Method and system for supervising maintenance operation risk of tug system based on Internet of Things

Abstract

The invention discloses a tug system maintenance operation risk supervision method and system based on the Internet of Things, and relates to the technical field of the Internet of Things, and the method comprises the steps: collecting multi-physical-quantity data, accessing the Internet of Things, carrying out the preprocessing of original data, and generating standard input; inputting the preprocessed data to a coupling resolver, capturing characteristics among physical quantities, and dividing a coupling state; generating risk factors based on abnormal coupling, and constructing a first safety potential energy surface and a second safety potential energy surface; comparing the two types of potential energy surface data, and outputting execution admission judgment of the maintenance step; and collecting the track of the maintenance personnel, comparing with the path template, and triggering the behavior lock if deviation occurs. According to the invention, through a cooperation mechanism of multi-physical quantity coupling analysis, potential energy surface admission determination and personnel trajectory supervision, real-time risk identification and safe controllable execution of the whole tug maintenance operation process are realized.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) technology, and in particular to a method and system for risk monitoring of mechanical maintenance operations of tugboat systems based on IoT. Background Technology

[0002] During long-term operation, tugboat systems continuously generate various physical effects, including heat, pressure, vibration, and electrical loads, from their main engine, auxiliary machinery, hydraulic systems, electrical control equipment, and air handling systems. These physical quantities exhibit dynamic coupling relationships. Existing tugboat maintenance risk monitoring technologies primarily rely on single-sensor alarms, experience-based judgment, and manual inspections, lacking effective means to identify the correlations and changes between multiple physical quantities. When equipment initially exhibits abnormalities, it often manifests as a shift in the coupling between multiple physical quantities. Traditional methods cannot decompose and identify these coupling relationships, frequently resulting in delayed detection of early risk warnings. Furthermore, current technologies typically do not discretize the engine room space, making it difficult to correlate the collected data with specific spatial locations and thus hindering accurate spatial risk assessment.

[0003] In terms of maintenance procedure execution management, existing systems mostly adopt fixed process control or simple threshold judgment methods, lacking the ability to dynamically determine access based on actual physical conditions. Single-variable threshold judgment methods cannot reflect the comprehensive effects of multiple sources of risk, such as thermal pressure superposition, combustible gas accumulation, or frictional heating, nor can they describe the changing trends of risk in spatial and temporal dimensions. Traditional technologies have not formed a risk quantification framework based on multiple physical quantity characteristics, and cannot determine the permissible, suspended, or prohibited states of maintenance procedures through scientific mechanisms. This results in the safety of maintenance procedure execution relying on human experience, exhibiting significant uncertainty and lag.

[0004] In terms of personnel behavior monitoring during maintenance, existing technologies mostly rely on video surveillance or visual inspections, which cannot obtain the real-time three-dimensional position of maintenance personnel within the cabin, nor can they perform path-level dynamic comparisons of personnel behavior. Therefore, traditional monitoring methods struggle to promptly detect personnel deviating from safe areas or entering high-risk areas, and cannot execute automated safety linkage measures when abnormal behavior occurs. Existing technologies lack a mechanism to compare maintenance path templates with real-time personnel trajectories, failing to achieve immediate identification of behavioral deviations, which can easily lead to personnel unknowingly entering dangerous areas.

[0005] Therefore, how to provide a method and system for risk monitoring of mechanical maintenance operations based on IoT tugboat systems is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] One objective of this invention is to propose a risk monitoring method and system for tugboat maintenance operations based on the Internet of Things (IoT). This invention utilizes technologies such as multi-physical quantity data acquisition and preprocessing, multi-state physical coupling decomposition, risk factor quantification, safety potential energy surface construction, and personnel trajectory comparison to form a comprehensive risk monitoring mechanism consisting of equipment status perception, maintenance step access determination, and personnel behavior monitoring. By decomposing the physical quantity coupling relationship into multi-state states, mapping risks to hierarchical safety potential energy surfaces and comparing differences, and using behavioral locks to dynamically constrain personnel paths, this invention achieves pre-identification of risks, scientific access to steps, and real-time safety control of behavior. This invention possesses advantages such as high security, scientific judgment criteria, timely risk response, and strong applicability, and can significantly improve the safety assurance level of tugboat maintenance operations.

[0007] A method for risk monitoring of mechanical maintenance operations based on an Internet of Things (IoT) tugboat system according to an embodiment of the present invention includes: Multi-physical quantity data is collected by various sensors deployed in the tugboat's engine room and connected to the Internet of Things. The multi-physical quantity data is preprocessed to obtain preprocessed multi-physical quantity data. The preprocessed multi-physical quantity data is input into the multi-state physical coupling decomposer to capture coupling features and classify the coupling relationship into three types of coupling states: steady-state coupling, transitional coupling, and abnormal coupling. The risk of coupling behavior is quantified by using abnormal coupling state, and abnormal coupling is converted into risk factor. The characteristic physical quantity parameters in the preprocessed multi-physical quantity data are mapped to the first safety potential energy surface, and the second safety potential energy surface is constructed based on the risk factor. The differences between the first and second safety potential energy surfaces are compared, and the entry conditions for maintenance steps are determined based on the separation, contact, or intersection relationship between the two. When the first and second safety potential energy surfaces remain separated, the maintenance steps are allowed to be executed. When the first and second safety potential energy surfaces are in partial contact, the maintenance steps are temporarily suspended. When the first and second safety potential energy surfaces form a continuous intersection area, the maintenance steps are prohibited from being executed. When the maintenance procedure is permitted to be executed, real-time 3D trajectory data of the maintenance personnel is collected and compared with the preset maintenance procedure safety path template. When the personnel's trajectory deviates from the safety path template or enters the preset restricted area, a behavior lock is triggered to stop the execution of the maintenance procedure and output a warning message.

[0008] Optionally, the multi-physical quantity data includes temperature data, pressure data, vibration data, air volume data, humidity data, combustible gas concentration data, voltage data, and current data.

[0009] Optionally, the preprocessing of the multi-physical quantity data includes performing denoising, filtering, time synchronization, and feature extraction on the multi-physical quantity data.

