A design optimization method and system for an expressway intelligent beam yard based on big data

Abstract

The application is suitable for the technical field of intelligent beam yard, and provides a design optimization method and system for an intelligent beam yard of an expressway based on big data. The method collects relevant data of transportation equipment through an Internet of Things and a real-time bidirectional data interaction channel, constructs and dynamically updates a three-dimensional digital twin model of the site; fuses the model and broadcast information of adjacent nodes, and uses an improved pheromone diffusion algorithm to output an optimal real-time driving path; sets a hierarchical processing protocol for different priority tasks, recalculates an optimal path sequence and adjusts task allocation when a resource conflict or path deadlock is detected; dynamically adjusts the path and equipment parameters in combination with frequency domain analysis to identify potential risks; and continuously optimizes the algorithm through offline training and online learning of a historical database. The system includes a model construction and updating module, a path selection module, a conflict detection module, a path adjustment module, and an operation optimization module, realizes efficient scheduling, risk early warning, and continuous optimization of the intelligent beam yard, and improves operation efficiency and safety.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent beam yard, and particularly relates to a design optimization method and system for an intelligent beam yard of an expressway based on big data. BACKGROUND

[0002] With the rapid development of expressway construction in China, the operation efficiency and safety of a beam yard, which is a core site for the production, storage and transportation of precast beams, directly affect the progress of a project. In recent years, the integration of technologies such as big data, the Internet of Things and artificial intelligence promotes the transformation of beam yards to intelligence, and intelligent beam yards realize intelligent scheduling and accurate decision-making through the digital management of information such as equipment, materials and environment, which becomes a development trend in the industry. The path optimization, task collaboration and risk control of transportation equipment are key links in the design of intelligent beam yards.

[0003] At present, there are schemes for scheduling beam yard equipment based on sensor data collection in the related art, which provide driving routes for transportation equipment through simple path planning algorithms and set a basic task priority mechanism. Some schemes introduce static digital models to assist management, and combine historical data to perform limited optimization on path selection, but mainly rely on preset rules or a single objective (such as the shortest path) for decision-making, and lack dynamic responses to real-time traffic flow, dynamic states of equipment and complex task conflicts.

[0004] The existing technology has the problems of static path planning, which cannot adapt to the movement of beam yard equipment, changes in traffic flow and distribution of obstacles in real time, resulting in path congestion or deadlock; an imperfect task priority processing mechanism, which lacks efficient dynamic resolution strategies when resource request conflicts occur, affecting overall scheduling efficiency; a lag in identifying potential risks (such as abnormal vibration of equipment and road hazards) during transportation, which mainly relies on manual inspection and is difficult to provide timely warnings; and a lack of continuous self-optimization ability of the algorithm, which gradually decreases adaptability during long-term operation, making it difficult to meet the efficient, safe and self-adaptive operation needs of intelligent beam yards. SUMMARY

[0005] The purpose of the present application is to provide a design optimization method for an intelligent beam yard of an expressway based on big data, which aims to solve the technical problems existing in the prior art identified in the background.

[0006] The present application is implemented as follows: a design optimization method for an intelligent beam yard of an expressway based on big data, the method comprising:

[0007] Through the Internet of Things network and the real-time bidirectional data interaction channel, real-time collection of transportation equipment related data is performed, and a site three-dimensional digital twin model reflecting the real-time physical state of the beam yard is constructed and dynamically updated;

[0008] Through a real-time bidirectional data interaction channel, broadcast information of adjacent transport nodes is received, and a three-dimensional digital twin model of the site is fused, an improved pheromone diffusion algorithm is used to calculate a real-time selection probability distribution of a connection path between nodes, and a local optimal path probability distribution obtained through screening is used to output an optimal real-time driving path for each transport device;

[0009] A hierarchical processing protocol is set for transport tasks of different priorities, when a new or changed task instruction is received, active scanning is triggered, each transport task is detected, when a resource request conflict or path deadlock is detected, a three-dimensional digital twin model of the site is called, an optimal path sequence of the affected transport device is recalculated in combination with the optimal real-time driving path, and the task allocation is adjusted;

[0010] In combination with the recalculated optimal path sequence, frequency domain analysis is performed on the collected transport device related data, potential risks are identified, and the optimal path sequence and device operating parameters are dynamically adjusted;

[0011] A historical operation database is established, transport device related data is continuously recorded, and the recorded transport device related data is offline trained and online learned to continuously optimize the improved pheromone diffusion algorithm.

[0012] As a further scheme of the present application, the three-dimensional digital twin model of the site reflecting the real-time physical state of the beam yard is constructed and dynamically updated, and specifically comprises:

[0013] Transport device related data is continuously transmitted through a real-time bidirectional data interaction channel at a sampling frequency of 5 times per second;

[0014] The transport device related data includes spatial coordinates, load information, residual energy state, running speed and direction parameters, material state information of the precast beam, and site environment index data;

[0015] An initial three-dimensional grid map of the site is generated, the real-time collected transport device related data is mapped to the grid vertex, the device position error is dynamically corrected and the obstacle distribution is updated, and a three-dimensional digital twin model of the site is generated;

[0016] A material flow state machine model is established, the pedestal occupation state is calculated in real time according to the precast beam hoisting progress, and the material flow logic relationship is synchronized to the topological structure of the three-dimensional digital twin model of the site.

