Method for joint scheduling and congestion early warning of waterway ship lock based on multi-source data fusion

Abstract

This invention discloses a method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion. The method includes: integrating multi-source sensing data with digital twin technology and combining it with 3D simulation to generate the optimal scheduling scheme, thereby improving lock chamber utilization and throughput efficiency; integrating deep learning, edge detection, and laser sensing to automatically identify various safety risks of vessels and trigger alarms and interlocking controls to ensure navigation safety; and employing a congestion assessment model with dynamic correction coefficients and differentiated weights to accurately quantify congestion levels and provide tiered early warning and guidance, supplemented by equipment monitoring and data analysis functions to comprehensively improve the intelligence, operational stability, and management level of lock scheduling. This invention addresses the problems of existing lock scheduling, which suffers from single data sources, disconnect between equipment status and scheduling, lack of full-process risk identification and congestion prediction capabilities, leading to biased decision-making, increased congestion, and various navigation safety hazards.

Description

Technical Field

[0001] This invention relates to the field of intelligent lock scheduling technology, specifically a method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion. Background Technology

[0002] As an important component of the integrated water and land transportation system, inland waterways rely on locks as core hub facilities for ensuring the continuity of navigation and regulating water level differences.

[0003] Currently, lock scheduling relies heavily on manual labor combined with a basic scheduling system, which has many drawbacks: First, the data source is singular, relying solely on vessel declaration information and simplified dispatch ledgers for scheduling, without integrating multi-dimensional data such as hydrology, meteorology, vessel dynamics, equipment status, and environmental monitoring, resulting in one-sided dispatch decisions; Secondly, the operation status of the lock equipment and facilities is not linked to the scheduling process. When the gates, hoists and other equipment malfunction, the scheduling plan cannot be adjusted in time, which further aggravates congestion in the lock area and the chaos of navigation order. Third, the lack of a full-process risk identification and congestion prediction mechanism means that safety hazards such as ships exceeding height or draft limits, illegally crossing lines, and mixed traffic of hazardous chemical vessels cannot be identified in advance. At the same time, congestion situations such as the gathering of ships in upstream and downstream channels and the backlog of ships waiting in the lock area can only be dealt with after the fact, making it difficult to achieve early warning and guidance.

[0004] Therefore, to meet existing needs, a joint scheduling and congestion early warning method for waterway locks based on multi-source data fusion is proposed. Summary of the Invention

[0005] The purpose of this invention is to provide a method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion. By integrating multi-source sensing data and digital twin technology, and combining three-dimensional simulation pre-drilling to generate the optimal blocking scheme, the utilization rate and passage efficiency of the lock chamber are improved. By integrating deep learning, edge detection, and laser sensing, various safety risks of ships are automatically identified and linked to alarms and interlocking control to ensure navigation safety. At the same time, a congestion judgment model with dynamic correction coefficients and differentiated weights is adopted to accurately quantify the congestion level and provide graded early warning and guidance. With the assistance of equipment monitoring and data analysis functions, the intelligent scheduling, operational stability and control level of locks are comprehensively improved, solving the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion includes the following steps: S1: Collect ship dynamic data, ship over-limit monitoring data, waterway environmental data, lock equipment monitoring data, upper-level dispatch system data and manually entered data through the field perception subsystem, and perform data cleaning, format standardization and time sequence alignment; S2: Perform feature fusion, correlation matching and redundancy removal on the processed multi-source data, and synchronize the fused data to the digital twin 3D simulation platform for real-time linkage mapping between the physical lock scene and the virtual twin scene. S3: Digital supervision of the entire process of ship passage through the lock is carried out through a digital twin 3D simulation platform. Ship scheduling queues are divided according to the lock operation scheduling rules. The lock passage simulation is carried out by combining lock chamber parameters and ship data to generate the optimal scheduling plan and dynamically optimize the queue. S4: Calculate the actual draft of the ship based on deep learning and edge detection technology, identify the ship's excessive draft in real time with laser beam equipment, and trigger graded alarms and linkage actions based on the identification results; S5: Based on historical operational data, combined with real-time number of vessels waiting to pass through the lock, vessel speed, traffic flow in adjacent channels, and equipment operating status, a congestion assessment model for the lock area is constructed. The congestion assessment model is used to conduct congestion classification early warning and proactive diversion. S6: Real-time monitoring of the main structure of the lock, gates, hoists and electrical control equipment. When an abnormal problem occurs in the equipment, it will be immediately reported to the intelligent dispatching platform for fault handling.

[0007] Furthermore, it also includes the following steps: S7: Based on the fused full data, regularly perform statistical analysis on the total number of vessels passing through the lock, cargo type structure, vessel tonnage distribution, proportion of overloaded / oversized vessels, lock operation efficiency, and vessel waiting time, and generate visual analysis reports. S8: Configure organizational structure, users and hierarchical permissions through the intelligent scheduling platform to distinguish and manage the operation permissions of different roles; record all user operation behaviors, operation modules and operation time throughout the process, and generate tamper-proof operation logs for fault diagnosis, responsibility tracing and system operation and maintenance.

