Advanced Engineering Informatics最新文献

{"title":"Generalized envelope nonlinear Gini index-gram guided two-stage chirp mode decomposition for shield machine main bearing fault diagnosis","authors":"Gang Shi,&nbsp;Chengjin Qin,&nbsp;Pengcheng Xia,&nbsp;Zhinan Zhang,&nbsp;Chengliang Liu","doi":"10.1016/j.aei.2026.104354","DOIUrl":"10.1016/j.aei.2026.104354","url":null,"abstract":"<div><div>The main bearing is the core mechanical equipment of shield machine, known as the heart of the shield machine, mainly applied to support the rotation of cutterhead system. The health status of main bearing has significant impact on the stable and safe construction for shield machine. However, main bearing has very low rotational speed of only 1 rpm to 5 rpm, and the structure is very complex, with a diameter even exceeding 10 m. The fault features for main bearing vibration signals are relatively weak. Therefore, the main bearing fault diagnosis is a challenging task. To tackle the problem, we develop a novel generalized envelope nonlinear Gini index-gram guided two-stage chirp mode decomposition (GENGI-TSCMD) for fault diagnosis of shield machine main bearing in this article. The proposed method consists of optimized demodulated frequency band (ODFB) selection and fault frequency extraction two key parts. We firstly developed GENGI-gram for ODFB extraction, which adopting stepwise increasing segmentation and signal frequency spectrum trend to divide the signal’s frequency bands, and obtain two grams. A new GENGI fault pulse identifier was constructed to select ODFB in these two grams. GENGI can availably characterize fault pulse features and suppress noise interference by combining generalized envelope and nonlinear weights based on Gini index. For fault frequency extraction, we proposed TSCMD by establishing bandwidth guided adaptive chirp mode decomposition (BACMD) and fault frequency extraction mode decomposition (FEMD). BACMD is adopted to extract sub-signals in each demodulation frequency band of GENGI-gram. FEMD is applied to decompose each order fault frequency sub-signals of ODFB signal’s envelope. In this way, the main bearing’s fault type can be precisely diagnosed. BACMD derives the bandwidth calculation formula for sub-signal, which can extract sub-signals with specific frequency bandwidth. FEMD constructs a novel adaptive signal decomposition model, which can effectively decompose each order fault frequency sub-signal with narrow frequency band. Then we use actual main bearing fault vibration signals to verify the proposed method. The experimental results show that proposed GENGI-TSCMD can availably filter out various types of strong noise disturbance, and precisely extract each order fault frequency component. Moreover, proposed method has superior performance than current signal processing fault diagnosis methods.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104354"},"PeriodicalIF":9.9,"publicationDate":"2026-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023135","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Implicit coordination through environment modification: Multi-agent RL approach for adaptive door placement in evacuation scenarios","authors":"Isabelle Fitkau,&nbsp;Timo Hartmann","doi":"10.1016/j.aei.2026.104359","DOIUrl":"10.1016/j.aei.2026.104359","url":null,"abstract":"<div><div>In evacuation modeling, architectural configurations are typically treated as fixed constraints, limiting the exploration of adaptive design strategies that respond dynamically to occupants’ movement patterns. This research investigates environment-mediated coordination in a prototypical multi-agent reinforcement learning (MARL) setting. Using Proximal Policy Optimization (PPO), a navigating agent learns escape trajectories, while a design agent modifies door placement solely based on observed movement, without explicit communication. The system was evaluated in a multi-room environment to explore how complementary strategies evolve when navigation performance and immediate floor plan changes are coupled. The results demonstrate successful implicit coordination, with substantial reductions in episode length as the door controller agent progressively shifted door placements towards positions that facilitated faster escape paths. Navigation trajectories converged on paths that maximized the reward of the navigating agent, while door placement strategies minimized evacuation times. Therefore, although allowing architectural changes during episode runtime, the results still show emerging adaptive behaviors of both agents and demonstrate the feasibility of coordinating navigation and design actions through dynamic environment modification. This approach paves the way for adaptive architectural design systems that can dynamically respond to user behavior patterns, with potential applications in performance-based building design tools and in emergency planning optimization.