Intelligent optimized marine loading and unloading arm parameter self-adaptive setting and control method

Abstract

The invention relates to an intelligent optimized marine loading and unloading arm parameter self-adaptive setting and control method, and belongs to the technical field of harbor machinery intelligent control. The method comprises the steps of establishing a unified spatio-temporal data platform, collecting time sequence, space and physical quantity multi-source heterogeneous data in a classified mode, deploying a clock synchronization mechanism to ensure time consistency, cleaning abnormal data, complementing missing values, and adding spatio-temporal labels to the data. Constructing a loading arm digital twinborn model, calculating the stress and vibration characteristics of an arm support, recording a hydraulic pressure transmission process, correcting the model by using a back propagation neural network, and adjusting parameters in real time; establishing a parameter and performance mapping model, training and optimizing, and verifying through a test set; the failure modes are sorted through failure mode and influence analysis, stress and vibration data are collected in real time, and the fatigue damage degree and the safety coefficient are calculated and updated. Unified management of multi-source data of the marine loading and unloading arm is achieved, the model precision and the parameter optimization accuracy are improved, and self-adaption, efficient and safe operation of the loading and unloading arm is supported.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of intelligent control of port machinery, and particularly relates to a parameter self-adaptive setting and control method for an intelligent and optimized ship loading and unloading arm. BACKGROUND

[0002] The ship loading and unloading arm is the core equipment for connecting sea and land transportation and realizing efficient loading and unloading of key media such as crude oil, liquefied natural gas and chemical products at the port, and its running state is directly related to the stability, safety and economic efficiency of the energy and chemical supply chain, and plays an important role in the intelligent construction of modern ports. However, the current parameter setting and control mode of the loading and unloading arm still has many technical bottlenecks, which is difficult to meet the complex and changeable operation requirements. At the same time, the existing data acquisition is scattered in independent systems such as hydraulic, structure and environment, and the multi-source heterogeneous data (time sequence, space and physical quantity) lack a unified space-time reference, which is difficult to realize correlation analysis. The digital twin model is mainly based on a single physical mechanism, which deviates from the actual running state. The nonlinear relationship between parameters and performance (energy consumption, stress, etc.) is not modeled, and the optimization lacks accurate basis. The safety evaluation relies on periodic detection, which is difficult to respond to the structure fatigue and failure risk under dynamic working conditions, and restricts the improvement of the intelligent control level of the loading and unloading arm. SUMMARY

[0003] To solve the above problems in the prior art, the application provides a parameter self-adaptive setting and control method for an intelligent and optimized ship loading and unloading arm, The object of the application can be achieved by the following technical solutions: S1: Collect multi-source heterogeneous data types by establishing a unified space-time data platform; deploy a clock synchronization mechanism, specify the collection node close to the dock control as the reference node, and the remaining nodes receive the clock calibration signal of the reference node at fixed intervals; process the abnormal value, and use the interpolation method to fill in the missing values in the data transmission gap; add space-time labels and spatial labels to each data; S2: Construct a digital twin model of the loading and unloading arm, calculate the stress of the arm support under different angles, simulate the vibration characteristics of the arm support under different medium flow rates, record the transmission process of the pressure from the hydraulic station to the actuator, collect historical operation data of the loading and unloading arm, divide it into training set and verification set, and use the back propagation neural network to correct the output results of the digital twin model of the loading and unloading arm; adjust the real-time data parameters in the model in real time; S3: Establish a parameter and performance mapping model, input parameters as model features and output performance as model labels, and substitute them into the training set for model training to adjust the parameter optimization model precision; verify the prediction results of the method through the test set, and if the prediction error exceeds the preset range, supplement the data collected under different operation scenarios to retrain the model; S4: Through failure mode and effects analysis, common failure modes of the loading and unloading arm are sorted out, and the influence factors and judgment basis of each failure mode are confirmed; according to the results of failure mode and effects analysis, the key monitoring data corresponding to each failure mode are recorded; the stress data and vibration data of the key parts are collected in real time, and the stress cycle times and the current fatigue damage degree are calculated in combination with the fatigue property curve of the material of the loading and unloading arm; the structural safety factor is calculated according to the formula of material mechanics; the fatigue damage degree and the safety factor are updated synchronously.

