arXiv subjects

Qiugang Lu

Publications and source records attributed to Qiugang Lu.

16 recordsLinked to original sources

arXiv2026

Physics-Informed Neural Network with Transfer Learning for State Estimation in Lithium-Ion Batteries using the Single Particle Model with Electrolyte

Physics-informed neural networks (PINNs) have emerged as a powerful tool for solving nonlinear partial differential equations (PDEs), including battery electrochemical models. They typically en-force conservation laws within the loss function to ensure physically consistent solutions. Tradi-tional numerical methods such as finite difference, finite volume, and finite element techniques, re-ly on discretization and can be computationally expensive for nonlinear systems. To address this challenge, PINNs offer improved scalability, particularly for reduced-order models like the single particle model with electrolyte (SPMe). The SPMe describes lithium-ion battery dynamics through coupled diffusion, transport, reaction kinetics, and voltage equations. Despite these advantages, training SPMe-based PINNs from scratch for different battery chemistries or operating conditions is demanding and often leads to slow convergence. To overcome this limitation, this work introduces a transfer learning framework for SPMe-PINNs. The model is first pretrained to learn general elec-trochemical dynamics and then adapted to a target battery by transferring weights, freezing se-lected layers, and fine tuning the remaining parameters, including estimating key electrochemical variables. Validation using PyBaMM demonstrates accurate voltage prediction, indicating that the proposed approach preserves electrochemical consistency while reducing training time and ena-bling efficient generalization across batteries.

cs.LG↗

arXiv2026

Multi-Level Temporal Graph Networks with Local-Global Fusion for Industrial Fault Diagnosis

Fault detection and diagnosis are critical for the optimal and safe operation of industrial processes. The correlations among sensors often display non-Euclidean structures where graph neural networks (GNNs) are widely used therein. However, for large-scale systems, local, global, and dynamic relations extensively exist among sensors, and traditional GNNs often overlook such complex and multi-level structures for various problems including the fault diagnosis. To address this issue, we propose a structure-aware multi-level temporal graph network with local-global feature fusion for industrial fault diagnosis. First, a correlation graph is dynamically constructed using Pearson correlation coefficients to capture relationships among process variables. Then, temporal features are extracted through long short-term memory (LSTM)-based encoder, whereas the spatial dependencies among sensors are learned by graph convolution layers. A multi-level pooling mechanism is used to gradually coarsen and learn meaningful graph structures, to capture higher-level patterns while keeping important fault related details. Finally, a fusion step is applied to combine both detailed local features and overall global patterns before the final prediction. Experimental evaluations on the Tennessee Eastman process (TEP) demonstrate that the proposed model achieves superior fault diagnosis performance, particularly for complex fault scenarios, outperforming various baseline methods.

cs.LG↗

arXiv2025

Equivalent-Circuit Thermal Model for Batteries with One-Shot Parameter Identification

Accurate state of temperature (SOT) estimation for batteries is crucial for regulating their temperature within a desired range to ensure safe operation and optimal performance. The existing measurement-based methods often generate noisy signals and cannot scale up for large-scale battery packs. The electrochemical model-based methods, on the contrary, offer high accuracy but are computationally expensive. To tackle these issues, inspired by the equivalentcircuit voltage model for batteries, this paper presents a novel equivalent-circuit electro-thermal model (ECTM) for modeling battery surface temperature. By approximating the complex heat generation inside batteries with data-driven nonlinear (polynomial) functions of key measurable parameters such as state-of-charge (SOC), current, and terminal voltage, our ECTM is simplified into a linear form that admits rapid solutions. Such simplified ECTM can be readily identified with one single (one-shot) cycle data. The proposed model is extensively validated with benchmark NASA, MIT, and Oxford battery datasets. Simulation results verify the accuracy of the model, despite being identified with one-shot cycle data, in predicting battery temperatures robustly under different battery degradation status and ambient conditions.

eess.SY↗

arXiv2025

Lithium-ion Battery Capacity Prediction via Conditional Recurrent Generative Adversarial Network-based Time-Series Regeneration