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104359"},"PeriodicalIF":9.9,"publicationDate":"2026-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146078381","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"DuoTransFormer for nonlinear seismic response prediction: Cover large, focus local, and enforce law","authors":"Xinyi Hu ,&nbsp;Congzhen Xiao ,&nbsp;Zhiqiang Zhang ,&nbsp;Jie Gao","doi":"10.1016/j.aei.2026.104375","DOIUrl":"10.1016/j.aei.2026.104375","url":null,"abstract":"<div><div>Deep learning-based methods have recently been applied to structural seismic response prediction, yet existing models still face challenges in capturing strongly nonlinear behaviors and exhibit limitations in multi-scale temporal modeling and inter-variable dependencies. Based on these observations, we propose <span><math><mi>DuoTransformer</mi></math></span>, a novel dual-view Transformer backbone that integrates a <span><math><mi>Trans</mi></math></span>posed Trans<span><math><mi>former</mi></math></span> Encoder treating each variable as a token and a Patch <span><math><mi>Transformer</mi></math></span> Encoder for local temporal patterns, thereby enabling multi-scale modeling of both coarse-grained and fine-grained temporal dependencies. Inspired by response characteristics such as period elongation and stiffness degradation, the Physical Module is designed following a phenomenon-driven design paradigm. Furthermore, DuoTransFormer decouples the prediction process into two stages: the Physical Module, which hard-codes structural dynamics and outputs three parallel data chains as physics-informed feature sequences, and the Transformer Module, which takes them as inputs. Two proof-of-concept experiments are first conducted to validate the effectiveness of the proposed method. Subsequently, extensive evaluations on two open-access structural models, including a 42-story case study building and a 632-m real-world building, demonstrate that DuoTransFormer outperforms state-of-the-art time-series forecasting models (e.g., PatchTST, iTransformer) while achieving substantially lower computational cost than finite element simulations. The model-agnostic Physical Module can be flexibly combined with all baselines to deliver substantial performance gains.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104375"},"PeriodicalIF":9.9,"publicationDate":"2026-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023133","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Data-driven modelling of unloading hours using explainable gradient boosting models","authors":"Celal Cakiroglu ,&nbsp;Najat Almasarwah ,&nbsp;Mehmet Hakan Özdemir ,&nbsp;Batin Latif Aylak ,&nbsp;Manjeet Singh ,&nbsp;Muhammet Deveci","doi":"10.1016/j.aei.2026.104353","DOIUrl":"10.1016/j.aei.2026.104353","url":null,"abstract":"<div><div>Unloading processes denote the extraction of finished goods and raw materials from transport units and their subsequent conveyance to designated locations. The efficiency of unloading processes is vital in supply chain and logistics management, regarded as an essential component. Delays in unloading operations result in numerous challenges, including heightened operational expenses, diminished labour efficiency, and supply chain bottlenecks. Consequently, it is essential to ascertain unloading times beforehand to mitigate these challenges, resulting in diminished idle time, enhanced overall efficiency, and optimized scheduling. Therefore, precise prediction of unloading times is critically significant. The novelty of this study lies in the application of machine learning techniques to improve operational efficiency by accurately predicting unloading time. To that end, this study employed LightGBM and XGBoost to predict the unloading time in a real case. The unloading time can be predicted with <span><math><msup><mrow><mi>R</mi></mrow><mo>2</mo></msup></math></span> score greater than 0.99 utilizing both models. Subsequently, the SHapley Additive exPlanations (SHAP) methodology was used to ascertain how each input feature contributed to the model’s output. The load of leg significantly influences the unloading time more than the gross weight of truck and the leg distance.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104353"},"PeriodicalIF":9.9,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023023","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"RLLM-SS: A knowledge-guided simplex search method integrating large language model and reinforcement learning for injection molding quality control","authors":"Haipeng Zou ,&nbsp;Xinyu Li ,&nbsp;Yongkuan Yang ,&nbsp;Ke Yao ,&nbsp;Xiangsong Kong ,&nbsp;Zhijiang Shao ,&nbsp;Furong Gao","doi":"10.1016/j.aei.2026.104372","DOIUrl":"10.1016/j.aei.2026.104372","url":null,"abstract":"<div><div>In injection molding (IM), product quality and process stability are highly dependent on the setting of key process parameters, making efficient parameter tuning essential for achieving reliable and consistent production. However, the tuning process is traditionally guided by expert experience and trial-and-error methods, which often lead to low efficiency and prolonged adjustment cycles. To address this challenge, we propose a knowledge-guided simplex search method that integrates a large language model (LLM) with the Soft Actor–Critic (SAC) reinforcement learning algorithm in a collaborative optimization framework, called RLLM-SS. In RLLM-SS, a quasi-gradient mechanism leverages historical data to dynamically estimate the step size and gradient compensation direction of the simplex search method. These estimated variables, integrated with domain knowledge, are encoded into structured prompts that guide the injection molding quality LLM in dynamically adjusting simplex coefficients through natural language reasoning. This enables the simplex search to overcome fixed-coefficient limitation and avoid local optima to the maximum extent. To mitigate the drawbacks of the LLM, such as its tendency to generate hallucinated outputs and lack of memory of past tuning adjustments, a SAC-based evaluation module is introduced. It assigns rewards based on optimization performance, thereby reinforcing effective strategies and fostering continuous policy improvement when similar conditions recur. Experimental evaluations first verified LLM-SS on standard high-dimensional benchmark functions, confirming its effectiveness in complex search spaces, and were then conducted on an injection molding quality simulation platform built on a neural network trained with practical IM process data. Results show that RLLM-SS outperforms several advanced methods, reducing the average number of iterations by 27.6% and the final Euclidean distance to the target quality curve by 68.3%. It also maintains strong robustness under Gaussian noise perturbations.