[0017] As a further scheme of the present application, the local optimal path probability distribution obtained through screening is used to output an optimal real-time driving path for each transport device, and specifically comprises:

[0018] Based on the path topology structure provided by the site three-dimensional digital twin model, a dynamic adjacency matrix is constructed to describe the relationship between the nodes of the transportation network, and the weight of the matrix elements is determined by the Euclidean distance between the nodes, the path turning angle cost and the real-time traffic flow density. Every 200ms, the adjacent transportation node broadcast information of the adjacent nodes within a radius of 50 meters is obtained through the real-time bidirectional data interaction channel;

[0019] The adjacent transportation node broadcast information includes: load information, destination coordinates, residual energy state information;

[0020] In the improved pheromone diffusion algorithm, a dynamic evaporation factor is introduced to calculate the real-time selection probability distribution of the connection path between the transportation nodes.

[0021] From the real-time selection probability distribution, candidate paths are selected, and the score of each path is calculated based on a multi-objective evaluation function, and the path with the highest score is taken as the optimal real-time driving path.

[0022] As a further scheme of the application, the optimal path sequence of the affected transportation device is recalculated based on the optimal real-time driving path, and the task allocation is adjusted, specifically including:

[0023] The priority rating of each transportation task is obtained, and a three-level task response mechanism is established.

[0024] The current motion state of each transportation node is continuously monitored, and when the speed of two or more transportation nodes is continuously lower than 0.5m / s within 200ms and the target path overlap degree is greater than 85%, it is determined as a deadlock state, and a conflict resolution flag is triggered.

[0025] The conflict resolution flag is analyzed, the optimal real-time driving path of the affected transportation device is discretized, the path update instruction set is obtained, and is transmitted to the affected transportation device through the real-time bidirectional data interaction channel as the recalculated optimal path sequence.

[0026] As a further scheme of the application, the three-level task response mechanism includes:

[0027] For priority <3, it is defined as a background task, and when the system idle rate is >40%, the delay queue task is activated by the central coordinator and the redundant path is allocated.

[0028] For 3≤priority <7, it is defined as a regular task, and a distributed negotiation protocol is triggered, and the optimal real-time driving path is allocated between nodes through a bidding game.

[0029] For priority ≥7, it is defined as an emergency task, and the central coordinator is activated to intervene and perform global resource matching.

[0030] As a further scheme of the present application, the identifying potential risks and dynamically adjusting the optimal path sequence and the device operation parameter specifically comprises:

[0031] The collected 0.5-200Hz vibration signals are subjected to wavelet denoising, and are converted to a frequency domain through fast Fourier transform;

[0032] The characteristic frequency amplitude is identified and a safety threshold is set, and a pre-warning is triggered when the characteristic frequency amplitude exceeds the safety threshold;

[0033] A speed reduction instruction is issued to the pre-warning device, and the area with a sudden increase in low-frequency energy displayed in the frequency domain analysis is avoided based on the three-dimensional digital twin model of the site, and a detour path is re-planned.

[0034] Another object of the present application is to provide a design optimization system for a highway intelligent beam yard based on big data, which comprises:

[0035] A model construction and updating module is configured to collect relevant data of the transportation device in real time through an Internet of Things network and a real-time bidirectional data interaction channel, construct and dynamically update a three-dimensional digital twin model of the site reflecting the real-time physical state of the beam yard;

[0036] A path selection module is configured to receive broadcast information of adjacent transportation nodes through the real-time bidirectional data interaction channel, and fuse the three-dimensional digital twin model of the site, calculate the real-time selection probability distribution of the connection path between nodes by using an improved pheromone diffusion algorithm, and filter the obtained local optimal path probability distribution to output an optimal real-time driving path for each transportation device;

[0037] A conflict detection module is configured to set a hierarchical processing protocol for transportation tasks of different priorities, trigger active scanning when receiving a new or changed task instruction, detect each transportation task, and when a resource request conflict or path deadlock is detected, call the three-dimensional digital twin model of the site, combine the optimal real-time driving path, recalculate the optimal path sequence of the affected transportation device, and adjust the task allocation;

[0038] A path adjustment module is configured to combine the recalculated optimal path sequence, perform frequency domain analysis on the collected transportation device related data, identify potential risks and dynamically adjust the optimal path sequence and the device operation parameter;

[0039] An operation optimization module is configured to establish a historical operation database, continuously record the transportation device related data, and perform offline training and online learning on the recorded transportation device related data to continuously optimize the improved pheromone diffusion algorithm.

