SEARCH · Engineering Papers

Results for “Predictive modelling”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 685 records · Page 38

All sources NASA NTRS DOE OSTI

DOE OSTI2021

Development of Explainable, Knowledge-Guided AI Models to Enhance the E3SM Land Model Development and Uncertainty Quantification

Focal Area(s): (2)Predictive modeling using AI techniques and AI-derived model components; use of AI and other tools to design a prediction system comprising of a hierarchy of models. (3) Insight gleaned from complex data (both observed and simulated) using AI, big data analytics, and other advanced methods, including explainable AI and physics- or knowledge- guided AI. Science Challenge: The Energy Exascale Earth System Model (E3SM) is a fully coupled, state-of-the-science Earth system model that uses code optimized for DOE's advanced computers to address the most critical scientific questions facing our nation and society (Golaz et al., 2019). The E3SM Land model (ELM) is designed to understand how the changes in terrestrial land surfaces will interact with other Earth system components and has been used to understand hydrologic cycles, biogeophysics, and ecosystem dynamics. In spite of great successes, the ELM has several known issues that restrain rapid improvements. For example, the ELM uses equilibrium models to simulate dynamic land-climate interactions and it requires long model spin-up time to identify suitable initial conditions for transient simulations. The ELM lacks built-in uncertainty mechanisms that can improve the robustness of model predictions. The ELM is a holistic, deterministic model system with a rigid design, and in many situations, it is hard to modify the ELM system to incorporate new theory/hypothesis and new data across scales to address emerging science problems (such as predicting the impacts of water cycle extremes). In addition, The ELM is technically optimized for traditional CPU-centric computers and it cannot fully utilize the current and incoming leadership computers for model simulations and uncertainty quantification (UQ). The success of artificial intelligence (AI) has inspired scientists to use AI models to discover intrinsic features from simulation data (Chattopadhyay et al., 2020) and observational data (Reichstein et al., 2019) to gain further process understanding of Earth science problems. However, autonomous AI model training through deep learning usually requires a huge amount of annotated data. To overcome the limitations from the data and computing resources, knowledge-guided AI models are necessary where human-knowledge is ingested in model construction (Banino et al., 2018) and training process (Silver et al., 2016) for efficient learning. Herein, we present a new way that leverages the process understanding from the ELM to guide AI model development for the ELM enhancement and UQ. We hope this study can inspire further Earth and environmental system model developments and transformations.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2023

Integration of Biorelevant Pediatric Dissolution Methodology into PBPK Modeling to Predict In Vivo Performance and Bioequivalence of Generic Drugs in Pediatric Populations: a Carbamazepine Case Study

This study investigated the impact of gastro-intestinal fluid volume and bile salt (BS) concentration on the dissolution of carbamazepine (CBZ) immediate release (IR) 100 mg tablets and to integrate these in vitro biorelevant dissolution profiles into physiologically based pharmacokinetic modelling (PBPK) in pediatric and adult populations to determine the biopredictive dissolution profile. Dissolution profiles of CBZ IR tablets (100 mg) were generated in 50–900 mL biorelevant adult fasted state simulated gastric and intestinal fluid (Ad-FaSSGF and Ad-FaSSIF), also in three alternative compositions of biorelevant pediatric FaSSGF and FaSSIF medias at 200 mL. This study found that CBZ dissolution was poorly sensitive to changes in the composition of the biorelevant media, where dissimilar dissolution (F2 = 46.2) was only observed when the BS concentration was changed from 3000 to 89 μM (Ad-FaSSIF vs Ped-FaSSIF 50% 14 BS). PBPK modeling demonstrated the most predictive dissolution volume and media composition to forecast the PK was 500 mL of Ad-FaSSGF/Ad-FaSSIF media for adults and 200 mL Ped-FaSSGF/FaSSIF media for pediatrics. A virtual bioequivalence simulation was conducted by using Ad-FaSSGF and/or Ad-FaSSIF 500 mL or Ped-FaSSGF and/or Ped-FaSSIF 200 mL dissolution data for CBZ 100 mg (reference and generic test) IR product. The CBZ PBPK models showed bioequivalence of the product. This study demonstrates that the integration of biorelevant dissolution data can predict the PK profile of a poorly soluble drug in both populations. Further work using more pediatric drug products is needed to verify biorelevant dissolution data to predict the in vivo performance in pediatrics.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2024