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104372"},"PeriodicalIF":9.9,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023024","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An adaptive spatial–temporal encoder with gated multi-convolutions for remaining useful life prediction","authors":"Wen Liu ,&nbsp;Jyun-You Chiang ,&nbsp;Yi Li ,&nbsp;Haobo Zhang","doi":"10.1016/j.aei.2026.104369","DOIUrl":"10.1016/j.aei.2026.104369","url":null,"abstract":"<div><div>Accurate estimation of Remaining Useful Life (RUL) is essential for safe and economical operation of turbofan engines. This paper introduces an Adaptive Spatial-Temporal Encoder with Gated Multi-Convolutions (GMC-ASTE), a novel approach that simultaneously models temporal dynamics and inter-sensor relationships to enhance RUL prediction accuracy. The methodology employs a multi-scale gated convolution module to extract refined features from raw multi-sensor data, effectively reducing noise while retaining transient signals. These features are subsequently processed through an adaptive spatial–temporal encoder, which utilizes graph attention mechanisms to dynamically adjust sensor connections and multi-head temporal attention to capture long-term dependencies parallelly. Extensive experiments on the Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) benchmark demonstrate that GMC-ASTE achieves superior performance, with the lowest Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Scoring metrics across all four sub-datasets. The results confirm the effectiveness and interpretability of the proposed model, providing an advanced framework that improves engine prognostic theory and offering airlines a practical tool to reduce downtime and maintenance costs.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104369"},"PeriodicalIF":9.9,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023134","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Building façade delamination quantification framework based on infrared instance segmentation and dual-modality vision calibration","authors":"Ziyu Wang ,&nbsp;Zhenfen Jin ,&nbsp;Xiaolan Zhuo ,&nbsp;Jiangpeng Shu ,&nbsp;Ziyue Zeng ,&nbsp;Rongrong Wei ,&nbsp;Lijun Ye","doi":"10.1016/j.aei.2026.104341","DOIUrl":"10.1016/j.aei.2026.104341","url":null,"abstract":"<div><div>Building façade delamination quantification is critical for damage severity assessment and maintenance. While infrared thermography combined with deep learning reduces manual inspection costs, it fails to achieve precise detection and quantification results due to noise and insufficient spatial details in infrared images. To address these issues, this paper proposes a comprehensive framework. A multi-stage iterative infrared delamination network (IR-DNet) centered on dual-weighted features is designed for delamination segmentation. To tackle infrared data scarcity and real data collection costs, IR-DNet is trained on laboratory and simulated data while ensuring generalization to field scenarios. Additionally, a dual-modality vision calibration method is developed: after image registration, visible images serve as complementary information to correct perspective distortion and calibrate the scale factor, with the resulting parameters shared between modalities, thus creating an information fusion pipeline. The corrected infrared images are then fed into IR-DNet for segmentation and quantification. Field test yields an average quantification error of 6.02%, verifying the practical value for damage severity judgment and maintenance decisions.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104341"},"PeriodicalIF":9.9,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023032","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Updating digital twin in supervisory control and data acquisition for sustainable laser beam micro machining","authors":"K. Venkata Rao ,&nbsp;K. Venkata Vaneesha","doi":"10.1016/j.aei.2026.104355","DOIUrl":"10.1016/j.aei.2026.104355","url":null,"abstract":"<div><div>In the Industry 4.0 era, improving energy efficiency and product quality is a key challenge for the manufacturing sector, with the aim of reducing both energy consumption and surface roughness. The present study proposed a supervisory control and data acquisition system to address these issues in laser beam machining of AISI 316 L stainless steel. A digital twin is integrated with supervisory control and data acquisition interface, forming a closed-loop connection between the physical system and its digital counterpart. The digital twin is a combination of a mathematical model and an artificial neural network.<!--> <!