[0008] Furthermore, in S5, congestion classification and early warning, as well as proactive traffic management, are conducted through a congestion assessment model, including the following steps: The system collects real-time data on the total number of vessels waiting to pass through the lock, the average waiting time for vessels, the average passage time per lock session, the effective operating rate of the lock equipment, and the vessel density in the upstream and downstream channels, forming five key parameters. Retrieve the pre-stored baseline warning thresholds and indicator weight coefficients for each indicator, and match dynamic correction coefficients according to the current air traffic period; The real-time comprehensive congestion index is obtained by calculating five index parameters using the comprehensive congestion index calculation formula. Using real-time operational data of the lock during a specified period, and combining seasonal water levels and navigation rules, we set first-level and second-level congestion thresholds. Based on the interval of the real-time comprehensive congestion index, combined with the preset first-level congestion threshold and second-level congestion threshold, the navigation status of the gate area is divided into four levels: normal, slight congestion, moderate congestion, and severe congestion, and threshold values ​​for each level are set. Based on the determined congestion level, the congestion level and differentiated warning signs are displayed on the dispatch screen; and warning information is pushed to the superior waterway dispatch system, on-site maintenance personnel and vessels in transit.

[0009] Furthermore, multidimensional congestion parameters are calculated using the comprehensive congestion index calculation formula, which is as follows: In the formula, This is expressed as a real-time comprehensive congestion index; This represents the total number of vessels waiting to pass through the lock in real time. This represents the industry benchmark threshold for the total number of vessels waiting to enter the lock; This is expressed as the average waiting time for ships in the locks; This is expressed as the industry benchmark threshold for average waiting time. This is expressed as the average passage time per gate. This is expressed as the design baseline threshold for single-gate passage time; This is expressed as the effective operating rate of the lock equipment; This represents the minimum guaranteed threshold for equipment effective operating rate; Expressed as vessel density in upstream and downstream waterways; This represents the normal threshold for vessel density in the waterway. , , , , These are respectively represented as the weight coefficients of the five indicators; This is represented as a dynamic correction factor for the navigation period.

[0010] Furthermore, in S5, the congestion classification, early warning, and proactive mitigation through the congestion assessment model also include the following steps: Establish a full-dimensional automatic data archiving mechanism to store the calculated value of the comprehensive congestion index (CI) of all gate areas, the real-time congestion level, the ship diversion strategies implemented by the system, the congestion relief time, and the monitoring data of the entire process of navigation recovery on a daily basis, forming a structured historical dataset. A quarterly weight iteration and annual threshold calibration mechanism is constructed. Each quarter, the newly added ship lock passage, congestion evolution, scheduling execution and equipment operation measured data are summarized. The weight coefficients of the five evaluation indicators are recalculated and iteratively replaced based on the entropy weight method, and the weight ratio is dynamically adjusted. The full threshold is calibrated at a fixed annual cycle to assess and analyze various influencing factors throughout the year. The benchmark thresholds for the total number of vessels waiting to pass through the lock, the average waiting time, the average passage time per lock session, the minimum guaranteed threshold for effective equipment operation, the normal threshold for vessel density in the waterway, as well as the thresholds for slight / moderate congestion and moderate / severe congestion, are reviewed, and the thresholds that deviate from the actual operating conditions are dynamically corrected. Tests were conducted using historical typical congestion scenario samples to compare the congestion identification accuracy, early warning lead time, level classification matching degree, and traffic management strategy adaptability before and after parameter updates. If the accuracy of the analysis does not meet the preset requirements, the parameters are fine-tuned again and a second verification is performed.

[0011] Furthermore, in S3, the optimal gear shifting scheme is generated and the queue is dynamically optimized, including the following steps: Dispatch queues are divided according to vessel type, cargo attributes, and priority authority, and vessels of the same type are sorted according to their arrival registration time. The digital twin 3D simulation platform was used to conduct a full-process simulation of the lock passage of ships that had completed the declaration and payment, based on the geometric dimensions of the lock chamber, the main dimensions of the ships, the number of ships and the distribution of ships waiting to pass through the lock. With the goals of maximizing lock chamber space utilization, optimizing single-lock passage efficiency, and minimizing the total waiting time for waiting vessels, multiple simulated gate allocation schemes are generated by combining lock chamber allocation and gate allocation algorithms. After verifying multiple simulated gear shifting schemes, the actual gear shifting scheme is determined and issued for implementation; By combining the flow rate of adjacent locks, changes in water level in the channel, and real-time navigation order, the order of ship queues and the allocation of lock chambers are dynamically adjusted.

[0012] Furthermore, in S4, the actual draft of the ship is calculated based on deep learning and edge detection technology, and the ship's excessive height behavior is identified in real time using laser beam detection equipment, including the following steps: Collect ship hull image data and use deep learning object detection algorithms to locate the feature positions of ship hull bollards and freeboard decks; The ship's waterline boundary and freeboard deck edge contours are extracted using edge detection technology; combined with the ship's built-in draft parameters, the ship's actual draft is calculated. By deploying laser beam equipment along the route, the overall height of vessels passing through the waterway is monitored in real time to determine whether the vessels exceed the height limit for navigation. A full-area three-dimensional scan of passing vessels is conducted using 3D LiDAR to collect actual vessel size data. The actual size data is then compared with the size data in the pre-stored vessel registration files to determine whether there is any concealment or violation. By integrating laser scanning data with video surveillance footage, the relative position of the vessel to the safety warning line of the lock area can be determined.