The use of digital thread for reconstruction of local fiber orientation in a compression molded pin bracket via deep learning

A deep convolutional neural network (DCNN) was used for microstructure reconstruction using artificial intelligence (MR-AI) by predicting local average fiber orientation distributions (FOD) in a 3D prepreg platelet molded composite (PPMC) pin bracket. To train the MR-AI model, surface strain fields from residual stresses simulated in PPMC plates were used as the input to the DCNN. A training dataset included PPMC plates with various degrees of global fiber alignment, based on the information obtained from high-fidelity flow simulation of a pin bracket. Further, the MR-AI model was then deployed to analyze FOD in the 3D pin bracket by conducting thermo-elastic residual stress analysis. Initially, the MR-AI model was established entirely on the synthetic simulation data. Then, a μCT scan of a physically molded pin bracket was used to create a finite element model that provided data for additional validation of the DCNN model. For the μCT scan finite element pin bracket the MR-AI model predicted the distribution of fiber orientation tensor components with MAE of 0.10 indicating a global prediction error of 10%. For the flow simulated pin bracket, the MR-AI model predicted the distribution of fiber orientation tensor components with a global prediction error of 11%. The MR-AI model showed the ability to predict regions of varying alignment in the base and flange of the pin bracket. The proposed MR-AI methodology allows for rapid prediction of FOD in geometrically complex parts and offers a promising path to detecting unique fiber orientation states in molded components.

42 ENGINEERING↗

DOE OSTI2020

Predictive Integrated Modeling of Heated Tokamak Plasmas (Final Technical Report)

The research carried out by the Lehigh University fusion group during this reporting period: (1) Self-consistent predictive core-pedestal ITER scenario modeling; (2) Microtearing instabilities and electron thermal transport in standard and spherical tokamak discharges; (3) Neural network model of the Multi-Mode anomalous transport module; (4) Anomalous transport in KSTAR, JET, EAST, Alcator C-Mod and DIII-D discharges; (5) Theory for transport in magnetized plasmas; (6) Control-oriented modeling of the q-profile dynamic response in preparation for advanced scenario control in KSTAR and in EAST; (7) Modeling of edge harmonic oscillation in DIII-D QH mode discharges; (8) Elucidating plasma dynamics in Hasegawa-Wakatani turbulence by information geometry; (9) Non-linear Langevin and Fractional Fokker–Planck equations for anomalous diffusion by Levy stable processes; and (10) Investigation of the plasma shaping effects on the H-mode pedestal structure.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2022

Accelerated screening of functional atomic impurities in halide perovskites using high-throughput computations and machine learning

The pressing need for novel materials that can serve rising demands in solar cell and optoelectronic technologies makes the nexus of halide perovskites, high-throughput computations, and machine learning, very promising. Ever increasing amounts of data on the structure, fundamental properties, and device performance of halide perovskites provide opportunities for learning chemical rules and design principles that make these materials attractive, and applying them across wide chemical spaces. In this work, we show that impurity properties of halide perovskites computed using density functional theory (DFT) can be combined with machine learning (ML) to deliver predictive models and quick identification of optoelectronically active impurity atoms. Our computation lead to the largest reported dataset of the formation energies and charge transition levels of Pb-site impurities in methylammonium lead halide (MAPbX 3 ) perovskites. Descriptors are defined to uniquely represent any impurity atom in any MAPbX 3 compound and mapped to the computed impurity properties using regression techniques such as Gaussian process regression, neural networks, and random forests. We use the best optimized predictive models to make predictions for hundreds of impurities across 9 MAPbX 3 compounds and create lists of dominating impurities, that is, impurities that can shift the equilibrium Fermi level in the perovskite as determined by native point defects. Finally, this accelerated screening powered by computations and machine learning can guide the identification of problematic impurities that may cause undesired recombination of charge carriers, as well as impurities that can be deliberately introduced to tune the perovskite conductivity and resulting photovoltaic absorption.