[0010] Optionally, the classification of coupling relationships into three types of coupling states—steady-state coupling, transitional coupling, and abnormal coupling—includes: The preprocessed multi-physical quantity data is input into the multi-state physical coupling decomposer. The decomposer is connected in series in the order of coupling sensing layer, state decomposition layer and risk indication layer. The coupling sensing layer is equipped with a dual-window differential scanning unit, an order relation encoding unit and a redundancy consistency check unit. For each physical quantity, the sign of change rate, the level of change amplitude and the order of occurrence of extreme values ​​are extracted in the short time window and the long time window respectively. The difference results of the short window and the long window are compared to generate start flag and continuous flag. After performing consistency check on the redundant sensor data of the same measurement point, a candidate coupling pair list and a sensing feature set are output. The state decomposition layer receives the sensing feature set and classifies the candidate coupling pairs according to the intensity criterion, the persistence criterion, and the order consistency criterion. Each coupling pair is classified as steady-state coupling, transitional coupling, or abnormal coupling. Couple pairs that meet two criteria but not the third criterion are assigned warning sub-labels. Adjacent coupling pairs are merged into coupling clusters and a coupling cluster list is generated based on the spatial proximity of the sensors in the cabin grid. The risk indication layer constructs a coupling strength matrix and a coupling label matrix based on the coupling cluster list. It references a fixed weight table and performs contextual cross-calibration by combining the current maintenance step identifier, ventilation status and related equipment on / off status. The calibrated coupling strength matrix and coupling label matrix are then output. Write the coupling strength matrix, coupling label matrix and early warning sub-label into the coupling status cache area to establish a coupling fingerprint ledger for each device. When a candidate coupling pair does not match the pattern library or the order code is not registered, it is marked as an unknown coupling and a blocking flag is added. The parameters of the polymorphic physical coupling decomposer are updated periodically. The baseline value is refreshed within the steady-state time window, the threshold is frozen within the abnormal time window and unfrozen within the recovery time window, the redundancy consistency threshold is refreshed by sensor group, and the latest version of the parameters is written back to the coupling state buffer after the parameter update is completed.

[0011] Optionally, the conversion of anomalous coupling into risk factors includes: Receive the coupling tag matrix and coupling strength matrix, combine them with the corresponding timestamp and cabin discrete space unit number, filter physical quantity pairs in abnormal coupling state, and generate an abnormal coupling candidate list; Based on the preset mechanism mapping rule library, each physical quantity pair in the abnormal coupling candidate list is assigned to at least one of the mechanisms of thermo-pressure superposition, combustible gas accumulation, frictional heating or electronically controlled damp-heat coupling according to the coupling mode. Within a specified time window, the coupling strength level, duration level, and spatial coverage level of each mechanism are calculated separately. The three levels are combined into the original risk score of the mechanism according to the weight table. The original risk score is quantified into a dimensionless mechanism risk value according to the classification threshold table. All mechanism risk values ​​are subjected to conflict, superposition and suppression judgment, and consistency processing is performed according to mutual exclusion rules, superposition rules and suppression rules. In addition, reliability correction and context correction are performed in combination with the redundancy consistency verification results, current maintenance step identification, ventilation condition status and related equipment on / off status to obtain the corrected mechanism risk factor value, which is limited to between zero and one. A list of risk factors is generated, including thermal pressure superposition risk factors, combustible gas accumulation risk factors, frictional heating risk factors, and electrically controlled damp heat risk factors. Each risk factor includes a numerical value, a set of discrete spatial units in the cabin that it affects, and an effective time window.

[0012] Optionally, the step of mapping the characteristic physical quantity parameters in the preprocessed multi-physical quantity data to a first safe potential energy surface and constructing a second safe potential energy surface based on risk factors includes: The preprocessed multi-physical quantity data is organized according to the discrete spatial unit of the cabin, and a feature set consisting of temperature gradient, pressure change rate, air volume attenuation rate, vibration offset, and voltage ripple amplitude is generated. Each feature is standardized based on the baseline value and scale factor of the historical stable period. For each cabin discrete space unit, a fixed bias value and a set of weight coefficients corresponding to each feature are set. The standardized features are multiplied by their corresponding weights one by one, and the results are summed and superimposed with the bias value to obtain the first safe potential energy surface value of the cabin discrete space unit at the current time. The data are then collected to form the first safe potential energy surface dataset. Based on the risk factors of thermal pressure superposition, combustible gas accumulation, frictional heating, and electrical control damp heat, the effective risk factors are allocated to the corresponding cabin discrete space units. For each cabin discrete space unit, a risk factor weight coefficient group and a fixed bias value are set. The risk factors are multiplied by the corresponding weights one by one, summed, and the bias value is superimposed to obtain the second safety potential energy surface value. The data is then compiled to form the second safety potential energy surface dataset. Align the first and second safety potential energy surface datasets with the same spatial grid and at the same time. For each cabin discrete spatial cell, calculate the difference between the second and first safety potential energy surface values ​​at each time, and collect the results to obtain the potential energy difference matrix.

[0013] Optionally, the step of comparing the differences between the first and second safety potential energy surfaces and determining the entry conditions for maintenance steps based on their separation, contact, or intersection relationships includes: Under the same time index and cabin discrete spatial unit, the first safety potential energy surface dataset and the second safety potential energy surface dataset are compared one-to-one to obtain the potential energy difference value of each spatial unit at each time. Set a separation threshold, a contact tolerance bandwidth, and a cross threshold for the potential energy difference value. When the potential energy difference value is greater than the separation threshold, the corresponding spatial cell is marked as a separated state. When the absolute value of the potential energy difference value is within the contact tolerance bandwidth, it is marked as a contact state. When the potential energy difference value is less than the cross threshold, it is marked as a cross state. Based on the face adjacency relationship, the spatial units marked as intersection are scanned for connectivity. The set of interconnected units that are connected to each other and whose number reaches a preset scale threshold or whose total volume reaches a preset volume threshold is defined as a continuous intersection region. For the spatial scope of the target maintenance steps, a minimum holding time and a statistical time window are set. Within the statistical time window, the separation state, contact state, and continuous intersection area are comprehensively evaluated according to the spatial coverage ratio and duration. If all spatial units remain separated within the minimum holding time, an admission decision is generated to allow execution. If a contact state exists and no consecutive cross-regions appear within the minimum holding time, a decision to temporarily suspend access is generated. If consecutive overlapping areas occur and the minimum holding time is maintained, an access prohibition decision is generated; Write the admission decision results, along with the thresholds, bandwidth, size threshold, volume threshold, minimum hold time, and statistical time window parameters used for the decision, along with their timestamps, into the admission result cache.