[0040] The present application has the following beneficial effects:

[0041] The scheme of the present application realizes the design optimization of the intelligent beam yard of the expressway through the coordinated linkage of multiple technical modules, and systematically realizes the design optimization of the intelligent beam yard of the expressway. Based on the Internet of Things and real-time two-way data interaction, a three-dimensional digital twin model of the site is constructed and dynamically updated. Through high-frequency collection of the spatial coordinates, load, energy state and other data of the transportation equipment, combined with the material flow state machine model, the physical state of the beam yard, the equipment movement and the material flow logic are mapped to the digital space in real time, providing accurate dynamic benchmarks for all subsequent optimization links, solving the problem of disconnection between physical scene and management decision in traditional beam yard design. In the path planning link, the digital twin model and the real-time broadcast information of the adjacent nodes are combined, and the improved pheromone diffusion algorithm is used to introduce the dynamic evaporation factor to adapt to the traffic flow changes. Through the multi-objective evaluation function, the path length, energy consumption and congestion cost are balanced, so that the path selection is upgraded from static preset to dynamic adaptation, avoiding the local inefficiency caused by single target planning. For task scheduling, a three-level response mechanism is established to differentiate between background, regular and emergency tasks according to priority. Combined with real-time deadlock detection (judged by speed and path overlap) and discrete path update, resource conflicts are quickly resolved to ensure efficient and orderly task execution rhythm and improve overall resource utilization. In terms of risk control, the vibration signals of the transportation equipment are analyzed in the frequency domain, and the abnormal characteristic frequencies are identified after wavelet denoising and Fourier transform, so as to early warn of equipment failure or path hazards, and adjust the speed and path through the digital twin model, so as to change the risk control from passive treatment to active prevention. At the same time, relying on the historical operation database, the algorithm core parameters are optimized through offline training, and the real-time changes are adapted through online learning, so that the improved pheromone diffusion algorithm continues to evolve and adapts to the dynamic scenes such as beam yard equipment aging and task upgrading. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 A flowchart of a design optimization method of an intelligent beam yard of an expressway based on big data provided for an embodiment of the present application;

[0043] Figure 2 A flowchart of constructing and dynamically updating a three-dimensional digital twin model of a site reflecting the real-time physical state of a beam yard provided for an embodiment of the present application;

[0044] Figure 3 A flowchart of outputting an optimal real-time driving path for each transportation equipment provided for an embodiment of the present application;

[0045] Figure 4 A flowchart of recalculating the optimal path sequence of the affected transportation equipment provided for an embodiment of the present application;

[0046] Figure 5 A flowchart of identifying potential risks and dynamically adjusting the optimal path sequence and equipment operation parameters provided for an embodiment of the present application;

[0047] Figure 6A structural block diagram of a design optimization system of a highway intelligent beam yard based on big data is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the present application clearer and more apparent, the present application will be further described in detail below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0049] Figure 1 A flowchart of a design optimization method of a highway intelligent beam yard based on big data is provided for an embodiment of the present application, as shown in Figure 1 The method comprises:

[0050] S100, real-time collection of transportation equipment related data through an Internet of Things network and a real-time bidirectional data interaction channel, construction and dynamic updating of a site three-dimensional digital twin model reflecting a real-time physical state of the beam yard;

[0051] The transportation equipment related data is continuously transmitted at a sampling frequency of 5 times per second. This frequency is selected in order to ensure data real-time while avoiding excessive consumption of network and computing resources. The moving speed of the transportation equipment in the beam yard is relatively stable. An excessively high sampling frequency will increase the data processing pressure, while an excessively low sampling frequency may miss key changes in the equipment state. The frequency of 5 times per second can accurately capture the dynamic changes of the equipment and provide timely data support for model updating. The collected transportation equipment related data is very comprehensive, covering spatial coordinates, load information, remaining energy state, running speed and direction parameters. These data directly reflect the position, load capacity, endurance and motion trend of the equipment. At the same time, it also includes material state information of the precast beam, such as pouring progress, curing state, etc., as well as site environment index data, such as temperature, humidity, wind speed, etc. These data together constitute the basic information of the model, ensuring that the model can completely reproduce the physical state of the beam yard from multiple dimensions of equipment, materials and environment.

[0052] An initial site three-dimensional grid map is generated, and the real-time collected data is mapped to the grid vertices. In this process, the equipment position error will be dynamically corrected. Since the positioning equipment may have a slight drift, the collected spatial coordinates are calibrated by an algorithm to ensure that the position of the equipment in the digital model is highly consistent with the physical position. At the same time, the real-time position of the equipment and static obstacles is updated to generate the site three-dimensional digital twin model. The purpose of this step is to achieve accurate correspondence between the digital model and the physical site in the spatial dimension, providing a reliable spatial reference for subsequent model-based path planning, conflict detection and other operations.

[0053] In addition, a material flow state machine model is established, which calculates the pedestal occupation state in real time according to the precast beam hoisting progress. When the precast beam of a pedestal is completed and hoisted and removed by the transportation equipment, the state machine model will immediately mark the pedestal as idle, and synchronize the logical relationship of the material flow to the topological structure of the three-dimensional digital twin model. In this way, the model not only reflects the physical space state of the beam field, but also integrates the dynamic information of the production process, so that the digital twin model becomes the link between the physical space and the production process, and realizes the comprehensive mapping of the beam field operation.