36 MATERIALS SCIENCE↗

DOE OSTI2024

Versatile stochastic model for predictive KMC simulation of fcc metal nanostructure evolution with realistic kinetics

Stochastic lattice-gas models provide the natural framework for analysis of the surface diffusion-mediated evolution of crystalline metal nanostructures on the appropriate time scale (often 10 1 –10 4 s) and length scale. Model behavior can be precisely assessed by kinetic Monte Carlo simulation, typically incorporating a rejection-free algorithm to efficiently handle the broad range of Arrhenius rates for hopping of surface atoms. The model should realistically prescribe these rates, or the associated barriers, for a diversity of local surface environments. However, commonly used generic choices for barriers fail, even qualitatively, to simultaneously describe diffusion for different low-index facets, for terrace vs step edge diffusion, etc. We introduce an alternative Unconventional Interaction–Conventional Interaction formalism to prescribe these barriers, which, even with few parameters, can realistically capture most aspects of behavior. Here, the model is illustrated for single-component fcc metal systems, mainly for the case of Ag. It is quite versatile and can be applied to describe both the post-deposition evolution of 2D nanostructures in homoepitaxial thin films (e.g., reshaping and coalescence of 2D islands) and the post-synthesis evolution of 3D nanocrystals (e.g., reshaping of nanocrystals synthesized with various faceted non-equilibrium shapes back to 3D equilibrium Wulff shapes).

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2021

Machine Learning Augmented Predictive and Generative Model for Rupture Life in Ferritic and Austenitic Steels

The Larson-Miller parameter (LMP) offers an efficient and fast scheme to estimate the creep rupture life of alloy materials for high temperature applications. However, owing to poor generalizability and dependence on the constant C, which is typically not known a-priori, estimations using the Larson-Miller parameter often result in suboptimal performance for a wide range of materials. At best it is useful in comparing alloys of similar composition. In this work, three machine learning (ML) schemes were developed for rupture life prediction for 9-12% Cr ferritic-martensitic steels and austenitic stainless steels, i.e., a hierarchical model to parameterize LMP using the LMP constant C to compute rupture life, a hierarchical model to parameterize both C and LMP to compute rupture life, and a direct prediction of rupture life. Specifically, we show that the third scheme, using a gradient boosting algorithm, can be used to train ML models for very accurate prediction of rupture life in a variety of alloys (Pear-son Correlation Coefficient > 0.9 for 9-12% Cr and > 0.8 for austenitic stainless steels). In addition, the Shapley value was used to quantify feature importance, making the model interpretable by identifying the effect of various features on the model performance. Furthermore, a variational autoencoder-based generative model was built by conditioning on the experimental dataset to sample hypothetical synthetic candidate alloys from the learnt joint distribution not existing in both 9-12 % Cr ferritic-martensitic steel and austenitic stainless steel datasets. Finally, based on the predictive and generative model, a reinforcement learning strategy has been proposed to guide experimentalists into designing better heat resistant alloys.

Mamun, Md Osman G.↗

DOE OSTI2025

Can protein expression be ‘solved’?

Recombinant protein expression is central to biotechnology’s application in academic exploration as well as human health, climate applications and the bioeconomy in general. However, not all proteins can be expressed in all organisms, and the field lacks a predictive model of soluble protein overexpression that could replace laborious experimental trial-and-error. Here, we discuss the state of the field and identify the lack of large, high-fidelity datasets as the primary bottleneck to progress. We review possible assays that could be used for data collection to identify a path toward an extensible experimental platform for collecting soluble recombinant protein overexpression data across organisms. We suggest that the resulting dataset should be used to train increasingly generalizable predictive models of protein expression to answer the question: “How can predictive protein expression be solved?”.