[0014] Optionally, the trigger behavior lock stops the execution of the maintenance steps and outputs a warning message, including: After the output access judgment is approved, the personnel positioning device is activated to obtain the three-dimensional coordinates of the maintenance personnel carrying the positioning tag in the cabin coordinate system at a fixed sampling interval and record the sampling time. The continuous three-dimensional coordinates are written into the trajectory data buffer in the order of sampling time to form a real-time three-dimensional trajectory sequence. The safety path template corresponding to the current maintenance step is read, which includes the sequence of allowed spatial units, the maximum stay time of each spatial unit, and the set of prohibited spatial units. For each coordinate point in the real-time three-dimensional trajectory sequence, a template comparison is performed. When the personnel coordinates fall into the prohibited entry space unit set, or the stay time in the allowed entry space unit exceeds the maximum stay time of the space unit, it is determined that the personnel trajectory deviates from the safe path template. When a deviation occurs, a behavior lock is immediately triggered, pausing the current maintenance step and generating a warning message containing personnel identification, trigger time, and trigger coordinates, which is simultaneously sent to the engine room alarm device and the shore-based monitoring center.

[0015] A risk monitoring system for mechanical maintenance operations of a tugboat system based on the Internet of Things (IoT) according to an embodiment of the present invention includes the following modules: The data acquisition and preprocessing module is used to acquire and preprocess multi-physical quantity data in the engine room of the tugboat to obtain preprocessed multi-physical quantity data. The multi-state physical coupling decomposition module is used to capture coupling features of preprocessed multi-physical quantity data and divide the coupling relationship into steady-state coupling, transitional coupling and abnormal coupling. The risk quantification and safety potential energy surface construction module is used to convert abnormal coupling into risk factors, map characteristic physical quantity parameters to the first safety potential energy surface, and construct the second safety potential energy surface based on the risk factors. The access determination module is used to compare the differences between the first safety potential energy surface and the second safety potential energy surface, and determine the access conditions for maintenance steps based on the separation, contact or intersection relationship. The trajectory monitoring and behavior lock module is used to collect real-time three-dimensional trajectory data of maintenance personnel when maintenance steps are allowed to be executed, compare it with the safety path template, and trigger the behavior lock and output warning information when the trajectory deviates or enters a restricted area.

[0016] The beneficial effects of this invention are: This invention achieves real-time acquisition of key features such as temperature gradient, pressure change rate, airflow attenuation rate, vibration offset, and voltage ripple amplitude by collecting and preprocessing multi-physical quantity data in the engine room of a tugboat. Furthermore, it utilizes a multi-state physical coupling decomposer to perform in-depth analysis of the changing trends, correlations, and hysteresis between these multi-physical quantities, enabling timely identification of coupling offsets in equipment operation. Existing technologies can only trigger alarms based on single-variable thresholds, while this invention can classify coupling offsets into steady-state coupling, transient coupling, and abnormal coupling, more accurately revealing early signs of equipment anomalies. This allows for proactive identification before risks escalate, improving maintenance safety assurance capabilities.

[0017] This invention constructs a first safety potential energy surface and a second safety potential energy surface, and uses the separation, contact, or intersection relationship between the two to form an access determination for maintenance steps, thus overcoming the shortcomings of traditional methods that rely on experience-based judgment or single-parameter judgment. By mapping characteristic physical quantity parameters to safety potential energy surfaces and transforming risk factors into spatially directional risk potential energy surfaces, this invention establishes a quantitative, calculable, and gradable risk expression method. This provides a clear physical basis and spatial justification for allowing, postponing, or prohibiting the execution of maintenance steps, effectively avoiding misoperations caused by inaccurate risk assessment, achieving scientific management of the maintenance step execution process, and improving the safety and reliability of the maintenance process.

[0018] This invention introduces a trajectory monitoring and behavior lock mechanism. After a maintenance step is permitted, the three-dimensional trajectory of the maintenance personnel is collected in real time and compared with a preset safe path template, thus achieving dynamic constraints on personnel behavior. Traditional methods cannot achieve fine-grained monitoring of personnel position and path deviation, while this invention can automatically trigger a behavior lock and output warning information when personnel deviate from the safe path or enter a restricted area, thereby effectively reducing the risk of personnel accidentally entering dangerous areas. By combining equipment-coupled risk identification, step access control, and personnel behavior supervision, this invention improves the overall safety, real-time performance, and controllability of tugboat mechanical maintenance operations, providing a more reliable and systematic risk supervision method for tugboat mechanical management. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a method for risk monitoring of mechanical maintenance operations of a tugboat system based on the Internet of Things, as proposed in this invention; Figure 2 This is a schematic diagram of the structure of a risk monitoring system for mechanical maintenance operations of a tugboat system based on the Internet of Things proposed in this invention. Detailed Implementation

[0020] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0021] refer to Figure 1 A method for risk monitoring of mechanical maintenance operations on tugboat systems based on the Internet of Things (IoT) includes: Multi-physical quantity data is collected by various sensors deployed in the tugboat's engine room and connected to the Internet of Things. The multi-physical quantity data is preprocessed to obtain preprocessed multi-physical quantity data. The preprocessed multi-physical quantity data is input into the multi-state physical coupling decomposer to capture coupling features and classify the coupling relationship into three types of coupling states: steady-state coupling, transitional coupling, and abnormal coupling. The risk of coupling behavior is quantified by using abnormal coupling state, and abnormal coupling is converted into risk factor. The characteristic physical quantity parameters in the preprocessed multi-physical quantity data are mapped to the first safety potential energy surface, and the second safety potential energy surface is constructed based on the risk factor. The differences between the first and second safety potential energy surfaces are compared, and the entry conditions for maintenance steps are determined based on the separation, contact, or intersection relationship between the two. When the first and second safety potential energy surfaces remain separated, the maintenance steps are allowed to be executed. When the first and second safety potential energy surfaces are in partial contact, the maintenance steps are temporarily suspended. When the first and second safety potential energy surfaces form a continuous intersection area, the maintenance steps are prohibited from being executed. When the maintenance procedure is permitted to be executed, real-time 3D trajectory data of the maintenance personnel is collected and compared with the preset maintenance procedure safety path template. When the personnel's trajectory deviates from the safety path template or enters the preset restricted area, a behavior lock is triggered to stop the execution of the maintenance procedure and output a warning message.

[0022] In this embodiment, the multi-physical quantity data includes temperature data, pressure data, vibration data, air volume data, humidity data, combustible gas concentration data, voltage data, and current data.

[0023] In this embodiment, the preprocessing of multi-physical quantity data includes performing noise reduction, filtering, time synchronization, and feature extraction on the multi-physical quantity data.

