Keke Huang’s research while affiliated with Central South University and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (130)


A weighted deep learning based predictive control for multimode nonlinear system with industrial applications
  • Article

January 2025

·

6 Reads

IEEE Transactions on Automation Science and Engineering

Keke Huang

·

Wenpu Cao

·

Yishun Liu

·

[...]

·

Weihua Gui

In response to the challenge of strongly nonlinear and multimode systems control, this paper introduces a weighted deep learning based adaptive predictive control method. This approach integrates LSTM networks for different operating modes using a set of weighting coefficients. These coefficients are dynamically updated during online control via an error-guided scheduling strategy to adapt to changing operation modes. Compared to offline identification based methods, the proposed method eliminates the need for mode recognition or model switching strategies and can adapt to drifted operation modes. In contrast to online methods, it achieves rapid model convergence and reduced computational cost, requiring only minimal data to update the weighting coefficients without necessitating the retraining of the LSTM networks. Theoretical convergence and stability analysis ensure the reliability of the proposed method. Numerical simulations and industrial control experiments demonstrate that the proposed approach exhibits favorable control performance across both known and drifted operation modes.


Global Information-Based Lifelong Dictionary Learning for Multimode Process Monitoring

December 2024

·

19 Reads

·

1 Citation

IEEE Transactions on Systems Man and Cybernetics Systems

Multimode process monitoring plays a significant role in ensuring the stable operation of industrial processes under changing conditions. Due to the continuous emergence of new modes, some adaptive model updating methods are proposed. However, the updated model may forget important features learned in previous modes, thus reducing the monitoring performance. To address this problem, this article proposes a global information-based lifelong dictionary learning (GI-LDL) method for multimode process monitoring. Specifically, this article adopts dictionary atoms as the fundamental units for representing process data and proposes a method for measuring the importance of dictionary atoms based on global information. Then, to ensure that the dictionary retains its representational ability for both new and historical mode data during the mode updating process, a surrogate quadratic loss considering the importance is further proposed to penalize changes of important atoms. Compared with dictionary constraints, finer-grained atomic constraints ensure that the dictionary preserves important features of previous modes while learning features of new modes. Finally, considering that the number of modes in multimode industrial processes is often unknown in advance, this article explicitly derives analytical solutions for dictionary updating, thus it is capable of accommodating ever-increasing modes in real industrial processes. To verify the effectiveness and advancement of the proposed method, extensive experiments are elaborately designed, and experimental results indicate that the proposed method has precise monitoring capabilities for both historical and new modes.


Interpretable Switching Deep Markov Model for Industrial Process Monitoring via Cloud-Edge Collaborative Framework

November 2024

·

27 Reads

IEEE Transactions on Instrumentation and Measurement

In modern manufacturing, process monitoring is crucial in ensuring the stability and safety of production processes. However, the frequent changes in industrial conditions necessitate timely updates and retraining of the on-site deployment of data-driven process monitoring methods, which is a task unattainable with the limited computational resources of edge devices. To address this issue, this paper proposes an interpretable switching deep Markov model (ISDMM)-based cloud-edge collaborative framework for industrial process monitoring. ISDMM defines discrete switching variables that represent working conditions and corresponding multiple transition networks. The transition networks are concurrently trained and switched based on the current working condition, enabling ISDMM to capture different dynamics of the system under different conditions. Before deployment, model simplification is performed to construct a model base where each model is tailored to individual working conditions. Moreover, in the edge layer, a model update strategy is designed to determine whether the model can be invoked from the model library based on the current working condition, alleviating the burden of unnecessary retraining. Finally, the effectiveness of ISDMM and its cloud-edge collaborative framework is validated through a numerical example and a real-world industrial process.


