Bulk cargo terminal operation and maintenance perception method and system based on device coding and dynamic threshold

Abstract

This application belongs to the field of port intelligent operation and maintenance and situational awareness technology, and relates to a method and system for operation and maintenance awareness of bulk cargo terminals based on equipment coding and dynamic thresholds. The method includes: constructing a unified coding system for bulk cargo terminal equipment and establishing an equipment attribute dictionary corresponding to the unified coding system; calculating dynamic thresholds corresponding to each operational data point; comparing real-time operational data with the corresponding dynamic thresholds, identifying and marking equipment exceeding the dynamic thresholds as abnormal state equipment; analyzing the scope and degree of impact of abnormal state equipment on the work process, and calculating the workability index of the process under the influence of abnormal state equipment; and evaluating the overall operational capability of the current bulk cargo terminal based on the scope, degree, and workability index of the impact, obtaining a corresponding process-level operation and maintenance situational awareness report.

Description

Technical Field

[0001] This application belongs to the field of port intelligent operation and maintenance and situational awareness technology, and in particular relates to a method and system for operation and maintenance awareness of bulk cargo terminals based on equipment coding and dynamic thresholds. Background Technology

[0002] As a crucial hub for port logistics, the operational efficiency of bulk cargo terminals directly impacts the overall throughput capacity and logistics smoothness of the port. Core terminal operations rely on a series of critical equipment, including belt conveyor systems, stacker-reclaimers, and loading / unloading equipment. This equipment typically operates under high loads, high salt spray conditions, and complex weather environments, resulting in a high failure rate and posing a potential threat to terminal operation safety and efficiency.

[0003] Currently, the operation and maintenance management of bulk cargo terminal equipment mainly relies on traditional manual inspections and planned maintenance, which has many shortcomings. First, the problem of data silos is serious: different types of equipment often rely on independent monitoring platforms, and data collection standards are inconsistent, making it difficult to achieve cross-equipment and cross-system status correlation analysis and comprehensive evaluation. Second, alarm threshold settings are rigid: existing systems mostly use fixed thresholds for anomaly detection, failing to consider dynamic factors such as equipment lifecycle, operating load, and environmental conditions, leading to frequent false alarms or missed potential faults. Third, there is a lack of predictive capabilities: traditional operation and maintenance methods are mainly reactive, unable to predict the development trend of equipment failures, resulting in frequent unplanned downtime and affecting the overall operational efficiency of the port. Finally, impact assessment is insufficient: the overall impact of single-point equipment anomalies on the terminal's operational processes is difficult to quantify, failing to provide effective reference for emergency decision-making and operational optimization.

[0004] In existing technologies, such as the bulk cargo terminal operation and maintenance situation awareness method and system disclosed in patent CN121389823A, although it can realize the operation monitoring of single-point equipment, it mainly focuses on the status of the equipment itself, without establishing a unified equipment coding system and a comprehensive operation and maintenance management framework, and also lacks intelligent perception and operation process-level impact analysis based on dynamic thresholds. Therefore, it is difficult to meet the needs of modern bulk cargo terminals for efficient, accurate and intelligent operation and maintenance.

[0005] In summary, how to achieve data fusion, dynamic monitoring and intelligent analysis of equipment in bulk cargo terminals, and how to quantitatively assess the impact of equipment malfunctions on operational processes and improve the efficiency of operation and maintenance decisions have become urgent technical problems that need to be solved. Summary of the Invention

[0006] To address the aforementioned technical issues, this application proposes a method and system for operation and maintenance perception of bulk cargo terminals based on equipment coding and dynamic thresholds. Through unified scheduling and intelligent collaboration of equipment coding management, real-time operation data collection, dynamic threshold calculation, and abnormal state analysis, the system achieves full-process situational awareness of equipment and operational processes, enabling bulk cargo terminals to achieve more accurate, reliable, and efficient intelligent operation and maintenance management in complex environments.

[0007] To achieve the above objectives, the first aspect of this application provides a method for operational awareness of bulk cargo terminals based on equipment coding and dynamic thresholds, comprising the following steps:

[0008] Based on the basic information of the equipment in the bulk cargo terminal, a unified coding system for the equipment in the bulk cargo terminal is constructed, and a dictionary of equipment attributes corresponding to the unified coding system is established.

[0009] Based on the device attribute dictionary and historical operating data statistical characteristics, combined with current environmental data, load status and device life cycle stage, the dynamic threshold corresponding to each operating data is calculated;

[0010] During equipment operation, real-time operating data of the equipment is continuously collected, and the real-time operating data is compared with the corresponding dynamic threshold. Equipment that exceeds the dynamic threshold is identified and marked as abnormal equipment.

[0011] Based on the node position of the abnormal state equipment in the pre-constructed bulk cargo terminal operation process model, analyze the scope and degree of influence of the abnormal state equipment on the operation process, and calculate the process operability index under the influence of the abnormal state equipment.

[0012] Based on the scope of impact, the degree of impact, and the process operability index, the overall operational capability of the current bulk cargo terminal is assessed, and a corresponding process-level operation and maintenance situation awareness report is obtained.

[0013] The calculation method for the dynamic threshold is as follows: Obtain the basic alarm threshold based on the current cumulative operating time of the monitored device, its environmental conditions, current actual operating load, and target operating data; obtain a lifecycle correction factor based on the proportional relationship between the current cumulative operating time of the monitored device and its designed lifespan; obtain an environmental correction factor based on the environmental conditions of the monitored device; obtain a load correction factor based on the deviation relationship between the current actual operating load and the rated operating load of the monitored device; weight and fuse the lifecycle correction factor, environmental correction factor, and load correction factor according to preset correction weight coefficients to obtain a comprehensive correction factor; multiply the basic alarm threshold by the comprehensive correction factor to obtain the dynamic threshold.

[0014] In some embodiments, the device attribute dictionary is organized using a four-level tree structure, including:

[0015] The equipment basic information layer is used to record the equipment's identification, classification, model, installation location, region, manufacturer, commissioning time, and other basic attribute information related to equipment identification, classification, and management.

[0016] The technical parameter layer is used to record the technical parameters that characterize the operation and performance of the equipment.

[0017] The maintenance record layer is used to record historical maintenance and repair information of the equipment.

[0018] The relationship layer is used to record the physical connection relationships, logical coordination relationships, and operation timing relationships between devices.

[0019] In some embodiments, the operating data includes bearing status data, equipment temperature data, load status data, hoisting mechanism wire rope tension data, equipment spatial location information and travel trajectory data, as well as operating condition parameter data generated during operation; the environmental data includes temperature, humidity, wind speed and salt spray concentration parameters of the bulk cargo terminal operating area.

[0020] In some embodiments, the method for constructing the bulk cargo terminal operation process model is as follows:

[0021] The various operational processes and key equipment of the bulk cargo terminal are abstracted as nodes, and the material flow direction or operational timing dependency between nodes is defined as an edge, thus constructing the basic topology of the operational process.

[0022] Configure each node with its corresponding equipment code, equipment type, rated processing capacity, current availability and health score, and configure each side with its corresponding material flow limit, conveying distance and energy consumption coefficient;

[0023] Based on the material flow direction or operation sequence, each node is connected in an orderly manner to obtain a complete logistics chain;

[0024] A directed acyclic graph is constructed based on the nodes and edges, where the direction of the edges is used to represent the material flow direction and the sequence of operation steps, resulting in a formalized bulk cargo terminal operation process model.

[0025] In the formalized bulk cargo terminal operation process model, the physical connection relationship, logical coordination relationship and operation sequence dependency relationship between key equipment are modeled to obtain a complete bulk cargo terminal operation process model.

[0026] In some embodiments, the method for analyzing the scope and degree of impact of abnormal state devices on the work process is as follows:

[0027] Locate the node position corresponding to the abnormal state equipment in the bulk cargo terminal operation process model, and determine its upstream and downstream correlation in the operation process;

[0028] Starting from the node corresponding to the abnormal device, perform traversal operations along the direction of the work process and in the opposite direction to identify the affected work path caused by the abnormal device.

[0029] Based on the affected work paths, determine the set of equipment and the set of work tasks affected by the abnormal equipment status;

[0030] Based on the set of affected equipment and the set of work tasks, calculate the scope of the impact of abnormal equipment on the work process, which is used to characterize the proportion of affected equipment or work tasks in the overall work process.

[0031] By combining the changes in the operational capabilities and operational delays of the affected objects, the impact of abnormal equipment on the operational process is quantitatively assessed, and the degree of impact of abnormal equipment on the operational process is obtained.

[0032] In some embodiments, the process workability index is calculated as follows:

[0033] Obtain the equipment availability list, the equipment importance weight list, and the process association matrix;

[0034] For each critical path in the process, calculate the path availability index based on the availability of each device within the path;

[0035] Based on the path availability metrics of all critical paths, the process integrity metrics are obtained.

[0036] For bottleneck equipment in the process, calculate the capacity matching index based on equipment availability and actual capacity;

[0037] Based on the preset index weighting coefficients, the process integrity index, capacity matching index, and overall equipment availability are weighted to obtain the process operability index.