59 BASIC BIOLOGICAL SCIENCES↗

DOE OSTI2023

Assessing turbulence and mixing parameterizations in the gray-zone of multiscale simulations over mountainous terrain during the METEX21 field experiment

Multiscale numerical weather prediction models transition from mesoscale, where turbulence is fully parameterized, to microscale, where the majority of highly energetic scales of turbulence are resolved. The turbulence gray-zone is situated between these two regimes and multiscale models must downscale through these resolutions. Here, we compare three multiscale simulations which vary by the parameterization used for turbulence and mixing within the gray-zone. The three parameterizations analyzed are the Mellor-Yamada Nakanishi and Niino (MYNN) Level 2.5 planetary boundary layer scheme, the TKE-1.5 large eddy simulation (LES) closure scheme, and a recently developed three-dimensional planetary boundary layer scheme based on the Mellor-Yamada model. The simulation domain includes complex (i.e., mountainous) terrain in Nevada that was instrumented with meteorological towers, profiling and scanning lidars, a tethered balloon, and a surface flux tower. Simulations are compared to each other and to observations, with assessment of model skill at predicting wind speed, wind direction and TKE, and qualitative evaluations of transport and dispersion of smoke from controlled releases. This analysis demonstrates that microscale predictions of transport and dispersion can be significantly influenced by the choice of turbulence and mixing parameterization in the terra incognita , particularly over regions of complex terrain and with strong local forcing. This influence may not be apparent in the analysis of model skill, and motivates future field campaigns involving controlled tracer releases and corresponding modeling studies of the turbulence gray-zone.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2020

Generalized and Mechanistic PV Module Performance Prediction From Computer Vision and Machine Learning on Electroluminescence Images

Electroluminescence (EL) imaging of photovoltiac (PV) modules offers high-speed, high-resolution information about device performance, affording opportunities for greater insight and efficiency in module characterization across manufacturing, research and development, and power plant operations and management. Predicting module electrical properties from EL image features is a critical step toward these applications. In this article, we demonstrate quantification of both generalized and performance mechanism-specific EL image features, using pixel intensity-based and machine learning classification algorithms. From EL image features, we build predictive models for PV module power and series resistance, using time-series current-voltage (I-V) and EL data obtained stepwise on five brands of modules spanning three Si cell types through two accelerated exposures: damp heat (DH) (85°C/85% RH) and thermal cycling (TC) (IEC 61215). Overall, 195 pairs of EL images and I-V characteristics were analyzed, yielding 11700 individual PV cell images. A convolutional neural network was built to classify cells by the severity of busbar corrosion with high accuracy (95%). Generalized power predictive models estimated the maximum power of PV modules from EL images with high confidence and an adjusted-R 2 of 0.88, across all module brands and cell types in extended DH and TC exposures. Mechanistic degradation prediction was demonstrated by quantification of busbar corrosion in EL images of three module brands in DH, and subsequent modeling of series resistance using these mechanism-specific EL image features. For modules exhibiting busbar corrosion, we demonstrated series resistance predictive models with adjusted-R 2 of up to 0.73.

14 SOLAR ENERGY↗

DOE OSTI2022

Search for coherent elastic neutrino-nucleus scattering at a nuclear reactor with CONNIE 2019 data

The Coherent Neutrino-Nucleus Interaction Experiment (CONNIE) is taking data at the Angra 2 nuclear reactor with the aim of detecting the coherent elastic scattering of reactor antineutrinos with silicon nuclei using charge-coupled devices (CCDs). In 2019 the experiment operated with a hardware binning applied to the readout stage, leading to lower levels of readout noise and improving the detection threshold down to 50 eV. The results of the analysis of 2019 data are reported here, corresponding to the detector array of 8 CCDs with a fiducial mass of 36.2 g and a total exposure of 2.2 kg-days. The difference between the reactor-on and reactor-off spectra shows no excess at low energies and yields upper limits at 95% confidence level for the neutrino interaction rates. In the lowest-energy range, 50 − 180 eV, the expected limit stands at 34 (39) times the standard model prediction, while the observed limit is 66 (75) times the standard model prediction with Sarkis (Chavarria) quenching factors.[graphic not available: see fulltext]

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

DOE OSTI2026

Enhanced matter power spectrum from axion kination after Big Bang nucleosynthesis

Despite stringent constraints from Big Bang Nucleosynthesis (BBN) and cosmic microwave background (CMB) observations, it is still possible for well-motivated particle physics models to substantially alter the cosmic expansion history between BBN and recombination. In this work we consider two different axion models that can realize a period of first matter domination, then kination, in this epoch. We perform fits to both primordial element abundances as well as CMB data and determine that up to a decade of late axion domination is allowed by these probes of the early universe. We establish the implications of late axion domination for the matter power spectrum on the scales 1/Mpc ≲ k ≲ 10 3 /Mpc. Our 'log' model predicts a relatively modest bump-like feature together with a small suppression relative to the standard ΛCDM predictions on either side of the enhancement. Our 'two-field' model predicts a larger, plateau-like feature that realizes enhancements to the matter power spectrum of up to two orders of magnitude. These features have interesting implications for structure formation at the forefront of current detection capabilities.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