[0054] As shown in Figure 2 , the construction and dynamic updating of the site three-dimensional digital twin model reflecting the real-time physical state of the beam field specifically includes:

[0055] S110, continuously transmitting transportation equipment related data through a real-time bidirectional data interaction channel at a sampling frequency of 5 times per second;

[0056] S120, the transportation equipment related data includes: spatial coordinates, load information, remaining energy state, running speed and direction parameters, material state information of precast beams, and site environment index data;

[0057] S130, generating an initial site three-dimensional grid map, mapping the real-time collected transportation equipment related data to the grid vertex, dynamically correcting the equipment position error and updating the obstacle distribution, and generating a site three-dimensional digital twin model;

[0058] S140, establishing a material flow state machine model, calculating the pedestal occupation state in real time according to the precast beam hoisting progress, and synchronizing the logical relationship of the material flow to the topological structure of the site three-dimensional digital twin model.

[0059] S200, through the real-time bidirectional data interaction channel, receiving the broadcast information of the adjacent transportation nodes, and fusing the site three-dimensional digital twin model, using an improved pheromone diffusion algorithm to calculate the real-time selection probability distribution of the connection path between nodes, and screening the obtained local optimal path probability distribution to output the optimal real-time driving path for each transportation equipment;

[0060] The dynamic adjacency matrix is constructed based on the path topology structure provided by the three-dimensional digital twin model of the site, to describe the relationship between nodes in the transportation network. The weight of the matrix element is not a fixed value, but is determined by the Euclidean distance between nodes, the path turning angle cost and the real-time traffic flow density. The Euclidean distance reflects the basic length of the physical path and is the basic reference for path selection. The path turning angle cost takes into account the special nature of the precast beam transportation: the difficulty of operation of the super-long or super-heavy precast beam when turning, the greater the turning angle, the higher the additional cost. This design can reduce the equipment damage or material safety risk caused by sharp turns. The real-time traffic flow density is used to measure the congestion degree of the path to avoid the device driving into the saturated road section.

[0061] To ensure that the matrix can reflect the changes in the site in real time, the broadcast information of the adjacent nodes within a radius of 50 meters is obtained through the real-time bidirectional data interaction channel every 200 ms. The setting of the frequency and range ensures the timeliness of the information (to avoid the disconnection between path planning and reality due to data lag) and controls the complexity of data processing (focusing on the local environment to improve computing efficiency), so that the matrix can dynamically adapt to the real-time movement state of the equipment in the site.

[0062] After obtaining the broadcast information of the adjacent transportation nodes (including load information, destination coordinates, remaining energy state, etc.), these information is fused with the three-dimensional digital twin model, and the real-time selection probability distribution of the connected paths between nodes is calculated through the improved pheromone diffusion algorithm.

[0063] The introduction of the dynamic evaporation factor is the core improvement point of this algorithm, which allows the pheromone concentration to be dynamically adjusted according to the traffic state. When the traffic flow density of a road section is high, the dynamic evaporation factor increases, accelerating the evaporation of the pheromone on that road section, thereby reducing the probability of subsequent equipment selecting that path. This design breaks the limitation of fixed evaporation rate in traditional pheromone algorithms, allowing path selection to actively avoid congestion and adapt to the dynamic changes of equipment flow in the beam field. At the same time, the calculation of path selection probability combines the pheromone concentration and the heuristic factor (inversely proportional to the Euclidean distance), and balances the influence of the two through a weight coefficient. In emergency tasks, the weight of the heuristic factor is increased, making the algorithm more inclined to refer to the historical optimal path, to improve the efficiency of path planning in emergency situations.

[0064] ​Subsequently, candidate paths are selected from the path selection probability distribution, and the comprehensive score of each path is calculated based on a multi-objective evaluation function. The path with the highest score (i.e., the overall optimal path) is then selected as the optimal real-time driving path. The multi-objective evaluation function comprehensively considers path length, energy cost, and congestion cost: the standardized path length ensures comparability between paths with different start and end points, avoiding misjudgments due to differences in basic distances; the standardized energy cost combines factors such as precast beam load, path length, and number of turns, reflecting considerations for transportation economy; and the standardized congestion cost, through weighted calculation of traffic flow density across each segment of the path, ensures that the path avoids congestion nodes. This multi-objective trade-off approach avoids the high energy consumption or high congestion risks that may result from a single objective (pursuing only the shortest path), making path selection more aligned with the actual operational needs of the beam yard.

[0065] like Figure 3 As shown, the locally optimal path probability distribution obtained through the screening provides the optimal real-time driving path for each transportation device, specifically including:

[0066] S210, Based on the path topology provided by the three-dimensional digital twin model of the site, a dynamic adjacency matrix is ​​constructed to describe the relationship between nodes in the transportation network. The weight of the matrix elements is determined by the Euclidean distance between nodes, the path turning angle cost, and the real-time traffic flow density. Every 200ms, the broadcast information of neighboring transportation nodes within a radius of 50 meters is obtained through the real-time bidirectional data interaction channel.