[0038] In some embodiments, the bulk cargo terminal operation and maintenance perception method further includes the following steps:

[0039] The real-time operating data of the abnormal equipment is input into a pre-built equipment health prediction model to obtain the health change trend of the abnormal equipment.

[0040] When the health status trend shows that the predicted health status is lower than the set health status threshold, the corresponding maintenance suggestions and maintenance work orders are automatically generated.

[0041] When the health status change trend shows that the predicted health status is not lower than the set health status threshold, the operation data of the abnormal state device continues to be monitored, and the dynamic threshold is adaptively updated based on the continuously collected real-time operation data.

[0042] The real-time operating data, the abnormal status equipment information, the health change trend, the maintenance suggestions, and the process-level operation and maintenance situation awareness report are mapped to the three-dimensional digital twin model of the bulk cargo terminal. The real-time display of equipment-level monitoring, regional-level statistics, and terminal-level panoramic situation is realized through a multi-level situation display interface.

[0043] In some embodiments, the device health prediction model is constructed based on a long short-term memory neural network architecture, comprising a feature temporal modeling layer, a feature mapping layer, and a health output layer connected in sequence, wherein:

[0044] The feature time series modeling layer is used to perform hierarchical and progressive time series correlation modeling on the time series of real-time running data. The feature time series modeling layer includes a first LSTM hidden layer, a second LSTM hidden layer, and a third LSTM hidden layer arranged in sequence. The first LSTM hidden layer receives the time series input of real-time running data and outputs a time-step feature representation. A first random deactivation layer is set at its output to suppress model overfitting. The second LSTM hidden layer extracts medium- and long-term dependency features based on the time-step feature representation output by the first LSTM hidden layer and outputs an enhanced time-step feature representation. A second random deactivation layer is set at its output to improve the model's generalization ability. The third LSTM hidden layer performs convergence processing on the sequence features output by the second LSTM hidden layer and outputs a fixed-length feature vector.

[0045] The feature mapping layer is a fully connected layer used to receive a fixed-length feature vector output by the third LSTM hidden layer, and to perform non-linear mapping on the fixed-length feature vector through the ReLU activation function;

[0046] The health output layer is a single-neuron structure, and its activation function is the Sigmoid function, which is used to map the model output to a preset health range to obtain the device health score.

[0047] A second aspect of this application provides a bulk cargo terminal operation and maintenance awareness system based on equipment coding and dynamic thresholds, used to implement the aforementioned bulk cargo terminal operation and maintenance awareness method, including:

[0048] The data acquisition layer is used to collect real-time operational data of bulk cargo terminal equipment and environmental data.

[0049] The intelligent analysis layer, based on the data output from the data processing layer, performs dynamic threshold calculation and work process status analysis to intelligently perceive the equipment operating status and work process.

[0050] The decision support layer generates maintenance plans, emergency response schemes, and resource optimization schemes based on the output information of the intelligent analysis layer, providing a basis for decision-making in operation and maintenance management.

[0051] Compared with the prior art, the advantages and positive effects of this application are as follows:

[0052] This invention achieves real-time, dynamic, quantitative, and visual management of bulk cargo terminal operations and maintenance by combining equipment coding, dynamic thresholds, and process-level situational analysis. This not only improves operation and maintenance efficiency and safety but also reduces operation and maintenance costs and enhances management decision support capabilities. Attached Figure Description

[0053] Figure 1 This is a structural block diagram of the bulk cargo terminal operation and maintenance perception system based on equipment coding and dynamic thresholds as described in the first aspect embodiment of the present invention;

[0054] Figure 2 This is a flowchart of the bulk cargo terminal operation and maintenance perception method based on equipment coding and dynamic thresholds according to the second aspect embodiment of the present invention;

[0055] Figure 3 This is a flowchart of the bulk cargo terminal operation and maintenance perception method based on equipment coding and dynamic thresholds as described in the third aspect embodiment of the present invention;

[0056] Figure 4 This is a structural block diagram of the equipment health prediction model in the embodiments of this application. Detailed Implementation

[0057] The present application will now be described in detail through exemplary embodiments. However, it should be understood that, without further description, elements, structures, and features in one embodiment may be advantageously incorporated into other embodiments.

[0058] In a broad embodiment of the present invention, the bulk cargo terminal operation and maintenance perception method based on equipment coding and dynamic thresholds includes the following steps:

[0059] Based on the basic information of the equipment in the bulk cargo terminal, a unified coding system for the equipment in the bulk cargo terminal is constructed, and a dictionary of equipment attributes corresponding to the unified coding system is established.

[0060] Based on the device attribute dictionary and historical operating data statistical characteristics, combined with current environmental data, load status and device life cycle stage, the dynamic threshold corresponding to each operating data is calculated;

[0061] During equipment operation, real-time operating data of the equipment is continuously collected, and the real-time operating data is compared with the corresponding dynamic threshold. Equipment that exceeds the dynamic threshold is identified and marked as abnormal equipment.

[0062] Based on the node position of the abnormal state equipment in the pre-constructed bulk cargo terminal operation process model, analyze the scope and degree of influence of the abnormal state equipment on the operation process, and calculate the process operability index under the influence of the abnormal state equipment.

[0063] Based on the scope of impact, the degree of impact, and the process operability index, the overall operational capability of the current bulk cargo terminal is assessed, and a corresponding process-level operation and maintenance situation awareness report is obtained.

[0064] It should be noted that the "basic information" mentioned in this application refers to the basic data set used for unified equipment coding, attribute mapping, classification management, location identification and operation and maintenance analysis, including equipment identification information, classification information, model information, installation location, region, manufacturer, commissioning time and other basic attribute information related to equipment identification, classification and management.

[0065] In some embodiments, the device attribute dictionary is organized using a four-level tree structure, including:

[0066] The equipment basic information layer is used to record the equipment's identification, classification, model, installation location, region, manufacturer, commissioning time, and other basic attribute information related to equipment identification, classification, and management.

[0067] The technical parameter layer is used to record the technical parameters that characterize the operation and performance of the equipment.

[0068] The maintenance record layer is used to record historical maintenance and repair information of the equipment.

[0069] The relationship layer is used to record the physical connection relationships, logical coordination relationships, and operation timing relationships between devices.

[0070] In some embodiments, the operating data includes bearing status data, equipment temperature data, load status data, hoisting mechanism wire rope tension data, equipment spatial location information and travel trajectory data, as well as operating condition parameter data generated during operation; the environmental data includes temperature, humidity, wind speed and salt spray concentration parameters of the bulk cargo terminal operating area.

[0071] In some embodiments, the dynamic threshold is calculated as follows:

[0072] Acquire the basic alarm thresholds for the current cumulative running time, environmental conditions, current actual operating load, and target operating data of the equipment to be monitored;

[0073] Based on the proportional relationship between the current cumulative operating time of the monitored equipment and its designed service life, a life cycle correction factor is obtained.

[0074] The environmental correction factor is obtained based on the environmental conditions of the equipment to be monitored;

[0075] The load correction factor is obtained based on the deviation between the current actual operating load and the rated operating load of the equipment under monitoring.

[0076] The life cycle correction factor, environmental correction factor, and load correction factor are weighted and fused according to the preset correction weight coefficients to obtain the comprehensive correction factor;

[0077] The dynamic threshold is obtained by multiplying the basic alarm threshold by the comprehensive correction factor.

[0078] In some embodiments, the method for constructing the bulk cargo terminal operation process model is as follows:

[0079] The various operational processes and key equipment of the bulk cargo terminal are abstracted as nodes, and the material flow direction or operational timing dependency between nodes is defined as an edge, thus constructing the basic topology of the operational process.

[0080] Configure each node with its corresponding equipment code, equipment type, rated processing capacity, current availability and health score, and configure each side with its corresponding material flow limit, conveying distance and energy consumption coefficient;

[0081] Based on the material flow direction or operation sequence, each node is connected in an orderly manner to obtain a complete logistics chain;

[0082] A directed acyclic graph is constructed based on the nodes and edges, where the direction of the edges is used to represent the material flow direction and the sequence of operation steps, resulting in a formalized bulk cargo terminal operation process model.

[0083] In the formalized bulk cargo terminal operation process model, the physical connection relationship, logical coordination relationship and operation sequence dependency relationship between key equipment are modeled to obtain a complete bulk cargo terminal operation process model.

[0084] In some embodiments, the method for analyzing the scope and degree of impact of abnormal state devices on the work process is as follows:

[0085] Locate the node position corresponding to the abnormal state equipment in the bulk cargo terminal operation process model, and determine its upstream and downstream correlation in the operation process;

[0086] Starting from the node corresponding to the abnormal device, perform traversal operations along the direction of the work process and in the opposite direction to identify the affected work path caused by the abnormal device.

[0087] Based on the affected work paths, determine the set of equipment and the set of work tasks affected by the abnormal equipment status;

[0088] Based on the set of affected equipment and the set of work tasks, calculate the scope of the impact of abnormal equipment on the work process, which is used to characterize the proportion of affected equipment or work tasks in the overall work process.