[0004] Specifically, the multi-source heterogeneous data type is used to construct a basic data source of a unified space-time data platform, and the specific process is as follows: collecting time series data, including real-time pressure of the hydraulic pipeline of the loading and unloading arm, vibration amplitude of each part of the boom, and medium conveying temperature, which are continuously collected at a fixed frequency; collecting spatial data, including berth identification of the loading and unloading arm, distance from the center of the column to the front of the wharf, and boom coordinates, which are determined according to the wharf surveying drawing; collecting physical quantity data, including actual telescopic length of the boom, rotation angle of each joint, hydraulic drive speed, and viscosity of the conveying medium, which are obtained by real-time reading of sensors and extraction of device parameter documents; and collecting multi-source heterogeneous data sets by summarizing the three types of data.

[0005] Specifically, the clock synchronization mechanism is used to ensure the time consistency of multi-source heterogeneous data collection, and the specific process is as follows: selecting a collection node close to the wharf central control as a reference node, which needs to meet the conditions of stable power supply and network signal; setting the interval of the calibration signal sent by the reference node in combination with the data collection frequency; after receiving the calibration signal, the rest of the nodes compare the time difference with the reference node through the local clock chip, record the time breakpoints in the correction process, and ensure the continuity of the time stamp through subsequent data supplement in the future; after all the nodes are corrected, the synchronization results are fed back to the reference node.

[0006] Specifically, the abnormal value is used to process multi-source heterogeneous data, and the specific process is as follows: when presetting the threshold value of the device parameter, the design standard of the loading and unloading arm out of the factory, the industry safety specification, and the statistical range of the historical normal operation data are comprehensively determined; the data inspection program is used to compare the collected data with the threshold value piece by piece, if the hydraulic pressure exceeds the design range or the vibration amplitude exceeds the device structure tolerance value, the abnormal data is directly marked and eliminated, and the abnormal occurrence time and the corresponding sensor position are recorded; for the missing values caused by data transmission gaps, the linear interpolation method is used to calculate; according to the numerical value and time interval ratio of the front and rear effective data, the numerical value of the missing time is calculated; after the filling is completed, the filled values are checked again, and if they exceed, they are recalculated.

[0007] Specifically, the space-time label and the space label are added to each piece of data for associating two types of data, and the specific process is as follows: after data cleaning, label addition is automatically triggered; the time label is collected from the clock signal synchronized by the reference node, and the current tidal period information is obtained by calling the real-time data interface of the terminal hydrological monitoring system; the berth number and the unloading arm equipment unique code of the space label are matched and called from the equipment basic information library, and the distance from the column center to the terminal front edge is correspondingly queried from the terminal mapping database; the data processing program converts the extracted time information and space information into labels in a unified format, binds each piece of valid data according to the data collection sequence, and finally forms a structured data set with corresponding labels.

[0008] Specifically, the S1 first classifies and collects multi-source heterogeneous data, wherein the time series data (hydraulic pipeline pressure, arm vibration amplitude, medium temperature, etc.) is collected by corresponding sensors according to the working state, the spatial data (berth identification, arm coordinates, etc.) is obtained according to the verified mapping paper and management system, the physical quantity data (arm length, joint angle, etc.) is read by the sensor and extracted in combination with the parameter document, and the basic data pool is formed by summarizing; then, a clock synchronization mechanism is deployed, a stable node close to the terminal control is selected as a reference node, a calibration signal is sent according to the collection frequency, the rest of the nodes are compared and corrected for time deviation and breakpoint data is supplemented, and the time stamps of all nodes are ensured to be consistent; then, data cleaning is performed, threshold values are set with reference to equipment standards and historical data to remove abnormal values, linear interpolation method is used to fill in missing values and secondary verification is performed; finally, time and space labels are added to the cleaned data, the time label includes the collection time and the tidal period (obtained from the reference clock and the hydrological system), the space label includes the berth number, the equipment code and the installation distance (called from the equipment library and the mapping library), and a structured data set is formed by program binding.