[0024] In this embodiment, the classification of coupling relationships into three types of coupling states—steady-state coupling, transitional coupling, and abnormal coupling—includes: Preprocessed multi-physical quantity data are input into a multi-state physical coupling decomposer. The decomposer is connected in series in the order of coupling sensing layer, state decomposition layer, and risk indication layer. The coupling sensing layer includes a dual-window differential scanning unit, an order relation encoding unit, and a redundancy consistency check unit. For each physical quantity, the sign of the rate of change, the level of change amplitude, and the order of extreme value occurrence are extracted within short and long time windows, respectively. The differential results of the short and long windows are compared to generate start-up and persistence markers. After performing consistency checks on redundant sensor data at the same measurement point, a candidate coupling pair list and a sensing feature set are output, where: Dual-window differential scanning unit: Based on the preprocessing of multi-physical quantity data with short time windows and long time windows, the sign of the rate of change, the level of change amplitude, and the order of occurrence of extreme values ​​of each physical quantity are differentially processed by the short window dataset and the long window dataset to form start markers and persistence markers. Order relation coding unit: Based on the start marker and the duration marker, the order of extreme values, the order of change and the relative temporal sequence of multiple physical quantities at the same spatial location are encoded to form the order relation coding; Redundancy Consistency Verification Unit: Performs consistency verification on redundant sensor data at the same measurement point location, compares the change trend, amplitude level and order relationship encoding, retains the data that meets the consistency conditions, and generates a candidate coupling pair list and a sensing feature set; The state decomposition layer receives the sensing feature set and classifies candidate coupling pairs according to the intensity criterion, persistence criterion, and order consistency criterion. Each coupling pair is classified as steady-state coupling, transitional coupling, or abnormal coupling. Coupling pairs that satisfy two criteria but not the third criterion are assigned warning sub-labels. Adjacent coupling pairs are merged into coupling clusters based on the spatial proximity of sensors in the cabin grid, and a coupling cluster list is generated. The state decomposition layer receives the sensing feature set and classifies candidate coupling pairs according to the intensity criterion, persistence criterion, and order consistency criterion, wherein: The intensity criterion is based on the combination of the magnitude level and the rate of change level of the candidate coupling pair within a specified time window, and the combined value falls into the low, medium and high range to achieve the classification of low intensity, medium intensity or high intensity level. The persistence criterion classifies candidates into short, medium, and long persistence levels based on the duration of the start and persistence markers that appear consecutively within short and long time windows. The order consistency criterion relies on the matching degree between the order of occurrence of extreme values ​​and the order of change of candidate coupling pairs, and classifies the matching degree into low, medium and high intervals to achieve low consistency, medium consistency or high consistency level classification. The risk indication layer constructs a coupling strength matrix and a coupling label matrix based on the coupling cluster list. It references a fixed weight table and performs contextual cross-calibration by combining the current maintenance step identifier, ventilation status, and related equipment on / off status. The calibrated coupling strength matrix and coupling label matrix are then output. Specifically, the contextual cross-calibration involves: Associate each item in the list of coupled clusters with the current maintenance step identifier, and map the association results to the corresponding rows and columns of the coupling strength matrix and the coupling label matrix; Based on the ventilation condition coefficients in the fixed weight table, the weights of each element in the coupling strength matrix are adjusted so that the matrix values ​​reflect the ventilation status at that time. Using the on / off status of relevant equipment as the verification condition, the corrected matrix is ​​compared element by element and inconsistent terms are adjusted to complete the contextual cross-calibration of the coupling strength matrix and the coupling label matrix; The fixed weight table is a weight mapping dataset established during the system deployment phase based on the risk mechanism of tugboat maintenance operations. The fixed weight table uses the coupling cluster number or coupling relationship type as an index to record the baseline coupling weight corresponding to each coupling relationship and the weight correction value related to maintenance steps, ventilation conditions and equipment on / off status. The weight allocation is determined based on the severity of risk, probability of occurrence and frequency of operation exposure, and is calibrated in combination with historical maintenance records and expert evaluation results. The baseline coupling weight is represented by a normalized numerical value, and each correction value is used to reflect the impact of different operation contexts on coupling risk. Finally, a weight data foundation that can be directly used for the calibration of coupling strength matrix and coupling label matrix is ​​formed. The coupling strength matrix, coupling label matrix, and early warning sub-label are written into the coupling status cache to establish a coupling fingerprint ledger for each device. When a candidate coupling pair does not match the pattern library or the sequence code is not registered, it is marked as an unknown coupling and a blocking flag is attached. The coupling fingerprint ledger is a coupling feature record book that is continuously archived according to the device dimension. The record content includes device number, spatial unit number, timestamp, coupling strength matrix slice, coupling label matrix slice, early warning sub-label, unknown coupling blocking flag, and release information. The parameters of the polymorphic physical coupling decomposer are updated periodically. The baseline value is refreshed within the steady-state time window, the threshold is frozen within the abnormal time window and unfrozen within the recovery time window, the redundancy consistency threshold is refreshed by sensor group, and the latest version of the parameters is written back to the coupling state buffer after the parameter update is completed.