[0089] By combining the changes in the operational capabilities and operational delays of the affected objects, the impact of abnormal equipment on the operational process is quantitatively assessed, and the degree of impact of abnormal equipment on the operational process is obtained.

[0090] In some embodiments, the process workability index is calculated as follows:

[0091] Obtain the equipment availability list, the equipment importance weight list, and the process association matrix;

[0092] For each critical path in the process, calculate the path availability index based on the availability of each device within the path;

[0093] Based on the path availability metrics of all critical paths, the process integrity metrics are obtained.

[0094] For bottleneck equipment in the process, calculate the capacity matching index based on equipment availability and actual capacity;

[0095] Based on the preset index weighting coefficients, the process integrity index, capacity matching index, and overall equipment availability are weighted to obtain the process operability index.

[0096] In some embodiments, the bulk cargo terminal operation and maintenance perception method further includes the following steps:

[0097] The real-time operating data of the abnormal equipment is input into a pre-built equipment health prediction model to obtain the health change trend of the abnormal equipment.

[0098] When the health status trend shows that the predicted health status is lower than the set health status threshold, the corresponding maintenance suggestions and maintenance work orders are automatically generated.

[0099] When the health status change trend shows that the predicted health status is not lower than the set health status threshold, the operation data of the abnormal state device continues to be monitored, and the dynamic threshold is adaptively updated based on the continuously collected real-time operation data.

[0100] The real-time operating data, the abnormal status equipment information, the health change trend, the maintenance suggestions, and the process-level operation and maintenance situation awareness report are mapped to the three-dimensional digital twin model of the bulk cargo terminal. The real-time display of equipment-level monitoring, regional-level statistics, and terminal-level panoramic situation is realized through a multi-level situation display interface.

[0101] In some embodiments, the device health prediction model is constructed based on a long short-term memory neural network architecture, comprising a feature temporal modeling layer, a feature mapping layer, and a health output layer connected in sequence, wherein:

[0102] The feature time series modeling layer is used to perform hierarchical and progressive time series correlation modeling on the time series of real-time running data. The feature time series modeling layer includes a first LSTM hidden layer, a second LSTM hidden layer, and a third LSTM hidden layer arranged in sequence. The first LSTM hidden layer receives the time series input of real-time running data and outputs a time-step feature representation. A first random deactivation layer is set at its output to suppress model overfitting. The second LSTM hidden layer extracts medium- and long-term dependency features based on the time-step feature representation output by the first LSTM hidden layer and outputs an enhanced time-step feature representation. A second random deactivation layer is set at its output to improve the model's generalization ability. The third LSTM hidden layer performs convergence processing on the sequence features output by the second LSTM hidden layer and outputs a fixed-length feature vector.

[0103] The feature mapping layer is a fully connected layer used to receive a fixed-length feature vector output by the third LSTM hidden layer, and to perform non-linear mapping on the fixed-length feature vector through the ReLU activation function;

[0104] The health output layer is a single-neuron structure, and its activation function is the Sigmoid function, which is used to map the model output to a preset health range to obtain the device health score.

[0105] This application also aims to provide a bulk cargo terminal operation and maintenance awareness system based on equipment coding and dynamic thresholds, used to implement the aforementioned bulk cargo terminal operation and maintenance awareness method, including:

[0106] The data acquisition layer is used to collect real-time operational data of bulk cargo terminal equipment and environmental data.

[0107] The intelligent analysis layer, based on the data output from the data processing layer, performs dynamic threshold calculation and work process status analysis to intelligently perceive the equipment operating status and work process.

[0108] The decision support layer generates maintenance plans, emergency response schemes, and resource optimization schemes based on the output information of the intelligent analysis layer, providing a basis for decision-making in operation and maintenance management.

[0109] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0110] like Figure 1 As shown in the first aspect of the present invention, a bulk cargo terminal operation and maintenance perception system based on equipment coding and dynamic thresholds is provided. The system adopts a layered architecture design, which includes, from bottom to top, a data acquisition layer, a data processing layer, an intelligent analysis layer, a decision support layer, and a human-computer interaction layer. The layers interact and collaborate with each other through standardized interfaces and data buses, thereby constructing a complete operation and maintenance management system covering equipment perception, status analysis, decision generation, and execution feedback.

[0111] 1) The data acquisition layer includes a sensor network and a data acquisition gateway, wherein:

[0112] Sensor networks are used to perceive the operational status of key equipment at bulk cargo terminals in multiple dimensions, specifically including:

[0113] Vibration monitoring module: Triaxial vibration sensors are installed at the belt conveyor drive drum, stacker-reclaimer slewing mechanism and ship loader traveling mechanism. The sampling frequency is 20 kHz and the range is ±50 g. They are used to collect bearing and structural vibration characteristic data during equipment operation.

[0114] Temperature monitoring module: PT100 temperature sensors are installed in motor bearings, hydraulic oil tanks, brakes and other key components, with a measurement accuracy of ±0.1 ℃, to obtain the temperature rise status of the equipment;

[0115] Load monitoring module: Current sensors are installed at the belt conveyor drive motor, stacker-reclaimer lifting mechanism and other key actuators to monitor changes in equipment operating load in real time;

[0116] Tension monitoring module: A tension sensor is installed at the wire rope of the stacker-reclaimer lifting mechanism to collect data on changes in wire rope tension during operation;

[0117] Location monitoring module: Configures GPS and Beidou dual-mode positioning system and laser ranging sensor on mobile devices, with positioning accuracy of ±10 cm, used to obtain spatial location and walking trajectory information of the device;

[0118] Environmental monitoring module: Temperature, humidity, wind speed and salt spray concentration sensors are deployed in the work area to acquire work environment data;

[0119] Operation data acquisition module: Through interfaces with the operation planning system and logistics management system, it collects cargo flow data, operation plan data and equipment utilization data in real time.

[0120] The data acquisition gateway is used to aggregate, convert protocols, and initially buffer data collected by various sensors. Its specific configuration is as follows:

[0121] Hardware platform: Industrial-grade ARM processor, supporting operation in working environments ranging from -40 degrees Celsius to 70 degrees Celsius, with an IP67 protection rating;

[0122] Communication interface: Supports industrial communication protocols such as Modbus RTU / TCP, OPC UA and MQTT to enable data interaction with the host system;

[0123] Local storage module: configured with a 32 GB industrial-grade SSD to cache no less than 7 days of historical operating data, including operating data, environmental data and job data.

[0124] 2) The data processing layer is used for unified management, preprocessing, and feature construction of the collected equipment operation data, including an equipment coding management module and a data preprocessing module, wherein:

[0125] The device coding management module is configured as follows:

[0126] To achieve unified identification and attribute mapping for bulk cargo terminal equipment, the coding rule adopts the format "WHF-BT-A1-001-MTR", which represents the terminal code, equipment type, area location, equipment serial number and subsystem code in sequence.

[0127] A device coding dictionary is constructed based on the aforementioned coding rules. The coding dictionary is organized in a four-level tree structure to support the rapid retrieval and location of hundreds of thousands of device codes.

[0128] A device attribute dictionary is established that corresponds one-to-one with the device code. The attribute dictionary contains more than 200 attribute fields, including basic device information, technical parameters, maintenance history, spatial location information, and inter-device relationships, to support subsequent data analysis and status assessment.

[0129] The data preprocessing module is used to improve the reliability and usability of raw monitoring data. Its processing includes:

[0130] The isolated forest algorithm is used to detect anomalies in the collected data to identify abnormal points in the process, with an anomaly detection accuracy of no less than 95%.

[0131] Outlier data and outliers are cleaned based on the 3σ criterion and box plot analysis to ensure data continuity and quality consistency.

[0132] After data cleaning, feature extraction is performed on the processed data. Statistical features such as mean, variance, and peak value are extracted from the time domain, and spectral features such as power spectral density and dominant frequency are extracted from the frequency domain.

[0133] Furthermore, data from different sensors are fused using a Kalman filter to improve the stability and overall accuracy of the measurement results, and to output a feature data vector for the intelligent analysis layer.

[0134] 3) The intelligent analysis layer is used to perform in-depth analysis of the feature data output by the data processing layer, and realize equipment anomaly judgment, health status prediction and work process status assessment, including dynamic threshold calculation engine, predictive maintenance engine and process status analysis engine.

[0135] The dynamic threshold calculation engine is used to dynamically and adaptively adjust the alarm thresholds of equipment operation data. Its threshold calculation includes a base threshold and a comprehensive correction factor, wherein:

[0136] The basic threshold is determined based on the equipment's factory technical parameters and the statistical results of historical operating data;

[0137] The comprehensive correction factors include life cycle correction factors, environmental correction factors, and load correction factors. The life cycle correction factor is used to characterize the performance degradation characteristics of equipment as its service time increases, and a differentiated threshold adjustment strategy is adopted for new equipment and aging equipment. The environmental correction factor is used to reflect the impact of environmental parameters such as temperature, humidity, salt spray concentration, and wind speed on the operating status of the equipment. The load correction factor is used to adaptively correct the alarm thresholds of operating data such as vibration and temperature based on the real-time load rate of the equipment.