Accurate capacity prediction is essential for the safe and reliable operation of batteries by anticipating potential failures beforehand. The performance of state-of-the-art capacity prediction methods is significantly hindered by the limited availability of training data, primarily attributed to the expensive experimentation and data sharing restrictions. To tackle this issue, this paper presents a recurrent conditional generative adversarial network (RCGAN) scheme to enrich the limited battery data by adding high-fidelity synthetic ones to improve the capacity prediction. The proposed RCGAN scheme consists of a generator network to generate synthetic samples that closely resemble the true data and a discriminator network to differentiate real and synthetic samples. Long shortterm memory (LSTM)-based generator and discriminator are leveraged to learn the temporal and spatial distributions in the multivariate time-series battery data. Moreover, the generator is conditioned on the capacity value to account for changes in battery dynamics due to the degradation over usage cycles. The effectiveness of the RCGAN is evaluated across six batteries from two benchmark datasets (NASA and MIT). The raw data is then augmented with synthetic samples from the RCGAN to train LSTM and gate recurrent unit (GRU) models for capacity prediction. Simulation results show that the models trained with augmented datasets significantly outperform those trained with the original datasets in capacity prediction.

eess.SY↗

arXiv2024

Transformer-based Capacity Prediction for Lithium-ion Batteries with Data Augmentation

Lithium-ion batteries are pivotal to technological advancements in transportation, electronics, and clean energy storage. The optimal operation and safety of these batteries require proper and reliable estimation of battery capacities to monitor the state of health. Current methods for estimating the capacities fail to adequately account for long-term temporal dependencies of key variables (e.g., voltage, current, and temperature) associated with battery aging and degradation. In this study, we explore the usage of transformer networks to enhance the estimation of battery capacity. We develop a transformer-based battery capacity prediction model that accounts for both long-term and short-term patterns in battery data. Further, to tackle the data scarcity issue, data augmentation is used to increase the data size, which helps to improve the performance of the model. Our proposed method is validated with benchmark datasets. Simulation results show the effectiveness of data augmentation and the transformer network in improving the accuracy and robustness of battery capacity prediction.

cs.LG↗

arXiv2024

Adaptive Safe Reinforcement Learning-Enabled Optimization of Battery Fast-Charging Protocols

Optimizing charging protocols is critical for reducing battery charging time and decelerating battery degradation in applications such as electric vehicles. Recently, reinforcement learning (RL) methods have been adopted for such purposes. However, RL-based methods may not ensure system (safety) constraints, which can cause irreversible damages to batteries and reduce their lifetime. To this end, this work proposes an adaptive and safe RL framework to optimize fast charging strategies while respecting safety constraints with a high probability. In our method, any unsafe action that the RL agent decides will be projected into a safety region by solving a constrained optimization problem. The safety region is constructed using adaptive Gaussian process (GP) models, consisting of static and dynamic GPs, that learn from online experience to adaptively account for any changes in battery dynamics. Simulation results show that our method can charge the batteries rapidly with constraint satisfaction under varying operating conditions.

eess.SY↗

arXiv2023

Fast Charging of Lithium-Ion Batteries Using Deep Bayesian Optimization with Recurrent Neural Network

Fast charging has attracted increasing attention from the battery community for electrical vehicles (EVs) to alleviate range anxiety and reduce charging time for EVs. However, inappropriate charging strategies would cause severe degradation of batteries or even hazardous accidents. To optimize fast-charging strategies under various constraints, particularly safety limits, we propose a novel deep Bayesian optimization (BO) approach that utilizes Bayesian recurrent neural network (BRNN) as the surrogate model, given its capability in handling sequential data. In addition, a combined acquisition function of expected improvement (EI) and upper confidence bound (UCB) is developed to better balance the exploitation and exploration. The effectiveness of the proposed approach is demonstrated on the PETLION, a porous electrode theory-based battery simulator. Our method is also compared with the state-of-the-art BO methods that use Gaussian process (GP) and non-recurrent network as surrogate models. The results verify the superior performance of the proposed fast charging approaches, which mainly results from that: (i) the BRNN-based surrogate model provides a more precise prediction of battery lifetime than that based on GP or non-recurrent network; and (ii) the combined acquisition function outperforms traditional EI or UCB criteria in exploring the optimal charging protocol that maintains the longest battery lifetime.

eess.SY↗

arXiv2022

A Novel Entropy-Maximizing TD3-based Reinforcement Learning for Automatic PID Tuning