[0025] In this embodiment, converting abnormal coupling into a risk factor includes: Receive the coupling tag matrix and coupling strength matrix, combine them with the corresponding timestamp and cabin discrete space unit number, filter physical quantity pairs in abnormal coupling state, and generate an abnormal coupling candidate list; Based on a pre-defined mechanism mapping rule base, each physical quantity pair in the abnormal coupling candidate list is assigned to at least one mechanism among thermo-pressure superposition, combustible gas accumulation, frictional heating, or electronically controlled hydrothermal coupling, according to the coupling mode. The pre-defined mechanism mapping rule base is as follows: A set of rules indexed by physical quantity combination characteristics and identified by coupling mechanism category. Each rule consists of physical quantity pair name, change direction combination, change amplitude range, temporal relationship combination, and corresponding mechanism number. It is used to describe the thermo-pressure superposition, combustible gas accumulation, frictional heating, or electronically controlled damp-heat coupling mechanism corresponding to typical physical quantity pairs such as temperature gradient and pressure change rate, temperature gradient and vibration offset, pressure change rate and vibration offset, air volume attenuation rate and combustible gas concentration gradient, voltage ripple amplitude and humidity gradient under specific amplitude ranges and temporal relationships. Within a specified time window, the coupling strength level, duration level, and spatial coverage level of each mechanism are statistically analyzed. These three levels are then combined according to a weighting table to form the original risk score for the mechanism. Based on a grading threshold table, the original risk score is quantified into a dimensionless mechanism risk value, where: The coupling strength level is obtained by dividing the numerical range of the corresponding coupling strength matrix entries of the mechanism into intervals. Values ​​falling into the lowest interval are marked as low strength, those falling into the middle interval are marked as medium strength, and those falling into the highest interval are marked as high strength. The duration level is obtained by computer processing the total duration of consecutive occurrences of the corresponding coupling state within a time window and dividing it into intervals. The shortest interval is marked as short duration, the middle interval as medium duration, and the longest interval as long duration. Spatial coverage level is obtained by statistically determining the proportion of the number of spatial units covered by the coupling state within a time window to the total number of monitored spatial units and dividing the area into intervals. The smallest interval is marked as low coverage, the middle interval as medium coverage, and the largest interval as high coverage. The grading threshold table is based on a pre-defined discrete grading rule and a matching relationship between the original risk score numerical range and the corresponding mechanism risk value. The original risk scores are uniformly converted to the range of 0 to 100, and four risk levels are set: Level 1 is low risk, corresponding to an original risk score greater than or equal to 0 and less than or equal to 25, and a mechanism risk value of 0.25; Level 2 is medium risk, corresponding to an original risk score greater than 25 and less than or equal to 50, and a mechanism risk value of 0.5; Level 3 is relatively high risk, corresponding to an original risk score greater than 50 and less than or equal to 75, and a mechanism risk value of 0.75; Level 4 is high risk, corresponding to an original risk score greater than 75 and less than or equal to 100, and a mechanism risk value of 1. The original risk scores obtained by each mechanism within a specified time window are quantified according to the risk level and mechanism risk value corresponding to the grading threshold table to form a dimensionless mechanism risk result in the range of 0 to 1. All mechanistic risk values ​​are subject to conflict, superposition, and suppression determination. Consistency processing is performed according to mutual exclusion, superposition, and suppression rules. Reliability and context corrections are then performed, combining the redundancy consistency check results with the current maintenance step identifier, ventilation status, and related equipment on / off status, to obtain corrected mechanistic risk factor values, limited to between zero and one. The consistency processing according to mutual exclusion, superposition, and suppression rules includes: The mutual exclusion rule states that when different mechanism risk values ​​originate from the same combination of physical quantities and are mutually exclusive, only the one with the higher risk value is retained, and the remaining risk values ​​are set to zero. The superposition rule refers to the practice that when the spatial units affected by risk values ​​of different mechanisms do not overlap or the causes of risk are independent, each risk value is retained and they are added together when combining the scores. The suppression rule refers to the fact that there is a primary and secondary relationship between different mechanism risk values, and when the primary mechanism risk value reaches a high-level threshold, the secondary mechanism risk value is reduced by a preset suppression coefficient. A list of risk factors is generated, including thermal pressure superposition risk factors, combustible gas accumulation risk factors, frictional heating risk factors, and electrically controlled damp heat risk factors. Each risk factor includes a numerical value, a set of discrete spatial units in the cabin that affect it, and an effective time window. Specifically, the generation of the risk factor list is as follows: Based on the corrected mechanism risk factor values, numerical entries that exceed the activation threshold are spatially marked, and their respective discrete cabin spatial units are aggregated into the scope of action. For each tagged risk factor, the time when its value first exceeds the activation threshold is taken as the effective start point, and the last moment when the value is continuously lower than the deactivation threshold and reaches the retention period is taken as the failure end point. The effective time window is defined by the interval between the two. Write the risk factor name, corresponding value, set of action space units, and effective time window into a unified format record, sort them by value, and summarize them to obtain a list of four risk factors: thermal pressure superposition, combustible gas accumulation, frictional heating, and electrical control damp heat coupling.

[0026] In this embodiment, mapping the characteristic physical quantity parameters in the preprocessed multi-physical quantity data to a first safety potential energy surface and constructing a second safety potential energy surface based on risk factors includes: The preprocessed multi-physical quantity data is organized according to the discrete spatial unit of the cabin, and a feature set consisting of temperature gradient, pressure change rate, air volume attenuation rate, vibration offset, and voltage ripple amplitude is generated. Each feature is standardized based on the baseline value and scale factor of the historical stable period. For each cabin discrete space unit, a fixed bias value and a set of weight coefficients corresponding to each feature are set. The standardized features are multiplied by their corresponding weights one by one, and the results are summed and superimposed with the bias value to obtain the first safe potential energy surface value of the cabin discrete space unit at the current time. The data are then collected to form the first safe potential energy surface dataset. Based on the risk factors of thermal pressure superposition, combustible gas accumulation, frictional heating, and electrical control damp heat, the effective risk factors are allocated to the corresponding cabin discrete space units. For each cabin discrete space unit, a risk factor weight coefficient group and a fixed bias value are set. The risk factors are multiplied by the corresponding weights one by one, summed, and the bias value is superimposed to obtain the second safety potential energy surface value. The data is then compiled to form the second safety potential energy surface dataset. Align the first and second safety potential energy surface datasets with the same spatial grid and at the same time. For each cabin discrete spatial cell, calculate the difference between the second and first safety potential energy surface values ​​at each time, and collect the results to obtain the potential energy difference matrix.

[0027] In this embodiment, the step of comparing the differences between the first and second safety potential energy surfaces and determining the entry conditions for maintenance steps based on their separation, contact, or intersection relationships includes: Under the same time index and cabin discrete spatial unit, the first safety potential energy surface dataset and the second safety potential energy surface dataset are compared one-to-one to obtain the potential energy difference value of each spatial unit at each time. A separation threshold, a contact tolerance bandwidth, and a crossover threshold are set for the potential energy difference value. When the potential energy difference value is greater than the separation threshold, the corresponding spatial cell is marked as separated. When the absolute value of the potential energy difference value is within the contact tolerance bandwidth, it is marked as contacted. When the potential energy difference value is less than the crossover threshold, it is marked as crossover. The separation threshold is used to determine whether physical contact is completely separated, with a set value between 0.1 and 1.0, based on the friction and inertial effects of the equipment; the contact tolerance bandwidth is used to handle small errors or fluctuations in contact, with a set value between 0.01 and 0.1, based on the contact accuracy of the equipment and the elastic properties of the material; the intersection threshold is used to determine whether objects intersect, with a set value between 0.05 and 0.5, based on the physical intersection situation in the equipment structure and working environment. A connectivity scan is performed on spatial units marked as intersecting based on their face adjacency. Sets of interconnected units whose number reaches a preset size threshold or whose total volume reaches a preset volume threshold are defined as continuous intersecting regions, where: The preset size threshold is used to define the minimum number of units in an intersection area, set between 10 and 100, and adjusted according to the size of the work space and the distribution of equipment; the preset volume threshold is used to define the minimum total volume of an intersection area, set between 1 and 50 cubic meters, and set according to the volume of equipment and the volume requirements of the work space. For the spatial scope of the target maintenance steps, a minimum holding time and a statistical time window are set. Within the statistical time window, the separation state, contact state, and continuous intersection area are comprehensively evaluated according to the spatial coverage ratio and duration. If all spatial units remain separated within the minimum holding time, an admission decision is generated to allow execution. If a contact state exists and no consecutive cross-regions appear within the minimum holding time, a decision to temporarily suspend access is generated. If consecutive overlapping areas occur and the minimum holding time is maintained, an access prohibition decision is generated; Write the admission decision results, along with the thresholds, bandwidth, size threshold, volume threshold, minimum hold time, and statistical time window parameters used for the decision, along with their timestamps, into the admission result cache.

