. CONCLUSION: Apart from the computational cost, we concluded that the method was easy to implement and robust.

Instruments & Instrumentation↗
DOE OSTI2023

Incorporating Operational Uncertainties into the Dispatch of an Integrated Solar and Storage System

The economic assessment of hybrid energy systems (HES) pairing battery energy storage systems (BESSs) and photovoltaics (PV) is highly important for advancing their deployment in power systems. This paper presents an innovative assessment framework, including an optimal control policy for dispatch under uncertainty and procedures for exploring control parameters that maximize economic benefits. The proposed dispatch policy consists of two steps using system forecast information. The first step is to determine whether a BESS will be used within an operational scheduling time frame based on the probability of events and their thresholds. Once the dispatch of BESS is triggered, a model predictive control (MPC) is carried out in the second step for scheduling using the expected value of system information. By exercising this policy with different thresholds, one can explore the trade-offs between short-term benefits and battery lifetime, and identify an optimal threshold that maximizes the total economic benefits within the battery lifetime. An evaluation study in a real-world HES project is presented to illustrate the proposed framework. Compared with traditional optimal dispatch algorithms, the proposed method can significantly improve the economic benefits of an HES scheduled under forecast uncertainties.

Ma, Xu↗
NASA NTRS2004

Selected Flight Test Results for Online Learning Neural Network-Based Flight Control System

The NASA F-15 Intelligent Flight Control System project team has developed a series of flight control concepts designed to demonstrate the benefits of a neural network-based adaptive controller. The objective of the team is to develop and flight-test control systems that use neural network technology to optimize the performance of the aircraft under nominal conditions as well as stabilize the aircraft under failure conditions. Failure conditions include locked or failed control surfaces as well as unforeseen damage that might occur to the aircraft in flight. This report presents flight-test results for an adaptive controller using stability and control derivative values from an online learning neural network. A dynamic cell structure neural network is used in conjunction with a real-time parameter identification algorithm to estimate aerodynamic stability and control derivative increments to the baseline aerodynamic derivatives in flight. This set of open-loop flight tests was performed in preparation for a future phase of flights in which the learning neural network and parameter identification algorithm output would provide the flight controller with aerodynamic stability and control derivative updates in near real time. Two flight maneuvers are analyzed a pitch frequency sweep and an automated flight-test maneuver designed to optimally excite the parameter identification algorithm in all axes. Frequency responses generated from flight data are compared to those obtained from nonlinear simulation runs. An examination of flight data shows that addition of the flight-identified aerodynamic derivative increments into the simulation improved the pitch handling qualities of the aircraft.

Williams, Peggy S.↗
DOE OSTI2021

TEAL Output Visualization

Tool for Economic Analysis (TEAL) is an open-source economics calculation package written and maintained by Idaho National Laboratory (INL). As a plugin of the Risk Analysis Virtual ENvironment (RAVEN), TEAL provides economics analysis models to RAVEN workflow users. In addition, TEAL is used in other RAVEN plugins such as LOGOS and the Holistic Energy Resource Optimization Network (HERON) to complement other analyses with economic calculations. This report documents the efforts in developing the capability of TEAL output visualization for users to better understand and investigate TEAL simulation results. In the past, outputs (e.g., various cash flows) were generated and printed to screen. According to the need and purpose for visualization, users can select various bar charts and donut charts to visualize the cash flows including inflows, outflows, and net cash flows in each year or in a selected time range, in addition to observing data stored in comma-separated value files (CSVs). The visualization capability is flexible and user-friendly by introducing several user-defined parameters, for instance, the start and the end project year of interest and the type of colors to be assigned to the cash flows.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗
NASA NTRS2003

Karhunen-Loeve Estimation of the Power Spectrum Parameters from the Angular Distribution of Galaxies in Early Sloan Digital Sky Survey Data

We present measurements of parameters of the three-dimensional power spectrum of galaxy clustering from 222 square degrees of early imaging data in the Sloan Digital Sky Survey (SDSS). The projected galaxy distribution on the sky is expanded over a set of Karhunen-Loeve (KL) eigenfunctions, which optimize the signal-to-noise ratio in our analysis. A maximum likelihood analysis is used to estimate parameters that set the shape and amplitude of the three-dimensional power spectrum of galaxies in the SDSS magnitude-limited sample with r* less than 21. Our best estimates are gamma = 0.188 +/- 0.04 and sigma(sub 8L) = 0.915 +/- 0.06 (statistical errors only), for a flat universe with a cosmological constant. We demonstrate that our measurements contain signal from scales at or beyond the peak of the three-dimensional power spectrum. We discuss how the results scale with systematic uncertainties, like the radial selection function. We find that the central values satisfy the analytically estimated scaling relation. We have also explored the effects of evolutionary corrections, various truncations of the KL basis, seeing, sample size, and limiting magnitude. We find that the impact of most of these uncertainties stay within the 2 sigma uncertainties of our fiducial result.