[0138] Furthermore, a Bayesian optimization algorithm is used to adaptively optimize the weight parameters of each correction factor based on the accuracy of historical alarm results, in order to improve the reliability of anomaly detection.

[0139] Predictive maintenance engines are used to predict and assess equipment health status and generate maintenance decisions. Their analysis process includes:

[0140] A device health status feature vector is constructed based on the feature data output by the data processing layer, and the feature vector contains no less than 200 feature parameters.

[0141] Long Short-Term Memory (LSTM) neural networks are used to predict the future operating status of equipment;

[0142] The predicted equipment health status is quantified into a health index from 0 to 100. A health index greater than 80 indicates that the equipment is operating well, a health index between 60 and 80 indicates that it needs to be closely monitored, and a health index less than 60 indicates that there is a significant risk of failure.

[0143] Based on the trend of equipment health changes and the probability of failure, maintenance suggestions are automatically generated and maintenance resources are optimized.

[0144] The process status analysis engine is used to assess the impact of equipment status changes on the overall work process. Its analysis methods include:

[0145] Construct a directed acyclic graph (DAG) workflow model covering the "unloading - belt conveyor - stockyard - material handling - loading" operation links;

[0146] Based on the aforementioned work process model, a three-dimensional relationship map is established, showing the physical connections, logical coordination, and work sequence dependencies between devices.

[0147] A graph neural network was used to analyze the propagation path and degree of impact of single-point equipment anomalies or failures in the work process;

[0148] The overall operability index of the bulk cargo terminal is calculated by combining equipment availability, operational process integrity, and capacity matching indicators to characterize the current operational status.

[0149] 4) The decision support layer is used to formulate executable maintenance and scheduling decisions based on the output of the intelligent analysis layer. It includes a maintenance strategy optimization module and an emergency response module, wherein:

[0150] The maintenance strategy optimization module is used to comprehensively optimize equipment maintenance activities. Its decision-making process includes:

[0151] Based on the equipment health assessment results and failure risk prediction information output by the predictive maintenance engine, and combined with maintenance resource constraints, the optimal maintenance plan for the equipment is automatically generated.

[0152] Based on the probability of equipment failure and historical consumption patterns, the demand for spare parts is predicted and analyzed, and the spare parts inventory structure and replenishment strategy are optimized accordingly.

[0153] By combining the skill requirements of maintenance tasks with information such as the skill level and available time of maintenance personnel, maintenance personnel can be intelligently matched and scheduled to improve maintenance execution efficiency.

[0154] The emergency response module provides rapid decision support when equipment malfunctions or fails. Its processing flow includes:

[0155] Based on the degree of equipment failure and its impact on the work process, the failure status is classified into three levels: Level I, Level II, and Level III. Level I failures correspond to immediate shutdown, Level II failures correspond to planned maintenance, and Level III failures correspond to continuous monitoring and handling.

[0156] For different fault levels and fault scenarios, corresponding emergency response plans are pre-constructed, including personnel dispatch plans, spare parts preparation plans, and work alternative plans.

[0157] During emergency response, the impact of equipment failure on production plans and work processes is assessed in real time, and the assessment results are fed back to the scheduling system to support production scheduling and resource allocation decisions.

[0158] 5) The human-computer interaction layer is used to provide a unified visual display and interaction entry for managers and on-site operators at different levels, so as to realize the intuitive presentation of the operation status and the effective issuance of business instructions. It includes a web-based management platform, a mobile inspection application, and a large-screen visualization system.

[0159] The web-based management platform provides administrators with comprehensive operation and maintenance and decision support, and its functions include:

[0160] The overall dashboard provides a centralized display of the terminal's overall operational status, the changing trends of key operational indicators, and statistical summary information on abnormal alarms.

[0161] It provides equipment monitoring functions to query the real-time operating status, historical operating data and alarm records of the equipment, and supports unified management of related information;

[0162] It provides maintenance management functions to support maintenance plan development, maintenance work order management, maintenance personnel scheduling, and maintenance material management;

[0163] It provides statistical analysis functions to analyze and display operational indicators such as equipment failure rate, maintenance cost, and equipment availability.

[0164] The mobile inspection app is used to support on-site inspection and maintenance operations, and its functions include:

[0165] It supports the receipt and execution of inspection tasks and maintenance work orders, and supports offline operations when the network is unavailable;

[0166] It supports collecting device status information and maintenance process records through methods such as taking photos, voice, and text, and sending the collected data back to the system;

[0167] By scanning the device's QR code using the mobile terminal's camera, the system obtains the corresponding device's real-time operating status, maintenance history, and operation guidance information, and provides augmented reality (AR) assisted display.

[0168] It provides expert support features, enabling remote video calls and image sharing to obtain technical guidance from remote experts.

[0169] The large-screen visualization system is used to centrally display the overall operational status of the terminal, and its functions include:

[0170] A 3D digital twin model of the dock was built using the Unity3D engine, and the real-time operating status of the equipment was mapped onto the 3D model.

[0171] The terminal's operational status is displayed in a multi-level view, supporting drill-down analysis by region and historical status review by time dimension;

[0172] The system uses visual methods such as color changes and flashing to intuitively present equipment malfunctions and warning information;

[0173] It dynamically displays the flow of goods, the collaborative operation status of equipment, and the progress of operations to assist in operation monitoring and scheduling decisions.

[0174] The aforementioned bulk cargo terminal operation and maintenance perception system based on equipment coding and dynamic thresholds, through a layered architecture and standardized interface design, achieves multi-dimensional perception and unified management of equipment operating status, environmental conditions, and operational processes. Relying on an equipment coding dictionary and a dynamic threshold adaptive mechanism, the system intelligently analyzes equipment health status, abnormal risks, and the impact of operational processes. Combined with predictive maintenance and emergency response strategies, it formulates executable operation and maintenance and scheduling decisions, thereby improving the safety, reliability, and overall operation and maintenance efficiency of bulk cargo terminals, reducing reliance on manual labor, and is suitable for intelligent operation and maintenance management applications under complex operating conditions.

[0175] like Figure 2 As shown in the second aspect embodiment of the present invention, a bulk cargo terminal operation and maintenance perception method based on equipment coding and dynamic thresholds is provided, applied to the bulk cargo terminal operation and maintenance perception system described in the first aspect of the present invention, and includes the following steps:

[0176] Based on the basic information of the equipment in the bulk cargo terminal, a unified coding system for the equipment in the bulk cargo terminal is constructed, and a dictionary of equipment attributes corresponding to the unified coding system is established.

[0177] Based on the device attribute dictionary and historical operating data statistical characteristics, combined with current environmental data, load status and device life cycle stage, the dynamic threshold corresponding to each operating data is calculated;

[0178] During equipment operation, real-time operating data of the equipment is continuously collected, and the real-time operating data is compared with the corresponding dynamic threshold. Equipment that exceeds the dynamic threshold is identified and marked as abnormal equipment.

[0179] Based on the node position of the abnormal state equipment in the pre-constructed bulk cargo terminal operation process model, analyze the scope and degree of influence of the abnormal state equipment on the operation process, and calculate the process operability index under the influence of the abnormal state equipment.

[0180] Based on the scope of impact, the degree of impact, and the process operability index, the overall operational capability of the current bulk cargo terminal is assessed, and a corresponding process-level operation and maintenance situation awareness report is obtained.

[0181] The aforementioned bulk cargo terminal operation and maintenance awareness method based on equipment coding and dynamic thresholds achieves accurate identification of abnormal equipment states through unified equipment coding and attribute mapping, adaptive calculation of dynamic thresholds, and comparison of real-time operational data. This method, combined with a bulk cargo terminal operation process model, analyzes the scope and degree of impact of abnormal equipment on the operation process, and assesses the operability of the process and overall operational capability accordingly, thereby generating process-level operation and maintenance situation awareness results, providing effective support for bulk cargo terminal operation management and operation and maintenance decisions.

[0182] Specifically, the unified coding system adopts a five-segment hierarchical structure, and its coding format is defined as: Terminal Code - Equipment Type - Area Location Code - Equipment Serial Number - Subsystem Code. Wherein: the terminal code uses a three-letter abbreviation to identify different terminal operation areas; the equipment type code is classified and identified according to the functional characteristics of the equipment, including but not limited to: BT for belt conveyors, SR for stacker-reclaimers, SL for ship loaders, UL for ship unloaders, GC for gantry cranes, etc.; the area location code uses a two-dimensional coordinate coding method combining letters and numbers to describe the spatial location information of the equipment in the terminal's plan layout; the equipment serial number uses a three-digit consecutive numeric code to sequentially number equipment of the same type located in the same area; and the subsystem code uses a three-letter abbreviation to identify the key functional subsystem corresponding to the equipment.

[0183] The unified equipment coding system described above, through a five-segment hierarchical structure design, uniformly identifies and systematically associates wharves, equipment types, spatial locations, equipment serial numbers, and subsystem functions, enabling precise positioning and classification management of bulk cargo terminal equipment. This coding system effectively supports equipment attribute mapping, status traceability, and cross-system data association, improving the standardization and consistency of equipment management and operation and maintenance analysis.

