Tunnel construction progress digital twinborn prediction and resource scheduling system and method thereof

Abstract

The invention discloses a tunnel construction progress digital twinborn prediction and resource scheduling system and method, relates to the technical field of tunnel construction, and provides the following scheme that the method comprises the steps that real-time monitoring data of a tunnel construction area are collected, the real-time monitoring data comprise equipment position information and process progress quantitative data, and a BIM model of the tunnel construction area is constructed; dividing the equipment position information into a plurality of continuous time windows according to time nodes of the BIM model, and obtaining equipment operation parameters and corresponding energy consumption values in each time window; when the effective energy consumption is calculated, the geological condition grade weight and the construction process reference value are introduced to dynamically correct the initial efficiency coefficient, and meanwhile, the geological condition fluctuation, the equipment performance attenuation and the process connection change in the construction conditions are reflected through the time attenuation coefficient, so that the nonlinear influence of the geological difference and the process adjustment on the energy consumption efficiency is effectively eliminated; the strong relevance between the energy consumption and the progress is established, and a reliable data association basis is provided for progress prediction.

Description

Technical Field

[0001] This invention relates to the field of tunnel construction technology, specifically to a digital twin prediction and resource scheduling system and method for tunnel construction progress. Background Technology

[0002] Tunnel construction is a complex systems engineering project involving the coordinated operation of various equipment, dynamic process connections, and changes in geological conditions. With the development of digital twin technology, real-time monitoring methods based on BIM models and the Internet of Things are gradually being applied to tunnel construction management, aiming to improve construction efficiency and the accuracy of resource allocation through data-driven approaches.

[0003] However, existing technologies for utilizing equipment energy consumption monitoring data still have limitations: the monitored values ​​contain a large amount of invalid standby energy consumption and lack an effective differentiation mechanism, making it difficult for real-time monitoring data to reflect actual construction input. Furthermore, the mapping relationship between energy consumption monitoring values ​​and actual construction progress is weakened by changes in geological conditions and construction procedures. These two factors combined result in equipment energy consumption monitoring values ​​failing to accurately map construction efficiency and progress status, making progress prediction and resource scheduling based on energy consumption monitoring values ​​lack a reliable basis. Therefore, we propose a digital twin prediction and resource scheduling system for tunnel construction progress and its methodology to address this problem. Summary of the Invention

[0004] To address the aforementioned technical problems, this paper provides a digital twin prediction and resource scheduling system and method for tunnel construction progress. This solution resolves the issues raised in the background, such as the presence of a large amount of invalid standby energy consumption in monitoring values, the lack of an effective differentiation mechanism, and the resulting difficulty in reflecting actual construction input in real-time monitoring data. Furthermore, the mapping relationship between energy consumption monitoring values ​​and actual construction progress is weakened due to changes in geological conditions and construction procedures. These combined factors result in equipment energy consumption monitoring values ​​failing to accurately map construction efficiency and progress status, thus lacking a reliable basis for progress prediction and resource scheduling based on energy consumption monitoring values.

[0005] To achieve the above objectives, the technical solution adopted by this invention is as follows: This invention provides a method for digital twin prediction and resource scheduling of tunnel construction progress, the method comprising:

[0006] Collect real-time monitoring data of the tunnel construction area, including equipment location information and quantitative data of process progress, and construct a BIM model of the tunnel construction area.

[0007] The equipment location information is divided into multiple continuous time windows according to the time nodes of the BIM model. Within each time window, the equipment operating parameters and corresponding energy consumption values ​​are obtained, the spatial coordinates of the equipment location information and the corresponding dwell time are extracted, and the movement trajectory is generated based on the temporal changes of the spatial coordinates.

[0008] The equipment operation type is determined by the combination of equipment operating parameters, energy consumption values ​​and movement trajectory characteristics. Based on the equipment operation type, the effective energy consumption value of the equipment is calculated in combination with changes in geological conditions and construction procedures, and an energy consumption heat map of the construction area is constructed.

[0009] By combining the characteristic parameters of the energy consumption thermogram with the quantitative data of the process progress, a progress prediction model is trained using the LSTM algorithm to generate progress prediction data.

[0010] Material storage points are generated based on energy consumption thermal maps. The BIM model is used to process the progress prediction data and generate structured material scheduling instructions.

[0011] Furthermore, a digital twin prediction and resource scheduling system for tunnel construction progress is proposed to implement the digital twin prediction and resource scheduling method for tunnel construction progress described above, including:

[0012] The data acquisition and model building module is used to collect real-time monitoring data of the tunnel construction area, including equipment location information and quantitative data of process progress, and to build a BIM model of the tunnel construction area.

[0013] The equipment spatiotemporal feature extraction module is used to divide the equipment location information into multiple continuous time windows according to the time nodes of the BIM model. Within each time window, the module obtains the equipment operating parameters and corresponding energy consumption values, extracts the spatial coordinates of the equipment location information and the corresponding dwell time, and generates a movement trajectory based on the temporal changes of the spatial coordinates.

[0014] The operation type determination and energy consumption analysis module is used to determine the equipment operation type based on the characteristic combination of equipment operating parameters, energy consumption values ​​and movement trajectory. Based on the equipment operation type, combined with changes in geological conditions and construction procedures, the module calculates the effective energy consumption value of the equipment and constructs an energy consumption heat map of the construction area.

[0015] The progress prediction module is used to combine the characteristic parameters of the energy consumption thermal map with the quantitative data of the process progress, train the progress prediction model through the LSTM algorithm, and generate progress prediction data.

