Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Lu Dong has a total of 85 co-authors. Here are the top co-authors by number of shared publications:

  • Hongyun Hu

    State Key Laboratory of Coal Combustion, Huazhong University of Science and Technology, Wuhan 430074, China

    6 shared publications
  • Chengyang Cao

    School of Resource & Safety Engineering, Wuhan Institute of Technology, Wuhan 430074, China

    2 shared publications
  • Hong Yao

    State Key Laboratory of Coal Combustion, School of Energy and Power Engineering, Huazhong University of Science and Technology, Wuhan, China

    2 shared publications
  • Hongyun Hu

    State Key Laboratory of Coal Combustion, School of Energy and Power Engineering, Huazhong University of Science and Technology, Wuhan, China

    2 shared publications
  • Jiewen Deng

    Hubei Key Laboratory of Multi-media Pollution Cooperative Control in Yangtze Basin, School of Environmental Science and Engineering, Huazhong University of Science and Technology 2 , Wuhan 430074

    2 shared publications
  • Liang Wang

    China Power Hua Chuang (Suzhou) Electricity Technology Research Company Co., Ltd., Suzhou 215125, China

    2 shared publications

Latest Publications

  • Open Access

    Hierarchical VMD parameter selection for short-term HVAC load forecasting: A compact TCN-LSTM implementation

    by Liang Yin, YouSong Zhang, Ruiqi Zhang, Cheng Xu, Jiewen Deng, Weiqi Zhu, Lu Dong

    Variational mode decomposition (VMD) separates heating, ventilation, and air-conditioning (HVAC) load into narrower-band signals, but the number of modes and the bandwidth-penalty parameter must be selected without using test-set outcomes. We evaluated a diagnostic-guided selection procedure using a multichannel temporal convolutional network (TCN) with long short-term memory (LSTM). The dataset comprised 2208 hourly observations from Wuhan Cigarette Factory in summer 2024. For each target, VMD used the preceding 168 h; the final 24 values from each mode served as TCN-LSTM input. Training-only diagnostics screened 64 parameter pairs using range-normalized reconstruction error, the coefficient of variation in mode-specific sample entropy, and the first-mode energy ratio. Sixteen pairs met the criteria. The eight lowest-scoring pairs formed a diagnostic shortlist, which was evaluated with seed 42 and then across seeds 42–44. The selected pair, K=7 and α=1500, yielded validation and test root-mean-square errors (RMSEs) of 2.113±0.079 and 2.253±0.161 kW across seeds 42–44 and 42–46, respectively. On the test segment, the reference 5,2000 and minimum-J admissible 8,1000 pairs yielded RMSEs of 2.243±0.094 and 2.285±0.161 kW, respectively. VMD-transformer with bidirectional LSTM (VMD-transformer-BiLSTM) and VMD-TCN-LSTM yielded 2.185±0.062 and 2.253±0.161 kW, respectively; Diebold–Mariano tests showed no consistent squared-loss difference across seeds. VMD-TCN-LSTM used 14 961 parameters, 73.8% fewer than VMD-transformer-BiLSTM. Excluding preprocessing and data handling, its single-sample network forward-pass time was 0.285±0.014 ms. Across 231 diagnostic-weight combinations, five pairs were selected, with seed-42 validation RMSE spanning 2.038–2.063 kW. Therefore, the procedure fixed VMD parameters without using test targets but did not establish an RMSE advantage over the reference pair.

  • Open Access

    Short-term photovoltaic power forecasting under issue-time constraints using validation-stabilized dynamic expert fusion

    by Lu Dong, Weiqi Zhu, Zhiyue Ding, Lingang Jiang, Yiyao Jin, Kaituo Jiao, Yudong Qiao, Jiewen Deng

