Learning heterogeneous reaction kinetics from X-ray videos pixel by pixel.

Reaction rates at spatially heterogeneous, unstable interfaces are notoriously difficult to quantify, yet are essential in engineering many chemical systems, such as batteries1 and electrocatalysts2. Experimental characterizations of such materials by operando microscopy produce rich image datasets3-6, but data-driven methods to learn physics from these images are still lacking because of the complex coupling of reaction kinetics, surface chemistry and phase separation7. Here we show that heterogeneous reaction kinetics can be learned from in situ scanning transmission X-ray microscopy (STXM) images of carbon-coated lithium iron phosphate (LFP) nanoparticles. Combining a large dataset of STXM images with a thermodynamically consistent electrochemical phase-field model, partial differential equation (PDE)-constrained optimization and uncertainty quantification, we extract the free-energy landscape and reaction kinetics and verify their consistency with theoretical models. We also simultaneously learn the spatial heterogeneity of the reaction rate, which closely matches the carbon-coating thickness profiles obtained through Auger electron microscopy (AEM). Across 180,000 image pixels, the mean discrepancy with the learned model is remarkably small (<7%) and comparable with experimental noise. Our results open the possibility of learning nonequilibrium material properties beyond the reach of traditional experimental methods and offer a new non-destructive technique for characterizing and optimizing heterogeneous reactive surfaces.

A cross-scale framework for evaluating flexibility values of battery and fuel cell electric vehicles.

Flexibility has become increasingly important considering the intermittency of variable renewable energy in low-carbon energy systems. Electrified transportation exhibits great potential to provide flexibility. This article analyzed and compared the flexibility values of battery electric vehicles and fuel cell electric vehicles for planning and operating interdependent electricity and hydrogen supply chains while considering battery degradation costs. A cross-scale framework involving both macro-level and micro-level models was proposed to compute the profits of flexible EV refueling/charging with battery degradation considered. Here we show that the flexibility reduction after considering battery degradation is quantified by at least 4.7% of the minimum system cost and enlarged under fast charging and low-temperature scenarios. Our findings imply that energy policies and relevant management technologies are crucial to shaping the comparative flexibility advantage of the two transportation electrification pathways. The proposed cross-scale methodology has broad implications for the assessment of emerging energy technologies with complex dynamics.

Realistic fault detection of li-ion battery via dynamical deep learning.

Accurate evaluation of Li-ion battery (LiB) safety conditions can reduce unexpected cell failures, facilitate battery deployment, and promote low-carbon economies. Despite the recent progress in artificial intelligence, anomaly detection methods are not customized for or validated in realistic battery settings due to the complex failure mechanisms and the lack of real-world testing frameworks with large-scale datasets. Here, we develop a realistic deep-learning framework for electric vehicle (EV) LiB anomaly detection. It features a dynamical autoencoder tailored for dynamical systems and configured by social and financial factors. We test our detection algorithm on released datasets comprising over 690,000 LiB charging snippets from 347 EVs. Our model overcomes the limitations of state-of-the-art fault detection models, including deep learning ones. Moreover, it reduces the expected direct EV battery fault and inspection costs. Our work highlights the potential of deep learning in improving LiB safety and the significance of social and financial information in designing deep learning models.

Matrix factorization based instrumental variable approach for simultaneous identification of Bi-directional path models.

This paper proposes a closed-loop identification approach to integrate matrix factorization algorithms with generalized instrumental variable (GIV) techniques to simultaneously identify the parameters and orders for both forward and backward path models. Aside from the technique of UD factorization, the QR factorization technique, which possesses good numerical property, is utilized for the proposed GIV-based method. The major difficulty and novelty of the proposed approach lies in how to properly construct instruments, number of instruments, and weighting matrices to obtain enhanced identification performance. To the end, the identification accuracy properties, in terms of the covariance matrix of the parameter estimates, are provided. In addition, a sufficient condition of consistent parameter estimates for the GIV-based approach is discussed. The effectiveness of the proposed identification method is demonstrated by simulation results.