[0016] The material scheduling module is used to generate material storage points based on energy consumption thermal maps, process progress prediction data using BIM models, and generate structured material scheduling instructions.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] By combining the characteristics of equipment operating parameters, energy consumption values, and movement trajectories, and using combination relationship judgment rules, the operation type is determined, and invalid standby energy consumption is accurately filtered out. This ensures that only the effective part of the energy consumption data is retained that is relevant to the actual construction, providing clean and reliable basic data for subsequent progress prediction and resource scheduling.

[0019] By calculating effective energy consumption, the initial efficiency coefficient is dynamically corrected by introducing the weight of geological conditions and the benchmark value of construction procedures. At the same time, the time decay coefficient reflects the fluctuation of geological conditions, the degradation of equipment performance, and the changes in the connection of procedures during construction. This effectively eliminates the nonlinear impact of geological differences and procedure adjustments on energy consumption efficiency, so that the effective energy consumption value accurately maps to the actual construction input, establishes a strong correlation between energy consumption and progress, and provides a reliable data correlation basis for progress prediction. Attached Figure Description

[0020] Figure 1 The flowchart shows the digital twin prediction and resource scheduling method for tunnel construction progress proposed in this invention.

[0021] Figure 2 This is a flowchart of the method for determining the operation type of equipment in this invention;

[0022] Figure 3 This is a flowchart of the method for calculating the effective energy consumption value of a time window in this invention;

[0023] Figure 4 This is a structural block diagram of the digital twin prediction and resource scheduling system for tunnel construction progress proposed in this invention. Detailed Implementation

[0024] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0025] In existing technologies, tunnel construction management generally adopts monitoring methods based on BIM models and the Internet of Things to assist construction decisions by collecting equipment data in real time. However, existing methods have obvious defects in processing equipment energy consumption data: the monitoring data contains a large amount of invalid standby energy consumption, lacks an effective filtering mechanism, and cannot accurately reflect the actual construction input. At the same time, changes in geological conditions and procedures weaken the correlation between energy consumption values ​​and construction efficiency, making the progress prediction model based on energy consumption data unreliable and causing deviations between resource scheduling instructions and actual needs.

[0026] To address these issues, it is necessary to establish a precise mapping relationship between equipment energy consumption data and actual construction efficiency. By analyzing the correlation between equipment operating parameters and movement trajectories, effective operation and ineffective standby states can be distinguished. Furthermore, it was discovered that the impact of geological condition differences and process adjustments on equipment energy efficiency exhibits non-linear characteristics, necessitating the introduction of a dynamic correction mechanism. In addition, the spatial distribution characteristics of energy consumption data have a potential correlation with construction progress; constructing heat maps can reveal the spatial distribution patterns of resource demand.

[0027] Reference Figure 1-3 As shown, this application proposes a digital twin prediction and resource scheduling method for tunnel construction progress, the method comprising:

[0028] Collect real-time monitoring data of the tunnel construction area, including equipment location information and quantitative data of process progress, and construct a BIM model of the tunnel construction area.

[0029] The equipment location information is divided into multiple continuous time windows according to the time nodes of the BIM model. Within each time window, the equipment operating parameters and corresponding energy consumption values ​​are obtained, the spatial coordinates of the equipment location information and the corresponding dwell time are extracted, and the movement trajectory is generated based on the temporal changes of the spatial coordinates.

[0030] The equipment operation type is determined by the combination of equipment operating parameters, energy consumption values ​​and movement trajectory characteristics. Based on the equipment operation type, the effective energy consumption value of the equipment is calculated in combination with changes in geological conditions and construction procedures, and an energy consumption heat map of the construction area is constructed.

[0031] By combining the characteristic parameters of the energy consumption thermogram with the quantitative data of the process progress, a progress prediction model is trained using the LSTM algorithm to generate progress prediction data.

[0032] Material storage points are generated based on energy consumption thermal maps. The BIM model is used to process the progress prediction data and generate structured material scheduling instructions.

[0033] Reference Figure 2 As shown, this application further proposes a method for determining the equipment operation type based on a combination of features including equipment operating parameters, energy consumption values, and movement trajectory, specifically including:

[0034] The equipment operating parameters are normalized to generate operating status feature values. Based on these feature values, the equipment operating status is divided into strong operating characteristics, weak operating characteristics, and non-operating characteristics. For example, the operating status feature values ​​can be calculated by weighting the equipment operating parameters, including working pressure, speed, and vibration frequency, in a 3:4:3 ratio. When the operating status feature value is ≥60, it is marked as a strong operating characteristic; when it is 30-59, it is marked as a weak operating characteristic; and when it is <30, it is marked as a non-operating characteristic.

[0035] Obtain historical valid operating data of the equipment under the same working conditions, calculate the energy consumption matching degree between the real-time energy consumption value and the historical valid operating data, and classify the energy consumption status of the equipment into high energy consumption matching, medium energy consumption matching and low energy consumption matching according to the matching degree; for example, when the matching degree is ≥70%, it is marked as high energy consumption matching, 40%-69% is marked as medium energy consumption matching, and <40% is marked as low energy consumption matching;

[0036] Dynamic features of the movement trajectory are extracted to generate trajectory activity feature values. Based on the trajectory activity feature values, the trajectory status of the device is divided into strong trajectory activity, weak trajectory activity, and no trajectory activity. For example, the trajectory activity index is calculated by weighting the movement speed fluctuation rate, the trajectory coverage area growth rate, and the dwell time in a ratio of 2:3:5. When the trajectory activity feature value is ≥60, it is marked as strong trajectory activity; 30-59 is marked as weak trajectory activity; and <30 is marked as no trajectory activity.