Measurement of the $H \rightarrow \gamma \gamma $ and $H \rightarrow ZZ^* \rightarrow 4 \ell $ cross-sections in $pp$ collisions at $\sqrt{s}=13.6$ TeV with the ATLAS detector

The inclusive Higgs boson production crosssection is measured in the di-photon and the Z Z* → 4$\ell$ decay channels using 31.4 and 29.0 fb –1 of pp collision data respectively, collected with the ATLAS detector at a centreof-mass energy of $\sqrt{s}$ = 13.6 TeV. To reduce the model dependence, the measurement in each channel is restricted to a particle-level phase space that closely matches the channel’s detector-level kinematic selection, and it is corrected for detector effects. These measured fiducial cross-sections are σ fid,γγ = $76^{+14}_{–13}$ fb, and $σ_{\text{fid,4}\ell}$ = 2.80 ± 0.74 fb, in agreement with the corresponding Standard Model predictions of 67.6±3.7 fb and 3.67±0.19 fb. Assuming Standard Model acceptances and branching fractions for the two channels, the fiducial measurements are extrapolated to the full phase space yielding total cross-sections of σ (pp → H) = $67^{+12}_{–11}$ pb and 46±12 pb at 13.6 TeV from the di-photon and Z Z* → 4$\ell$ measurements respectively. The two measurements are combined into a total cross-section measurement of σ (pp → H) = 58.2±8.7 pb, to be compared with the Standard Model prediction of σ (pp → H) SM = 59.9 ± 2.6 pb.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2023

Total kinetic energy release in the fast neutron induced fission of actinide nuclei

The total kinetic energy release and fission mass distributions for the fast neutron (En = 3–100 MeV) induced fission of 232 Th, 233 U, 235 U, 237 Np, 239 Pu, 240 Pu, and 242 Pu have been measured using the LANSCE facility. The neutron energies were deduced from time-of- flight measurements. The fission fragments were detected using Si PIN diode detectors, giving us the fragment energies. The actinide targets were made by vapor deposition leading to high-quality targets, that were thin and uniform with reduced impurities. Corrections were made to the data for pulse height defect and the fragment energy loss in the target and its backing. The TKE distributions were Gaussian in shape and their mean value as a function of incoming neutron energy could be fitted with second order polynomials. In the case of 233 U and 235 U, our measurements agree with prior work. Our measurements for 232 Th are unique. Our data agree with Viola scaling. The constant position of the heavy mass peak is interpreted as being due to the influence of the N = 88 and Z = 50 shells. The GEF model predictions agree with the data in general as do the CGMF model predictions.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2021

Reliable modeling and prediction of precipitation & radiation for mountainous hydrology

This White Paper focuses on data-driven atmospheric process model emulation and atmospheric process surrogate model development. It proposes leveraging recent AI advances in these approaches to fill in unavoidable observational gaps and enable high-fidelity modeling/predictability of the atmosphere and land-surface interactions in mountainous watersheds. This approach will support studies and predictability of water cycle extremes.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2024

Exclusive η Electro-Production Beam Spin Asymmetry Measurements using CLAS12 at Jefferson Lab

The exploration of nucleon structure and electromagnetic transitions from ground-state to excited-state is a cornerstone of nuclear physics research. Meson electro-production experiments have opened new avenues for investigating these phenomena, particularly in the 12 GeV era at Jefferson Lab with the CLAS12 spectrometer. The ?N final states, accessible only through isospin resonances I = 1/2, provide a unique tool for studying nucleon excitations. By simplifying the analysis and enabling a cleaner extraction of resonance properties compared to the extensively studied ?N final states, ? electroproduction offers a complementary approach to unraveling the structure of excited nucleons. This work presents the first-ever measurement of the beam spin asymmetry (BSA) in exclusive ? electroproduction, covering a previously unexplored kinematic region with 1.6 ? W ? 2.2 GeV. The BSA is extracted from the CLAS12 data using a comprehensive analysis framework that carefully considers the statistical limitations of the data set. The results are compared to predictions from theoretical models, such as the Jülich-Bonn-Washington (JBW) and MAID, as well as compared to previously published cross-section and spin observable results from CLAS and SLAC. Notably, the extracted BSA exhibits discrepancies with the model predictions, highlighting the potential for refining theoretical descriptions of nucleon resonances and their electromagnetic couplings through the incorporation of these new data. The high precision data obtained in this previously unmeasured kinematic region now serve as valuable input for theorists to refine their models.