[0067] For constructing a dynamic adjacency matrix:

[0068] Let the set of nodes in the transportation network be... Dynamic adjacency matrix The element is defined as:

[0069] ;

[0070] in:

[0071] Represents a node To the node The path weight value is used to describe the connection cost between two nodes;

[0072] Represents a node With nodes The Euclidean distance (unit: meters) between them reflects the physical path length, which in the beam yard scenario corresponds to the straight-line distance between the current positions of two transport devices or between fixed nodes such as pedestals and hoisting points;

[0073] Represents the path turning angle cost (in radians), with values ​​taken from nodes. To the node The cost of deviation between the steering angle and the standard straight path, such as the additional cost of steering difficulty caused by the excessive length of precast beams when transport equipment turns in the beam yard.

[0074] It represents the real-time traffic flow density (unit: vehicles / 100 meters), reflecting the current congestion level of the route, and is calculated based on the number of transportation equipment operating within 50 meters of the road segment during the same period in the three-dimensional digital twin model of the site.

[0075] These are weighting coefficients, dynamically adjusted by the beam yard scheduling strategy, increasing during heavy-load transportation. To reduce sharp turns and increase congestion periods Prioritize routes with good traffic flow.

[0076] The broadcast information from the nearby transportation nodes includes: load information, destination coordinates, and remaining energy status information;

[0077] S220 introduces a dynamic evaporation factor into the improved pheromone diffusion algorithm to calculate the real-time selection probability distribution of connection paths between transport nodes:

[0078] Pheromon concentration update formula:

[0079] set up for Time Node To the node Pheromones concentration, after the introduction of dynamic volatile factors:

[0080] ;

[0081] Path selection probability formula:

[0082] ;

[0083] in:

[0084] Represents the dynamic volatility factor (0 < <1), the calculation formula is: ,in For real-time traffic flow density, To achieve the maximum carrying density of the path, pheromone volatilization is accelerated during congestion to reduce the probability of path selection;

[0085] Indicates the first Taiwan transport equipment route The increase in pheromones left behind is positively correlated with the remaining energy of the equipment.

[0086] Indicates the first Station device selection path Probability of the path;

[0087] is the heuristic factor, inversely proportional to the Euclidean distance;

[0088] represents the Station device can select the set of adjacent nodes (filtered based on a 50-meter communication radius);

[0089] is the weight coefficient of pheromone and heuristic factor (default is 0.5 respectively, and is increased during emergency tasks Priority reference to historical optimal path);

[0090] represents Node Pheromone concentration from node to node reflects the degree of advantage of the path from node to node in historical scheduling, the higher the pheromone concentration, the better the path performed in historical scheduling; represents Heuristic factor from node to node is the real-time basis for guiding path selection, the closer the distance, the greater the heuristic factor, the higher the basic probability of selecting this path.

[0091] S230, filtering candidate paths from the real-time selection probability distribution, calculating the score of each path based on the multi-objective evaluation function, and taking the path with the highest score as the optimal real-time driving path:

[0092] Let the candidate path set be , then the score of the th path is:

[0093] ;

[0094] Where:

[0095] represents the comprehensive score of path (the smaller the value, the better);

[0096] represents the normalized value of path length, reflecting the physical distance cost of the path, where, represents the The actual physical length (unit: meter) of the strip candidate path, calculated by the path topology data of the three-dimensional digital twin model of the site, that is, the sum of the Euclidean distances between all consecutive nodes in the path, The shortest path length (unit: meter) between the starting point and the end point, based on the dynamic adjacency matrix The theoretical minimum path cost of the transportation task is calculated by Dijkstra algorithm;

[0097] The normalized value of energy consumption cost, the calculation formula is , wherein The load of the precast beam (tons), The path length, The turning energy consumption coefficient, The number of turns, The reference energy consumption;

[0098] The normalized value of congestion cost, calculated based on the traffic flow density of each road segment in the path, , The length of the road section;

[0099] The actual physical length of the road section between node and node in the transportation network;

[0100] The weight.

[0101] S300, set a hierarchical processing protocol for transportation tasks of different priorities, when receiving a new or changed task instruction, trigger active scanning, detect each transportation task, when detecting resource request conflict or path deadlock, call the three-dimensional digital twin model of the site, combine the optimal real-time driving path to recalculate the optimal path sequence of the affected transportation equipment, and adjust the task allocation;

[0102] The priority rating of each transportation task is obtained, and a three-level task response mechanism is established accordingly. Based on the difference between the task urgency and the resource demand, differentiated allocation of resources is realized to avoid the decline of overall efficiency due to resource contention. For tasks with a priority lower than 3, which are usually less time-sensitive (such as daily maintenance material transportation), the system will put them into a delay queue, and only when the system idle rate exceeds 40%, will the central coordinator activate and allocate redundant paths. The purpose of this is to make full use of idle resources without affecting core tasks, and to avoid resource waste. For regular tasks with a priority between 3 and 7 (such as daily precast beam transfer), the system will trigger a distributed negotiation protocol, allowing nodes involved in the task to allocate optimal real-time driving paths through a bidding game. This approach can reduce the pressure on central coordination, improve decision-making efficiency through autonomous negotiation between nodes, and ensure relative fairness in path allocation. For urgent tasks with a priority not lower than 7 (such as emergency precast beam supply), the central coordinator is activated to intervene in global resource matching. This is because urgent tasks often affect the overall construction progress and require global resource allocation to ensure task priority and avoid suboptimal results from local decisions.