[0037] A combination relationship determination rule is set to determine the job type based on the combination relationship between operating status, energy consumption status and trajectory status. The job type includes valid job and invalid standby.

[0038] For example, the rule for determining the combination relationship is: when the combination includes strong operational characteristics and high energy consumption matching, it is determined to be a valid operation regardless of the trajectory activity;

[0039] When the combination consists of weak operational characteristics, medium energy consumption matching, and strong trajectory activity, supplementary verification is triggered: if the current position of the equipment belongs to the process operation area in the BIM model, it is determined to be a valid operation;

[0040] When the combination includes strong operational characteristics and low energy consumption matching, the trajectory activity is determined: if the trajectory activity index is ≥40, it is determined to be a valid operation; otherwise, it is marked as an invalid standby.

[0041] When a combination exhibits non-operational characteristics or lacks trajectory activity and is matched with low energy consumption, it is determined to be an invalid standby.

[0042] The remaining combinations will activate a weighted compensation mechanism: the comprehensive score will be calculated by weighting the operating status, energy consumption matching, and trajectory activity in a 5:3:2 ratio. A score of ≥50 will be considered a valid operation, otherwise it will be considered an invalid standby.

[0043] Compared to existing technologies, traditional methods typically rely solely on a single energy consumption value or equipment operating status to determine the job type, failing to consider the impact of historical valid job data matching and dynamic trajectory characteristics. This leads to confusion between invalid standby energy consumption and valid job energy consumption. This solution introduces historical valid job data matching degree calculation, combined with trajectory activity feature value analysis, to construct a multi-dimensional state combination judgment rule. This effectively distinguishes the actual operating status of equipment under the same energy consumption level, reducing misjudgments caused by invalid states such as equipment idling or temporary shutdown.

[0044] By comprehensively considering the three parameters, the above technical solution can fully and flexibly address various complex combinations of situations, avoid misjudging the job type based on only one or two parameters, improve the accuracy and reliability of the judgment, and thus provide a reliable data foundation for subsequent construction progress prediction and resource allocation.

[0045] Reference Figure 3 As shown, this application further proposes to calculate the effective energy consumption value of equipment based on the equipment operation type, combined with changes in geological conditions and construction procedures, specifically including:

[0046] The time window is divided according to the type of work, and the time segments of effective work are selected. These time segments are then divided into several time sub-units based on the stability of energy consumption values. The stability of energy consumption values ​​is obtained by analyzing the standard deviation. When the standard deviation is less than a preset stability threshold, the time sub-units are divided equally at fixed time intervals. When the standard deviation is greater than or equal to the stability threshold, the time sub-units are dynamically divided using the peak and trough times of energy consumption values ​​as dividing points. This ensures that the fluctuation range of energy consumption values ​​within each time sub-unit does not exceed a preset fluctuation threshold, guaranteeing that the energy consumption values ​​within a time sub-unit represent the energy consumption level of that time sub-unit. The stability threshold is determined by statistically analyzing the standard deviation distribution of energy consumption for similar equipment under similar operating conditions, and selecting the 80%–90% quantile from historical data.

[0047] The geological condition level and corresponding weight are preset, and the benchmark value of the process is associated with the time sub-unit of the construction process. For example, the preset geological condition level can be Class IV surrounding rock, with corresponding weight coefficients of 0.8, 0.9, 1.0, 1.2, and 1.5, respectively. The higher the level, the greater the weight, reflecting the influence of geology on energy consumption effectiveness. The construction process includes excavation, support, and lining. The excavation stage is usually a high-intensity operation, and the equipment needs to continuously output maximum power. The energy consumption is high and linearly related to the progress. Therefore, it is set as the benchmark value of 1.0. Energy consumption values ​​and actual progress data under similar geological conditions and construction processes are collected, and the energy consumption per unit progress of the process and the energy consumption per unit progress of excavation are used as the corresponding process benchmark values.

[0048] Historical energy consumption values ​​of equipment under similar geological conditions and construction procedures, along with corresponding construction progress data, are obtained. A multi-dimensional outlier filtering mechanism is employed to remove outliers. The first dimension uses the Z-score algorithm to remove energy consumption values ​​that exceed the normal fluctuation range. The second dimension combines abnormal events marked in the construction logs, such as equipment failures and temporary shutdowns, to filter energy consumption data for the corresponding time periods. The third dimension uses the isolated forest algorithm to cluster historical data and filters out isolated points with a confidence level >90%. Invalid data is removed from the statistical, event, and logical levels to ensure the purity of historical energy consumption values ​​and provide a reliable foundation for mapping relationships.

[0049] For the filtered historical energy consumption values ​​and corresponding construction progress data, a polynomial curve fitting method is used to establish a mapping relationship between energy consumption values ​​and initial efficiency coefficients. The grade weight corresponding to the current geological conditions and the process benchmark value corresponding to the construction process are used as correction factors to dynamically correct the initial efficiency coefficients. For example, the polynomial curve can be a quadratic function. It captures the nonlinear relationship between energy consumption values ​​and efficiency coefficients, and fits the actual law of diminishing marginal efficiency of energy consumption in tunnel construction.

[0050] An attenuation coefficient is configured for each time sub-unit. This attenuation coefficient is a function value related to the interval between that time sub-unit and the current time, and the longer the interval, the smaller the attenuation coefficient. For example, the formula for calculating the attenuation coefficient can be: ,in, The attenuation coefficient is... This represents the interval between the current time unit and the current time. This is the attenuation rate coefficient, which is dynamically set based on geological conditions. It is 0.15 for Class IV surrounding rock sections and 0.1 for Class III surrounding rock sections. The more complex the geology, the higher the coefficient. The larger the value, the faster the decay.