Illari, Isabella↗

DOE OSTI2021

AI-Batt (Autonomous Identification of Battery Life Models) [SWR 21-36]

Autonomous Identification of Battery Life Models (AI-Batt) AI-Batt is a MATLAB code base for developing lifetime models for batteries from accelerated aging data. The code base provides many functions for processing, visualizing, and modeling battery aging data, making the data processing, exploration, and modeling workflow substantially faster. These tools are tailored for working with battery aging data sets, which usually consist of many separate time-series for each cell, with many test conditions and possible replicates at each condition, which makes it difficult to simply process or visualize the data set. Complex modeling tasks, such as cross-validation, sensitivity analysis, and uncertainty quantification have been implemented to enable thorough statistical investigation of model predictions. Additionally, several machine-learning algorithms are implemented to autonomously identify suitable models via symbolic regression. Data processing functions automatically cast data from the struct data type, which is commonly used to store experimental data, but is not an acceptable input for most algorithms, to the table data type, which can be easily used as input to any optimization algorithm. Also, the data can be separated into time-invariant and time-variant data tables, which is helpful for exploring the data set as well as developing separate models for time-variant and time-invariant aging mechanisms. For example, in aging tests with constant temperature, temperature is a time-invariant experimental condition. Visualization tools enable plotting of data, model fits, and model simulations possible with single-line function calls, empowering data exploration of complex data sets with both time-varying and time-invariant trends. Plots can be automatically generated for the whole data set, or separated by data group (groups of test replicates) or individual data series. Data points or data series can be automatically colored by the value of a variable with a variety of color maps, and model predictions can also be colored by the value of a fit statistic. Comparisons between data sets and the predictions/simulations of different models on the same data set can be easily plotted as well. Distributions of parameter values from bootstrap resampling can be plotted to visualize the reliability of parameter estimation, or determine any correlations between parameters. Modeling tools handle the complex task of creating and parsing symbolic equations for modeling battery lifetime. Equations are parsed to grab relevant data variables, parameter values, or specified sub-models for input into optimization, evaluation, or simulation functions. Models can be optimized locally (one set of parameters for each data series), bi-level (some parameters shared across the data set), or globally (single set of parameters for all data). Functions implementing symbolic regression algorithms help users to discover effective model equations, even in poorly sampled, high-dimensional data.

Smith, Kandler↗

DOE OSTI2024

A Generation-Storage Coordination Dispatch Strategy for Power System Based on Causal Reinforcement Learning

In the backdrop of global energy transformation, power systems integrating high proportions of renewable energy sources are facing unprecedented challenges in operational stability and dispatch efficiency. To address these challenges, this study introduces a generation-storage coordination real-time dispatch strategy based on Causal Power System Dynamic Reinforcement Learning (CPSDRL). Diverging from traditional reinforcement learning approaches, CPSDRL innovatively incorporates causal inference within the state prediction model - the crux of model-based reinforcement learning - thereby establishing the Power Causal Dynamic Model (PCDM). Assisted by the prior knowledge of power systems, the model significantly enhances prediction accuracy and reliability through a two-stage training process. Utilizing PCDM, this study further applies a direct policy search algorithm to optimize the real-time dispatch strategy. Experimental results indicate that the proposed method improves the stability of generation-storage coordination real-time dispatch and exhibits competitive advantages in sample efficiency and computational speed, compared to traditional model-based and model-free reinforcement learning algorithms. This method is expected to enhance the practicality and adaptability of causal reinforcement learning techniques in power system scheduling and control.

causal reinforcement learning↗