[0184] Specifically, the device attribute dictionary is organized using a four-level tree structure, including:

[0185] The equipment basic information layer is used to record the equipment's identification, classification, model, installation location, region, manufacturer, commissioning time, and other basic attribute information related to equipment identification, classification, and management.

[0186] The technical parameter layer is used to record the technical parameters that characterize the operation and performance of the equipment.

[0187] The maintenance record layer is used to record historical maintenance and repair information of the equipment.

[0188] The relationship layer is used to record the physical connection relationships, logical coordination relationships, and operation timing relationships between devices.

[0189] In some embodiments, the device attribute dictionary may also include basic alarm threshold configuration items corresponding to different device types and various operating data, or include reference parameters, standard limits and historical statistical baselines for generating basic alarm thresholds, so as to support dynamic threshold calculation for different devices and different operating data.

[0190] The device attribute dictionary described in this invention organizes device information hierarchically through a four-layer tree structure, unifying the modeling and management of static device attributes, operating technical parameters, maintenance history records, and inter-device relationships. This structure enables the systematic expression and efficient retrieval of device lifecycle information, providing complete and reliable data support for device status analysis, correlation impact assessment, and operation and maintenance decisions.

[0191] Specifically, the method for extracting the statistical features of the historical operating data is as follows:

[0192] Acquire time-series data of various operating data during the historical operation of the equipment, and classify the operating data according to the rate of change of the operating data;

[0193] Based on the classification results of the operational data, determine the corresponding statistical time windows for different types of operational data;

[0194] Within the statistical time window, time-domain statistical analysis is performed on the running data to extract time-domain statistical features that characterize the data fluctuation characteristics and trends.

[0195] For monitoring data with periodic characteristics, frequency domain transformation is performed on its time domain data to extract the corresponding frequency domain statistical features, and further calculation of the correlation statistical features between different operational data is performed.

[0196] The obtained time-domain statistical features, frequency-domain statistical features, and correlation statistical features are fused to form the historical operation data statistical features.

[0197] The historical operation data statistical feature extraction method of this invention classifies, windows, and performs multi-domain feature analysis on the time series of equipment operation data to systematically characterize the fluctuation characteristics and trends of equipment operation. This method integrates time domain, frequency domain, and parameter correlation features to fuse and model historical operation data, providing a stable and reliable data foundation for dynamic threshold calculation, status assessment, and anomaly detection.

[0198] Specifically, the operational data includes: bearing status data monitored by vibration sensors, equipment temperature data monitored by temperature sensors, load status data monitored by current sensors, wire rope tension data of the hoisting mechanism monitored by tension sensors, equipment spatial location information and travel trajectory data obtained by position sensors and satellite positioning systems, as well as operating condition parameter data generated during operation; the environmental data includes temperature, humidity, wind speed, and salt spray concentration parameters of the bulk cargo terminal operating area; and the operational data includes cargo flow data, operation plan data, and equipment utilization rate data.

[0199] Furthermore, the calculation method for the dynamic threshold is as follows:

[0200] The basic alarm thresholds for obtaining the current cumulative operating time, environmental conditions, current actual operating load, and target operating data of the device under monitoring are obtained. ;

[0201] Based on the current cumulative running time of the equipment to be monitored Compared with its designed service life The proportional relationship between them is used to calculate the life cycle correction factor. The calculation formula is as follows: ,in, This represents the life cycle correction factor;

[0202] Calculate the environmental correction factor based on the environmental conditions of the equipment to be monitored. The calculation formula is as follows: ,in, Indicates the humidity effect coefficient; This represents the relative humidity parameter; This indicates the salt spray concentration parameter; Indicates the salt spray effect coefficient; Indicates the influence coefficient of temperature deviation; Indicates ambient temperature; The standard reference ambient temperature for the equipment;

[0203] Based on the current actual operating load of the equipment to be monitored With rated operating load The deviation relationship between them is used to calculate the load correction factor. The calculation formula is as follows:

[0204]

[0205] in, Indicates the load correction factor; This represents the reference load ratio threshold for load correction, used to define the baseline point at which the load effect begins to have a significant impact.

[0206] The life cycle correction factor is adjusted according to the preset correction weight coefficient. Environmental correction factors and load correction factor A weighted fusion is performed to obtain the comprehensive correction factor. The calculation formula is as follows: ,in, , , These are the lifecycle weight coefficient, environment weight coefficient, and load weight coefficient, respectively, with values ​​ranging from 0 to 1, and satisfying the following conditions: ;

[0207] Based on the aforementioned basic alarm threshold and the aforementioned comprehensive correction factor Calculate and output dynamic threshold The calculation formula is as follows: .

[0208] The basic alarm threshold refers to the initial alarm baseline value corresponding to the target operating data of the monitored equipment under standard or reference operating conditions, which serves as the basis for dynamic threshold calculation. The target operating data is not limited to vibration parameters, but also includes one or more other monitoring parameters that can characterize the equipment's operating status and fault risk. For different equipment and different operating data, the basic alarm threshold can be set based on one or more of the following: equipment factory technical parameters, industry standards, design specifications, operation and maintenance procedures, historical healthy operating data statistics, and fault experience thresholds.

[0209] The dynamic threshold calculation method of this invention achieves adaptive dynamic adjustment of the alarm threshold by comprehensively incorporating multiple influencing factors such as equipment lifecycle, environmental conditions, and operating load on top of the basic alarm threshold. This method uses lifecycle correction factors, environmental correction factors, and load correction factors to quantitatively characterize the equipment's operating conditions, and then forms a comprehensive correction factor through weighted fusion, thereby obtaining a dynamic threshold that matches the actual operating state of the equipment, effectively improving the accuracy and adaptability of anomaly detection.

[0210] Furthermore, the method for constructing the bulk cargo terminal operation process model is as follows:

[0211] The various operational processes and key equipment of the bulk cargo terminal are abstracted as nodes, and the material flow direction or operational timing dependency between nodes is defined as an edge, thus constructing the basic topology of the operational process.

[0212] Configure each node with its corresponding equipment code, equipment type, rated processing capacity, current availability and health score, and configure each side with its corresponding material flow limit, conveying distance and energy consumption coefficient;

[0213] Based on the material flow direction or operation sequence, each node is connected in an orderly manner to obtain a complete logistics chain. The stacker-reclaimer completes the cargo storage operation in the yard and sends the cargo back into the belt conveyor system during the loading operation.

[0214] A directed acyclic graph is constructed based on the nodes and edges, where the direction of the edges is used to represent the material flow direction and the sequence of operation steps, resulting in a formalized bulk cargo terminal operation process model.

[0215] In the formalized bulk cargo terminal operation process model, constraint rules are introduced to model the physical connection relationships, logical coordination relationships, and operation timing dependencies between key equipment, resulting in a complete bulk cargo terminal operation process model. Among them, physical connection relationships are used to limit the reachable paths between nodes and follow the interlocking logic of starting with the coal flow and stopping with the coal flow; logical coordination relationships are used to define the parallel or mutual exclusion rules for multi-equipment collaborative operations; and operation timing dependencies are used to constrain the execution order of nodes.

[0216] For example, the operational steps corresponding to ship berthing and unloading are taken as the starting node of the process. After the ship completes berthing, the bulk cargo is unloaded by the ship unloader, and the unloaded material is fed into the belt conveyor. The material is continuously transferred by the belt conveyor according to the preset conveying path. During the transfer process, it can be allocated to different conveying branches according to the conveying distance and path topology. Subsequently, the material is transported to the storage operation area, where the stacker-reclaimer completes the storage or retrieval and allocation of the material to realize the transfer buffer and inventory management of the material. When it is necessary to carry out external transportation operations, the stacker-reclaimer feeds the material back into the belt conveyor and finally transports it to the ship loader, which completes the external loading operation. In the above process, each operational step corresponds to a different node, and the nodes are connected in an orderly manner through the material flow direction, thereby forming a continuous logistics chain structure covering the entire process of unloading, conveying, storage and loading.

[0217] The aforementioned bulk cargo terminal operation process model of this invention abstracts operational links and key equipment as nodes, and material flow and operational sequence relationships as directed edges, constructing a directed acyclic operational topology covering the entire process of unloading, conveying, storage, and loading. Based on unified equipment coding and capacity parameter configuration, this model introduces constraints such as physical connections, logical coordination, and operational sequence, achieving a formal expression of the operational process structure and operating mechanism. This provides a clear and computable process foundation for equipment anomaly impact analysis, process operability assessment, and maintenance decisions.

[0218] Furthermore, the method for analyzing the scope and degree of impact of abnormal equipment on the work process is as follows:

[0219] Locate the node position corresponding to the abnormal state equipment in the bulk cargo terminal operation process model, and determine its upstream and downstream correlation in the operation process;

[0220] Starting from the node corresponding to the abnormal device, perform traversal operations along the direction of the work process and in the opposite direction to identify the affected work path caused by the abnormal device.

[0221] Based on the affected work paths, determine the set of equipment and the set of work tasks affected by the abnormal equipment status;

[0222] Based on the set of affected equipment and the set of work tasks, calculate the scope of the impact of abnormal equipment on the work process, which is used to characterize the proportion of affected equipment or work tasks in the overall work process.