[0009] Specifically, the force condition of the arm of the unloading arm at different angles is calculated to construct a digital twin model of the unloading arm, and the specific process is as follows: according to the torque balance formula in rigid body dynamics, the material parameters and structural dimensions of the arm are input; when considering three types of loads, the wind load is calculated according to the long-term meteorological data provided by the meteorological department of the terminal location, the medium weight is calculated according to the product of the medium density and the pipeline volume, and the arm self weight is calculated according to the product of the material density and the arm volume; the arm model is constructed, the grid at the joint is encrypted, the rest of the parts is divided into grids according to the conventional size, the three types of loads are applied to the corresponding positions of the model, the rotation process of the arm from horizontal to vertical is simulated, and the axial force, the radial force of each joint, and the maximum stress value and stress distribution of the middle section of the arm are output at different rotation angles.

[0010] Specifically, the simulation of the vibration characteristics of the arm under different medium flow rates is used to reflect the dynamic characteristics of the loading and unloading arm digital twin model, and the specific process is as follows: based on the fluid-structure coupling theory, a loading and unloading arm pipeline model is constructed, and the roughness of the inner wall of the pipeline is set according to the actual machining precision; the fluid-structure coupling simulation software is imported, the medium flow rate variable covering the design flow rate range of the loading and unloading arm is set, and the viscosity parameters of different media are input; the boundary conditions are set, and the impact pressure and impact force of the medium on the inner wall of the pipeline under different flow rates are calculated; the impact force is applied to the arm structure model as a load, modal analysis is performed, and the natural vibration frequency of the arm, the vibration frequency and amplitude under different flow rates are output; the simulation results are classified and arranged according to the type and flow rate of the medium.

[0011] Specifically, the recording of the transmission process of the pressure from the hydraulic station to the actuator is used to supplement the hydraulic system characteristics of the loading and unloading arm digital twin model, and the specific process is as follows: pressure sensors are installed at the outlet main pipeline of the hydraulic station, the middle section of the pipeline and the inlet of the actuator; the pressure values and pressure transmission time of each point are recorded simultaneously, and the time difference from the time when the pressure at the outlet of the hydraulic station rises to the time when the pressure at the middle section and the inlet of the actuator rises is recorded; when calculating the pressure transmission delay, the extreme values are removed after multiple measurements and the average value is taken; the pressure loss values of each section are calculated according to the pipeline resistance formula; the transmission delay and the pressure loss data of each section are recorded, and the corresponding relationship between oil temperature and pressure loss is associated.

[0012] Specifically, the back propagation neural network is used to correct the output results of the loading and unloading arm digital twin model, and the specific process is as follows: a multi-layer back propagation neural network is constructed, the input layer neurons correspond to the stress values and vibration amplitudes output by the physical model, and the output layer neurons correspond to the corrected stress values and vibration amplitudes; when screening historical operation data, remove the data during equipment failure, cover different media and different tidal periods, and divide the operation scenarios into training set and validation set in proportion; adjust the network weight by gradient descent method, set the initial learning rate according to the model convergence speed, adjust according to the validation set error after a certain number of iterations, and iterate until the validation set error reaches the preset condition; embed the trained neural network into the calculation model of the digital twin model through the interface, and real-time correct the stress and vibration data output by the physical model.

[0013] Specifically, the real-time data parameters in the real-time adjustment model are used to maintain the consistency of the loading and unloading arm digital twin model and the actual equipment, and the specific process is as follows: select the real-time parameters to be adjusted; real-time collect the current wind speed of the wharf, the actual flow rate of the medium and the oil temperature of the hydraulic pipeline through the edge node; dynamically adjust the corresponding parameter values according to the collected data; complete parameter update according to the fixed period set according to the frequency of change of the actual operation environment of the equipment, and compare the deviation between the model output and the actual monitoring data after each update.

