Search

This paper developed a novel online optimization deep learning framework for accurate estimation and control of tunnel boring machine's (TBM) attitude and position. The proposed spatio-temporal deep learning model integrates graph convolutional network (GCN) and long short-term memorial (LSTM) neural network which can provide an accurate estimation of TBM's trajectory deviations. This model is later utilized as the meta-model and integrated with an online non-dominated sorting genetic algorithm II (NSGA-II) method to optimize and minimize the trajectory deviations in real-time. A tunnel project from Singapore is utilized as a case study for demonstration purposes. TBM's key operational parameters, articulation displacements, and TBM's trajectory deviations are processed and utilized for real-time prediction. The results indicate that (1) The developed deep learning models can predict TBM's trajectory deviations with high accuracy. (2) TBM's trajectory deviations can be effectively reduced by the proposed optimization method. (3) The proposed spatio-temporal deep learning methods outperform the state-of-the-art methods. (4) The proposed online optimization method is more practical and creates better overall improvement than the conventional multi-objective optimization (MOO) approach. The novelty of this paper includes (a) developing a hybrid framework for accurate trajectory deviation estimation and TBM operating optimization, and (b) investigating the influence of time and distance lag between the TBM operation and position control by considering an online optimization. • A novel online multi-objective optimization (MOO) method is proposed. • Deep learning model is established for reliable estimation on TBM's trajectory deviations. • The developed model is utilized as meta-models and hybrids with online NSGA-II method. • A tunnel project from Singapore is utilized as a case study • TBM's overall trajectory deviations can be reduced by 46.4% by the proposed method. [ABSTRACT FROM AUTHOR]

Evacuation is critical for safety management due to the highly overcrowded passengers in the metro stations. A simulation-based approach integrating Random Forest (RF) and Non-dominated Sorting Genetic Algorithm III (NSGA-III) is proposed to perform the evacuation evaluation and optimization at metro stations. A 3D model of the metro station is built to simulate the dynamic process of passenger evacuation in metro stations. A framework consisting of 9 influential factors and 3 objectives is developed to model the input-output relationship in passenger evacuation. An RF-based meta-model is used to construct the relationship between influential factors and objectives. At last, NSGA-III is applied to seeking the optimal solutions for the station renovation in order to achieve a safe evacuation. A station model simulating a real metro station in Singapore is constructed to test the effectiveness and applicability of the proposed approach. It is found that (1) A safe evacuation could be achieved for the station, but along with the increasing passenger volume and panic level, the requirement of evacuation objectives, the evacuation time and density, may not be met. Especially under the high passenger volume conditions, the passenger density could reach up to 6.2 unit/m2 (extremely dangerous); (2) An average improvement degree, 7.5%, can be achieved for the optimization of 20 test cases, and a maximum improvement degree, 22.5%, can be achieved for the evacuation optimization at metro stations; (3) It could be difficult to keep both of the evacuation time and density within the standards if one major exit is closed, even after the optimization. But a larger average improvement degree, 10.8%, can be achieved by the proposed optimization approach, which indicates the optimal solutions still could reduce the risk to a great extent. The novelty of this research lies in that (a) An RF algorithm is incorporated to build the meta-model that can properly represent the relationship between influential factors and objectives, despite the complexity and even conflicting between them; (b) Optimal measures for the evacuation improvement are provided from the MOO perspective by integrating NSGA-III. This hybrid approach can be used as a decision tool to assist regulatory authorities in developing effective emergency evacuation evaluation and optimization plans with adequate consideration of the complexity and multi-objective nature under evacuation events. • A 3D model of the metro station is built to simulate the dynamic process of passenger evacuation. • A framework with 9 factors and 3 objectives is developed to model passenger evacuation. • A random forest is used to build the meta-model and NSGA III is used for the optimization. • A real metro station in Singapore is constructed to test its effectiveness and applicability. • An average 7.5% and a maximum 22.5% of the evacuation optimization degree is presented. [ABSTRACT FROM AUTHOR]

• Presents a combined approach of lifting path planning and assembly sequencing. • Introduces a new BIM-based intelligent assembly scheduler for precast construction. • The scheduler generates the optimal micro-schedule from inbuilt data in BIM. • Proposes a multi-level elitist genetic algorithm to compute the optimal sequence. • A case study is conducted on a real-world 4D BIM model using a tower crane. This research addresses the problem of assembly scheduling in crane-assisted precast construction while considering issues of building layout interference and optimal crane lifting. Traditionally, assembly scheduling and lifting path planning are treated as two separate issues due to their distinct natures. The current work introduces an approach that combines them for precast construction planning to achieve a comprehensive and cost-effective solution. A BIM4D-based Intelligent Assembly Scheduler (BIAS) is designed in conjunction with the Computer-Aided Lifting Planner developed at Nanyang Technological University, Singapore. BIM4D is the Building Information Modeling (BIM) with the time dimension (i.e., scheduling information). Our scheduler takes an inbuilt timeframe for selected precast elements from BIM4D as input and outputs the micro-schedule in terms of the assembly sequence for these precast elements. This problem is solved using multi-objective optimization. Given a group of precast elements and their BIM4D timeframe, the micro-scheduling is determined based on (1) the relative importance of the elements' physical properties, (2) the interference (neighbouring relation) among the elements' positions, and (3) collision-free lifting paths of the elements. A Multi-level Elitist Genetic Algorithm (MEGA) is proposed to determine the optimal sequence taking into consideration of both assembling and lifting for the elements. A case study is performed with the BIM4D data of a residential building. The results of the case study demonstrate BIAS's efficiency and effectiveness for BIM4D based construction scheduling. [ABSTRACT FROM AUTHOR]