[0223] By combining the changes in the operational capabilities and operational delays of the affected objects, the impact of abnormal equipment on the operational process is quantitatively assessed, and the degree of impact of abnormal equipment on the operational process is obtained.

[0224] The present invention provides a method for analyzing the impact of abnormal equipment on work processes. This method locates abnormal equipment nodes within the work process model and identifies affected work links along upstream and downstream paths to determine the scope of the abnormal propagation. By combining the proportion of affected equipment to work tasks, the attenuation of work capacity, and delays, the method quantitatively assesses the impact of abnormal equipment on the overall work process, providing a clear basis for workability calculations and maintenance decisions.

[0225] Furthermore, the calculation method for the process operability index is as follows:

[0226] Get device availability list Equipment Importance Weight List and process association matrix ,in, This represents the number of devices involved in the calculation process.

[0227] For each critical path in the process Based on the availability of each device within the path Calculate path availability index This can be expressed as a formula: ;

[0228] Based on the path availability metrics of all critical paths Calculation process integrity index This can be expressed as a formula: ;

[0229] Bottleneck equipment in the process Based on equipment availability Compared with actual production capacity Calculate the capacity matching index This can be expressed as a formula: ;

[0230] Based on the set weights, the process integrity indicators Capacity matching index and overall equipment availability The process operability index is obtained by weighting. This can be expressed as a formula: ,in, , , These are the weights for process integrity indicators, capacity matching indicators, and overall equipment availability indicators, respectively, and they must satisfy the following conditions: + + =1.

[0231] The method for calculating the process operability index of this invention quantitatively assesses the overall operational capability of bulk cargo terminal operations by comprehensively considering critical path availability indicators, process integrity indicators, and bottleneck equipment capacity matching. Based on equipment availability, equipment importance weights, and process correlations, this method identifies critical paths and bottlenecks affecting process continuity and forms a process operability index through weighted fusion. This objectively reflects the executable level of the operation process under the current equipment status, providing an intuitive and comparable quantitative basis for operation and maintenance decisions and production scheduling.

[0232] Furthermore, the method for assessing the overall operational capacity of the bulk cargo terminal is as follows:

[0233] Construct a multi-dimensional evaluation index system for the overall operational capacity of bulk cargo terminals. The evaluation index system shall include at least throughput capacity, operational efficiency, equipment utilization rate, energy efficiency and safety level, and set target values ​​and warning values ​​for each evaluation index.

[0234] Real-time operating data, operation process data, energy consumption data, and safety-related data of bulk cargo terminal equipment are collected under a unified time reference.

[0235] Based on the aforementioned work process model, the processing capacity of each work step is calculated, the bottleneck work step with the lowest processing capacity is identified, and the actual processing capacity of the bottleneck work step is used as the current actual throughput capacity of the terminal.

[0236] Based on the operational process data, ship time efficiency, berth utilization rate and equipment utilization rate are calculated, and operational efficiency and equipment utilization rate are analyzed in combination with operational interruption time and equipment load.

[0237] Based on the energy consumption data, the energy consumption index per unit cargo throughput is calculated, and based on the safety data, the safety risk index is calculated to analyze energy efficiency and safety level.

[0238] The analysis results of throughput capacity, operational efficiency, equipment utilization, energy efficiency and safety level are normalized and comprehensively evaluated. Combined with the impact range and degree of abnormal equipment and the process operability index, the overall operational capacity of the bulk cargo terminal is assessed.

[0239] The present invention provides a method for assessing the overall operational capacity of bulk cargo terminals. This method constructs a multi-dimensional evaluation index system encompassing throughput capacity, operational efficiency, equipment utilization, energy efficiency, and safety levels to systematically analyze the terminal's operational status. By combining operational data under a unified time series, operational process models, and anomaly impact analysis, this method identifies bottlenecks and comprehensively evaluates the performance of each index. Based on the fusion of process operability indices, it generates an overall operational capacity assessment result, providing a comprehensive and quantitative decision-making basis for bulk cargo terminal operation optimization and maintenance decisions.

[0240] Furthermore, the process-level operation and maintenance situation awareness report is automatically generated using a structured template, including six modules: execution summary, equipment status overview, anomaly analysis details, process impact assessment, maintenance suggestion plan, and trend prediction analysis.

[0241] The execution summary module is used to summarize and present key status information, including overall health score, process operability index, number of abnormal devices, number of pending maintenance work orders, and estimated daily operational capacity;

[0242] The Equipment Status Overview module displays the health status distribution of various operating systems and equipment types using categorized statistical charts, and enables quick identification through color coding.

[0243] The anomaly analysis details module provides diagnostic information for abnormal or faulty equipment, including equipment identification, anomaly time, parameter exceeding limits, trend characteristics, possible fault modes, and maintenance records.

[0244] The process impact assessment module uses network topology diagrams and comparative charts to quantitatively demonstrate the impact of abnormal equipment on work processes and operational capabilities.

[0245] The maintenance suggestion module proposes tiered handling recommendations based on the scope and urgency of the anomaly.

[0246] The trend prediction analysis module identifies potential future anomalies and failure risks based on equipment state evolution and prediction model results, providing decision support for preventive maintenance.

[0247] The aforementioned process-level operation and maintenance situation awareness report of this invention is automatically generated through a structured template, which uniformly summarizes and expresses equipment status, anomaly information, and process impact results. With an execution summary at its core, and combined with modules such as equipment status overview, anomaly analysis, process impact assessment, maintenance recommendations, and trend prediction, this report provides a clear and actionable basis for decision-making by operation and maintenance management personnel.

[0248] like Figure 3 As shown in the third aspect embodiment of the present invention, a bulk cargo terminal operation and maintenance perception method based on equipment coding and dynamic thresholds is provided, applied to the bulk cargo terminal operation and maintenance perception system described in the first aspect of the present invention, and further includes the following steps:

[0249] The real-time operating data of the abnormal equipment is input into a pre-built equipment health prediction model to obtain the health change trend of the abnormal equipment.

[0250] When the health status trend shows that the predicted health status is lower than the set health status threshold, the corresponding maintenance suggestions and maintenance work orders are automatically generated.

[0251] When the health status change trend shows that the predicted health status is not lower than the set health status threshold, the operation data of the abnormal state device continues to be monitored, and the dynamic threshold is adaptively updated based on the continuously collected real-time operation data.

[0252] The real-time operating data, the abnormal status equipment information, the health change trend, the maintenance suggestions, and the process-level operation and maintenance situation awareness report are mapped to the three-dimensional digital twin model of the bulk cargo terminal. The real-time display of equipment-level monitoring, regional-level statistics, and terminal-level panoramic situation is realized through a multi-level situation display interface.

[0253] The aforementioned bulk cargo terminal operation and maintenance perception method based on equipment coding and dynamic thresholds combines real-time data from equipment in abnormal states with an equipment health prediction model to achieve continuous assessment and tiered handling of equipment health trends. This method automatically generates maintenance suggestions and work orders while continuously monitoring and adaptively updating thresholds for equipment that has not reached maintenance thresholds. It also maps equipment status, prediction results, and process-level situational information to a three-dimensional digital twin model, enabling multi-level, visualized operation and maintenance status display, thereby improving the intelligence level and response efficiency of bulk cargo terminal operation and maintenance management.

[0254] like Figure 4 As shown, the device health prediction model is built on a long short-term memory neural network architecture, including a feature temporal modeling layer, a feature mapping layer, and a health output layer connected in sequence, wherein:

[0255] The feature time series modeling layer is used to perform hierarchical and progressive time series correlation modeling on the time series of device operation data. The feature time series modeling layer includes a first LSTM hidden layer, a second LSTM hidden layer, and a third LSTM hidden layer set in sequence. The first LSTM hidden layer receives the time series input of device operation data and sets the return_sequences parameter to True to output a time-step feature representation. A first random deactivation layer is set at its output, and the dropout rate is set to 0.2 to suppress model overfitting. The second LSTM hidden layer extracts medium- and long-term dependency features based on the time-step feature representation output by the first LSTM hidden layer. It also sets return_sequences to True to output an enhanced time-step feature representation. A second random deactivation layer is set at its output to improve the model's generalization ability. The third LSTM hidden layer performs convergence processing on the sequence features output by the second LSTM hidden layer and outputs a fixed-length feature vector, which is then connected to a feature mapping layer. The activation function uses ReLU to introduce non-linear transformation capability.

[0256] The feature mapping layer is a fully connected layer used to receive a fixed-length feature vector output by the third LSTM hidden layer;

[0257] The health output layer is a single-neuron structure, and its activation function is the Sigmoid function, which is used to map the model output to a range of 0 to 1, and convert it into a health score by multiplying it by 100.

[0258] The first and second LSTM hidden layers each have 128 neuron units, the third LSTM hidden layer has 64 neuron units, and the fully connected layer has 32 neuron units.