[0014] Specifically, the adjustment parameter optimization model precision is used to improve the prediction accuracy of the parameter and performance mapping model, and the specific process is: determining the model parameters to be adjusted; adjusting the parameters in the reasonable interval preset in combination with the algorithm characteristics; after each adjustment, the test set data is input into the model to calculate the prediction error, and if the error exceeds the preset threshold, the adjustment is continued, and if the error is within the preset threshold, the adjustment is stopped; the adjustment trajectory of each parameter is recorded during the optimization process.

[0015] Specifically, the common failure modes of the combing loading and unloading arm are used to confirm the influencing factors and determination basis of each failure mode, and the specific process is: common failure modes of the combing loading and unloading arm; for hydraulic system leakage, the influencing factors include pressure decay rate and sealing wear degree, and the determination basis is that the pressure decay rate exceeds the normal operation range or the sealing wear reaches the set degree affecting the sealing effect; for arm frame structure fatigue, the influencing factors include stress cycle number and vibration frequency, and the determination basis is that the stress cycle number approaches the material fatigue limit number or the vibration frequency continuously stays in the material fatigue characteristic frequency range; for rotary joint jamming, the influencing factors include angle change rate and lubricating grease viscosity, and the determination basis is that the angle change rate drops below the normal operation range or the lubricating grease viscosity exceeds the equipment adaptation range.

[0016] The beneficial effects of the present application are: (1) By setting the unified space-time data platform and the matching data collection, synchronization, cleaning and labeling mechanism, the problem of chaotic management of multi-source heterogeneous data of the ship loading and unloading arm can be solved; the platform first collects time sequence data (hydraulic pressure, vibration amplitude, etc.), spatial data (berth identification, arm frame coordinates, etc.) and physical quantity data (arm frame length, medium viscosity, etc.), then ensures the consistency of the time stamps of the data of each node through the clock synchronization mechanism, avoids the invalidation of data association caused by time misplacement, at the same time, improves the data quality through outlier elimination and interpolation method to complete the missing value, and finally adds a unified format space-time label to each data to realize association, forms a structured data set, and provides accurate and usable basic data support for subsequent digital twin modeling and parameter mapping analysis; (2) By setting up a loading arm digital twin model, parameter-performance mapping model and dynamic safety assessment mechanism that integrates physical mechanism and data driving, the loading arm parameters can be accurately and adaptively set and the operation safety can be controlled in real time. The digital twin model is based on rigid body dynamics and fluid-structure interaction theory to build a physical foundation. It combines backpropagation neural network to correct deviations and adjust parameters in real time to ensure that the model matches the actual equipment state. The parameter-performance mapping model clarifies the nonlinear relationship between input parameters (boom angle, hydraulic pressure, etc.) and output performance (operation time, energy consumption, etc.) through training, providing a quantitative basis for parameter optimization. The dynamic safety assessment mechanism combines FMEA to sort out failure modes, collects stress and vibration data in real time to calculate fatigue damage degree and safety factor, responds to structural risks in a timely manner, and avoids problems such as low operation efficiency and equipment damage caused by improper parameter adaptation or delayed safety assessment. Attached Figure Description