[0028] In this embodiment, the triggering behavior lock to stop the execution of the maintenance step and output a warning message includes: After the output access judgment is approved, the personnel positioning device is activated to obtain the three-dimensional coordinates of the maintenance personnel carrying the positioning tag in the cabin coordinate system at a fixed sampling interval and record the sampling time. The continuous three-dimensional coordinates are written into the trajectory data buffer in the order of sampling time to form a real-time three-dimensional trajectory sequence. The safety path template corresponding to the current maintenance step is read, which includes the sequence of allowed spatial units, the maximum stay time of each spatial unit, and the set of prohibited spatial units. For each coordinate point in the real-time three-dimensional trajectory sequence, a template comparison is performed. When the personnel coordinates fall into the prohibited entry space unit set, or the stay time in the allowed entry space unit exceeds the maximum stay time of the space unit, it is determined that the personnel trajectory deviates from the safe path template. When a deviation occurs, a behavior lock is immediately triggered, pausing the current maintenance step and generating a warning message containing personnel identification, trigger time, and trigger coordinates, which is simultaneously sent to the engine room alarm device and the shore-based monitoring center.

[0029] refer to Figure 2 A risk monitoring system for tugboat maintenance operations based on the Internet of Things (IoT) includes the following modules: The data acquisition and preprocessing module is used to acquire and preprocess multi-physical quantity data in the engine room of the tugboat to obtain preprocessed multi-physical quantity data. The multi-state physical coupling decomposition module is used to capture coupling features of preprocessed multi-physical quantity data and divide the coupling relationship into steady-state coupling, transitional coupling and abnormal coupling. The risk quantification and safety potential energy surface construction module is used to convert abnormal coupling into risk factors, map characteristic physical quantity parameters to the first safety potential energy surface, and construct the second safety potential energy surface based on the risk factors. The access determination module is used to compare the differences between the first safety potential energy surface and the second safety potential energy surface, and determine the access conditions for maintenance steps based on the separation, contact or intersection relationship. The trajectory monitoring and behavior lock module is used to collect real-time three-dimensional trajectory data of maintenance personnel when maintenance steps are allowed to be executed, compare it with the safety path template, and trigger the behavior lock and output warning information when the trajectory deviates or enters a restricted area.

[0030] Example 1: To verify the feasibility of this invention in practice, it was applied to a tugboat operating in a port. The tugboat's engine room has a complex structure and dense equipment, resulting in coupled changes in multiple physical quantities during operation, including temperature, pressure, vibration, electrical fluctuations, and flammable gas fluctuations. In traditional maintenance operations, engineers rely on pressure gauges, thermometers, and alarms for monitoring, which cannot identify abnormal coupling relationships between these multiple physical quantities in advance. For example, risks such as superposition of heat and pressure, abnormal temperature rise due to insufficient cooling, or simultaneous aggravation of oil temperature and vibration cannot be identified. Furthermore, it is impossible to scientifically determine whether maintenance procedures are suitable for immediate execution, posing a high operational risk.

[0031] During this overhaul, the method of this invention continuously collected various physical quantities such as temperature, pressure, airflow, humidity, voltage ripple, and vibration through 23 IoT sensors deployed in the engine room. Twenty minutes before the maintenance began, the system was activated to perform data preprocessing, coupling decomposition, risk factor generation, and safety potential surface calculation. The system first divided the three-dimensional space of the engine room into 136 discrete spatial units using a 0.5-meter grid. Each unit received sensor data in real time and generated corresponding characteristic physical quantity parameters. These parameters were then input into a multi-state physical coupling decomposer, which identified the coupling state between physical quantities through mechanisms such as differential trend, correlation, time delay, and redundancy check. Between 08:57 and 09:05 before the start of the operation, the system detected a synchronous increase in the rate of temperature change and pressure change in the upper left region of the main engine cylinder block, accompanied by a significant increase in vibration offset. The system identified this as abnormal coupling and automatically generated a thermo-pressure superposition risk factor with a risk value of 0.71, designating spatial units C22, C23, and C30.

[0032] After generating risk factors, the system constructs a second safety potential energy surface based on these factors, and simultaneously constructs a first safety potential energy surface based on characteristic physical quantities. The difference between the two shows that in each spatial unit, the potential energy difference between C22 and C23 is below the permissible threshold, forming a continuous overlapping region with a continuous area of ​​approximately 1.3 square meters. The system automatically determines that this region does not meet the conditions for maintenance procedure access. The method of this invention outputs a prohibition decision at 09:05:16, while traditional manual inspection did not detect the risk in this area. At this time, the temperature near the manual temperature measurement point only increased by 3°C, which would not trigger any alarm under the traditional empirical model.

[0033] Table 1. Measured data of changes in multiple physical quantities in key monitoring areas of the engine room.

[0034] As shown in Table 1, during the period from 08:57 to 08:58, although the temperature gradient, pressure change rate, and vibration offset of space units C22 and C23 were at relatively low levels, their changes were consistent and gradually increased, exhibiting typical early coupling characteristics. The multi-state physical coupling decomposer of this invention can promptly identify the coupling trend during this stage, marking the state as a transitional state and generating risk factor values ​​of 0.32 and 0.41. This process demonstrates that even if the monitored quantities have not yet triggered traditional threshold alarms, this invention can capture potential risks through trends and synchronicity.