[0013] Furthermore, in S4, the actual draft of the ship is calculated based on deep learning and edge detection technology, and the ship's excessive height behavior is identified in real time using laser beam detection equipment. This also includes the following steps: When a vessel is found to be in violation of regulations regarding excessive draft or excessive height, an alarm notification window will immediately pop up in the dispatch center, and the VHF maritime voice equipment will be used simultaneously to issue a warning to the vessel involved. When it is determined that a vessel has crossed the warning line and entered a dangerous area, a lockout command is immediately sent to the gate valve control system to prohibit the equipment from receiving and performing any opening or closing operations.

[0014] Further, after calculating the actual draft of the ship, the following steps are included: The actual draft is compared with the maximum permissible draft threshold preset by the lock to determine whether the vessel has violated the draft limit. If the vessel is determined to have exceeded its draft limit, the AIS data and vessel records will be linked to identify the vessel and trigger multi-level alarms. Depending on the degree of exceeding the limits, different control measures will be taken for the vessels; and the draft of the vessels after rectification will be re-verified, and they will be allowed to pass through the lock normally only after passing the verification.

[0015] Furthermore, in S3, the digital twin 3D simulation platform updates scene data synchronously in real time, with a data refresh frequency of no less than 1 time per second.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. By using a congestion assessment model that integrates multiple indicators to assess congestion in the lock area, and combining dynamic correction coefficients for navigation periods with differentiated weighting ratios, the degree of congestion can be accurately quantified. Based on measured operational data and on-site conditions, a graded threshold is scientifically set, which can accurately classify four types of navigation states and provide visual early warnings and multi-terminal information pushes. This effectively improves the accuracy of congestion identification, the timeliness of early warnings, and the adaptability of diversion strategies, ensuring the long-term stable and efficient operation of the lock.

[0017] 2. By collecting multi-source sensing data and using a digital twin 3D simulation platform for simulation and pre-running, the optimal ship berthing scheme and scheduling queue are generated, effectively improving the lock chamber space utilization and ship passage efficiency. Through deep learning, edge detection and laser sensing technologies, safety risks such as ship draft, over-height, size concealment, and crossing of lines are identified, graded, and equipment is linked and interlocked, enabling timely handling of equipment failures and visualized analysis of operational data, thus comprehensively improving the intelligence and safety of the joint scheduling of the waterway lock. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] To address the technical problems in existing lock scheduling, such as limited data sources, disconnect between equipment status and scheduling, lack of end-to-end risk identification and congestion prediction capabilities, which easily lead to biased decision-making, exacerbated congestion, and various navigation safety hazards, please refer to [the relevant documentation]. Figure 1 This embodiment provides the following technical solution: A method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion includes the following steps: S1: The field sensing subsystem collects dynamic data of ships, overload monitoring data of ships, waterway environmental data, lock equipment monitoring data, data from the superior dispatch system, and manually entered data, and performs data cleaning, format standardization, and time-series alignment. Specifically, the dynamic data of ships is collected by connecting to the AIS system and BeiDou ground-based augmentation base stations, including ship number, ship identification number, MMSI code, ship size, registered tonnage, draft, cargo type, cargo capacity, navigation trajectory, real-time location, credit records, and safety records. Simultaneously, it accesses ship lock passage declaration data, payment data, and priority lock passage application data. The overload monitoring data of ships is collected through 3D lidar, laser beam equipment, and high-definition video surveillance, including actual ship dimensions, actual draft, ship height, and the relative position of the ship to the lock area safety warning line. The waterway environmental data is collected by accessing hydrological and meteorological monitoring equipment, including environmental parameters such as water level inside and outside the lock chamber, water flow velocity, temperature, visibility, and wind force. The monitoring data for the lock equipment includes data on gate and valve opening, structural deformation, cracks, and component stress; simultaneously, it collects data on hoist oil temperature, level, pressure, current, voltage, and operating speed, as well as operational status data from the industrial control network and PLC equipment. Data from the higher-level dispatch system is obtained by connecting to the Beijing-Hangzhou Grand Canal operation dispatch and monitoring system, acquiring data on the overall lock operation plan, historical schedule records, vessel flow at adjacent locks, and vessel distribution across the entire waterway. Manually entered data is supplemented by personnel inputting vessel accident records, lists of key monitored vessels, and temporary navigation control instructions.