[0017] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0018] Fig. 1 This is an architecture diagram of an intelligent optimized adaptive setting and control method for marine loading and unloading boom parameters according to the present invention. Fig. 2 This is a data flow diagram of an intelligent optimized adaptive setting and control method for marine loading and unloading boom parameters according to the present invention. Fig. 3 This is a schematic diagram of the backpropagation neural network in this invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0020] Please see Figs. 1-3 An intelligent and optimized adaptive setting and control method for marine loading and unloading boom parameters. S1: By establishing a unified spatiotemporal data platform, collect multi-source heterogeneous data types; deploy a clock synchronization mechanism, designate the collection node closest to the dock control center as the reference node, and the remaining nodes receive the clock calibration signal from the reference node at fixed intervals; process outliers, and use interpolation to fill missing values ​​in the data transmission gaps; add spatiotemporal tags and spatial tags to each data item; S2: Construct a digital twin model of the loading and unloading boom, calculate the force on the boom at different angles; simulate the boom vibration characteristics under different medium flow velocities; record the pressure transmission process from the hydraulic station to the actuator; collect historical operation data of the loading and unloading boom, divide it into training set and validation set, and use a backpropagation neural network to correct the output results of the digital twin model of the loading and unloading boom; adjust the real-time data parameters in the model in real time. S3: Establish a parameter-performance mapping model, using input parameters as model features and output performance as model labels. Substitute these into the training set to train the model, adjust the parameters to optimize model accuracy, and verify the prediction results of the method through the test set. If the prediction error exceeds the preset range, supplement the data under different work scenarios and retrain the model. S4: Through failure mode and effect analysis, identify common failure modes of loading arms and confirm the influencing factors and judgment criteria for each failure mode; based on the failure mode and effect analysis results, record the key monitoring data corresponding to each failure mode; collect stress and vibration data of key parts in real time, and calculate the stress cycle number and current fatigue damage degree by combining the fatigue characteristic curve of the loading arm material; calculate the structural safety factor according to the material mechanics formula; and update the fatigue damage degree and safety factor simultaneously.

[0021] Specifically, the multi-source heterogeneous data type is used as the basic data source for constructing a unified spatiotemporal data platform. The specific process is as follows: Time-series data is collected, including real-time pressure of the hydraulic pipeline of the loading and unloading boom, vibration amplitude of various parts of the boom, and temperature of the medium being transported, continuously collected at a fixed frequency; Spatial data includes the berth identification of the loading and unloading boom, distance from the center of the column to the front edge of the wharf, and boom coordinates, with basic values ​​determined based on wharf survey drawings; Physical quantity data includes the actual extension length of the boom, rotation angle of each joint, hydraulic drive speed, and viscosity of the transported medium, obtained through real-time reading from sensors and extraction from equipment parameter documents. The three types of data are then summarized to obtain a multi-source heterogeneous dataset.

[0022] Specifically, the clock synchronization mechanism is used to ensure the time consistency of multi-source heterogeneous data acquisition. The specific process is as follows: select the acquisition node close to the dock control center as the reference node. This node must meet the conditions of stable power supply and network signal; the interval at which the reference node sends calibration signals is set in conjunction with the data acquisition frequency; after receiving the calibration signal, the other nodes compare the time difference with the reference node through the local clock chip, record the time breakpoints during the correction process, and subsequently ensure the continuity of timestamps through data supplementation; after all nodes have completed the correction, they feed back the synchronization result to the reference node.

[0023] Specifically, the outliers are used to process multi-source heterogeneous data. The specific process is as follows: When setting the equipment parameter thresholds, the thresholds are determined by referring to the factory design standards of the loading and unloading arm, industry safety specifications, and the statistical range of historical normal operation data; the collected data is compared with the thresholds one by one through the data inspection program. If the hydraulic pressure exceeds the design range or the vibration amplitude exceeds the structural tolerance value of the equipment, it is directly marked as abnormal data and removed. At the same time, the time of occurrence of the abnormality and the corresponding sensor location are recorded; for missing values ​​caused by data transmission gaps, linear interpolation is used to calculate them; the value at the time of the missing moment is calculated based on the ratio of the values ​​of the valid data before and after to the time interval; after filling is completed, the filled value is checked a second time, and if it exceeds the limit, it is recalculated.

[0024] Specifically, the relevant calculation formula for filling missing data values ​​using the linear interpolation method is as follows: , Where y is the missing data value to be filled (such as hydraulic pressure, vibration amplitude); x is the acquisition time of the missing data; y1 is the most recent valid data value before the missing time, and x1 is the acquisition time of the corresponding valid data; y2 is the most recent valid data value after the missing time, and x2 is the acquisition time of the corresponding valid data.