[0035] After 09:00, the coupling offset intensified significantly. Multiple physical characteristics of point C22 simultaneously and significantly increased within a short period, with temperature gradient, pressure change rate, and vibration offset amplifying synchronously, causing the coupling state to jump directly to an abnormal state, increasing the risk factor value to 0.66. Subsequently, point C30 also exhibited the same trend, particularly with a significant decrease in airflow attenuation rate, indicating that the equipment's thermal pressure condition and insufficient ventilation were being coupled and amplified. This invention can accurately capture this multi-point, multi-quantity synchronously enhanced risk pattern, marking the relevant spatial units as abnormal areas, with a risk factor value reaching 0.71, further enhancing the accuracy and spatial orientation of risk location.

[0036] Between 09:04 and 09:05, the parameters of C22 and C23 both reached their peak values ​​during this monitoring, exhibiting typical thermo-pressure coupling characteristics. Traditional inspection methods, because the absolute values ​​of temperature and pressure had not yet exceeded limits, typically struggle to identify such early-stage risks. This invention, through real-time analysis of the change rates and coupling characteristics of multiple physical quantities, successfully identified this as a stable abnormal state, providing risk factor values ​​of 0.71 and 0.70 respectively. This invention can detect risk trends 5 to 8 minutes earlier than traditional methods, achieving accurate identification of hazardous areas, effectively preventing maintenance personnel from misoperating in high-coupling-risk areas, and improving the safety and controllability of tugboat maintenance operations.

[0037] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for risk monitoring of mechanical maintenance operations on tugboats based on the Internet of Things (IoT), characterized in that, include: Multi-physical quantity data is collected by various sensors deployed in the tugboat's engine room and connected to the Internet of Things. The multi-physical quantity data is preprocessed to obtain preprocessed multi-physical quantity data. The preprocessed multi-physical quantity data is input into the multi-state physical coupling decomposer to capture coupling features and classify the coupling relationship into three types of coupling states: steady-state coupling, transitional coupling, and abnormal coupling. The risk of coupling behavior is quantified by using abnormal coupling state, and abnormal coupling is converted into risk factor. The characteristic physical quantity parameters in the preprocessed multi-physical quantity data are mapped to the first safety potential energy surface, and the second safety potential energy surface is constructed based on the risk factor. The differences between the first and second safety potential energy surfaces are compared, and the entry conditions for maintenance steps are determined based on the separation, contact, or intersection relationship between the two. When the first and second safety potential energy surfaces remain separated, the maintenance steps are allowed to be executed. When the first and second safety potential energy surfaces are in partial contact, the maintenance steps are temporarily suspended. When the first and second safety potential energy surfaces form a continuous intersection area, the maintenance steps are prohibited from being executed. When the maintenance procedure is permitted to be executed, real-time 3D trajectory data of the maintenance personnel is collected and compared with the preset maintenance procedure safety path template. When the personnel's trajectory deviates from the safety path template or enters the preset restricted area, a behavior lock is triggered to stop the execution of the maintenance procedure and output a warning message.

2. The method for risk monitoring of mechanical maintenance operations based on an IoT-based tugboat system according to claim 1, characterized in that, The multi-physical quantity data includes temperature data, pressure data, vibration data, air volume data, humidity data, combustible gas concentration data, voltage data, and current data.

3. The method for risk monitoring of mechanical maintenance operations based on an IoT-based tugboat system according to claim 1, characterized in that, The preprocessing of multi-physical quantity data includes performing denoising, filtering, time synchronization, and feature extraction on the multi-physical quantity data.

4. The method for risk monitoring of mechanical maintenance operations based on an IoT tugboat system according to claim 1, characterized in that, The coupling relationship is divided into three types of coupling states: steady-state coupling, transitional coupling, and abnormal coupling, including: The preprocessed multi-physical quantity data is input into the multi-state physical coupling decomposer. The decomposer is connected in series in the order of coupling sensing layer, state decomposition layer and risk indication layer. The coupling sensing layer is equipped with a dual-window differential scanning unit, an order relation encoding unit and a redundancy consistency check unit. For each physical quantity, the sign of change rate, the level of change amplitude and the order of occurrence of extreme values ​​are extracted in the short time window and the long time window respectively. The difference results of the short window and the long window are compared to generate start flag and continuous flag. After performing consistency check on the redundant sensor data of the same measurement point, a candidate coupling pair list and a sensing feature set are output. The state decomposition layer receives the sensing feature set and classifies the candidate coupling pairs according to the intensity criterion, the persistence criterion, and the order consistency criterion. Each coupling pair is classified as steady-state coupling, transitional coupling, or abnormal coupling. Couple pairs that meet two criteria but not the third criterion are assigned warning sub-labels. Adjacent coupling pairs are merged into coupling clusters and a coupling cluster list is generated based on the spatial proximity of the sensors in the cabin grid. The risk indication layer constructs a coupling strength matrix and a coupling label matrix based on the coupling cluster list. It references a fixed weight table and performs contextual cross-calibration by combining the current maintenance step identifier, ventilation status and related equipment on / off status. The calibrated coupling strength matrix and coupling label matrix are then output. Write the coupling strength matrix, coupling label matrix and early warning sub-label into the coupling status cache area to establish a coupling fingerprint ledger for each device. When a candidate coupling pair does not match the pattern library or the order code is not registered, it is marked as an unknown coupling and a blocking flag is added. The parameters of the polymorphic physical coupling decomposer are updated periodically. The baseline value is refreshed within the steady-state time window, the threshold is frozen within the abnormal time window and unfrozen within the recovery time window, the redundancy consistency threshold is refreshed by sensor group, and the latest version of the parameters is written back to the coupling state buffer after the parameter update is completed.

5. The method for risk monitoring of mechanical maintenance operations based on an IoT tugboat system according to claim 1, characterized in that, The process of converting anomalous coupling into risk factors includes: Receive the coupling tag matrix and coupling strength matrix, combine them with the corresponding timestamp and cabin discrete space unit number, filter physical quantity pairs in abnormal coupling state, and generate an abnormal coupling candidate list; Based on the preset mechanism mapping rule library, each physical quantity pair in the abnormal coupling candidate list is assigned to at least one of the mechanisms of thermo-pressure superposition, combustible gas accumulation, frictional heating or electronically controlled damp-heat coupling according to the coupling mode. Within a specified time window, the coupling strength level, duration level, and spatial coverage level of each mechanism are calculated separately. The three levels are combined into the original risk score of the mechanism according to the weight table. The original risk score is quantified into a dimensionless mechanism risk value according to the classification threshold table. All mechanism risk values ​​are subjected to conflict, superposition and suppression judgment, and consistency processing is performed according to mutual exclusion rules, superposition rules and suppression rules. In addition, reliability correction and context correction are performed in combination with the redundancy consistency verification results, current maintenance step identification, ventilation condition status and related equipment on / off status to obtain the corrected mechanism risk factor value, which is limited to between zero and one. A list of risk factors is generated, including thermal pressure superposition risk factors, combustible gas accumulation risk factors, frictional heating risk factors, and electrically controlled damp heat risk factors. Each risk factor includes a numerical value, a set of discrete spatial units in the cabin that it affects, and an effective time window.