[0259] The equipment health prediction model was trained using the Adam optimizer with an initial learning rate of 0.001. The mean squared error (MSE) was used as the loss function to measure the deviation between the predicted and actual values, and the mean absolute error (MAE) was used as the evaluation metric to quantify the prediction accuracy. The training dataset contained equipment status feature vectors with more than 200 dimensions, covering time-domain statistical features such as mean, variance, peak value, and peak-to-peak value, frequency-domain features such as power spectral density, dominant frequency components, and harmonic energy distribution, as well as contextual information such as equipment operating time, load history, and maintenance records.

[0260] The device health prediction model supports health prediction for three time windows: 7 days, 15 days, and 30 days. The prediction accuracy reaches over 85% on the validation set. The health score ranges from 0 to 100 points, where a score greater than 80 indicates that the device is in a healthy state, a score between 60 and 80 indicates a state requiring attention, and a score less than 60 indicates a warning state where there is an anomaly and timely maintenance is needed.

[0261] The equipment health prediction model of this invention is based on a Long Short-Term Memory (LSTM) neural network. It progressively models the time-series features of equipment operation data using a multi-layer LSTM and combines a fully connected mapping with a health output structure to achieve quantitative prediction of equipment health status. This model comprehensively utilizes multi-dimensional time-domain features, frequency-domain features, and operation and maintenance context information to make short- to medium-term predictions of equipment health trends. It can uniformly map the prediction results into an intuitive health score, providing a reliable basis for anomaly warning, maintenance decisions, and operational status analysis.

[0262] Furthermore, the method for generating maintenance suggestions and maintenance work orders is as follows:

[0263] Devices whose predicted health level is lower than a set health level threshold are marked as target devices, and their operating status feature information is extracted. At the same time, the operating status feature information is matched with a preset fault mode library to determine the fault type of the target device.

[0264] The severity level of the target equipment's failure is determined by combining the decline in health prediction results with the degree of deviation of key operating parameters.

[0265] Based on the fault type and fault severity level, the corresponding maintenance handling strategy is matched from the preset maintenance strategy library to generate maintenance suggestions corresponding to the operating status of the target equipment;

[0266] Based on the maintenance recommendations, maintenance work orders are automatically generated. The maintenance work orders include at least the target equipment code, fault type, fault severity level, estimated fault time, and resource requirements.

[0267] Maintenance work orders are prioritized based on the severity of the fault, and the execution time window and execution resources for maintenance work orders are scheduled and configured in conjunction with the work plan, equipment operating status, spare parts inventory, and maintenance personnel capabilities.

[0268] The maintenance suggestion and maintenance work order generation method of this invention combines health prediction results with fault mode recognition to achieve intelligent determination of the fault type and severity level of the target equipment. Based on a matched maintenance strategy, this method automatically generates targeted maintenance suggestions and produces maintenance work orders containing equipment identification, fault information, and resource requirements. Simultaneously, it prioritizes and configures resource scheduling for work orders according to fault level and operational constraints, thereby improving the timeliness of maintenance response and resource utilization efficiency.

[0269] Furthermore, the method for constructing a three-dimensional digital twin model of a bulk cargo terminal is as follows:

[0270] Acquire spatial real-view data and equipment structure data of bulk cargo terminals. The spatial real-view data is used to characterize the overall spatial layout of the terminal, and the equipment structure data is used to describe the structural characteristics of the equipment within the terminal.

[0271] A holistic spatial scene model of the wharf is constructed based on real-world spatial data to form the spatial foundation for the digital twin model;

[0272] Based on the equipment structure data, a three-dimensional solid model of the equipment in the dock is constructed, and the equipment is arranged in the overall spatial scene model according to its corresponding spatial position.

[0273] Multi-level detail models are set for the overall spatial scene model and the 3D solid model of the equipment so as to dynamically adjust the rendering accuracy according to the observation conditions;

[0274] For working equipment with motion characteristics, establish a corresponding equipment motion model and define its motion degrees of freedom and motion range;

[0275] Acquire real-time operating status information of the equipment and map it to the corresponding equipment motion model to achieve synchronization between the physical equipment and the three-dimensional digital twin model;

[0276] The system integrates equipment operation information, historical information, and predictive information into a 3D digital twin model and provides an interactive interface for the visualization of equipment status.

[0277] An environmental simulation and situation presentation mechanism is introduced to dynamically simulate changes in the dock environment, thereby enhancing the expressive power of the three-dimensional digital twin model.

[0278] The present invention discloses a method for constructing a 3D digital twin model of a bulk cargo terminal. By integrating real-world data of the terminal space with equipment structure data, a 3D visualization model covering the overall scene and key equipment is built. Combined with multi-level detailed control and equipment motion modeling, a realistic mapping of the terminal's operational status is achieved. This method synchronously maps real-time equipment operating information, historical data, and prediction results into the 3D model and introduces environmental simulation and situational awareness mechanisms. It supports multi-dimensional interaction and dynamic display, providing intuitive and efficient visualization support for bulk cargo terminal operation and maintenance monitoring, situational awareness, and decision support.

[0279] To verify the effectiveness of the bulk cargo terminal operation and maintenance perception method based on equipment coding and dynamic threshold disclosed in this application under actual working conditions, the following uses the BC3-1 belt conveyor of a large coal terminal as a case study.

[0280] In March 2024, the BC3-1 belt conveyor (equipment code TSN-A1-BC-3-1) at a large coal terminal experienced intermittent abnormal noises during operation. Conventional manual inspections and planned maintenance failed to identify any clear anomalies. To prevent the potential for further escalation of the fault, a port intelligent integrated monitoring platform was constructed based on the technical solution of this application. This platform is used for systematic diagnosis, data collection, and evaluation of the equipment's operating status, enabling comprehensive perception and early warning of equipment anomalies.

[0281] First, the BC3-1 belt conveyor was quickly located and its information associated using an equipment coding dictionary. This revealed its model as B1600, design capacity as 5000 t / h, commissioning date as June 2019, cumulative operating hours as 41850, and key component as the SKF22230CC / W33 drive roller bearing. Historically, it underwent 6 planned maintenance and 2 fault repairs. Based on this, multi-source operational data was collected and managed uniformly, including key parameters such as the three-dimensional vibration of the drive roller bearing, motor temperature, real-time load level, and ambient humidity and salt spray concentration, providing comprehensive data support for subsequent analysis.

[0282] During the dynamic threshold calculation process, the basic alarm thresholds corresponding to various operating data of the BC3-1 belt conveyor are obtained respectively, and the basic alarm thresholds are dynamically adjusted based on the cumulative operating time, environmental conditions and actual operating load level to generate dynamic thresholds corresponding to various operating data.

[0283] To verify the effectiveness of the dynamic threshold calculation method, the vibration parameters of the BC3-1 belt conveyor were selected as an example for illustration: Using the basic vibration alarm threshold of 7.1 mm / s given in ISO 10816-3 as a benchmark, a multi-factor correction mechanism was introduced, incorporating equipment lifecycle, environmental conditions, and load status, to calculate the lifecycle correction factor. Environmental correction factors and load correction factor And obtain the comprehensive correction factor through a weighted method. The value is 1.076, thus establishing a dynamic threshold of 7.64 mm / s corresponding to the vibration parameters adapted to the current operating conditions of the equipment. Although the current vibration amplitude of the equipment does not exceed this dynamic threshold, trend analysis identifies a continuous upward trend in the vibration level, triggering further health prediction and fault diagnosis processes.

[0284] Based on the equipment's historical operating data over the past 60 days, a health prediction model was used to model and analyze the state evolution of the BC3-1 belt conveyor. The model predicts that its health score will decrease from the current 72 points to approximately 54 points within 30 days. Combined with the spectral feature identification results, the equipment was determined to have a bearing inner ring defect with a diagnostic confidence level of 84%. The typical characteristic of this defect is a significant sideband at 1.3 times the rotational frequency. The fault is expected to become apparent within 25 to 35 days.

[0285] Based on this, the impact of this potential failure on the overall terminal operation process was further assessed. The analysis results show that if the BC3-1 conveyor belt fails, it will directly lead to the interruption of material handling operations in area A1, and cause an increase of approximately 15% in the load of the BC3-2 and BC3-3 conveyor belts, resulting in a capacity loss of approximately 300 t / h. The overall process operability index will decrease from 96% to 89%. Although a backup conveyor belt can be used as a replacement, it still requires approximately 4 hours of switchover time, significantly impacting continuous operation.

[0286] Based on the above diagnostic and impact assessment results, predictive maintenance decision recommendations were automatically generated, suggesting maintenance operations be carried out during the low-intensity period of the coming week, specifying the required spare parts, personnel allocation, and operation duration. During the actual maintenance, a crack of approximately 0.8 mm was confirmed in the bearing inner ring, highly consistent with the prediction. The maintenance took 5.5 hours, better than originally planned. After maintenance, the equipment vibration level was significantly reduced, the operating temperature returned to normal, and the equipment health score improved to 94 points.

[0287] The effectiveness of the predictive maintenance scheme was verified during the subsequent three months of operation. The deviation between the predicted fault type and occurrence time was controlled within ±3 days, equipment availability increased from 94.2% to 98.1%, maintenance time was reduced by 25%, and the risk of unplanned downtime was significantly reduced. Based on maintenance feedback, adaptive optimization of model parameters, updates to the fault feature library, and adjustments to maintenance strategies were completed simultaneously, optimizing the predictive maintenance cycle for this model of equipment from quarterly to monthly.