[0025] In this embodiment, for the No. 2 oil conveying arm (model RC / E, diameter DN250) of a 100,000-ton crude oil terminal, the missing hydraulic pipeline pressure value is calculated during data processing in S1. The specific process is as follows: When the data acquisition system experiences network fluctuations and misses hydraulic pressure data at the time of data acquisition (x), the system queries the valid data before and after the missing data, obtaining the pressure value at the time of acquisition of the valid data; this value is then substituted into the formula to calculate y; after filling the missing data, a second verification is performed, and this value is within the normal operating pressure range of the oil conveying arm (12-16MPa), confirming its validity; the data source is real-time data collected by the pressure sensor in the middle section of the hydraulic pipeline of the oil conveying arm.

[0026] Specifically, the calculation formula related to the pressure loss along the hydraulic pipeline is as follows: , Where ΔP is the pressure loss along the hydraulic pipeline (unit: MPa); λ is the friction coefficient (related to the roughness of the pipeline inner wall, the viscosity of the medium, and the flow velocity; λ decreases as the oil temperature increases); L is the length of the hydraulic pipeline (unit: m); d is the inner diameter of the hydraulic pipeline (unit: m); and ρ is the density of the hydraulic oil (unit: kg / m³). 3 (v varies with oil temperature); v is the flow velocity of hydraulic oil in the pipeline (unit: m / s); this formula is used in S2 to record the transmission process of hydraulic pressure from the hydraulic station to the actuator and to calculate pressure loss.

[0027] In this embodiment, for the No. 3 loading and unloading arm of an LNG terminal, the pressure loss along the pipeline is calculated when recording the hydraulic pressure transmission process in S2. The specific process is as follows: First, the parameters are determined: pipeline inner wall roughness Ra, hydraulic oil density ρ (at 40℃, refer to the hydraulic oil property manual), flow velocity v (calculated by the hydraulic pump outlet flow sensor), and friction coefficient λ (determined by referring to the Moody diagram based on roughness and flow velocity). Substitute these parameters into the formula to calculate ΔP. This pressure loss value will be entered into the hydraulic system sub-model of the digital twin model.

[0028] In this embodiment, for the boarding ladder (model EBL-E3, applicable to 10,000-80,000-tonnage terminals) of a 30,000-tonnage refined oil terminal, when adding spatiotemporal tags to the data in S1, the actual operation process is followed. The specific process is as follows: First, determine the data source—the "collection time" of the time tag is extracted from the clock signal synchronized from the reference node near the terminal's central control (this node is connected to the terminal's central control UPS power supply system to ensure clock stability), accurate to milliseconds; the "current tidal period" is obtained by calling the API of the terminal's hydrological monitoring system. The system obtains information through an interface connected to the local marine bureau, updating high tide, slack tide, and low tide levels in real time. The second step involves acquiring spatial tag information: the "berth number" is retrieved from the wharf berth management system, corresponding to "B3"; the "unique code for the boarding ladder equipment" is matched against the equipment's basic information database, resulting in "EBL-E3-017"; and the "distance from the center of the column to the wharf's edge" is queried from the wharf surveying database. The third step is tag binding: through a data processing program, time and spatial tags are automatically bound to the currently collected physical quantity data of the boarding ladder, namely "ladder pitch angle 15°," forming structured data entries. These entries are then stored in the relational database of a unified spatiotemporal data platform for subsequent use in digital twin modeling.

[0029] Specifically, the process of adding spatiotemporal and spatial tags to each data point for data association involves the following steps: After data cleaning, tag addition is automatically triggered; the time tag acquisition time information is extracted from the clock signal synchronized by the reference node, and the current tidal period information is obtained by calling the real-time data interface of the wharf hydrological monitoring system; the spatial tag berth number and the unique code of the loading and unloading boom equipment are matched and retrieved from the equipment basic information database, and the distance from the center of the column to the wharf front edge is queried from the wharf surveying database; the data processing program converts the extracted time and spatial information into tags in a unified format, binds them to each valid data point according to the data acquisition order, and finally forms a structured dataset with corresponding tags.