6. The method for risk monitoring of mechanical maintenance operations based on an Internet of Things (IoT) tugboat system according to claim 1, characterized in that, The process of mapping the characteristic physical quantity parameters in the preprocessed multi-physical quantity data to a first safety potential energy surface and constructing a second safety potential energy surface based on risk factors includes: The preprocessed multi-physical quantity data is organized according to the discrete spatial unit of the cabin, and a feature set consisting of temperature gradient, pressure change rate, air volume attenuation rate, vibration offset, and voltage ripple amplitude is generated. Each feature is standardized based on the baseline value and scale factor of the historical stable period. For each cabin discrete space unit, a fixed bias value and a set of weight coefficients corresponding to each feature are set. The standardized features are multiplied by their corresponding weights one by one, and the results are summed and superimposed with the bias value to obtain the first safe potential energy surface value of the cabin discrete space unit at the current time. The data are then collected to form the first safe potential energy surface dataset. Based on the risk factors of thermal pressure superposition, combustible gas accumulation, frictional heating, and electrical control damp heat, the effective risk factors are allocated to the corresponding cabin discrete space units. For each cabin discrete space unit, a risk factor weight coefficient group and a fixed bias value are set. The risk factors are multiplied by the corresponding weights one by one, summed, and the bias value is superimposed to obtain the second safety potential energy surface value. The data is then compiled to form the second safety potential energy surface dataset. Align the first and second safety potential energy surface datasets with the same spatial grid and at the same time. For each cabin discrete spatial cell, calculate the difference between the second and first safety potential energy surface values ​​at each time, and collect the results to obtain the potential energy difference matrix.

7. The method for risk monitoring of mechanical maintenance operations based on an IoT tugboat system according to claim 1, characterized in that, The comparison of the first and second safety potential energy surfaces to determine the entry conditions for maintenance steps based on their separation, contact, or intersection relationships includes: Under the same time index and cabin discrete spatial unit, the first safety potential energy surface dataset and the second safety potential energy surface dataset are compared one-to-one to obtain the potential energy difference value of each spatial unit at each time. Set a separation threshold, a contact tolerance bandwidth, and a cross threshold for the potential energy difference value. When the potential energy difference value is greater than the separation threshold, the corresponding spatial cell is marked as a separated state. When the absolute value of the potential energy difference value is within the contact tolerance bandwidth, it is marked as a contact state. When the potential energy difference value is less than the cross threshold, it is marked as a cross state. Based on the face adjacency relationship, the spatial units marked as intersection are scanned for connectivity. The set of interconnected units that are connected to each other and whose number reaches a preset scale threshold or whose total volume reaches a preset volume threshold is defined as a continuous intersection region. For the spatial scope of the target maintenance steps, a minimum holding time and a statistical time window are set. Within the statistical time window, the separation state, contact state, and continuous intersection area are comprehensively evaluated according to the spatial coverage ratio and duration. If all spatial units remain separated within the minimum holding time, an admission decision is generated to allow execution. If a contact state exists and no consecutive cross-regions appear within the minimum holding time, a decision to temporarily suspend access is generated. If consecutive overlapping areas occur and the minimum holding time is maintained, an access prohibition decision is generated; Write the admission decision results, along with the thresholds, bandwidth, size threshold, volume threshold, minimum hold time, and statistical time window parameters used for the decision, along with their timestamps, into the admission result cache.

8. The method for risk monitoring of mechanical maintenance operations based on an Internet of Things (IoT) tugboat system according to claim 1, characterized in that, The trigger behavior lock stops the execution of the maintenance steps and outputs a warning message, including: After the output access judgment is approved, the personnel positioning device is activated to obtain the three-dimensional coordinates of the maintenance personnel carrying the positioning tag in the cabin coordinate system at a fixed sampling interval and record the sampling time. The continuous three-dimensional coordinates are written into the trajectory data buffer in the order of sampling time to form a real-time three-dimensional trajectory sequence. The safety path template corresponding to the current maintenance step is read, which includes the sequence of allowed spatial units, the maximum stay time of each spatial unit, and the set of prohibited spatial units. For each coordinate point in the real-time three-dimensional trajectory sequence, a template comparison is performed. When the personnel coordinates fall into the prohibited entry space unit set, or the stay time in the allowed entry space unit exceeds the maximum stay time of the space unit, it is determined that the personnel trajectory deviates from the safe path template. When a deviation occurs, a behavior lock is immediately triggered, pausing the current maintenance step and generating a warning message containing personnel identification, trigger time, and trigger coordinates, which is simultaneously sent to the engine room alarm device and the shore-based monitoring center.

9. A risk monitoring system for mechanical maintenance operations of an IoT-based tugboat system, comprising the method for risk monitoring of mechanical maintenance operations of an IoT-based tugboat system as described in any one of claims 1 to 8, characterized in that, Includes the following modules: The data acquisition and preprocessing module is used to acquire and preprocess multi-physical quantity data in the engine room of the tugboat to obtain preprocessed multi-physical quantity data. The multi-state physical coupling decomposition module is used to capture coupling features of preprocessed multi-physical quantity data and divide the coupling relationship into steady-state coupling, transitional coupling and abnormal coupling. The risk quantification and safety potential energy surface construction module is used to convert abnormal coupling into risk factors, map characteristic physical quantity parameters to the first safety potential energy surface, and construct the second safety potential energy surface based on the risk factors. The access determination module is used to compare the differences between the first safety potential energy surface and the second safety potential energy surface, and determine the access conditions for maintenance steps based on the separation, contact or intersection relationship. The trajectory monitoring and behavior lock module is used to collect real-time three-dimensional trajectory data of maintenance personnel when maintenance steps are allowed to be executed, compare it with the safety path template, and trigger the behavior lock and output warning information when the trajectory deviates or enters a restricted area.

Images