An artificial intelligence-based shield TBM big data base cloud analysis method and system

Abstract

The application relates to the technical field of tunnel boring machine intelligent construction and big data analysis, and discloses a shield TBM big data base cloud analysis method and system based on artificial intelligence. The application aims to solve the technical problems that in the prior art, data collection is scattered and standards are different, forming a data island; data management efficiency is low, precision is insufficient, and reliable support cannot be provided; the analysis mode is lagging, the intelligent degree is low, and real-time intelligent decision cannot be realized; cloud adaptation is poor, deployment flexibility is low, and the massive data processing demand cannot be met. The application comprises the following steps: multi-modal data collection, artificial intelligence data management, cloud storage and calculation, artificial intelligence analysis, application output and closed-loop optimization. The application realizes the deep integration of big data base cloud and AI intelligent analysis, improves the intelligent level, safety and efficiency of shield TBM construction, and has very high engineering application value and popularization prospect.

Description

Technical Field

[0001] This invention relates to the field of intelligent construction and big data analysis technology for tunnel boring machines (TBMs), and particularly to a cloud-based analysis method and system for the big data base of shield TBMs based on artificial intelligence. Background Technology

[0002] During operation, TBMs (tunnel boring machines) generate massive amounts of multi-source heterogeneous data, encompassing equipment operating parameters (such as cutterhead torque, thrust, rotational speed, and earth pressure), geological environment information (such as lithology, groundwater pressure, and ground deformation), and construction operation records. This data contains the patterns of equipment state evolution, geological response characteristics, and correlations with construction efficiency. It forms the core foundation for achieving intelligent optimization of tunneling parameters, monitoring of equipment health status and prediction of remaining life, and proactive early warning and prevention of construction risks.

[0003] Meanwhile, the rapid development of new-generation information technologies such as cloud computing, the Internet of Things, big data, and artificial intelligence has provided new technical support for the real-time acquisition, flexible storage, efficient computing, and intelligent decision-making of TBM construction data. How to build a unified TBM big data platform with cloud-native elastic scalability, and establish a closed-loop system encompassing data acquisition, data governance, feature extraction, AI modeling, online inference, and decision feedback, breaking through the inherent limitations of low data transmission efficiency, lagging model updates, and low levels of intelligence in traditional stand-alone or LAN modes, and achieving a shift from experience-driven to data-driven + AI-empowered construction paradigms, has become a major technical challenge and an urgent industry need in the field of intelligent tunnel construction.

[0004] Chinese patent document CN121388577A discloses an artificial intelligence-based method for predicting the total deformation of shield tunnels. This method acquires geological sensor data streams, construction operation data streams, and a historical deformation case library; it performs multimodal data fusion on the geological sensor data streams and construction operation data streams to generate a spatiotemporally consistent construction environment map; it retrieves case sequences from the historical deformation case library that match the current construction environment map, evaluates the feature saliency of the case sequences, and identifies the dominant feature group for total deformation prediction; it constructs an algorithm optimization engine based on the dominant feature group; it activates the algorithm optimization engine to process the construction environment map in parallel, generating an adaptability metric set for each algorithm; it integrates the adaptability metric set with real-time construction constraints, and uses a trade-off decision mechanism to select the best prediction algorithm; it generates a total deformation prediction model and integrates this model into a shield tunnel construction monitoring platform.

[0005] However, the above solutions suffer from at least the following technical problems during implementation: data collection is fragmented and lacks standardized criteria, creating data silos; data governance is inefficient and lacks precision, failing to provide reliable support; analysis models are outdated and lack intelligence, making real-time intelligent decision-making difficult; and cloud adaptability is poor, deployment flexibility is low, and the needs of massive data processing cannot be met. Existing related technologies only focus on single scenarios or fail to achieve deep integration of cloudification and intelligence, thus failing to address the above pain points. There is an urgent need for a method and system to achieve unified big data collection, efficient governance, intelligent analysis, and cloud deployment. Summary of the Invention

[0006] In view of the above technical problems, this disclosure provides a cloud-based analysis method and system for shield tunneling TBM big data base based on artificial intelligence, which solves the technical problems in the prior art, such as scattered and inconsistent data collection, forming data silos; low data governance efficiency and insufficient accuracy, which cannot provide reliable support; lagging analysis mode and low level of intelligence, which makes it difficult to achieve real-time intelligent decision-making; and poor cloud adaptability and low deployment flexibility, which cannot meet the needs of massive data processing.