[0021] S2: Perform feature fusion, correlation matching, and redundancy removal on the processed multi-source data, and synchronize the fused data to the digital twin 3D simulation platform for real-time linkage mapping between the physical lock scene and the virtual twin scene; including: based on the spatial coordinate transformation algorithm, mapping the real-time position and navigation trajectory of the vessel underway to the digital twin 3D simulation platform to achieve a one-to-one correspondence between the virtual vessel and the real vessel's position and attitude; overlaying water level, flow velocity, and meteorological information of the waterway and lock chamber into the virtual scene of the waterway and lock chamber, and dynamically updating the environmental visualization status; rendering the gate and valve opening, equipment operating status, and main structure monitoring data in real time to ensure that the operating parameters of the virtual equipment are completely synchronized with the physical equipment; retrieving vessel file information and lock passage declaration information, binding the corresponding virtual vessel tags in the 3D scene, and realizing one-click traceability query of vessel information.

[0022] S3: Digital monitoring of the entire ship passage process through the lock is achieved using a digital twin 3D simulation platform. The platform updates scene data in real-time with a refresh rate of at least once per second, ensuring no significant time delay between the virtual scene and the physical site. Ship scheduling queues are divided according to the lock operation scheduling rules. By combining lock chamber parameters and ship data, a lock passage simulation is performed to generate the optimal scheduling plan and dynamically optimize the queues. This includes the following steps: The scheduling queues are divided according to vessel type, cargo attributes, and priority authority, into general cargo ship queues, fleet queues, priority passage vessel queues, and dangerous goods transport vessel queues. Vessels of the same type are sorted according to their arrival registration time. Dangerous goods transport vessels are arranged in separate queues and are prohibited from being assigned to the same lock session as passenger ships and general cargo ships. A digital twin 3D simulation platform is used to simulate the entire lock passage process for vessels that have completed declaration and payment, based on the lock chamber geometry, vessel main dimensions, number of vessels, and waiting distribution status. The simulation content includes: displaying the vessel berthing positions for the current and next lock sessions in a 3D scene; using a blurred display when a vessel has not reached the designated berthing position, and removing the blurred display after the vessel is in position; and digitally monitoring the entire lock passage process, including opening the upper valve, opening the upper gate, vessel entry and exit from the lock, closing the valve and gate, opening the lower valve, and opening the lower gate, and displaying the equipment status and environmental parameters of each process in real time.

[0023] With the goals of maximizing lock chamber space utilization, optimizing single-lock passage efficiency, and minimizing the total waiting time for waiting vessels, multiple simulated queuing schemes are generated by combining lock chamber allocation and queuing algorithms. After verifying the multiple simulated queuing schemes, the actual queuing scheme is determined and issued for execution. The vessel queue order and lock chamber allocation scheme are dynamically fine-tuned by combining the flow of adjacent locks, changes in water level in the channel, and real-time navigation order, taking into account both navigation fairness and overall passage efficiency.

[0024] S4: Calculate the ship's actual draft based on deep learning and edge detection technology, combine this with laser beam detection equipment to identify the ship's excessive draft in real time, and trigger tiered alarms and coordinated responses based on the identification results, including the following steps: The system collects ship hull image data using high-definition video equipment at the front end, and uses deep learning object detection algorithms to locate the ship's hull bollards and freeboard deck features. Edge detection technology is used to extract the ship's waterline boundary and the edge contour of the freeboard deck. Combined with the ship's built-in draft parameters, the actual draft of the ship is calculated. The actual draft is compared with the lock's preset maximum allowable draft threshold to determine if the ship has exceeded the draft limit. If the ship is found to have exceeded the draft limit, the system associates AIS data and the ship's file to identify the ship involved, triggering multi-level alarms including pop-up windows, audible and visual warnings, VHF voice announcements, and SMS push notifications. The monitoring data, image data, and alarm information are archived and stored to complete the evidence collection for violations. Depending on the degree of exceeding the limit, temporary docking and rectification or lock passage prohibition measures are taken for the ship. The draft of the rectified ship is re-verified, and normal passage through the lock is only allowed after the verification is passed.

[0025] Laser beam monitoring equipment deployed along the route is used to monitor the overall height of vessels in real time and determine whether they exceed the navigation height limit. For risks related to excessive draft and excessive height, alarm levels are assigned according to the severity of the violation, providing a basis for differentiated handling. When a vessel is found to be in violation of excessive draft or excessive height, an alarm notification window immediately pops up in the dispatch center, and VHF maritime voice equipment is simultaneously used to issue a warning to the vessel, informing it of the violation and requesting its cooperation with control measures. In addition, the system automatically retrieves the vessel's registered contact information and sends SMS alarm messages to the ship owner's mobile phone, achieving multi-dimensional and comprehensive risk warnings.

[0026] 3D LiDAR is used to perform full-area three-dimensional scanning of passing vessels, collecting actual dimensional data of the vessels, including overall length and beam. The actual dimensional data is compared with the dimensional data in the pre-stored vessel registration file to determine whether there is any concealment or violation, such as the violation of a large vessel with a different registration file. If a violation is confirmed, the on-site terminal immediately triggers an audible and visual alarm, and simultaneously pushes the information of the violating vessel, the captured image, and the reason for the alarm to the on-site staff at the lock remote control station and the core dispatch system.