[0051] The duration of each time sub-unit, the corresponding attenuation coefficient, the energy consumption value, and the corrected efficiency coefficient are multiplied sequentially. The product results of all time sub-units are then summed to obtain the effective energy consumption value of the corresponding time window.

[0052] As a preferred embodiment, the specific implementation of calculating the effective energy consumption value of the time window is as follows: For example, an excavation operation is carried out in a Class V surrounding rock section of a tunnel, the equipment is an excavator, the time window is 9:00-11:00, and the calculation process includes S1-S5;

[0053] S1. Core Input Data:

[0054] The first time sub-unit is 9:00-10:00, lasting for 60 minutes, with an energy consumption of 70kWh, a geological grade weight of 1.5, a process baseline value of 1.0, and an interval of 1 hour from the current time.

[0055] The second time sub-unit is 10:00-10:40, lasting for 40 minutes, with an energy consumption of 50kWh, a geological grade weight of 1.5, a process baseline value of 1.0, and an interval of 0.33 hours from the current time.

[0056] S2. Initial efficiency coefficient calculation, based on the polynomial fitted for Class V surrounding rock excavation: :

[0057] First time unit: x = 70 kWh, y = -0.0005 × 70² + 0.012 × 70 + 0.3 = 0.71;

[0058] Second time sub-unit: x=50kWh, y=-0.0005×50²+0.012×50+0.3=0.65;

[0059] S3. Calculation of the corrected efficiency coefficient:

[0060] First time unit: 1.0 × 1.5 × 0.71 = 1.065;

[0061] Second time sub-unit: 1.0 × 1.5 × 0.65 = 0.975;

[0062] S4. Attenuation coefficient calculation:

[0063] First Time Subunit: ;

[0064] Second time subunit: ;

[0065] S5. Calculation of effective energy consumption value for the time window:

[0066] First time sub-unit: 1×70kWh×1.065×0.82≈61.5kWh;

[0067] Second time sub-unit: 40 / 60×50kWh×0.975×0.93≈30.1kWh;

[0068] The effective energy consumption value of the time window = 61.5 + 30.1 = 91.6 kWh. By linking the process baseline value and geological grade weight with the corrected efficiency coefficient, and combining the attenuation coefficient to highlight the weight of recent data, 91.6 kWh of effective energy consumption is finally extracted from the original total energy consumption of 120 kWh, and invalid standby energy consumption is eliminated.

[0069] Specifically, effective work time segments are screened by task type to eliminate invalid standby energy consumption interference. Then, based on energy consumption stability, these time segments are divided into time sub-units. Geological condition level weights and process baseline values ​​are dynamically linked to these time sub-units. An initial efficiency coefficient is established by combining historical data on energy consumption and progress relationships, and polynomial fitting enhances the model's adaptability to nonlinear relationships. Current geological conditions and construction process changes are dynamically adjusted using weights and baseline values ​​to make energy consumption calculations more closely reflect actual working conditions. The attenuation coefficient gradually decreases in weight based on the interval between the time sub-unit and the current moment, making recent data contribute more to the effective energy consumption value, thus reflecting dynamic changes in construction conditions such as geological condition fluctuations, equipment performance degradation, and process transitions. Finally, through the multiplication and summation of time sub-units, the influence of geological conditions, process changes, and time attenuation is integrated to form a dynamically optimized effective energy consumption calculation model.

[0070] Through the above technical solution, this application can accurately distinguish between effective operation and ineffective standby energy consumption, eliminate the impact of geological conditions and dynamic changes in processes on energy consumption efficiency, and improve the calculation accuracy of effective energy consumption value through dynamic correction and time decay mechanism, so as to provide a reliable data foundation for subsequent construction progress prediction and resource scheduling.

[0071] This application further proposes an energy consumption thermal map of the construction area, specifically including:

[0072] A dynamic adaptive mechanism is used to divide the construction area into several grid units. The initial grid is of a preset size. If the effective energy consumption value of an initial grid is higher than the first density threshold, it is automatically refined into a grid of a smaller preset size. If the effective energy consumption value of adjacent initial grids is lower than the second density threshold for two consecutive time windows, they are automatically merged into a grid of a larger preset size. For example, the basic grid is 10m×10m, the first density threshold can be the average of the effective energy consumption values ​​of all grid units, the smaller preset size can be 5m×5m, the second density threshold can be 1 / 4 of the average, and the larger preset size can be 20m×20m.

[0073] It should be noted that the dynamic adaptive mechanism refers to an algorithm that dynamically adjusts the grid size based on energy consumption density. Specifically, a preset density threshold can be used to trigger grid splitting or merging operations. When the effective energy consumption value of a grid cell exceeds the density threshold, grid refinement is triggered. When the energy consumption value of adjacent grid cells is lower than the density threshold for multiple consecutive time windows, grid merging is triggered.

[0074] The spatial coordinates of the equipment are matched to the corresponding grid cells, and the boundaries of the grid cells are aligned with the construction process zones to ensure the spatial correspondence between the energy consumption analysis unit and the actual construction activity area. The energy consumption heat value of each grid cell is calculated based on the effective energy consumption.

