Analysis and prediction of diaphragm wall deflection induced by deep braced excavations using finite element method and artificial neural network optimized by metaheuristic algorithms

Reliability Engineering & System Safety, Год журнала: 2022, Номер 221, С. 108335 - 108335

Опубликована: Янв. 14, 2022

Язык: Английский

---

Application of SVR models built with AOA and Chaos mapping for predicting tunnel crown displacement induced by blasting excavation

Applied Soft Computing, Год журнала: 2023, Номер 147, С. 110808 - 110808

Опубликована: Сен. 4, 2023

Язык: Английский

---

Prediction of Rock Strain Using Hybrid Approach of Ann and Optimization Algorithms

Geotechnical and Geological Engineering, Год журнала: 2022, Номер 40(9), С. 4617 - 4643

Опубликована: Май 27, 2022

Язык: Английский

---

Hybrid forecasting system considering the influence of seasonal factors under energy sustainable development goals

Measurement, Год журнала: 2023, Номер 211, С. 112607 - 112607

Опубликована: Фев. 17, 2023

Язык: Английский

---

Assessment of Machine Learning Models for the Prediction of Rate-Dependent Compressive Strength of Rocks

Minerals, Год журнала: 2022, Номер 12(6), С. 731 - 731

Опубликована: Июнь 8, 2022

The prediction of rate-dependent compressive strength rocks in dynamic compression experiments is still a notable challenge. Four machine learning models were introduced and employed on dataset 164 to achieve an accurate the rocks. Then, relative importance seven input features was analyzed. results showed that compared with extreme (ELM), random forest (RF), original support vector regression (SVR) models, correlation coefficient R2 hybrid model combines particle swarm optimization (PSO) algorithm SVR highest both training set test set, exceeding 0.98. PSO-SVR obtained higher accuracy smaller error than other three terms evaluation metrics, which possibility as tool. Additionally, besides static strength, stress rate most important influence factor rock among listed parameters. Moreover, strain has positive effect strength.

Язык: Английский

---

Prediction for segment strain and opening of underwater shield tunnel using deep learning method

Transportation Geotechnics, Год журнала: 2023, Номер 39, С. 100928 - 100928

Опубликована: Янв. 5, 2023

Язык: Английский

---

Multi-Channel Assessment Policies for Energy-Efficient Data Transmission in Wireless Underground Sensor Networks

Energies, Год журнала: 2023, Номер 16(5), С. 2285 - 2285

Опубликована: Фев. 27, 2023

Wireless Underground Sensor Networks (WUGSNs) transmit data collected from underground objects such as water substances, oil soil contents, and others. In addition, the sensor nodes to surface regarding irregularities, earthquake, landslides, military border surveillance, other issues. The channel difficulties of WUGSNs create uncertain communication barriers. Recent research works have proposed different types assessment techniques security approaches. Moreover, existing are inadequate learn real-time attributes in order build reactive transmission models. system implements Deep Learning-based Multi-Channel Learning Protection Model (DMCAP) using optimal set attribute classification techniques. model uses Ensemble Model, Multi-Layer Perceptron (EMLP) Classifiers, Nonlinear Channel Regression models Entropy Analysis Support Vector Machine (ENLSVM) for evaluating conditions. Additionally, Variable Generative Adversarial Network (VGAN) engine makes intrusion detection routines under distributed environment. According principles, WUGSN channels classified based on characteristics acoustic channels, ground station channels. On behaviors, EMLP ENLSVM operated extract Signal Noise Interference Ratio (SNIR) entropy distortions multiple Furthermore, nonlinear regression was trained understanding predicting link (channel behaviors). DMCAP has extreme difficulty finding differences impacts due issues malicious attacks. this regard, VGAN-Intrusion Detection System (VGAN-IDS) configured monitor instabilities against nodes. Thus, deeply analyzes multi-channel qualities improve throughput WUGSN. testbed created parameters (acoustic air) with network parameters; uncertainties considered failures, noise distortions, interference, node number retransmissions. Consequently, experimental results show that attains 10% 15% better performance than systems through throughput, minimum retransmission rate, delay, energy consumption rate. (SVM) Random Forest (RF)-based Classification (SMC), Optimal Energy-Efficient Transmission (OETN), channel-aware multi-path routing principles Reinforcement (CRLR) identified suitable experiments.

Язык: Английский

---

Prediction of surrounding rock grade for ultra-deep roadway engineering using BPNN combined with optimal algorithms

Engineering Computations, Год журнала: 2025, Номер unknown

Опубликована: Фев. 26, 2025

Purpose This paper attempts to combine the application of artificial intelligence in predicting and evaluating classification surrounding rock grades provides guidance for subsequent support design reinforcement operations. Design/methodology/approach discusses use BPNN as primary tool, combined with three swarm bionic optimization algorithms (GA, PSO, GWO), solve stability evaluation grade prediction ultra-deep roadway excavation. Findings Taking Great Wall ore group core Shanghaimiao mining area extension, optimal model is applied engineering. Prediction results show that performance models excellent. Research limitations/implications Due limitations geological conditions construction environment deep coal mines, period excavation too long, resulting less data collection. Practical implications The can provide method, scheme correction mine Social It (the premise stability), so ensure economic safety benefits enterprises. Originality/value neural network mechanics a site first time, which used direction evaluation. index input layer determined by combining “three high one disturbance” on-site situation, closer actual project. intelligent are selected optimize hyperparameters back propagation network, improve accuracy models. system constructed, northwest China, guiding dynamic adjustment

Язык: Английский

---

Advancements in Machine Learning‐Enhanced Green Wireless Sensor Networks: A Comprehensive Survey on Energy Efficiency, Network Performance, and Future Directions

Journal of Sensors, Год журнала: 2025, Номер 2025(1)

Опубликована: Янв. 1, 2025

Wireless sensor networks (WSNs) are a collection of nodes that collect data from the environment using wireless technology. WSNs have many applications in various domains, such as public utilities, industrial monitoring and control, defense military activities. However, limited energy, short network lifetime, high bandwidth requirements, low throughput (TP), unreliable connections. Green (GWSNs) approaches optimize energy consumption enhance sustainable networks. Despite these advancements, nonadaptability to dynamic conditions use static historical necessitates introducing machine learning (ML) techniques address challenges. GWSNs aim reduce environmental impact, while ML will improve processing performance. This paper surveys recent advances ML‐based GWSNs, covering different aspects structure, exchange, location information, quality service (QoS), multiple path support. We also present performance metrics, implementation issues, future trends GWSNs. The introduces new taxonomy categorizing based on architecture, sharing, data, multipath support, QoS. survey findings show can achieve up 50% savings, 30% TP improvement, 40% delay reduction (DR) compared conventional WSNs.

Язык: Английский

---

Application of Artificial Intelligence in Rock Tunnel Engineering: A Survey on Where and How

Expert Systems, Год журнала: 2025, Номер 42(7)

Опубликована: Май 23, 2025

ABSTRACT Rock tunnel engineering (RTE) plays a crucial role in modern infrastructure development. The development of artificial intelligence (AI) is able to drive transformative advances RTE. This review provides an in‐depth analysis the AI application Through comprehensive examination existing literature, we explore how technologies have revolutionised various aspects RTE, including construction methodology, rock parameter estimation, hazard disaster management during construction, and operation. In addition, provide study synergies between algorithms related open datasets. work also outlines promising future research directions for aiming inspire further advancements this emerging field. conclusion, underscores positive influence on emphasising its capacity elevate efficiency, accuracy, safety standards throughout phases projects. convergence with RTE holds immense promise advancing field ensuring success sustainability endeavours.

Язык: Английский

---