Proportional-integral-derivative (PID) controllers have been widely used in the process industry. However, the satisfactory control performance of a PID controller depends strongly on the tuning parameters. Conventional PID tuning methods require extensive knowledge of the system model, which is not always known especially in the case of complex dynamical systems. In contrast, reinforcement learning-based PID tuning has gained popularity since it can treat PID tuning as a black-box problem and deliver the optimal PID parameters without requiring explicit process models. In this paper, we present a novel entropy-maximizing twin-delayed deep deterministic policy gradient (EMTD3) method for automating the PID tuning. In the proposed method, an entropy-maximizing stochastic actor is employed at the beginning to encourage the exploration of the action space. Then a deterministic actor is deployed to focus on local exploitation and discover the optimal solution. The incorporation of the entropy-maximizing term can significantly improve the sample efficiency and assist in fast convergence to the global solution. Our proposed method is applied to the PID tuning of a second-order system to verify its effectiveness in improving the sample efficiency and discovering the optimal PID parameters compared to traditional TD3.

eess.SY↗

arXiv2022

Enhanced CNN with Global Features for Fault Diagnosis of Complex Chemical Processes

Convolutional neural network (CNN) models have been widely used for fault diagnosis of complex systems. However, traditional CNN models rely on small kernel filters to obtain local features from images. Thus, an excessively deep CNN is required to capture global features, which are critical for fault diagnosis of dynamical systems. In this work, we present an improved CNN that embeds global features (GF-CNN). Our method uses a multi-layer perceptron (MLP) for dimension reduction to directly extract global features and integrate them into the CNN. The advantage of this method is that both local and global patterns in images can be captured by a simple model architecture instead of establishing deep CNN models. The proposed method is applied to the fault diagnosis of the Tennessee Eastman process. Simulation results show that the GF-CNN can significantly improve the fault diagnosis performance compared to traditional CNN. The proposed method can also be applied to other areas such as computer vision and image processing.

eess.SY↗

arXiv2022

Improving Convolutional Neural Networks for Fault Diagnosis by Assimilating Global Features

Deep learning techniques have become prominent in modern fault diagnosis for complex processes. In particular, convolutional neural networks (CNNs) have shown an appealing capacity to deal with multivariate time-series data by converting them into images. However, existing CNN techniques mainly focus on capturing local or multi-scale features from input images. A deep CNN is often required to indirectly extract global features, which are critical to describe the images converted from multivariate dynamical data. This paper proposes a novel local-global CNN (LG-CNN) architecture that directly accounts for both local and global features for fault diagnosis. Specifically, the local features are acquired by traditional local kernels whereas global features are extracted by using 1D tall and fat kernels that span the entire height and width of the image. Both local and global features are then merged for classification using fully-connected layers. The proposed LG-CNN is validated on the benchmark Tennessee Eastman process (TEP) dataset. Comparison with traditional CNN shows that the proposed LG-CNN can greatly improve the fault diagnosis performance without significantly increasing the model complexity. This is attributed to the much wider local receptive field created by the LG-CNN than that by CNN. The proposed LG-CNN architecture can be easily extended to other image processing and computer vision tasks.

cs.CV↗

arXiv2022

Stability-Preserving Automatic Tuning of PID Control with Reinforcement Learning

PID control has been the dominant control strategy in the process industry due to its simplicity in design and effectiveness in controlling a wide range of processes. However, traditional methods on PID tuning often require extensive domain knowledge and field experience. To address the issue, this work proposes an automatic PID tuning framework based on reinforcement learning (RL), particularly the deterministic policy gradient (DPG) method. Different from existing studies on using RL for PID tuning, in this work, we consider the closed-loop stability throughout the RL-based tuning process. In particular, we propose a novel episodic tuning framework that allows for an episodic closed-loop operation under selected PID parameters where the actor and critic networks are updated once at the end of each episode. To ensure the closed-loop stability during the tuning, we initialize the training with a conservative but stable baseline PID controller and the resultant reward is used as a benchmark score. A supervisor mechanism is used to monitor the running reward (e.g., tracking error) at each step in the episode. As soon as the running reward exceeds the benchmark score, the underlying controller is replaced by the baseline controller as an early correction to prevent instability. Moreover, we use layer normalization to standardize the input to each layer in actor and critic networks to overcome the issue of policy saturation at action bounds, to ensure the convergence to the optimum. The developed methods are validated through setpoint tracking experiments on a second-order plus dead-time system. Simulation results show that with our scheme, the closed-loop stability can be maintained throughout RL explorations and the explored PID parameters by the RL agent converge quickly to the optimum.

eess.SY↗

arXiv2022

Deep Learning-based Predictive Control of Battery Management for Frequency Regulation