[0075] The kernel density estimation method is used to transform the discrete grid cell energy consumption and heat values ​​into a continuous spatial distribution. A Gaussian kernel is used to weight and average the energy consumption and heat values ​​of neighboring grid cells to generate a smooth energy consumption and heat map, which is updated at a preset period. The kernel density estimation method is an algorithm for spatial interpolation of discrete grid data based on the Gaussian kernel function. Specifically, it can calculate the influence weight of each grid cell on the surrounding area, generate a continuous heat value surface through weighted averaging, and eliminate abrupt errors in the spatial distribution of discrete grid data through spatial smoothing.

[0076] Specifically, during tunnel construction, the dynamic adaptive mechanism first divides the initial grid with a preset size. When the effective energy consumption value of a certain initial grid exceeds the first density threshold, it indicates that the energy consumption density in that area is high and requires more refined monitoring. At this point, the grid unit is refined into smaller sub-grids. In this way, the grid unit division can dynamically adapt to changes in energy consumption distribution during construction. After the grid unit division is completed, the equipment coordinates are matched to the corresponding grid units, and the thermal value is calculated based on the effective energy consumption within the grid. Furthermore, a Gaussian kernel function is used to weight the energy consumption values ​​of neighboring grids. For example, with each grid as the center, the Gaussian weighted average value is calculated within a 3×3 grid area around it, thereby generating a continuous and smooth energy consumption thermal map. The preset periodic update mechanism can be set to recalculate the thermal map every 30 minutes to ensure data real-time performance.

[0077] Through the above technical solution, this application solves the problem of low accuracy of thermal maps caused by fixed grid division for energy consumption monitoring in construction areas. It achieves dynamic adjustment of grid size according to energy consumption density, ensuring monitoring accuracy in high-energy-consumption areas while reducing computational resource consumption in low-energy-consumption areas. The alignment design between grid boundaries and work process zones enhances the correlation between thermal values ​​and actual construction activities. The application of kernel density estimation eliminates spatial abrupt errors in discrete grid data, and the generated thermal map accurately reflects the spatial gradient distribution characteristics of energy consumption in the construction area, providing a reliable basis for subsequent progress prediction and resource scheduling.

[0078] This application further proposes combining energy consumption thermal map feature parameters with process progress quantification data, training a progress prediction model using the LSTM algorithm, and generating progress prediction data, specifically including:

[0079] The energy consumption heat map is spatially aligned with the tunnel construction BIM model so that the effective energy consumption value of each grid cell in the energy consumption heat map corresponds to the process link in the BIM model. The spatial coordinate alignment of the BIM model refers to the precise matching of the three-dimensional coordinates of the grid cells in the energy consumption heat map with the process link in the BIM model. Specifically, coordinate transformation algorithms or spatial projection mapping technology can be used to ensure that the energy consumption data and the construction process are physically spatially associated.

[0080] Extract the feature parameters of the energy consumption heat map, including the effective energy consumption value of each grid cell, the energy consumption spatial distribution gradient and equipment distribution information, and associate the process progress quantification data of the corresponding time nodes of the process links to generate a feature dataset;

[0081] The construction area is divided into several prediction sub-units according to the time nodes of the process steps. For each prediction sub-unit, energy consumption heat map characteristic parameters and construction progress data under similar geological conditions and construction procedures are selected as training samples. After removing abnormal samples, a sample library is constructed. The prediction sub-unit refers to the sub-module of the construction area divided according to the time nodes of the process steps. Specifically, the spatiotemporal slicing technology can be used to divide the construction area into independent analysis units with the same process stages and geological conditions, which are used to construct a training sample library under local working conditions.

[0082] The LSTM algorithm is used, with the feature dataset of the prediction sub-unit as input and the time-series progress completion percentage of the corresponding prediction sub-unit as output. The model is trained through a sample library to obtain the progress prediction model. The LSTM algorithm refers to the Long Short-Term Memory Neural Network model, which can be implemented by using a time-series data training framework combined with the backpropagation algorithm to capture the non-linear change pattern of construction progress in the time dimension.

[0083] The energy consumption heat map feature parameters and process progress quantification data at the current time node are combined to form a feature dataset, which is then input into the progress prediction model to generate the progress completion percentage of each prediction sub-unit with a timestamp.

[0084] Specifically, by aligning the energy consumption heatmap with the BIM model's spatial coordinates, the effective energy consumption value of each grid cell can be accurately associated with the corresponding construction process, forming a spatial mapping relationship between energy consumption data and construction activities. The division of prediction sub-cells is based on process time nodes, and historical data samples are selected by combining similar geological conditions and construction processes to effectively eliminate the interference of working condition differences on model training. The application of the LSTM algorithm, through memory cells and gating mechanisms, processes time-series data, enabling dynamic learning of the complex correlation between construction progress and energy consumption characteristics. The final generated timestamped progress completion percentage provides an accurate time dimension reference for progress prediction in different construction areas.

[0085] Through the above technical solutions, this application enhances the reliability of the mapping between energy consumption data and construction efficiency by establishing a spatial correlation between energy consumption characteristic parameters and process links, improves the adaptability of the progress prediction model to complex working conditions through multi-dimensional feature extraction and dynamic sample screening, and realizes dynamic and accurate prediction of construction progress through LSTM algorithm processing, providing a reliable time dimension basis for resource scheduling.

[0086] This application further proposes the generation of material storage locations based on energy consumption thermal maps, specifically including:

[0087] Obtain the energy consumption heat map corresponding to the energy consumption heat value matrix, and calculate the gradient difference values ​​of the matrix in the horizontal and vertical directions respectively to obtain the energy consumption gradient magnitude. The energy consumption gradient magnitude refers to the intensity of energy consumption change in the spatial dimension reflected by matrix difference operation. Specifically, it can be achieved by using the Sobel operator to perform convolution calculation in the horizontal and vertical directions, which is used to identify the energy consumption abrupt boundary of the grid cell.