[0007] According to one aspect of this disclosure, an artificial intelligence-based cloud-based analysis method for shield tunneling machine (TBM) big data base is provided, comprising the following steps:

[0008] S1: Multimodal data acquisition: The multimodal data acquisition module collects multi-source data from the shield tunneling machine across all dimensions. After preliminary preprocessing, the data is transmitted to the artificial intelligence data governance module. The preliminary preprocessing includes format conversion, preliminary outlier removal, and data compression.

[0009] S2: Artificial intelligence data governance, which uses artificial intelligence algorithms to clean, deduplicatize, complete, standardize, enhance and extract features from the data output by the multimodal data acquisition module described in S1, and outputs standardized data; among them, data cleaning uses a hybrid model to identify abnormal data and repair missing data, data enhancement uses algorithms to supplement the sample size, and feature extraction uses an attention mechanism model to extract key features and construct mechanistic features;

[0010] S3: Cloud-based storage and computing, which transmits the standardized data described in S2 to the cloud-native big data base module to realize cloud-based distributed storage and elastic computing of data; the cloud-native big data base module adopts a hybrid storage architecture to realize layered data storage and establish an index system, realizes elastic scaling of computing resources based on container orchestration technology, and completes data processing using a stream-batch integrated computing framework;

[0011] S4: Artificial intelligence analysis, based on the storage and computing capabilities of the cloud-native big data infrastructure module described in S3, calls a pre-trained artificial intelligence analysis model to perform real-time analysis on the standardized data and outputs analysis results; the artificial intelligence analysis model includes an equipment health early warning model, a tunneling parameter optimization model, and a construction risk prediction model;

[0012] S5: Application output and closed-loop optimization. The analysis results described in S4 are output through the application output module, and the feedback data is returned to the cloud-native big data base module for iterative updates of the artificial intelligence analysis model, forming a closed loop.

[0013] In S1, the multimodal data acquisition module integrates multiple acquisition methods through multiple types of acquisition units to acquire equipment operation data, geological data, environmental data, operation and maintenance data, and multimodal sensing data; the data transmission adopts a dual-mode transmission method to achieve real-time encrypted data transmission and support breakpoint resume transmission.

[0014] In S2, the data standardization establishes unified coding rules to achieve unified management of data from multiple sources.

[0015] In S3, the cloud-native big data foundation module implements data scheduling through a distributed scheduling algorithm and establishes a security protection unit to ensure data security.

[0016] In S4, the artificial intelligence analysis model uses incremental training to achieve iterative model updates and supports online deployment and rolling updates.

[0017] In S5, the application output module provides decision support through a visual interface, early warning prompts, and command issuance. The visual interface is displayed on both ends, the early warning prompt unit issues early warnings and marks relevant information through various methods, and the command issuance unit supports remote command issuance and records execution status.

[0018] A cloud-based analysis system for a shield tunneling machine (TBM) big data base, based on artificial intelligence, includes a multimodal data acquisition module, an artificial intelligence data governance module, a cloud-native big data base module, an artificial intelligence analysis module, and an application output module connected in sequence.

[0019] The multimodal data acquisition module is used to collect multi-source data from the tunnel boring machine (TBM) across all dimensions. After preliminary preprocessing, the data is transmitted to the artificial intelligence data governance module. The artificial intelligence data governance module uses artificial intelligence algorithms to clean, deduplicatize, complete, standardize, enhance, and extract features from the received data, and outputs standardized data to the cloud-native big data infrastructure module. The cloud-native big data infrastructure module receives the standardized data, realizes cloud-based distributed storage, elastic computing, and scheduling, and provides storage and computing capabilities to support the artificial intelligence analysis module. The artificial intelligence analysis module uses a pre-trained artificial intelligence analysis model to perform real-time analysis on the standardized data and outputs the analysis results to the application output module. The application output module receives the analysis results and provides decision support through a visual interface, early warning prompts, and command issuance.

[0020] The multimodal data acquisition module includes multiple acquisition units, a preliminary data preprocessing unit, and a data transmission unit. The multiple acquisition units integrate various acquisition methods to acquire equipment operation data, geological data, environmental data, maintenance data, and multimodal sensing data. The preliminary data preprocessing unit performs format conversion, preliminary outlier removal, and data compression on the raw data. The data transmission unit adopts a dual-mode transmission method to achieve real-time encrypted data transmission and support breakpoint resume.

[0021] The beneficial effects of this invention are as follows:

[0022] Unified management of multi-source data significantly improves data integrity and governance efficiency; equipment fault early warning accuracy reaches 96.2%, and the fault occurrence rate is reduced by 35%; tunneling efficiency is increased by 12%, saving construction time; construction risk prediction accuracy reaches 91.5%, and no safety accidents occur; cloud deployment improves resource utilization and reduces deployment costs and operation and maintenance difficulty.