Multimodel Self-Learning Predictive Control Method With Industrial Application

November 2024

·

44 Reads

IEEE Transactions on Industrial Electronics

In industrial sites, system operation conditions fluctuate due to changes in raw material and equipment status, making it critical to identify the operation conditions and obtain appropriate controllers accurately. Additionally, even for a specific operation condition, fixed control strategies may result in mismatches due to varying operational stages. To address the accurate control of industrial processes across multiple operation conditions, this article proposes a multimodel self-learning predictive control (MSLPC) method to simultaneously improve the accuracy of offline condition partition and online control performance. Specifically, in the offline stage, for complex and multidimensional industrial data, condition indicators are selected based on expert systems and data analytics, and a “presetting precise-fusion” two-stage operation condition learning (TSOCL) algorithm is proposed to accurately identify the operation conditions of the system. In the online stage, a self-learning predictive control algorithm is proposed, which improves adaptability and control performance by adjusting controllers. This maintains a high match between the control strategy and system state. Simulation experiments demonstrate that the MSLPC method achieves higher control accuracy and faster control rate in the presence of varying operation conditions. Finally, the proposed method is deployed in a real industrial roaster to validate its effectiveness and excellent control performance.




One Network Fits All: A Self-Organizing Fuzzy Neural Network Based Explicit Predictive Control Method for Multimode Process

September 2024

·

25 Reads

·

3 Citations

IEEE Transactions on Fuzzy Systems

Modern industrial processes often exhibit complex and uncertain operating state fluctuations due to the diversification of production materials, the complexity of production processes, and the harsh production environment. To address the real-time control challenge in multimode processes, this article proposes a learning framework for fuzzy neural network explicit control for full operation conditions. It includes: the fuzzy neural network-based explicit control law (FNNECL) and the neural network-based predictive model (NNPM). By adjusting the structure and parameters of FNNECL, precise control of full operation conditions is achieved. Specifically, a novel truncated radial basis activation function is first presented, and the operation condition change measurement index of data coverage is proposed to identify the mismatch degree between current and historical operation conditions. Then, for small-scale operation condition changes, an elastic weight consolidation (EWC) mechanism is introduced to ensure that FNNECL can learn control strategies for new operation conditions while maintaining control performance for historical conditions. Finally, to address large-scale operation condition changes, a radial basis function (RBF) neuron growth mechanism based on data coverage is proposed. This mechanism calculates the data coverage of fuzzy rules and selectively adds new neurons to the FNNECL. This enables the FNNECL to fit large-scale changes and effectively learn control strategies for new operation conditions. It is worth noting that without knowing the current condition, accurate control sequences can be rapidly obtained based on a single FNNECL. Extensive experiments including numerical simulation and industrial roasting process verified the feasibility and effectiveness of the proposed method.





Citations (64)


... HE primary goal of violent surveillance is to monitor abnormal events in the real world to prevent violent behavior and maintain social order [1]. With the rapid advancement of artificial intelligence, researchers are exploring the use of deep learning technologies in surveillance systems to replace the inefficiencies and high costs associated with traditional manual monitoring [2] [3]. Weakly-supervised violence surveillance (WSVM) has emerged as an important research area in recent years. ...

Reference:

From Explicit Rules to Implicit Reasoning in an Interpretable Violence Monitoring System
Global Information-Based Lifelong Dictionary Learning for Multimode Process Monitoring
  • Citing Article
  • December 2024

IEEE Transactions on Systems Man and Cybernetics Systems

... I NDUSTRIAL defect detection aiming to identify and localize various defects is an indispensable link in industrial manufacturing chain [1], which guarantees the quality and safety of products and is used for fault analysis [2] and production optimization. Deep learning methods [3], [4] have played a pivot role in defect detection recently. ...

Process Manufacturing Intelligence Empowered by Industrial Metaverse: A Survey
  • Citing Article
  • July 2024

IEEE Transactions on Cybernetics

... Resilience in this system is achieved through automated inspection and remanufacturing, optimizing resource use and minimizing waste. Robust ML models, such as the Trusted Connection Dictionary Learning (TCDL) method proposed by Huang et al. [204], enhance fault detection and operational safety in IS by addressing label noise and ensuring reliable condition monitoring. Energy resilience is another critical area, as Wang et al. [205] demonstrated through a robust demand response (DR) framework for industrial microgrids, which enhances flexibility and reduces costs under fluctuating electricity prices. ...