[0103] During task execution, the current motion state of each transportation node is continuously monitored to determine whether there is a resource request conflict or path deadlock. When it is detected that the speed of two or more transportation nodes is continuously lower than 0.5 m / s within 200 ms, and the target path overlap degree exceeds 85%, the system will determine that it is in a deadlock state and trigger a conflict resolution flag. The basis for this determination logic is that a very low speed of a transportation node indicates that its motion is blocked, and a high path overlap degree suggests that the nodes may be blocked by each other, causing them to stop moving. If not handled in time, it will cause a chain reaction, leading to more nodes stopping. A 200 ms monitoring window can ensure timely detection of potential deadlocks, while not being misjudged due to transient fluctuations. A speed threshold of 0.5 m / s and a path overlap degree of 85% accurately define the boundary of substantial blockage, ensuring the accuracy of deadlock determination.

[0104] After triggering the conflict resolution flag, the system will analyze the flag to determine the affected transportation devices and the conflict range. Then, the optimal real-time driving paths of these devices are discretized, the continuous path is broken down into a series of executable short path segment instructions, forming a path update instruction set, and transmitted to the affected devices through a real-time bidirectional data interaction channel as the optimal path sequence for recalculation. The purpose of this is to allow devices to quickly receive and execute new paths, avoiding prolonged downtime due to the complexity of path adjustment, while discretization can reduce the error of instruction execution and ensure accurate movement of devices along the new path.

[0105] As Figure 4As shown, the optimal real-time driving path re-computes the optimal path sequence of the affected transportation equipment, and adjusts the task allocation, specifically including:

[0106] S310, obtain the priority rating of each transportation task, and establish a three-level task response mechanism;

[0107] S320, continuously monitor the current motion state of each transportation node, and when the speed of two or more transportation nodes is continuously lower than 0.5 m / s within 200 ms and the target path overlap degree is greater than 85%, it is determined that a deadlock state exists, and a conflict resolution flag is triggered;

[0108] S330, analyze the conflict resolution flag, discretize the optimal real-time driving path of the affected transportation equipment, obtain a path update instruction set, and pass it to the affected transportation equipment through the real-time bidirectional data interaction channel as the re-computed optimal path sequence.

[0109] In this step, the three-level task response mechanism includes:

[0110] For priority <3, it is defined as a background task, and when the system idle rate is >40%, the central coordinator activates the delay queue task and allocates a redundant path;

[0111] For 3≤priority<7, it is defined as a regular task, and a distributed negotiation protocol is triggered, and the optimal real-time driving path is allocated between nodes through a bidding game;

[0112] For priority ≥7, it is defined as an emergency task, and the central coordinator is activated to intervene and perform global resource matching.

[0113] S400, in combination with the re-computed optimal path sequence, frequency domain analysis is performed on the collected transportation equipment related data, potential risks are identified, and the optimal path sequence and equipment operating parameters are dynamically adjusted;

[0114] For the vibration signals of 0.5-200 Hz in the collected transportation equipment related data, the selection of this frequency band is based on the characteristics of the beam field transportation scene. The mechanical vibration during equipment operation, the bumping caused by uneven road surface, and the structural vibration during the precast beam bearing are mostly concentrated in this range. Analysis of this range can effectively capture equipment failure precursors or road condition abnormalities. In the processing process, wavelet denoising technology is used, the purpose is to eliminate the influence of environmental interference on effective signals, and to retain the vibration characteristics directly related to equipment state and path safety, providing a pure data basis for subsequent analysis.

[0115] Then the time-domain vibration signal is converted to the frequency domain by fast Fourier transform, which has the value of decomposing the complex time-domain waveform into different frequency components, making the characteristic frequency hidden in the time domain clear, and more accurately locating the problem source than time-domain analysis.

[0116] On the basis of frequency domain analysis, the amplitude of the characteristic frequency is identified and a safety threshold is set. The setting of the threshold needs to combine the factory parameters of the equipment, the transportation safety standards of the precast beam and the historical fault data of the beam field to ensure that both the abnormality can be captured in time and the false judgment caused by excessive sensitivity can be avoided. When the amplitude of the characteristic frequency exceeds the safety threshold, the system immediately triggers an early warning to control the risk in the embryonic stage.

[0117] Vibration anomalies are often early signals of equipment failure or path risks, and timely warning can prevent small problems from evolving into major failures. After the warning, the system will issue a speed reduction instruction to the relevant equipment. Speed reduction not only reduces the vibration load of the equipment itself and reduces the probability of failure, but also buys time for subsequent path adjustment; at the same time, based on the three-dimensional digital twin model of the site, the area with a sudden increase in low-frequency energy displayed by the frequency domain analysis is avoided. These areas usually correspond to more serious road unevenness or potential structural hazards, and detouring can isolate the risk source in space. The re-planned path will consider safety and efficiency to ensure that the transportation schedule is not significantly affected while avoiding risks.