    Operational short-term photovoltaic (PV) forecasting requires a strict separation between information available when a forecast is issued and measurements observed only at the target time. This study evaluates validation-stabilized DynamicExpertFusion at an industrial 20 MW PV plant and tests the issue-time protocol on public data. The industrial models use causal histories of power, weather, and inverter state together with deterministic calendar variables. Validation reliability and issue-time predictions of expert error produce sample-specific, shrinkage-controlled weights. Training, validation, and testing use May-November 2020, December 2020-January 2021, and February-May 2021, respectively, at 15, 30, and 60 min horizons. DynamicExpertFusion obtains active-period RMSE values of 1238.739 ± 2.513, 1611.893 ± 7.205, and 2018.581 ± 8.018 kW. On matched 96-step windows, it has lower RMSE than DLinear and the tested Hugging Face PatchTST implementation. Retrospective high-power reductions relative to fixed weighting are 25.487-75.905 kW. An exploratory issue-time policy applies DynamicExpertFusion when [[EQUATION]] and otherwise uses fixed weighting. It reduces active-period RMSE relative to fixed weighting by 9.887 and 16.523 kW at 15 and 30 min (95% confidence intervals [4.042, 16.179] and [7.952, 25.091] kW); the 60 min interval crosses zero, and comparisons with always-on DynamicExpertFusion are inconclusive. The policy requires independent prospective confirmation. NREL PVDAQ and HKUST leave-one-station-out tests show mixed transfer performance. The evidence supports plant-calibrated, state-conditional reliability allocation, not universal model superiority or realized economic benefit.

  • Open Access

    Co-Pyrolysis of Waste Tires and Beech Sawdust: Comprehensive Analysis of Thermal Behavior, Synergistic Effect, and Interaction Mechanisms

    by Guangyao Zheng, Chengyang Cao, Qiming Zhang, Pei Jia, Lu Dong, Hongyun Hu

    Against the backdrop of the global search for alternatives to fossil fuels, waste tires have attracted attention as a significant resource due to their enormous production volume and considerable energy potential. However, the application of tar derived from waste tires alone is limited by its poor stability and other deficiencies. This study systematically investigates the co-pyrolysis behavior and synergistic mechanisms of waste tires and beech sawdust at various blending ratios. Thermogravimetric analysis indicates that the addition of beech sawdust reduces the decomposition temperature of the blend and induces a synergistic effect that promotes waste tire pyrolysis within the temperature range of 384–440 °C. Pyrolysis experiments results show that tar yield of the blends reached 64.45 wt.%, while the char yield decreased from 40.67 wt.% to 24.83 wt.%. Also, the presence of beech sawdust synergistically enhanced the formation of aromatic hydrocarbons in the tar of waste tires, with the total yield of aromatics increasing synergistically by up to 54.8%. Specifically, the yields of stable alkylbenzenes such as toluene and xylene were consistently promoted, whereas the yields of unsaturated aromatics such as allylbenzene and 2,4-dimethylstyrene were enhanced at low beech sawdust ratios but suppressed at higher ratios. Based on these findings, the interaction mechanisms underlying the co-pyrolysis process were elucidated, providing theoretical guidance for the high-value utilization of waste tires.

  • Open Access

    One-Step Synthesis of Bamboo Biochar for Efficiency Adsorption of Tetracycline: Characterization, Kinetics and Cost–Benefit Analysis

    by Qi Liao, Chengyang Cao, Qiming Zhang, Pei Jia, Lu Dong

    Tetracycline has been widely used as an efficient broad-spectrum antibiotic, while its long-term environmental pollution characteristics have gradually gained recognition and attention, highlighting the urgent need to identify a low-cost and effective method for removing tetracycline pollutants. This study aims to develop a one-step bamboo-based biochar preparation method based on a KCl-ZnCl2 molten salt system; the potential application of obtained bamboo-based biochar as a tetracycline adsorbent was characterized and analyzed. Results show that the biochar prepared at 900 °C possesses abundant microporous and mesoporous structures, with abundant surface functional groups. Also, it exhibits a composite type I/IV isotherm, with a specific surface area of 1305.91 m2·g−1, a total pore volume of 0.944 cm3·g−1, demonstrating excellent tetracycline adsorption capacity of 298.93 mg·g−1. XRD analysis confirmed that increasing the activation temperature significantly enhanced the graphitization degree of the biochar, which is a key factor influencing its tetracycline adsorption performance. Kinetic studies indicated that the adsorption kinetic process was better described by the Elovich model and Freundlich isotherm. Furthermore, cost-effectiveness analysis revealed that the cyclic preparation cost of biochar via this technique could be reduced to 18.25 USD per kilogram owing to the low consumption characteristics of the KCl-ZnCl2 molten salt, which represents a 93.4% reduction compared with conventional preparation methods, underscoring the economic applicability of this technology in the field of tetracycline removal. The findings of this study are expected to lay a foundation for the industrial preparation of low-cost, high-performance bamboo-based biochar for tetracycline removal.