[0027] By integrating laser scanning data and video surveillance footage, the system comprehensively determines the relative position of the vessel to the safety warning line of the lock area. When it is determined that a vessel has crossed the warning line and entered a dangerous area, a locking command is immediately issued to the lock valve control system, prohibiting the equipment from receiving and executing any opening or closing operations, thus avoiding the safety hazard of the vessel colliding with the lock gate from a hardware perspective. At the same time, the on-site audible and visual warning device is activated to remind the crew to adjust the vessel's position in a timely manner. The lock valve control system can only resume normal operation after the vessel has returned to the safe area and the alarm has been cleared. In addition, for vessels carrying hazardous chemicals, key vessels with safety accident records, or those with credit deductions, full-process accompanying monitoring is initiated, and real-time access to monitoring footage and navigation data along the route is obtained, enabling full traceability of key vessels passing through the lock.

[0028] S5: Based on historical operational data, combined with real-time data on the number of vessels waiting to pass through the lock, vessel speed, traffic flow in adjacent channels, and equipment operating status, a congestion assessment model for the lock area is constructed. This model is used for graded congestion warnings and proactive traffic management, including the following steps: The system collects real-time data on the total number of vessels waiting to pass through the lock, the average waiting time for vessels, the average passage time per lock session, the effective operating rate of the lock equipment, and the vessel density in the upstream and downstream channels, forming five indicator parameters. It also retrieves the pre-stored baseline warning thresholds and indicator weight coefficients for each indicator and matches dynamic correction coefficients based on the current navigation period.

[0029] The real-time comprehensive congestion index is obtained by calculating five indicator parameters using the comprehensive congestion index calculation formula. The calculation formula is as follows: In the formula, This is expressed as a real-time comprehensive congestion index; This represents the total number of vessels waiting to pass through the lock in real time. This represents the industry benchmark threshold for the total number of vessels waiting to enter the lock; This is expressed as the average waiting time for ships in the locks; This is expressed as the industry benchmark threshold for average waiting time. This is expressed as the average passage time per gate. This is expressed as the design baseline threshold for single-gate passage time; This is expressed as the effective operating rate of the lock equipment; This represents the minimum guaranteed threshold for equipment effective operating rate; Expressed as vessel density in upstream and downstream waterways; This represents the normal threshold for vessel density in the waterway. , , , , These are respectively represented as the weight coefficients of the five indicators; This is represented as a dynamic correction factor for the navigation period.

[0030] Using real-time operational data from the locks during designated periods, and considering seasonal water levels and navigation rules, a primary congestion threshold and a secondary congestion threshold were established. The primary congestion threshold... Slight to moderate congestion; Level 2 congestion threshold The congestion level is classified as moderate to severe. Based on the real-time comprehensive congestion index range and the preset primary and secondary congestion thresholds, the navigation status of the gate area is divided into four levels: normal, slight congestion, moderate congestion, and severe congestion, with threshold values ​​set for each level. The normal state is defined as follows: All indicators were below the warning threshold, and the relative deviations of all indicators did not exceed the benchmark threshold. The number of vessels waiting to pass through the lock was reasonable, and vessel passage was orderly throughout the entire process with no delays. Minor congestion: The overall congestion index rose slightly, the number of vessels waiting to pass through the locks exceeded the standard slightly, and the average waiting time for vessels increased slightly, but overall navigation order was not significantly affected. Moderate congestion: The overall congestion index continues to rise, resulting in a persistent backlog of vessels waiting to pass through the locks, a significant decrease in single-lock turnaround efficiency, and a slowdown in the pace of traffic flow within the lock area. Severe congestion: The overall congestion index exceeded the upper limit of moderate congestion, resulting in large-scale vessel delays in the lock area and simultaneous vessel aggregation in upstream and downstream waterways, significantly reducing navigation capacity.

[0031] Based on the determined congestion level, the congestion level and differentiated early warning signs are displayed on the dispatch screen, along with real-time CI values, five basic indicators, and detailed deviations of these indicators. Early warning information is also pushed to the superior waterway dispatch system, on-site maintenance personnel, and vessels in transit. When congestion is determined to be minor or moderate, the intelligent dispatch platform dynamically optimizes vessel scheduling based on digital twin simulation results. While adhering to lock dispatch rules, priority is given to large-tonnage vessels and vessels with priority passage permits to pass through the locks in groups, reducing the interval between lock passes and increasing the lock turnaround speed to gradually alleviate the backlog of waiting vessels. When congestion is determined to be severe, the cross-lock linkage diversion mechanism is immediately activated, requesting the superior dispatch system to coordinate with adjacent locks to divert vessels, splitting the waiting vessel fleet and dispersing them to surrounding lock chambers. Simultaneously, a waterway navigation advisory is issued across the entire area, guiding distant vessels to temporarily dock or detour via diversion routes, reducing the influx of vessels into the lock area from the source and quickly alleviating large-scale congestion.

[0032] In one embodiment, it is assumed that the real-time indicators collected during the morning peak period from 8:00 to 10:00 include: the total number of vessels waiting to pass through the lock. =45 ships (basic) =25), average waiting time =2.6h (baseline) =1.5h), average passage time per gate =52min (benchmark) =40min), equipment effective operating rate =82% (Minimum Guarantee Threshold) =90%), vessel density in upstream and downstream waterways =18 ships / km (normal threshold) =12 ships / km), weighting coefficient - =[0.3,0.2,0.2,0.15,0.15], peak hour correction factor =1.2.