[0118] As shown in Figure 5 , the identification of potential risks and dynamic adjustment of the optimal path sequence and equipment operation parameters specifically includes:

[0119] S410, wavelet denoising is performed on the collected 0.5-200Hz vibration signal, and the signal is converted to the frequency domain by fast Fourier transform;

[0120] S420, identify the amplitude of the characteristic frequency and set a safety threshold. When the amplitude of the characteristic frequency exceeds the safety threshold, an early warning is triggered;

[0121] S430, issue a speed reduction instruction to the warning equipment, and based on the three-dimensional digital twin model of the site, avoid the area with a sudden increase in low-frequency energy displayed by the frequency domain analysis, and re-plan the detour path.

[0122] S500, establish a historical operation database, continuously record the transportation equipment related data, and perform offline training and online learning on the recorded transportation equipment related data to continuously optimize and improve the pheromone diffusion algorithm.

[0123] The establishment of the historical operation database provides solid data support for the optimization process and avoids blind adjustment dominated by subjective experience; offline training can deeply mine the rules in long-term operation, optimize the overall framework of the algorithm, and improve its adaptability in complex scenarios; online learning ensures the rapid response of the algorithm to short-term dynamic changes and maintains the efficiency in real-time operation. The closed loop of data accumulation-algorithm optimization-performance improvement enables the design of the beam field to change from static initial planning to dynamic continuous iteration, which can continuously adapt to various changes such as equipment aging, task upgrading, and site expansion, and ultimately achieve the comprehensive optimization effect of improving transportation efficiency, reducing equipment wear and tear, and reducing safety risks, truly embodying the core characteristics of the intelligent beam field.

[0124] Figure 6 The structural block diagram of the design optimization system of the intelligent highway beam field based on big data provided by the embodiment of the application is shown in Figure 6 The system comprises:

[0125] The model construction and updating module 100 is configured to collect transportation equipment related data in real time through the Internet of Things network and the real-time bidirectional data interaction channel, construct and dynamically update a site three-dimensional digital twin model reflecting the real-time physical state of the beam field.

[0126] The path selection module 200 is configured to receive adjacent transportation node broadcast information through the real-time bidirectional data interaction channel, and fuse the site three-dimensional digital twin model, calculate the real-time selection probability distribution of the connection path between nodes by using the improved pheromone diffusion algorithm, and filter the obtained local optimal path probability distribution to output the optimal real-time driving path for each transportation equipment.

[0127] The conflict detection module 300 is configured to set a hierarchical processing protocol for transportation tasks of different priorities, trigger active scanning when receiving a new or changed task instruction, detect each transportation task, and when detecting resource request conflicts or path deadlocks, call the site three-dimensional digital twin model, combine the optimal real-time driving path, recalculate the optimal path sequence of the affected transportation equipment, and adjust the task allocation.

[0128] The path adjustment module 400 is configured to combine the recalculated optimal path sequence to perform frequency domain analysis on the collected transportation equipment related data, identify potential risks, and dynamically adjust the optimal path sequence and equipment operation parameters.

[0129] The operation optimization module 500 is configured to establish a historical operation database, continuously record transportation equipment related data, and perform offline training and online learning on the recorded transportation equipment related data to continuously optimize the improved pheromone diffusion algorithm.

[0130] Any technical features in the above-described embodiments can be combined in any manner, and for the sake of brevity, not all possible combinations are described, but it is understood that the scope of the present specification includes all possible combinations.

[0131] The above-described embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for those skilled in the art, some modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

[0132] The above-described embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for those skilled in the art, some modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims. The above-described embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the patent scope of the present application. It should be pointed out that, for those skilled in the art, some modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