Robust condition identification against label noise in industrial processes based on trusted connection dictionary learning
  • Citing Article
  • April 2024

Reliability Engineering & System Safety

... The problem of vulnerability analysis of novel NR IQA models to adversarial attacks was widely discussed in previous works: (Yang et al., 2024a), (Leonenkova et al., 2024), (Kashkarov et al., 2024), (Deng et al., 2024), (Konstantinov et al., 2024), (Yang et al., 2024b), (Zhang et al., 2024), (Ran et al., 2025), (Meftah et al., 2023), (Siniukov et al., 2023), (Shumitskaya et al., 2024b), (Shumitskaya et al., 2024a). Some works have been conducted as part of the MediaEval task: "Pixel Privacy: Quality Camouflage for Social Images" (MediaEval, 2020), where participants aimed to improve image quality while reducing the predicted quality score. ...

Sparse Adversarial Video Attack Based on Dual-Branch Neural Network on Industrial Artificial Intelligence of Things
  • Citing Article
  • Full-text available
  • July 2024

IEEE Transactions on Industrial Informatics

... Currently, building a DT model requires a large amount of data as an input, and ensuring the accuracy and completeness of the data is the challenge faced by DTs. The DT, serving as a synchronized replica of the real equipment in the digital space [14], with the development of modeling, sensing, data analysis, data mining algorithms, and cyber physical systems (CPS) technology [15], extensive research has been conducted on its fundamental theories, DT consistency, and DT applications. DT thoroughly depicts the high-dimensional operational states of equipment. ...

Digital twin driven soft sensing for key variables in zinc rotary kiln
  • Citing Article
  • April 2024

IEEE Transactions on Industrial Informatics

... Numerous studies have focused on the development of mathematical models to represent and analyze manufacturing and engineering processes. For example, studies [13][14][15][16][17][18][19] explore the use of modeling techniques such as IDEF and UML to describe and optimize industrial processes. Other relevant works in this context include those by [20][21][22], who investigate the influence of additive manufacturing technologies on the machinability of titanium parts, and [23][24][25], who propose a hybrid approach for multi-response optimization of micro-milling parameters. ...

Variable partition based parallel dictionary learning for linearity and nonlinearity coexisting dynamic process monitoring
  • Citing Article
  • January 2024

Control Engineering Practice

... For example, laser synchronization has been a focus of physics research [6], while biology explores synchronous phenomena ranging from circadian rhythms to collective behaviors [7]. In engineering, synchronization plays a pivotal role in communication and power systems, with an increasing emphasis on synchronization control and prediction issues [8][9][10]. Moreover, synchronization phenomena have also been extensively studied in economics [11], chemistry [12], and sociology [13]. ...

Knowledge-Informed Neural Network for Nonlinear Model Predictive Control With Industrial Applications
  • Citing Article
  • January 2023

IEEE Transactions on Systems Man and Cybernetics Systems

... For example, laser synchronization has been a focus of physics research [6], while biology explores synchronous phenomena ranging from circadian rhythms to collective behaviors [7]. In engineering, synchronization plays a pivotal role in communication and power systems, with an increasing emphasis on synchronization control and prediction issues [8,9,10]. Moreover, synchronization phenomena have also been extensively studied in economics [11], chemistry [12], and sociology [13]. ...

EaLDL: Element-Aware Lifelong Dictionary Learning for Multimode Process Monitoring
  • Citing Article
  • December 2023

IEEE Transactions on Neural Networks and Learning Systems

... Consequently, interest in employing ANNs for AEFP modeling has grown. Yang et al. 10 developed a convolutional neural network (CNN)-based model for superheat detection in electrolyzers. Similarly, Lundby et al. 11 proposed a sparse neural network to model the aluminum electrolysis process, while Yue et al. 12 utilized a CNN to identify the occurrence of the anode effect in aluminum electrolysis. ...

Physical-Knowledge Embedded Convolutional Neural Network for Aluminum Electrolysis Superheat Degree Identification
  • Citing Article
  • January 2023

IEEE Transactions on Industrial Electronics

... Although the mentioned method above provide functional and efficient control performance, there is still a possibility that instability problem will emerge in practical system since excitation of the unmodeled, high-frequency vibration modes, which is called as spillover effects [11]. These controllers derived on truncated system model is conservative to system critical modes and controller with high accuracy is impractical to implement [12][13][14]. ...

Physical Informed Sparse Learning for Robust Modeling of Distributed Parameter System and Its Industrial Applications
  • Citing Article
  • January 2023

IEEE Transactions on Automation Science and Engineering