[0023] This invention integrates equipment operation data, geological data, environmental data, maintenance data, and multimodal sensing data through a multimodal data acquisition module, breaking through the limitations of traditional single-source data acquisition and constructing a data sensing system covering all dimensions of TBM (Tunnel Boring Machine) construction. Simultaneously, it employs a dual-mode transmission method to achieve real-time encrypted data transmission and support breakpoint resumption, effectively solving the data loss problem caused by network instability in the tunnel construction environment. This ensures the integrity, security, and timeliness of data transmission, providing a high-quality data source guarantee for subsequent analysis.

[0024] This invention introduces artificial intelligence algorithms to automatically clean, deduplicate, complete, standardize, enhance, and extract features from multi-source heterogeneous data. Specifically, the hybrid model intelligently identifies abnormal data and repairs missing values, effectively solving common problems of noise interference and data incompleteness in shield tunneling construction data; the data augmentation algorithm supplements the sample size, alleviating insufficient model training caused by scarce samples under extreme working conditions; and the attention mechanism model automatically extracts key features and constructs mechanistic features, avoiding the subjectivity and limitations of traditional manual feature engineering. The synergistic effect of these technologies significantly improves the degree of data standardization and feature representation capabilities, providing high-quality, highly representative data input for upper-level intelligent analysis models.

[0025] This invention employs a cloud-native big data infrastructure module, achieving layered data storage and establishing an efficient indexing system based on a hybrid storage architecture. It utilizes container orchestration technology to enable on-demand elastic scaling of computing resources, adapting to the real-world scenario of large data volume fluctuations during tunnel boring machine (TBM) construction and avoiding resource waste or insufficient computing power. The integrated stream-batch computing framework unifies the computing links for real-time stream processing and offline batch processing, simplifying the data processing architecture and improving overall data processing efficiency and real-time performance. The introduction of distributed scheduling algorithms and security protection units further ensures the efficiency of data scheduling and the security of data assets.

[0026] This invention pre-trains three core artificial intelligence analysis models: an equipment health early warning model, a tunneling parameter optimization model, and a construction risk prediction model. This forms a comprehensive intelligent analysis capability encompassing equipment status assessment, tunneling parameter optimization, and construction risk prevention. The equipment health early warning model can identify equipment degradation trends in advance, enabling predictive maintenance and avoiding sudden downtime. The tunneling parameter optimization model dynamically recommends optimal tunneling parameters based on real-time geological and equipment data, reducing reliance on manual experience and improving tunneling efficiency and construction quality. The construction risk prediction model provides early warnings for typical risks such as sudden surges, collapses, and machine jams, providing a basis for safe construction decisions. The three models work together to achieve intelligent and precise management and control of TBM (Tunnel Boring Machine) construction.

[0027] This invention employs incremental training to iteratively update the artificial intelligence analysis model, supporting online deployment and rolling updates. Model upgrades can be completed without system downtime, ensuring construction continuity. Simultaneously, the application output module feeds back the analysis results to the cloud-native big data foundation module, forming a closed-loop data chain encompassing data collection, intelligent governance, cloud computing, AI analysis, application output, and feedback iteration. This allows the model to continuously adapt to changes in data distribution under different geological conditions and construction stages, achieving self-learning and self-optimization capabilities. System performance continuously improves with data accumulation.

[0028] The application output module of this invention displays the analysis results on both ends through a visual interface, taking into account the usage needs of different scenarios; the early warning unit issues early warnings and marks relevant information through multiple methods to ensure that risk information reaches relevant personnel in a timely manner; the instruction issuance unit supports remote instruction issuance and records the execution status, realizing a rapid closed loop of analysis-decision-execution-feedback, which significantly improves construction management efficiency and emergency response capabilities.

[0029] The present invention comprises a multimodal data acquisition module, an artificial intelligence data governance module, a cloud-native big data base module, an artificial intelligence analysis module, and an application output module connected in sequence, which facilitates subsequent functional expansion and module replacement. It can be widely applied to various shield tunneling TBM construction scenarios and has broad prospects for industrialization and promotion. Attached Figure Description

[0030] Figure 1 A flowchart illustrating the cloud-based analysis method for big data base of tunnel boring machines (TBMs) based on artificial intelligence. Detailed Implementation

[0031] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0032] Example 1

[0033] This example discloses an artificial intelligence-based cloud-based analysis method for TBM (Tunnel Boring Machine) big data base. See [link to relevant documentation]. Figure 1 It includes the following steps:

[0034] S1: Multimodal data acquisition. The multimodal data acquisition module collects multi-source data from the shield tunneling machine in all dimensions. After preliminary preprocessing, the data is transmitted to the artificial intelligence data governance module. The preliminary preprocessing includes format conversion, preliminary removal of outliers, and data compression.