[0088] The energy diffusion coefficient is calculated based on the difference in energy consumption thermal values ​​between adjacent grid cells. The product of the energy diffusion coefficient and the energy gradient modulus is used as the regional energy consumption fluctuation rate. The energy diffusion coefficient refers to the propagation rate of energy consumption differences between adjacent grid cells. Specifically, it can be calculated by the ratio of the standard deviation to the mean of the thermal values ​​of adjacent grid cells, which is used to capture the dynamic trend of energy consumption changes during construction.

[0089] The resource replenishment priority index is obtained by weighting the regional energy consumption fluctuation rate with the energy consumption heat value of the corresponding grid cell. Based on the current progress completion percentage and the next process step, the time sequence demand of resources is judged. The resource replenishment priority is determined by comprehensively calculating with the current priority index. The resource replenishment priority index is a comprehensive evaluation index that takes into account real-time energy consumption fluctuation and future process demand. Specifically, a linear weighting method can be used to superimpose the fluctuation rate and heat value according to a preset ratio to form a priority evaluation mechanism for resource distribution.

[0090] Based on energy consumption heatmaps and resource replenishment priorities, a grid clustering and priority matching algorithm is used to generate material storage points. By analyzing the energy consumption characteristics and priority correlation of grid cells, the material storage points are divided into independent storage points and shared storage points. The grid clustering and priority matching algorithm refers to a method of grouping grid cells based on spatial proximity and priority similarity. Specifically, it can be implemented by combining K-means clustering with priority threshold judgment to generate a storage point layout that conforms to the resource distribution characteristics.

[0091] The environmental parameters of the construction area are obtained, and the spatial constraints of the material storage points are extracted. The spatial constraints include path accessibility, space occupancy limit, and safety compliance range setting. The spatial location of the material storage points is adjusted according to the spatial constraints. The spatial constraints refer to the physical environmental restrictions that the layout of the storage points must meet. Specifically, path planning algorithms can be used to detect accessibility, 3D scanning can be used to obtain space occupancy data, and safety specifications can be used to set buffer ranges to ensure that the layout plan meets the actual construction conditions.

[0092] Compared with existing technologies, the layout of existing material storage points is mostly based on static energy consumption data or experience judgment, lacking dynamic analysis of changes in energy consumption spatial gradient and temporal demand, resulting in low matching degree between storage points and construction hotspot areas.

[0093] Through the above technical solutions, this application can improve the spatial matching degree between the layout of material storage points and the energy-intensive zones of construction, ensuring the shortest possible resource scheduling path. A dynamic priority evaluation mechanism enables the storage point layout to adaptively adjust with the construction progress, avoiding secondary handling due to changes in work processes. The refined configuration of independent and shared storage points optimizes storage space utilization and reduces redundant material storage. Automatic verification of spatial constraints effectively avoids path obstacles and safety hazards within the construction area, enhancing the feasibility of the resource scheduling plan.

[0094] This application further proposes using BIM models to process schedule prediction data and generate structured material scheduling instructions, specifically including:

[0095] The progress completion percentage of each predicted sub-unit with timestamps is imported into the tunnel construction BIM model. Combined with the material consumption quota parameters of each process step, the material requirements of each predicted sub-unit under the corresponding timestamp are calculated. The progress completion percentage with timestamps refers to dynamic data that binds the construction progress prediction results to specific time nodes. This can be achieved using time series marking methods to establish a precise correspondence between material requirements and construction sequence. The material consumption quota parameters refer to the material quantity required per unit progress of various processes pre-set in the BIM model. This can be obtained through statistics from historical project data and is used to convert the remaining progress percentage into specific material quantities.

[0096] Based on the spatial coordinate correspondence between the predicted sub-units, grid units and the construction area, the material requirements of each predicted sub-unit are divided according to the spatial proportion of the grid unit to obtain the material requirements of each grid unit at the corresponding timestamp.

[0097] Based on the correspondence between grid cells and material storage points, the material requirements of each grid cell are allocated to the corresponding material storage points;

[0098] The material dispatch instructions are generated by integrating the allocated demand of each material storage point under different timestamps, according to material type, time sequence and storage point number. The material dispatch instructions include material type, dispatch quantity, delivery time and target storage point information.

[0099] Specifically, by combining timestamped progress data with process parameters in the BIM model, a synchronization relationship between material demand and construction pace is first established in the time dimension. The material demand of each predicted sub-unit is broken down into finer-grained grid units, achieving coupled calculation of construction progress and spatial distribution. Based on the grid mapping relationship of storage point service range, material demand is accurately allocated to the nearest or optimal storage point, forming a dynamic adaptation of spatial distribution and storage layout. Finally, through time-dimensional orchestration and storage point classification coding, structured instructions containing information on material type, scheduling quantity, delivery time, and target storage point are generated, ensuring that material scheduling simultaneously meets the requirements of time accuracy and spatial accessibility.

[0100] Compared to existing technologies, traditional material scheduling methods typically estimate total demand based on overall progress, lacking analysis of spatial demand differences within the construction area. This can easily lead to localized material shortages or redundancies. Furthermore, existing technologies often use fixed-location planning for material storage points, failing to dynamically adapt to changes in spatial demand as construction progresses. This solution establishes spatial service relationships between grid cells and material storage points through coupled calculations of progress forecasting and spatial grids. This improves the rationality and accessibility of material delivery paths, enabling scheduling instructions to match real-time spatial demand shifts during construction.