[0033] Substituting into the above formula, we get: .

[0034] normal: <0.8; Slight congestion: 0.8-1.2; Moderate congestion: 1.2-1.8; Severe congestion: ≥1.8. The current value of 1.73 corresponds to moderate congestion. An orange indicator will be displayed on the dispatch screen, and a warning will be pushed to the higher-level dispatch platform. It is recommended that the upstream lock restrict the release frequency and guide 10 small vessels to a temporary anchorage for diversion. Vessels waiting to pass through the lock will be notified via SMS / VHF: A delay of 2 hours is expected; staggered waiting times are recommended.

[0035] Establish a full-dimensional automatic data archiving mechanism to store the calculated values ​​of the comprehensive congestion index (CI) for all gate areas, the real-time congestion level, the ship diversion strategies implemented by the system, the congestion relief time, and the monitoring data of the entire process of navigation recovery on a daily basis, forming a structured historical dataset; and store it in encrypted form to ensure data integrity and traceability.

[0036] A quarterly weighted iteration and annual threshold calibration mechanism is established. Each quarter, newly added data on vessel passage through locks, congestion evolution, scheduling execution, and equipment operation are summarized. After removing abnormal interference data, the weight coefficients of the five evaluation indicators are recalculated and iteratively replaced based on the entropy weight method. The weight ratio is dynamically adjusted to adapt to the phased characteristics of vessel traffic fluctuations, changes in mainstream vessel types, and changes in local waterway traffic conditions within the quarter, ensuring that the contribution of the indicators matches the actual causes of congestion. The full threshold is calibrated at a fixed period each year to identify and analyze various influencing factors throughout the year. Influencing factors include: waterway water level changes throughout the year, water level fluctuation characteristics during the flood season / dry season, updates to regional navigation control policies, adjustments to priority passage rules, distribution of newly added vessel types and tonnages, changes in total vessel traffic volume, and frequency of extreme weather events.

[0037] The system reviews the baseline thresholds for the total number of vessels waiting to pass through the lock, the average waiting time, the average passage time per lock session, the minimum guaranteed threshold for effective equipment operation, the normal threshold for vessel density in the waterway, as well as the thresholds for slight / moderate congestion and moderate / severe congestion. Thresholds deviating from actual operating conditions are dynamically corrected, and verification records and explanations are retained during the correction process. Historical typical congestion samples are used for testing, comparing the congestion identification accuracy, early warning lead time, level classification matching degree, and adaptability of traffic management strategies before and after parameter updates. If the judgment accuracy does not meet the preset requirements, parameters are fine-tuned again and a second verification is performed. Through data archiving, periodic parameter iteration, and effect verification, the congestion judgment model achieves continuous adaptive evolution, effectively improving the comprehensive capabilities of lock area congestion identification, graded early warning, and intelligent traffic management in the long term.

[0038] The beneficial effects achieved by the above are as follows: By using a congestion assessment model that integrates multiple indicators to assess congestion in the lock area, and by combining dynamic correction coefficients for navigation periods and differentiated weighting ratios to accurately quantify the degree of congestion, and by scientifically setting grading thresholds based on measured operational data and on-site conditions, four types of navigation states can be accurately classified and visualized for early warning and multi-terminal information push; this effectively improves the accuracy of congestion identification, the timeliness of early warning, and the adaptability of diversion strategies, ensuring the long-term stable and efficient operation of the lock.

[0039] S6: Real-time monitoring of the main structure of the lock, gates, hoists and electrical control equipment. When abnormal problems occur in the equipment, such as abnormal stress, excessive oil temperature or fault alarm, the system will immediately report to the intelligent dispatching platform for fault handling, such as temporarily adjusting the lock schedule, suspending the corresponding lock chamber operation, prioritizing equipment maintenance, and avoiding large-scale congestion and safety accidents caused by equipment failure. After the equipment is restored to normal, the normal dispatching order will be gradually restored.

[0040] S7: Based on the fused full data, it regularly performs statistical analysis on the total number of vessels passing through the lock, cargo type structure, vessel tonnage distribution, proportion of overloaded / oversized vessels, lock operation efficiency, and vessel waiting time, and generates visual analysis reports to provide data support for overall waterway planning, scheduling rule optimization, and lock operation and maintenance plan formulation.

[0041] S8: Configure organizational structure, users, and hierarchical permissions through the intelligent scheduling platform to distinguish the operation permissions of different roles such as administrators, dispatchers, and inspectors; record all user operation behaviors, operation modules, and operation times throughout the process to generate tamper-proof operation logs for fault diagnosis, responsibility tracing, and system maintenance.