-->The mathematical models estimate the surface roughness and specific machining energy using the experimental data collected from the physical system. When the estimated surface roughness and the specific machining energy exceeded their target values, the digital twin optimizes and adjusts the input parameters such as laser power, laser frequency, and cutting speed. The process parameters were also optimized using the Taguchi method and compared with supervisory control and data acquisition. The supervisory control and data acquisition system showed superior performance over the Taguchi method, achieving 47.44 %, 54.67 %, and 36.99 % reductions in surface roughness and 34.92 %, 22.61 %, and 83.66 % reductions in specific machining energy for 4, 6, and 8 mm thick sheets, respectively. The artificial neural network within the digital twin is continuously retrained using rich, real-time vibration signals, enabling instantaneous adaptation to the dynamic behaviour of the process, resulting in reduced surface roughness and specific machining energy, and enhanced overall sustainability.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104355"},"PeriodicalIF":9.9,"publicationDate":"2026-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023136","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A general and efficient approach for uncertainty quantification in neural networks: Identifying risky decisions in AI systems","authors":"Zhao Zhang,&nbsp;Senlin Luo,&nbsp;Xiaolong Wu,&nbsp;Xikai Gao,&nbsp;Jiawei Pi,&nbsp;Limin Pan","doi":"10.1016/j.aei.2026.104343","DOIUrl":"10.1016/j.aei.2026.104343","url":null,"abstract":"<div><div>Uncertainty quantification in neural networks enables the assessment of predictive reliability in artificial intelligence systems, thereby reducing the risk of unsafe decisions. Existing approaches rely heavily on ensemble construction to sample the model parameter space and capture decision variability. However, under realistic resource constraints, small-scale sampling leads to insufficient evidence sources and inaccurate uncertainty estimates. In addition, the design of uncertainty metrics significantly influences estimation accuracy and may limit applicability across different types of machine learning (ML) tasks. In this paper, a Systematic Reusable Ensemble (SRE) framework is proposed for uncertainty quantification. The approach reuses and shares neural network components during retraining to efficiently generate multiple model instances within a single training process. Furthermore, a compounded ensemble pruning strategy is introduced to promote more uniform sampling in parameter space. A general fusion metric is then developed based on evidence theory with a redesigned trust allocation mechanism. Experimental results demonstrate that the proposed framework systematically reduces ensemble construction overhead while improving the reliability of uncertainty estimation. The generalization capability of the SRE is further validated through its effectiveness in identifying high-risk decisions across at least five categories of ML tasks.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104343"},"PeriodicalIF":9.9,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023035","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A two-phase cost sensitive-based domain adversarial neural network for anomaly detection in mass customized production","authors":"Yuan Gao ,&nbsp;Thong Ngee Goh ,&nbsp;Qingan Cui ,&nbsp;Qing Zhang ,&nbsp;Zhen He","doi":"10.1016/j.aei.2026.104371","DOIUrl":"10.1016/j.aei.2026.104371","url":null,"abstract":"<div><div>In the context of globalization and intense market competition, detecting anomalies in products is vital for enterprises to improve quality, enhance efficiency, and reduce costs. Enterprises are also pursuing mass customization to meet personalized customer needs. However, customized products are typically produced in small batches. Furthermore, modern manufacturing processes typically exhibit low defective rates. Additionally, quality results are only available after the inspections. These characteristics lead to the challenges of small sample sizes, imbalanced data, and delayed label acquisition for production data, potentially undermining the effectiveness of existing anomaly detection models. To cope with these challenges, this study proposes a two-phase cost sensitive domain adversarial neural network framework that leverages transfer learning to apply knowledge from similar mass-standardized products to customized products. When process parameters are available but inspection results are not, an unsupervised domain adversarial neural network integrated with cost-sensitive learning is employed to address the issues of small sample sizes and unlabeled data. Within this network, the cost-sensitive learning component simultaneously tackles data imbalance by biasing the label predictor towards the minority class through higher loss assignment compared to the majority class. Once inspection results become available for some customized products, the label predictor loss for these labeled samples is incorporated to further enhance anomaly detection. The experimental results demonstrate that the proposed method outperforms other state-of-the-art anomaly detection methods in practical smartphone speaker and laptop solid-state drive manufacturing processes.</div></div>","PeriodicalId":50941,"journal":{"name":"Advanced Engineering Informatics","volume":"71 ","pages":"Article 104371"},"PeriodicalIF":9.9,"publicationDate":"2026-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023033","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}