Claims

1. A design optimization method for a big data-based intelligent highway beam yard, characterized in that, The method comprises: Real-time collection of transportation equipment related data through an Internet of Things network and a real-time bidirectional data interaction channel, construction and dynamic updating of a site three-dimensional digital twin model reflecting the real-time physical state of the beam field; Through the real-time bidirectional data interaction channel, receiving adjacent transportation node broadcast information and fusing the site three-dimensional digital twin model, using an improved pheromone diffusion algorithm to calculate the real-time selection probability distribution of the connection path between nodes, and screening the obtained local optimal path probability distribution to output the optimal real-time driving path for each transportation equipment; Setting a hierarchical processing protocol for transportation tasks of different priorities, triggering active scanning when receiving a new or changed task instruction, detecting each transportation task, calling the site three-dimensional digital twin model when detecting resource request conflicts or path deadlocks, recalculating the optimal path sequence of the affected transportation equipment in combination with the optimal real-time driving path, and adjusting the task allocation; Combining the recalculated optimal path sequence, performing frequency domain analysis on the collected transportation equipment related data, identifying potential risks and dynamically adjusting the optimal path sequence and equipment operating parameters; Establishing a historical operation database, continuously recording transportation equipment related data, and performing offline training and online learning on the recorded transportation equipment related data to continuously optimize the improved pheromone diffusion algorithm; Wherein: The construction and dynamic updating of the site three-dimensional digital twin model reflecting the real-time physical state of the beam field specifically comprises: Continuously transmitting transportation equipment related data through the real-time bidirectional data interaction channel at a sampling frequency of 5 times per second; Generating an initial site three-dimensional grid map, mapping the real-time collected transportation equipment related data to the grid vertex, dynamically correcting the equipment position error and updating the obstacle distribution, and generating the site three-dimensional digital twin model; Establishing a material flow state machine model, real-time solving the pedestal occupation state according to the precast beam hoisting progress, and synchronizing the material flow logic relationship to the topological structure of the site three-dimensional digital twin model; The screening of the local optimal path probability distribution to output the optimal real-time driving path for each transportation equipment specifically comprises: Based on the path topological structure provided by the site three-dimensional digital twin model, constructing a dynamic adjacency matrix to describe the transportation network node relationship, and the matrix element weight is determined by the Euclidean distance between nodes, the path turning angle cost and the real-time traffic flow density, and every 200ms, the adjacent node broadcast information of the adjacent nodes within a radius of 50 meters is obtained through the real-time bidirectional data interaction channel; Introducing a dynamic evaporation factor in the improved pheromone diffusion algorithm to calculate the real-time selection probability distribution of the connection path between transportation nodes; Screening candidate paths from the path selection probability distribution, calculating the score of each path based on a multi-objective evaluation function, and taking the highest score path as the optimal real-time driving path; The recalculating of the optimal path sequence of the affected transportation equipment in combination with the optimal real-time driving path and the adjusting of the task allocation specifically comprises: Obtaining the priority rating of each transportation task and establishing a three-level task response mechanism; Continuously monitor the current motion state of each transport node, and when the speed of two or more transport nodes is continuously below 0.5 m / s within 200 ms and the target path overlap degree is greater than 85%, it is determined that a deadlock state exists, and a conflict resolution flag is triggered; The conflict resolution flag is analyzed, the optimal real-time driving path of the affected transport equipment is discretized, a path update instruction set is obtained, and the affected transport equipment is transmitted through the real-time bidirectional data interaction channel as a recalculated optimal path sequence.

2. The method of claim 1, wherein, The transport equipment related data includes spatial coordinates, load information, residual energy state, running speed and direction parameters, material state information of the precast beam, and site environment index data.

3. The method of claim 2, wherein, The adjacent transport node broadcast information includes load information, destination coordinates, and residual energy state information.

4. The method of claim 3, wherein, The three-level task response mechanism includes: For priority < 3, defined as a background task, when the system idle rate > 40%, the central coordinator activates the delay queue task and assigns a redundant path; For 3 ≤ priority < 7, defined as a regular task, a distributed negotiation protocol is triggered, and the optimal real-time driving path is allocated between nodes through a bidding game; For priority ≥ 7, defined as an emergency task, the central coordinator is activated to intervene and perform global resource matching.

5. The method of claim 4, wherein, The identification of potential risks and dynamic adjustment of the optimal path sequence and equipment operating parameters specifically includes: Wavelet denoising is performed on the collected 0.5-200 Hz vibration signal, and fast Fourier transform is used to convert it to the frequency domain; Identify the characteristic frequency amplitude and set a safety threshold. When the characteristic frequency amplitude exceeds the safety threshold, a warning is triggered; A speed reduction instruction is issued to the warning device, and based on the site three-dimensional digital twin model, the area with a sudden increase in low-frequency energy is avoided, and a detour path is re-planned.

6. The method of claim 1, wherein, The system implementing the design optimization method of the highway intelligent beam yard based on big data includes: A model construction and update module for real-time collection of transport equipment related data through an Internet of Things network and a real-time bidirectional data interaction channel to construct and dynamically update a site three-dimensional digital twin model reflecting the real-time physical state of the beam yard; A path selection module for receiving adjacent transport node broadcast information through the real-time bidirectional data interaction channel, and combining the site three-dimensional digital twin model, using an improved pheromone diffusion algorithm to calculate the real-time selection probability distribution of the connection path between nodes, and selecting the local optimal path probability distribution to output the optimal real-time driving path for each transport equipment; A conflict detection module for setting a hierarchical processing protocol for transport tasks of different priorities, triggering active scanning when receiving a new or changed task instruction, detecting each transport task, and when a resource request conflict or path deadlock is detected, calling the site three-dimensional digital twin model, combining the optimal real-time driving path to recalculate the optimal path sequence of the affected transport equipment, and adjusting the task allocation; A path adjustment module for combining the recalculated optimal path sequence to perform frequency domain analysis on the collected transport equipment related data, identify potential risks, and dynamically adjust the optimal path sequence and equipment operating parameters; An operation optimization module is configured to establish a historical operation database, continuously record the transportation equipment related data, and perform offline training and online learning on the recorded transportation equipment related data to continuously optimize and improve the pheromone diffusion algorithm.

Images