[0042] The beneficial effects achieved by the above are as follows: By collecting multi-source sensing data and using a digital twin 3D simulation platform for simulation and pre-running, the optimal ship berthing scheme and scheduling queue are generated, effectively improving the utilization rate of lock chamber space and the efficiency of ship passage; by using deep learning, edge detection and laser sensing technologies to identify, classify and alarm, and link equipment interlocks for safety risks such as ship draft, over-height, size concealment and crossing of lines, timely handling of equipment failures and visualization analysis of operational data are achieved, comprehensively improving the intelligence and safety of joint scheduling of waterway locks.

[0043] Working principle: By collecting various types of data and using a digital twin platform to link virtual and real scenarios, the system generates and dynamically optimizes ship locking schemes through simulation and pre-visualization. It integrates technologies such as deep learning, edge detection, and laser sensing to automatically identify violations such as excessive draft, over-height, size concealment, and crossing of lines, and simultaneously executes tiered alarms and equipment linkage control. It constructs a congestion assessment model, calculates a comprehensive congestion index, determines the congestion level based on tiered thresholds, and issues warnings and implements diversion strategies. It also monitors equipment operating status in real time, visualizes and analyzes operational data, and configures tiered permissions and operation logs to achieve intelligent operation of lock scheduling, safety supervision, congestion warning, and maintenance management.

[0044] It should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0045] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion, characterized in that, Includes the following steps: S1: Collect ship dynamic data, ship over-limit monitoring data, waterway environmental data, lock equipment monitoring data, upper-level dispatch system data and manually entered data through the field perception subsystem, and perform data cleaning, format standardization and time sequence alignment; S2: Perform feature fusion, correlation matching and redundancy removal on the processed multi-source data, and synchronize the fused data to the digital twin 3D simulation platform for real-time linkage mapping between the physical lock scene and the virtual twin scene. S3: Digital supervision of the entire process of ship passage through the lock is carried out through a digital twin 3D simulation platform. Ship scheduling queues are divided according to the lock operation scheduling rules. The lock passage simulation is carried out by combining lock chamber parameters and ship data to generate the optimal scheduling plan and dynamically optimize the queue. S4: Calculate the actual draft of the ship based on deep learning and edge detection technology, identify the ship's excessive draft in real time with laser beam equipment, and trigger graded alarms and linkage actions based on the identification results; S5: Based on historical operational data, combined with real-time number of vessels waiting to pass through the lock, vessel speed, traffic flow in adjacent channels, and equipment operating status, a congestion assessment model for the lock area is constructed. The congestion assessment model is used to conduct congestion classification early warning and proactive diversion. S6: Real-time monitoring of the main structure of the lock, gates, hoists and electrical control equipment. When an abnormal problem occurs in the equipment, it will be immediately reported to the intelligent dispatching platform for fault handling.

2. The method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to claim 1, characterized in that, It also includes the following steps: S7: Based on the fused full data, regularly perform statistical analysis on the total number of vessels passing through the lock, cargo type structure, vessel tonnage distribution, proportion of overloaded / oversized vessels, lock operation efficiency, and vessel waiting time, and generate visual analysis reports. S8: Configure organizational structure, users and hierarchical permissions through the intelligent scheduling platform to distinguish and manage the operation permissions of different roles; record all user operation behaviors, operation modules and operation time throughout the process, and generate tamper-proof operation logs for fault diagnosis, responsibility tracing and system operation and maintenance.

3. The method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to claim 1, characterized in that, In S5, congestion classification and early warning, and proactive traffic management are carried out through a congestion assessment model, including the following steps: The system collects real-time data on the total number of vessels waiting to pass through the lock, the average waiting time for vessels, the average passage time per lock session, the effective operating rate of the lock equipment, and the vessel density in the upstream and downstream channels, forming five key parameters. Retrieve the pre-stored baseline warning thresholds and indicator weight coefficients for each indicator, and match dynamic correction coefficients according to the current air traffic period; The real-time comprehensive congestion index is obtained by calculating five index parameters using the comprehensive congestion index calculation formula. Using real-time operational data of the lock during a specified period, and combining seasonal water levels and navigation rules, we set first-level and second-level congestion thresholds. Based on the interval of the real-time comprehensive congestion index, combined with the preset first-level congestion threshold and second-level congestion threshold, the navigation status of the gate area is divided into four levels: normal, slight congestion, moderate congestion, and severe congestion, and threshold values ​​for each level are set. Based on the determined congestion level, the congestion level and differentiated warning signs are displayed on the dispatch screen; and warning information is pushed to the superior waterway dispatch system, on-site maintenance personnel and vessels in transit.

4. The method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to claim 3, characterized in that, The multidimensional congestion parameters are calculated using the comprehensive congestion index formula, which is as follows: In the formula, This is expressed as a real-time comprehensive congestion index; This represents the total number of vessels waiting to pass through the lock in real time. This represents the industry benchmark threshold for the total number of vessels waiting to enter the lock; This is expressed as the average waiting time for ships in the locks; This is expressed as the industry benchmark threshold for average waiting time. This is expressed as the average passage time per gate. This is expressed as the design baseline threshold for single-gate passage time; This is expressed as the effective operating rate of the lock equipment; This represents the minimum guaranteed threshold for equipment effective operating rate; Expressed as vessel density in upstream and downstream waterways; This represents the normal threshold for vessel density in the waterway. , , , , These are respectively represented as the weight coefficients of the five indicators; This is represented as a dynamic correction factor for the navigation period.