  • Open Access

    QPSO-optimized VMD-SE-transformer-BiLSTM for short-term air conditioning load forecasting in industrial buildings

    by Yingfeng Wei, Hui Chen, Ye Tian, Chilong Wang, Yulei Hu, Zhiheng Xie, Yuchan Du, Jianping Yang, Jiaqi Qian, Huimin Liu, et al.

    The air-conditioning loads of industrial buildings exhibit considerable nonlinearity, nonstationarity, and multi-scale variability due to the interaction between production processes and external weather conditions. This study proposes a three-stage hybrid forecasting framework that integrates Variational Mode Decomposition and Sample Entropy (VMD–SE) with a transformer–Bidirectional Long Short-Term Memory (BiLSTM) network architecture, optimized using Quantum Particle Swarm Optimization (QPSO). The VMD–SE module reconstructs the load signal by entropy-based component screening, significantly reducing high-frequency noise while maintaining essential dynamic features. The dual-path transformer–BiLSTM network concurrently captures long-term global dependencies and short-term local changes, facilitating an extensive temporal representation. By automatically adjusting its core parameters, QPSO achieves more efficient convergence, increased stability, and enhanced global search. Two real-world datasets obtained from industrial facilities in Hubei, China, reflecting distinct seasons and load levels, were utilized for validation using ablation and comparative experiments. The proposed model outperformed benchmark models such as SVR, GRU, and CNN–informer, achieving a coefficient of determination (R2) of 0.981 23 and a root mean square error of 0.181 32. The findings indicate that the proposed model is effective and versatile across various situations, offering both theoretical understanding and practical guidance for intelligent HVAC scheduling, dynamic demand response, and low-carbon energy management in industrial facilities. Our results show that the model is robust and flexible, providing a practical and theoretically grounded approach for intelligent HVAC scheduling, demand response, and industrial carbon reduction.


  • Open Access

    A Review of Machine Learning-Based Image Segmentation and Recognition Techniques with Applications in the Tobacco Industry

    by Jian Wang, Linlin Li, Youzhi Shi, Zaiyu Tan, Meng Sun, Yang Chen, Yong Pan, Hongyun Hu, Lu Dong, Qin Zhang


  • Application of machine learning in prediction of Pb2+ adsorption of biochar prepared by tube furnace and fluidized bed

    by Wei Huang, Liang Wang, JingJing Zhu, Lu Dong, Hongyun Hu, Hong Yao, LinLing Wang, Zhong Lin

    Data mining by machine learning (ML) has recently come into application in heavy metals purification from wastewater, especially in exploring lead removal


  • Open Access

    The Effect of Nitrogen Functional Groups on Pb0, PbO, and PbCl2 Adsorption over a Carbonaceous Surface

    by Liang Wang, Huaizhou Wen, Lei Guo, Ancheng Liang, Tingan Liu, Dongxu Zhao, Lu Dong

    Lead (Pb) pollution, especially from the incineration of municipal solid waste (MSW), poses a significant threat to the environment. Among all the effective methods, activated carbon (AC) injection serves as an effective approach for lead removal from flue gas, while the modification of ACs emerges as a crucial pathway for enhancing Pb adsorption capacities. Density functional theory (DFT) is employed in this study to investigate the mechanisms underlying the enhanced adsorption of Pb species (Pb0, PbO, and PbCl2) on nitrogen-functionalized carbonaceous surfaces. The results show that nitrogen-containing groups substantially enhance lead adsorption capacity, with adsorption energies ranging from −526.18 to −288.31 kJ/mol on nitrogen-decorated carbonaceous surfaces, much higher than those on unmodified surfaces (−310.35 to −260.96 kJ/mol). Additionally, electrostatic potential and density-of-states analyses evidence that pyridinic nitrogen atoms remarkably expand charge distribution and strengthen orbital hybridization, thereby augmenting lead capture. This research elucidates the role of nitrogen-containing functional groups in lead adsorption, offering valuable insights for the development of highly efficient biomass-derived activated carbon sorbents for lead removal.


  • The migration and transformation mechanisms of heavy metals during molten salt cyclic thermal treatment of MSWI fly ash

    by Yicheng Wang, Lu Dong, Hongyun Hu, Dahai Yan, Sihua Xu, Chan Zou, Yongda Huang, Guangzhao Guo, Hong Yao