Szalay, Alexander S.↗
NASA NTRS2014

Reply to Comment by Laprise on 'the Added Value to Global Model Projections of Climate Change by Dynamical Downscaling: a Case Study over the Continental U.S. Using the GISS-ModelE2 and WRF Models'

In his comment, Laprise raises several points that we agree merit consideration. His primary critique is that our study [Racherla et al., 2012] tested the ability of the WRF regional climate model to reproduce historical temperature and precipitation change relative to the driving global climate model (GCM) using only a single simulation rather than an ensemble. He asserts that the observed changes are smaller than the internal variability in the climate system (i.e., not statistically significant) and that thus a single simulation should not necessarily be able to capture the observations. Laprise points out that the statistical signal is reduced for a multi-decadal trend such as the one we analyzed in comparison with mean climatology and cites two studies showing that for particular climate parameters it can take any years for a signal to be discerned over internal variability. He states that The results of theexperiment as designed were strongly influenced by the presence of internal variability and sampling errors,which masked the rather small climate changes that may have occurred as a consequence of changes inforcing during the period considered. While Laprise discusses statistics in general terms at some length, for the actual climate trends examined in our study, he offers no evidence that the forced signal was smallcompared with internal variability. The two studies he cites [de Ela et al., 2013; Maraun, 2013] do not provide convincing evidence as they concern climate variables averaged over different times and areas. One in fact examines extreme precipitation events, which by definition are rare and thus have a lower significance level. We accept the general point that it is important to consider internal variability, and as noted in our paper we agree that an ensemble of simulations is in principle an optimal, though computationally expensive, approach. While we did not present the statistical significance of the observations in our original paper, we have now evaluated those for the regional temperature trends used in our study to evaluate the added value of WRF and thus can analyze data as to the magnitude of the trends with respect to internal variability.

climate↗
DOE OSTI2024

Increasing the power of weak lensing data with multipole-based intrinsic alignment estimators

ABSTRACT It has long been known that galaxy shapes align coherently with the large-scale density field. Characterizing this effect is essential to interpreting measurements of weak gravitational lensing, the deflection of light from distant galaxies by matter overdensities along the line of sight, as it also produces coherent galaxy alignments that we wish to interpret in terms of a cosmological model. Existing direct measurements of intrinsic alignments using galaxy samples with high-quality shape and redshift measurements typically use well-understood but sub-optimal projected estimators, which do not make good use of the information in the data when comparing those estimators to theoretical models. We demonstrate a more optimal estimator, based on a multipole expansion of the correlation functions or power spectra, for direct measurements of galaxy intrinsic alignments. We show that even using the lowest order multipole alone increases the significance of inferred model parameters using simulated and real data, without any additional modelling complexity. We apply this estimator to measurements of parameters of the non-linear alignment model using data from the Sloan Digital Sky survey, demonstrating consistent results with a factor of ∼2 greater precision in parameter fits to intrinsic alignments models. This result is functionally equivalent to quadrupling the survey area, but without the attendant costs – thereby demonstrating the value in using this new estimator in current and future intrinsic alignments measurements using spectroscopic galaxy samples.

Astronomy & Astrophysics↗
DOE OSTI2025

Machine Learning for Automated Weld Quality Monitoring and Control

Resistance Spot Welding (RSW) is a critical process in the automotive industry, valued for its cost-effectiveness, short cycle time, and robustness. However, achieving consistent high-quality joints remains challenging due to the complex interplay of various factors, like materials, processes, and manufacturing uncertainties, etc. Under the collaborative project between Oak Ridge National Laboratory (ORNL) and General Motors (GM), we have developed a robust and expansible machine learning (ML) framework aimed at enhancing quality control in RSW. By harnessing the power of machine learning, we have developed the ability to ensure every aspect of the welding process, from the initial process design stage to the final weld joint quality. The framework operates by analyzing a variety of data streams, including in-line process signals, process parameters, materials, and postprocessed weld joint data. Through this analysis, the models have been trained to detect deviations from optimal quality standards, leveraging their ability to identify signature data patterns and anomalies within in-line signals and construct complex correlations between these signals and weld quality parameters. Meanwhile, the machine learning framework is designed to adapt to a variety of materials, including high strength steels and aluminum alloys, etc. Its flexible architecture facilitates the incorporation of diverse data sources and features, enabling precise modeling and prediction across a broad range of material properties and weld quality variables. The expansible ML frameworks represent a promising transformation in weld quality monitoring and control, empowering industry to achieve high levels of efficiency, consistency, and reliability in manufacturing.

99 GENERAL AND MISCELLANEOUS↗
DOE OSTI2025

Machine Learning for Automated Weld Quality Monitoring and Control

Resistance Spot Welding (RSW) is a critical process in the automotive industry, valued for its cost-effectiveness, short cycle time, and robustness. However, achieving consistent high-quality joints remains challenging due to the complex interplay of various factors, like materials, processes, and manufacturing uncertainties, etc. Under the collaborative project between Oak Ridge National Laboratory (ORNL) and General Motors (GM), we have developed a robust and expansible machine learning (ML) framework aimed at enhancing quality control in RSW. By harnessing the power of machine learning, we have developed the ability to ensure every aspect of the welding process, from the initial process design stage to the final weld joint quality. The framework operates by analyzing a variety of data streams, including in-line process signals, process parameters, materials, and postprocessed weld joint data. Through this analysis, the models have been trained to detect deviations from optimal quality standards, leveraging their ability to identify signature data patterns and anomalies within in-line signals and construct complex correlations between these signals and weld quality parameters. Meanwhile, the machine learning framework is designed to adapt to a variety of materials, including high strength steels and aluminum alloys, etc. Its flexible architecture facilitates the incorporation of diverse data sources and features, enabling precise modeling and prediction across a broad range of material properties and weld quality variables. The expansible ML frameworks represent a promising transformation in weld quality monitoring and control, empowering industry to achieve high levels of efficiency, consistency, and reliability in manufacturing.

42 ENGINEERING↗