[0035] S2: Artificial intelligence data governance, which uses artificial intelligence algorithms to clean, deduplicatize, complete, standardize, enhance and extract features from the data output by the multimodal data acquisition module in S1, and outputs standardized data. Among them, data cleaning uses a hybrid model to identify abnormal data and repair missing data, data enhancement uses algorithms to supplement the sample size, and feature extraction uses an attention mechanism model to extract key features and construct mechanistic features.

[0036] S3: Cloud-based storage and computing, which transmits standardized data from S2 to the cloud-native big data infrastructure module to achieve cloud-based distributed storage and elastic computing of data; the cloud-native big data infrastructure module adopts a hybrid storage architecture to achieve layered data storage and establish an index system, realizes elastic scaling of computing resources based on container orchestration technology, and completes data processing using a stream-batch integrated computing framework;

[0037] S4: Artificial intelligence analysis, based on the storage and computing capabilities of the cloud-native big data infrastructure module in S3, calls pre-trained artificial intelligence analysis models to perform real-time analysis on standardized data and output analysis results; the artificial intelligence analysis models include equipment health early warning models, tunneling parameter optimization models and construction risk prediction models;

[0038] S5: Application Output and Closed-Loop Optimization. The analysis results in S4 are output through the application output module, and the feedback data is returned to the cloud-native big data base module for iterative updates of the artificial intelligence analysis model, forming a closed loop.

[0039] In S1, the multimodal data acquisition module integrates multiple acquisition methods through multiple types of acquisition units to acquire equipment operation data, geological data, environmental data, operation and maintenance data, and multimodal sensing data; data transmission adopts a dual-mode transmission method to achieve real-time encrypted data transmission and support breakpoint resume.

[0040] In S2, data standardization establishes unified coding rules to achieve unified management of data from multiple sources.

[0041] In S3, the cloud-native big data foundation module implements data scheduling through a distributed scheduling algorithm and establishes a security protection unit to ensure data security.

[0042] In S4, the artificial intelligence analysis model uses incremental training to achieve iterative model updates and supports online deployment and rolling updates.

[0043] In S5, the application output module provides decision support through a visual interface, early warning prompts, and command issuance. The visual interface is displayed on both ends, the early warning prompt unit issues early warnings and marks relevant information in multiple ways, and the command issuance unit supports remote command issuance and records execution status.

[0044] A cloud-based analysis system for a shield tunneling machine (TBM) big data base, based on artificial intelligence, includes a multimodal data acquisition module, an artificial intelligence data governance module, a cloud-native big data base module, an artificial intelligence analysis module, and an application output module connected in sequence.

[0045] The multimodal data acquisition module collects multi-source data from the tunnel boring machine (TBM) across all dimensions. After preliminary preprocessing, the data is transmitted to the AI ​​data governance module. The AI ​​data governance module uses AI algorithms to clean, deduplicatize, complete, standardize, enhance, and extract features from the received data, outputting standardized data to the cloud-native big data infrastructure module. The cloud-native big data infrastructure module receives the standardized data, enabling cloud-based distributed storage, elastic computing, and scheduling, and providing storage and computing capabilities to support the AI ​​analysis module. The AI ​​analysis module performs real-time analysis of the standardized data using a pre-trained AI analysis model, outputting the analysis results to the application output module. The application output module receives the analysis results and provides decision support through a visual interface, early warning prompts, and command issuance.

[0046] The multimodal data acquisition module includes multiple acquisition units, a preliminary data preprocessing unit, and a data transmission unit. The multiple acquisition units integrate various acquisition methods to acquire equipment operation data, geological data, environmental data, maintenance data, and multimodal sensing data. The preliminary data preprocessing unit performs format conversion, preliminary outlier removal, and data compression on the raw data. The data transmission unit adopts a dual-mode transmission method to achieve real-time encrypted data transmission and support breakpoint resume.

[0047] Although some preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention.