5. The method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to claim 3, characterized in that, In S5, the congestion assessment model is used to conduct congestion classification, early warning, and proactive traffic management, which also includes the following steps: Establish a full-dimensional automatic data archiving mechanism to store the calculated value of the comprehensive congestion index (CI) of all gate areas, the real-time congestion level, the ship diversion strategies implemented by the system, the congestion relief time, and the monitoring data of the entire process of navigation recovery on a daily basis, forming a structured historical dataset. A quarterly weight iteration and annual threshold calibration mechanism is constructed. Each quarter, the newly added ship lock passage, congestion evolution, scheduling execution and equipment operation measured data are summarized. The weight coefficients of the five evaluation indicators are recalculated and iteratively replaced based on the entropy weight method, and the weight ratio is dynamically adjusted. The full threshold is calibrated at a fixed annual cycle to assess and analyze various influencing factors throughout the year. The benchmark thresholds for the total number of vessels waiting to pass through the lock, the average waiting time, the average passage time per lock session, the minimum guaranteed threshold for effective equipment operation, the normal threshold for vessel density in the waterway, as well as the thresholds for slight / moderate congestion and moderate / severe congestion, are reviewed, and the thresholds that deviate from the actual operating conditions are dynamically corrected. Tests were conducted using historical typical congestion scenario samples to compare the congestion identification accuracy, early warning lead time, level classification matching degree, and traffic management strategy adaptability before and after parameter updates. If the accuracy of the analysis does not meet the preset requirements, the parameters are fine-tuned again and a second verification is performed.

6. The method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to claim 1, characterized in that, In S3, the optimal gear shifting scheme is generated and the queue is dynamically optimized, including the following steps: Dispatch queues are divided according to vessel type, cargo attributes, and priority authority, and vessels of the same type are sorted according to their arrival registration time. The digital twin 3D simulation platform was used to conduct a full-process simulation of the lock passage of ships that had completed the declaration and payment, based on the geometric dimensions of the lock chamber, the main dimensions of the ships, the number of ships and the distribution of ships waiting to pass through the lock. With the goals of maximizing lock chamber space utilization, optimizing single-lock passage efficiency, and minimizing the total waiting time for waiting vessels, multiple simulated gate allocation schemes are generated by combining lock chamber allocation and gate allocation algorithms. After verifying multiple simulated gear shifting schemes, the actual gear shifting scheme is determined and issued for implementation; By combining the flow rate of adjacent locks, changes in water level in the channel, and real-time navigation order, the order of ship queues and the allocation of lock chambers are dynamically adjusted.

7. The method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to claim 1, characterized in that, In S4, the actual draft of the ship is calculated based on deep learning and edge detection technology, and the ship's excessive draft is identified in real time using laser beam detection equipment, including the following steps: Collect ship hull image data and use deep learning object detection algorithms to locate the feature positions of ship hull bollards and freeboard decks; The ship's waterline boundary and freeboard deck edge contours are extracted using edge detection technology; combined with the ship's built-in draft parameters, the ship's actual draft is calculated. By deploying laser beam equipment along the route, the overall height of vessels passing through the waterway is monitored in real time to determine whether the vessels exceed the height limit for navigation. A full-area three-dimensional scan of passing vessels is conducted using 3D LiDAR to collect actual vessel size data. The actual size data is then compared with the size data in the pre-stored vessel registration files to determine whether there is any concealment or violation. By integrating laser scanning data with video surveillance footage, the relative position of the vessel to the safety warning line of the lock area can be determined.

8. The method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to claim 7, characterized in that, In S4, the actual draft of the ship is calculated based on deep learning and edge detection technology, and the ship's excessive draft is identified in real time using laser beam detection equipment. The process also includes the following steps: When a vessel is found to be in violation of regulations regarding excessive draft or excessive height, an alarm notification window will immediately pop up in the dispatch center, and the VHF maritime voice equipment will be used simultaneously to issue a warning to the vessel involved. When it is determined that a vessel has crossed the warning line and entered a dangerous area, a lockout command is immediately sent to the gate valve control system to prohibit the equipment from receiving and performing any opening or closing operations.

9. The method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to claim 7, characterized in that, After calculating the actual draft of the ship, the following steps are included: The actual draft is compared with the maximum permissible draft threshold preset by the lock to determine whether the vessel has violated the draft limit. If the vessel is determined to have exceeded its draft limit, the AIS data and vessel records will be linked to identify the vessel and trigger multi-level alarms. Depending on the degree of exceeding the limits, different control measures will be taken for the vessels; and the draft of the vessels after rectification will be re-verified, and they will be allowed to pass through the lock normally only after passing the verification.

10. The method for joint scheduling and congestion early warning of waterway locks based on multi-source data fusion according to claim 1, characterized in that, In S3, the digital twin 3D simulation platform updates scene data in real time, with a data refresh frequency of no less than once per second.

Images