[0101] See Figure 4 As shown, this solution proposes a digital twin prediction and resource scheduling system for tunnel construction progress, used to implement the aforementioned digital twin prediction and resource scheduling method for tunnel construction progress, including:

[0102] The data acquisition and model building module is used to collect real-time monitoring data of the tunnel construction area, including equipment location information and quantitative data of process progress, and to build a BIM model of the tunnel construction area.

[0103] The equipment spatiotemporal feature extraction module is used to divide the equipment location information into multiple continuous time windows according to the time nodes of the BIM model. Within each time window, the module obtains the equipment operating parameters and corresponding energy consumption values, extracts the spatial coordinates of the equipment location information and the corresponding dwell time, and generates a movement trajectory based on the temporal changes of the spatial coordinates.

[0104] The operation type determination and energy consumption analysis module is used to determine the equipment operation type based on the characteristic combination of equipment operating parameters, energy consumption values ​​and movement trajectory. Based on the equipment operation type, combined with changes in geological conditions and construction procedures, the module calculates the effective energy consumption value of the equipment and constructs an energy consumption heat map of the construction area.

[0105] The progress prediction module is used to combine the characteristic parameters of the energy consumption thermal map with the quantitative data of the process progress, train the progress prediction model through the LSTM algorithm, and generate progress prediction data.

[0106] The material scheduling module is used to generate material storage points based on energy consumption thermal maps, process progress prediction data using BIM models, and generate structured material scheduling instructions.

[0107] An electronic device, comprising:

[0108] processor;

[0109] A memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to execute the aforementioned digital twin prediction and resource scheduling method for tunnel construction progress.

[0110] A computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the aforementioned method for digital twin prediction and resource scheduling of tunnel construction progress.

[0111] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for digital twin prediction and resource scheduling of tunnel construction progress, characterized in that, The method includes: Collect real-time monitoring data of the tunnel construction area, including equipment location information and quantitative data of process progress, and construct a BIM model of the tunnel construction area. The equipment location information is divided into multiple continuous time windows according to the time nodes of the BIM model. Within each time window, the equipment operating parameters and corresponding energy consumption values ​​are obtained, the spatial coordinates of the equipment location information and the corresponding dwell time are extracted, and the movement trajectory is generated based on the temporal changes of the spatial coordinates. The equipment operation type is determined by the combination of equipment operating parameters, energy consumption values ​​and movement trajectory characteristics. Based on the equipment operation type, combined with changes in geological conditions and construction procedures, the effective energy consumption value of the equipment is calculated, and an energy consumption heat map of the construction area is constructed. By combining the characteristic parameters of the energy consumption thermogram with the quantitative data of the process progress, a progress prediction model is trained using the LSTM algorithm to generate progress prediction data. Material storage points are generated based on energy consumption thermal maps. The BIM model is used to process the progress prediction data and generate structured material scheduling instructions.

2. The method according to claim 1, characterized in that, The method of determining the equipment operation type based on a combination of features including equipment operating parameters, energy consumption values, and movement trajectory specifically includes: The equipment operating parameters are normalized to generate operating status feature values. Based on the operating status feature values, the equipment operating status is divided into strong operating features, weak operating features, and non-operating features. Obtain historical valid operating data of the equipment under the same working conditions, calculate the energy consumption matching degree between the real-time energy consumption value and the historical valid operating data, and classify the energy consumption status of the equipment into high energy consumption matching, medium energy consumption matching and low energy consumption matching according to the matching degree. Extract the dynamic features of the movement trajectory, generate trajectory activity feature values, and classify the device trajectory status into strong trajectory activity, weak trajectory activity, and no trajectory activity based on the trajectory activity feature values; A combination relationship determination rule is set to determine the job type by the combination relationship between the operating status, energy consumption status and trajectory status. The job type includes valid job and invalid standby.

3. The method according to claim 2, characterized in that, The calculation of the effective energy consumption value of the equipment based on the equipment operation type, combined with changes in geological conditions and construction procedures, specifically includes: The time window is divided according to the type of work, and the time segments of effective work are filtered out. The time segments are then divided into several time sub-units according to the stability of energy consumption values. Preset geological condition levels and corresponding level weights, and associate process benchmark values ​​with time sub-units according to construction procedures; The historical energy consumption values ​​of the equipment under similar geological conditions and construction procedures are obtained along with the corresponding construction progress data. An outlier is removed by a multi-dimensional outlier filtering mechanism. For the selected historical energy consumption values ​​and the corresponding construction progress data, a polynomial curve fitting method is used to establish the mapping relationship between energy consumption values ​​and initial efficiency coefficients. The grade weight corresponding to the current geological conditions and the process benchmark value corresponding to the construction process are used as correction factors to dynamically correct the initial efficiency coefficients. An attenuation coefficient is configured for each time sub-unit. The attenuation coefficient is a function value related to the interval between the time sub-unit and the current time. The longer the interval, the smaller the value of the attenuation coefficient. The duration of each time sub-unit, the corresponding attenuation coefficient, the energy consumption value, and the corrected efficiency coefficient are multiplied sequentially. The product results of all time sub-units are then summed to obtain the effective energy consumption value of the corresponding time window.