[0048] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this application and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A cloud-based analysis method for big data base of tunnel boring machines (TBMs) based on artificial intelligence, characterized in that, Includes the following steps: S1: Multimodal data acquisition: The multimodal data acquisition module collects multi-source data from the shield tunneling machine across all dimensions. After preliminary preprocessing, the data is transmitted to the artificial intelligence data governance module. The preliminary preprocessing includes format conversion, preliminary outlier removal, and data compression. S2: Artificial intelligence data governance, which uses artificial intelligence algorithms to clean, deduplicatize, complete, standardize, enhance and extract features from the data output by the multimodal data acquisition module described in S1, and outputs standardized data; among them, data cleaning uses a hybrid model to identify abnormal data and repair missing data, data enhancement uses algorithms to supplement the sample size, and feature extraction uses an attention mechanism model to extract key features and construct mechanistic features; S3: Cloud-based storage and computing, which transmits the standardized data described in S2 to the cloud-native big data base module to achieve cloud-based distributed storage and elastic computing of data; the cloud-native big data base module adopts a hybrid storage architecture to achieve layered data storage and establish an index system, realizes elastic scaling of computing resources based on container orchestration technology, and completes data processing using a stream-batch integrated computing framework; S4: Artificial intelligence analysis, based on the storage and computing capabilities of the cloud-native big data infrastructure module described in S3, calls a pre-trained artificial intelligence analysis model to perform real-time analysis on the standardized data and outputs analysis results; the artificial intelligence analysis model includes an equipment health early warning model, a tunneling parameter optimization model, and a construction risk prediction model; S5: Application output and closed-loop optimization. The analysis results described in S4 are output through the application output module, and the feedback data is returned to the cloud-native big data base module for iterative updates of the artificial intelligence analysis model, forming a closed loop.

2. The cloud-based analysis method for big data base of shield tunneling TBM based on artificial intelligence as described in claim 1, characterized in that, In S1, the multimodal data acquisition module integrates multiple acquisition methods through multiple types of acquisition units to acquire equipment operation data, geological data, environmental data, operation and maintenance data, and multimodal sensing data; the data transmission adopts a dual-mode transmission method to achieve real-time encrypted data transmission and support breakpoint resume transmission.

3. The cloud-based analysis method for big data base of shield tunneling TBM based on artificial intelligence as described in claim 1, characterized in that, In S2, the data standardization establishes unified coding rules to achieve unified management of data from multiple sources.

4. The cloud-based analysis method for big data base of shield tunneling TBM based on artificial intelligence as described in claim 1, characterized in that, In S3, the cloud-native big data foundation module implements data scheduling through a distributed scheduling algorithm and establishes a security protection unit to ensure data security.

5. The cloud-based analysis method for big data base of shield tunneling TBMs based on artificial intelligence as described in claim 1, characterized in that, In S4, the artificial intelligence analysis model uses incremental training to achieve iterative model updates and supports online deployment and rolling updates.

6. The cloud-based analysis method for big data base of shield tunneling TBM based on artificial intelligence as described in claim 1, characterized in that, In S5, the application output module provides decision support through a visual interface, early warning prompts, and command issuance. The visual interface is displayed on both ends, the early warning prompt unit issues early warnings and marks relevant information in multiple ways, and the command issuance unit supports remote command issuance and records execution status.

7. An AI-based shield tunneling TBM big data base cloud analysis system, applicable to the AI-based shield tunneling TBM big data base cloud analysis method as described in any one of claims 1 to 6, characterized in that, It includes a multimodal data acquisition module, an artificial intelligence data governance module, a cloud-native big data infrastructure module, an artificial intelligence analysis module, and an application output module connected in sequence; The multimodal data acquisition module is used to collect multi-source data from the tunnel boring machine (TBM) across all dimensions. After preliminary preprocessing, the data is transmitted to the artificial intelligence data governance module. The artificial intelligence data governance module uses artificial intelligence algorithms to clean, deduplicatize, complete, standardize, enhance, and extract features from the received data, and outputs standardized data to the cloud-native big data infrastructure module. The cloud-native big data infrastructure module receives the standardized data, realizes cloud-based distributed storage, elastic computing, and scheduling, and provides storage and computing capabilities to support the artificial intelligence analysis module. The artificial intelligence analysis module uses a pre-trained artificial intelligence analysis model to perform real-time analysis on the standardized data and outputs the analysis results to the application output module. The application output module receives the analysis results and provides decision support through a visual interface, early warning prompts, and command issuance.

8. The cloud-based analysis system for big data base of shield tunneling TBM based on artificial intelligence as described in claim 7, characterized in that, The multimodal data acquisition module includes multiple acquisition units, a preliminary data preprocessing unit, and a data transmission unit. The multiple acquisition units integrate various acquisition methods to acquire equipment operation data, geological data, environmental data, maintenance data, and multimodal sensing data. The preliminary data preprocessing unit performs format conversion, preliminary outlier removal, and data compression on the raw data. The data transmission unit adopts a dual-mode transmission method to achieve real-time encrypted data transmission and support breakpoint resume.

Images