[0288] In summary, the technical solution presented in this application, through a unified equipment coding dictionary, multi-factor dynamic threshold algorithm, and process-level situational awareness analysis, achieves a shift from single-point equipment monitoring to system-level operation and maintenance decision-making. Compared with existing technologies, it demonstrates significant advantages in equipment location efficiency, alarm accuracy, and process impact assessment capabilities. Verified through application at multiple bulk cargo terminals, this technical solution has achieved significant technical and economic benefits in improving equipment availability, reducing operation and maintenance costs, optimizing management decisions, and enhancing overall operational efficiency.

[0289] The above embodiments are used to explain this application, not to limit it. Any modifications and changes made to this application within the spirit and scope of the claims shall fall within the protection scope of this application.

Claims

1. A bulk cargo terminal operation and maintenance perception method based on equipment coding and dynamic thresholds, characterized in that, Includes the following steps: Based on the basic information of the equipment in the bulk cargo terminal, a unified coding system for the equipment in the bulk cargo terminal is constructed, and a dictionary of equipment attributes corresponding to the unified coding system is established. Based on the device attribute dictionary and historical operating data statistical characteristics, combined with current environmental data, load status and device life cycle stage, the dynamic threshold corresponding to each operating data is calculated; During equipment operation, real-time operating data of the equipment is continuously collected, and the real-time operating data is compared with the corresponding dynamic threshold. Equipment that exceeds the dynamic threshold is identified and marked as abnormal equipment. Based on the node position of the abnormal state equipment in the pre-constructed bulk cargo terminal operation process model, analyze the scope and degree of influence of the abnormal state equipment on the operation process, and calculate the process operability index under the influence of the abnormal state equipment. Based on the scope of impact, the degree of impact, and the process operability index, the overall operational capability of the current bulk cargo terminal is assessed, and a corresponding process-level operation and maintenance situation awareness report is obtained. The calculation method for the dynamic threshold is as follows: Obtain the basic alarm threshold based on the current cumulative operating time of the monitored device, its environmental conditions, current actual operating load, and target operating data; obtain a lifecycle correction factor based on the proportional relationship between the current cumulative operating time of the monitored device and its designed lifespan; obtain an environmental correction factor based on the environmental conditions of the monitored device; obtain a load correction factor based on the deviation relationship between the current actual operating load and the rated operating load of the monitored device; weight and fuse the lifecycle correction factor, environmental correction factor, and load correction factor according to preset correction weight coefficients to obtain a comprehensive correction factor; multiply the basic alarm threshold by the comprehensive correction factor to obtain the dynamic threshold.

2. The bulk cargo terminal operation and maintenance awareness method according to claim 1, characterized in that, The device attribute dictionary is organized using a four-level tree structure, including: The equipment basic information layer is used to record the equipment's identification, classification, model, installation location, region, manufacturer, commissioning time, and other basic attribute information related to equipment identification, classification, and management. The technical parameter layer is used to record the technical parameters that characterize the operation and performance of the equipment. The maintenance record layer is used to record historical maintenance and repair information of the equipment. The relationship layer is used to record the physical connection relationships, logical coordination relationships, and operation timing relationships between devices.

3. The bulk cargo terminal operation and maintenance sensing method according to claim 1, characterized in that, The operational data includes bearing status data, equipment temperature data, load status data, hoisting mechanism wire rope tension data, equipment spatial location information and travel trajectory data, as well as operating condition parameter data generated during operation; the environmental data includes temperature, humidity, wind speed and salt spray concentration parameters of the bulk cargo terminal operation area.

4. The bulk cargo terminal operation and maintenance awareness method according to claim 1, characterized in that, The method for constructing the bulk cargo terminal operation process model is as follows: The various operational processes and key equipment of the bulk cargo terminal are abstracted as nodes, and the material flow direction or operational timing dependency between nodes is defined as an edge, thus constructing the basic topology of the operational process. Configure each node with its corresponding equipment code, equipment type, rated processing capacity, current availability and health score, and configure each side with its corresponding material flow limit, conveying distance and energy consumption coefficient; Based on the material flow direction or operation sequence, each node is connected in an orderly manner to obtain a complete logistics chain; A directed acyclic graph is constructed based on the nodes and edges, where the direction of the edges is used to represent the material flow direction and the sequence of operation steps, resulting in a formalized bulk cargo terminal operation process model. In the formalized bulk cargo terminal operation process model, the physical connection relationship, logical coordination relationship and operation sequence dependency relationship between key equipment are modeled to obtain a complete bulk cargo terminal operation process model.

5. The bulk cargo terminal operation and maintenance sensing method according to claim 1, characterized in that, The method for analyzing the scope and degree of impact of abnormal equipment on the work process is as follows: Locate the node position corresponding to the abnormal state equipment in the bulk cargo terminal operation process model, and determine its upstream and downstream correlation in the operation process; Starting from the node corresponding to the abnormal device, perform traversal operations along the direction of the work process and in the opposite direction to identify the affected work path caused by the abnormal device. Based on the affected work paths, determine the set of equipment and the set of work tasks affected by the abnormal equipment status; Based on the set of affected equipment and the set of work tasks, calculate the scope of the impact of abnormal equipment on the work process, which is used to characterize the proportion of affected equipment or work tasks in the overall work process. By combining the changes in the operational capabilities and operational delays of the affected objects, the impact of abnormal equipment on the operational process is quantitatively assessed, and the degree of impact of abnormal equipment on the operational process is obtained.

6. The bulk cargo terminal operation and maintenance awareness method according to claim 1, characterized in that, The calculation method for the process operability index is as follows: Obtain the equipment availability list, the equipment importance weight list, and the process association matrix; For each critical path in the process, calculate the path availability index based on the availability of each device within the path; Based on the path availability metrics of all critical paths, the process integrity metrics are obtained. For bottleneck equipment in the process, calculate the capacity matching index based on equipment availability and actual capacity; Based on the preset index weighting coefficients, the process integrity index, capacity matching index, and overall equipment availability are weighted to obtain the process operability index.

7. The bulk cargo terminal operation and maintenance awareness method according to claim 1, characterized in that, The bulk cargo terminal operation and maintenance perception method also includes the following steps: The real-time operating data of the abnormal equipment is input into a pre-built equipment health prediction model to obtain the health change trend of the abnormal equipment. When the health status trend shows that the predicted health status is lower than the set health status threshold, the corresponding maintenance suggestions and maintenance work orders are automatically generated. When the health status change trend shows that the predicted health status is not lower than the set health status threshold, the operation data of the abnormal state device continues to be monitored, and the dynamic threshold is adaptively updated based on the continuously collected real-time operation data. The real-time operating data, the abnormal status equipment information, the health change trend, the maintenance suggestions, and the process-level operation and maintenance situation awareness report are mapped to the three-dimensional digital twin model of the bulk cargo terminal. The real-time display of equipment-level monitoring, regional-level statistics, and terminal-level panoramic situation is realized through a multi-level situation display interface.

8. The bulk cargo terminal operation and maintenance awareness method according to claim 7, characterized in that, The device health prediction model is built on a long short-term memory neural network architecture, comprising a feature temporal modeling layer, a feature mapping layer, and a health output layer connected in sequence, wherein: The feature time series modeling layer is used to perform hierarchical and progressive time series correlation modeling on the time series of real-time running data. The feature time series modeling layer includes a first LSTM hidden layer, a second LSTM hidden layer, and a third LSTM hidden layer arranged in sequence. The first LSTM hidden layer receives the time series input of real-time running data and outputs a time-step feature representation. A first random deactivation layer is set at its output to suppress model overfitting. The second LSTM hidden layer extracts medium- and long-term dependency features based on the time-step feature representation output by the first LSTM hidden layer and outputs an enhanced time-step feature representation. A second random deactivation layer is set at its output to improve the model's generalization ability. The third LSTM hidden layer performs convergence processing on the sequence features output by the second LSTM hidden layer and outputs a fixed-length feature vector. The feature mapping layer is a fully connected layer used to receive a fixed-length feature vector output by the third LSTM hidden layer, and to perform non-linear mapping on the fixed-length feature vector through the ReLU activation function; The health output layer is a single-neuron structure, and its activation function is the Sigmoid function, which is used to map the model output to a preset health range to obtain the device health score.

9. A bulk cargo terminal operation and maintenance perception system based on equipment coding and dynamic thresholds, used to implement the bulk cargo terminal operation and maintenance perception method as described in any one of claims 1-8, characterized in that, include: The data acquisition layer is used to collect real-time operational data of bulk cargo terminal equipment and environmental data. The intelligent analysis layer, based on the data output from the data acquisition layer, performs dynamic threshold calculation and work process status analysis to intelligently perceive the equipment operating status and work process. The decision support layer generates maintenance plans, emergency response schemes, and resource optimization schemes based on the output information of the intelligent analysis layer, providing a basis for decision-making in operation and maintenance management.

Images