4. The method according to claim 1, characterized in that, The energy consumption thermal map of the construction area specifically includes: A dynamic adaptive mechanism is used to divide the construction area into several grid units. The initial grid is of a preset size. If the effective energy consumption value of an initial grid is higher than the first density threshold, it is automatically refined into a grid of a smaller preset size. If the effective energy consumption value of adjacent initial grids is lower than the second density threshold for two consecutive time windows, they are automatically merged into a grid of a larger preset size. The spatial coordinates of the equipment are matched to the corresponding grid cells, the boundaries of the grid cells are aligned with the construction process zones, and the energy consumption heat value of each grid cell is calculated based on the effective energy consumption. The kernel density estimation method is used to convert the discrete grid cell energy consumption and heat values ​​into a continuous spatial distribution. The energy consumption and heat values ​​of neighboring grid cells are weighted and averaged using a Gaussian kernel to generate a smooth energy consumption and heat map, which is updated according to a preset period.

5. The method according to claim 4, characterized in that, The step of combining energy consumption thermal map feature parameters with process progress quantification data, and training a progress prediction model using the LSTM algorithm to generate progress prediction data, specifically includes: Align the energy consumption heat map with the tunnel construction BIM model in terms of spatial coordinates, so that the effective energy consumption value of each grid unit in the energy consumption heat map corresponds to the process link in the BIM model. Extract the feature parameters of the energy consumption heat map, including the effective energy consumption value of each grid cell, the energy consumption spatial distribution gradient and equipment distribution information, and associate the process progress quantification data of the corresponding time nodes of the process links to generate a feature dataset; The construction area is divided into several prediction sub-units according to the time nodes of the process. For each prediction sub-unit, the energy consumption heat map characteristic parameters and construction progress data under similar geological conditions and construction processes are selected as training samples. After removing abnormal samples, a sample library is constructed. The LSTM algorithm is used, with the feature dataset of the prediction sub-unit as input and the time-series progress completion percentage of the corresponding prediction sub-unit as output. The model is trained through a sample database to obtain the progress prediction model. The energy consumption heat map feature parameters and process progress quantification data of the current time node are combined to form a feature dataset, which is then input into the progress prediction model to generate the progress completion percentage of each prediction sub-unit with a timestamp.

6. The method according to claim 5, characterized in that, The material storage points generated based on the energy consumption thermal map specifically include: Obtain the energy consumption thermodynamic value matrix corresponding to the energy consumption thermodynamic map, and calculate the gradient difference values ​​of the matrix in the horizontal and vertical directions respectively to obtain the energy consumption gradient magnitude. The energy diffusion coefficient is calculated based on the difference in energy and thermal values ​​between adjacent grid cells, and the product of the energy diffusion coefficient and the energy gradient magnitude is used as the regional energy fluctuation rate. The resource replenishment priority index is obtained by weighting the regional energy consumption fluctuation rate with the energy consumption heat value of the corresponding grid cell. The timing demand for resources is determined based on the current progress completion percentage and the next process step. The resource replenishment priority is determined by comprehensively calculating with the current priority index. Based on the energy consumption heat map and resource replenishment priority, a grid clustering and priority matching algorithm is used to generate material storage points. By analyzing the energy consumption characteristics and priority correlation of grid cells, the material storage points are divided into independent storage points and shared storage points. The environmental parameters of the construction area are obtained, and the spatial constraints of the material storage points are extracted. The spatial constraints include path accessibility, space occupancy limit and safety compliance range setting. The spatial location of the material storage points is adjusted according to the spatial constraints.

7. The method according to claim 6, characterized in that, The process of using the BIM model to process the schedule prediction data and generate structured material scheduling instructions specifically includes: Import the progress completion percentage of each predicted sub-unit with timestamps into the tunnel construction BIM model, and combine it with the material consumption quota parameters of each process step to calculate the material requirements of each predicted sub-unit under the corresponding timestamp. Based on the spatial coordinate correspondence between the predicted sub-units, grid units and the construction area, the material requirements of each predicted sub-unit are divided according to the spatial proportion of the grid unit to obtain the material requirements of each grid unit at the corresponding timestamp. Based on the correspondence between grid cells and material storage points, the material requirements of each grid cell are allocated to the corresponding material storage points; The material dispatch instructions are generated by integrating the allocated demand of each material storage point under different timestamps, according to material type, time sequence and storage point number. The material dispatch instructions include material type, dispatch quantity, delivery time and target storage point information.

8. A digital twin prediction and resource scheduling system for tunnel construction progress, characterized in that: The method for implementing digital twin prediction and resource scheduling of tunnel construction progress as described in any one of claims 1-7 includes: The data acquisition and model building module is used to collect real-time monitoring data of the tunnel construction area, including equipment location information and quantitative data of process progress, and to build a BIM model of the tunnel construction area. The equipment spatiotemporal feature extraction module is used to divide the equipment location information into multiple continuous time windows according to the time nodes of the BIM model. Within each time window, the module obtains the equipment operating parameters and corresponding energy consumption values, extracts the spatial coordinates of the equipment location information and the corresponding dwell time, and generates a movement trajectory based on the temporal changes of the spatial coordinates. The operation type determination and energy consumption analysis module is used to determine the equipment operation type based on the characteristic combination of equipment operating parameters, energy consumption values ​​and movement trajectory. Based on the equipment operation type, combined with changes in geological conditions and construction procedures, the module calculates the effective energy consumption value of the equipment and constructs an energy consumption heat map of the construction area. The progress prediction module is used to combine the characteristic parameters of the energy consumption thermal map with the quantitative data of the process progress, train the progress prediction model through the LSTM algorithm, and generate progress prediction data. The material scheduling module is used to generate material storage points based on energy consumption thermal maps, process progress prediction data using BIM models, and generate structured material scheduling instructions.

9. An electronic device, characterized in that, include: processor; A memory for storing processor-executable instructions; wherein the processor is configured to invoke the instructions stored in the memory to perform the method according to any one of claims 1-7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1-7.

Images