This paper proposes a deep learning-based optimal battery management scheme for frequency regulation (FR) by integrating model predictive control (MPC), supervised learning (SL), reinforcement learning (RL), and high-fidelity battery models. By taking advantage of deep neural networks (DNNs), the derived DNN-approximated policy is computationally efficient in online implementation. The design procedure of the proposed scheme consists of two sequential processes: (1) the SL process, in which we first run a simulation with an MPC embedding a low-fidelity battery model to generate a training data set, and then, based on the generated data set, we optimize a DNN-approximated policy using SL algorithms; and (2) the RL process, in which we utilize RL algorithms to improve the performance of the DNN-approximated policy by balancing short-term economic incentives and long-term battery degradation. The SL process speeds up the subsequent RL process by providing a good initialization. By utilizing RL algorithms, one prominent property of the proposed scheme is that it can learn from the data generated by simulating the FR policy on the high-fidelity battery simulator to adjust the DNN-approximated policy, which is originally based on low-fidelity battery model. A case study using real-world data of FR signals and prices is performed. Simulation results show that, compared to conventional MPC schemes, the proposed deep learning-based scheme can effectively achieve higher economic benefits of FR participation while maintaining lower online computational cost.

eess.SY↗

arXiv2021

Image-Based Model Predictive Control via Dynamic Mode Decomposition

We present a data-driven model predictive control (MPC) framework for systems with high state-space dimensionalities. This work is motivated by the need to exploit sensor data that appears in the form of images (e.g., 2D or 3D spatial fields reported by thermal cameras). We propose to use dynamic mode decomposition (DMD) to directly build a low-dimensional model from image data and we use such model to obtain a tractable MPC controller. We demonstrate the scalability of this approach (which we call DMD-MPC) by using a 2D thermal diffusion system. Here, we assume that the evolution of the thermal field is captured by 50x50 pixel images, which results in a 2500-dimensional state-space. We show that the dynamics of this high-dimensional space can be accurately predicted by using a 40-dimensional DMD model and we show that the field can be manipulated satisfactorily by using an MPC controller that embeds the low-dimensional DMD model. We also show that the DMD-MPC controller significantly outperforms a standard MPC controller that uses data from a finite set of spatial locations (proxy locations) to manipulate the high-dimensional thermal field. This comparison illustrates the value of information embedded in image data.

eess.SY↗

arXiv2021

MPC Controller Tuning using Bayesian Optimization Techniques

We present a Bayesian optimization (BO) framework for tuning model predictive controllers (MPC) of central heating, ventilation, and air conditioning (HVAC) plants. This approach treats the functional relationship between the closed-loop performance of MPC and its tuning parameters as a black-box. The approach is motivated by the observation that evaluating the closed-loop performance of MPC by trial-and-error is time-consuming (e.g., every closed-loop simulation can involve solving thousands of optimization problems). The proposed BO framework seeks to quickly identify the optimal tuning parameters by strategically exploring and exploiting the space of the tuning parameters. The effectiveness of the BO framework is demonstrated by using an MPC controller for a central HVAC plant using realistic data. Here, the BO framework tunes back-off terms for thermal storage tanks to minimize year-long closed-loop costs. Simulation results show that BO can find the optimal back-off terms by conducting 13 year-long simulations, which significantly reduces the computational burden of a naive grid search. We also find that the back-off terms obtained with BO reduce the closed-loop costs.

eess.SY↗

arXiv2020

Characterizing the Predictive Accuracy of Dynamic Mode Decomposition for Data-Driven Control

Dynamic mode decomposition (DMD) is a versatile approach that enables the construction of low-order models from data. Controller design tasks based on such models require estimates and guarantees on predictive accuracy. In this work, we provide a theoretical analysis of DMD model errors that reveals impact of model order and data availability. The analysis also establishes conditions under which DMD models can be made asymptotically exact. We verify our results using a 2D diffusion system.

eess.SY↗

arXiv2020

Unifying Theorems for Subspace Identification and Dynamic Mode Decomposition

This paper presents unifying results for subspace identification (SID) and dynamic mode decomposition (DMD) for autonomous dynamical systems. We observe that SID seeks to solve an optimization problem to estimate an extended observability matrix and a state sequence that minimizes the prediction error for the state-space model. Moreover, we observe that DMD seeks to solve a rank-constrained matrix regression problem that minimizes the prediction error of an extended autoregressive model. We prove that existence conditions for perfect (error-free) state-space and low-rank extended autoregressive models are equivalent and that the SID and DMD optimization problems are equivalent. We exploit these results to propose a SID-DMD algorithm that delivers a provably optimal model and that is easy to implement. We demonstrate our developments using a case study that aims to build dynamical models directly from video data.

eess.SY↗