[0030] Specifically, S1 first classifies and collects multi-source heterogeneous data. Time-series data (hydraulic pipeline pressure, boom vibration amplitude, medium temperature, etc.) is collected by adjusting the frequency according to the working state through corresponding sensors. Spatial data (berth identification, boom coordinates, etc.) is obtained based on the calibrated survey drawings and management system. Physical quantity data (boom length, joint angle, etc.) is combined by sensor reading and parameter document extraction to form a basic data pool. Then, a clock synchronization mechanism is deployed, selecting a stable node close to the terminal control center as the reference node, sending calibration signals according to the collection frequency, and comparing and correcting time deviations and supplementing broken data at other nodes to ensure that the timestamps of all nodes are consistent. Subsequently, data cleaning is performed, setting thresholds based on equipment standards and historical data to remove outliers, filling in missing transmission values ​​using linear interpolation, and performing secondary verification. Finally, spatiotemporal tags are added to the cleaned data. The time tag includes the collection time and tidal period (obtained from the reference clock and hydrological system), and the spatial tag includes the berth number, equipment code, and installation distance (retrieved from the equipment library and survey library). The data is then bound by the program to form a structured dataset.

[0031] Specifically, the calculation of the force on the boom at different angles is used to construct a digital twin model of the boom. The specific process is as follows: based on the moment balance formula in rigid body dynamics, the boom material parameters and structural dimensions are input; when considering three types of loads, the wind load is calculated based on the long-term meteorological data provided by the meteorological department of the wharf location, the medium weight is calculated based on the product of the density of the transported medium and the pipeline volume, and the boom self-weight is calculated based on the product of the material density and the boom volume; the boom model is constructed, the joints are meshed with finer mesh, and the remaining parts are meshed according to conventional dimensions; the three types of loads are applied to the corresponding positions of the model to simulate the boom's rotation process from horizontal to vertical, and the axial force and radial force of each joint, as well as the maximum stress value and stress distribution of the middle section of the boom are output at different rotation angles.

[0032] Specifically, the simulated boom vibration characteristics under different medium flow velocities are used to reflect the dynamic characteristics of the loading boom digital twin model. The specific process is as follows: Based on fluid-structure interaction theory, a loading boom pipeline model is constructed, and the roughness of the pipeline inner wall is set according to the actual machining accuracy; the model is imported into fluid-structure interaction simulation software, and a medium flow velocity variable covering the design flow range of the loading boom is set, and the viscosity parameters of different media are input; boundary conditions are set, and the impact pressure and impact force of the medium on the pipeline inner wall under different flow velocities are calculated; the impact force is applied as a load to the boom structure model, modal analysis is performed, and the natural vibration frequency of the boom, the vibration frequency and vibration amplitude under different flow velocities are output; the simulation results are sorted and organized according to medium type and flow velocity.

[0033] Specifically, the recorded pressure transmission process from the hydraulic station to the actuator is used to supplement the hydraulic system characteristics of the loading and unloading arm digital twin model. The specific process is as follows: pressure sensors are installed at the main outlet pipe of the hydraulic station, the middle section of the pipeline, and the inlet of the actuator; the pressure value and pressure transmission time at each point are recorded synchronously, taking the moment when the pressure rises at the outlet of the hydraulic station as the starting point, and recording the time difference between that moment and the moment when the pressure rises at the middle section and the inlet of the actuator; when calculating the pressure transmission delay, extreme values ​​are removed after multiple measurements and the average value is taken; the pressure loss value of each section is calculated according to the pipeline friction resistance formula; at the same time, the transmission delay and the pressure loss data of each section are recorded, and the correspondence between oil temperature and pressure loss is correlated.

[0034] Specifically, the backpropagation neural network is used to correct the output of the loading arm digital twin model. The specific process is as follows: a multi-layer backpropagation neural network is constructed, with input layer neurons corresponding to the stress values ​​and vibration amplitudes output by the physical model, and output layer neurons corresponding to the corrected stress values ​​and vibration amplitudes; when filtering historical operation data, data from equipment failure periods are removed, covering operation scenarios of different media and different tidal periods, and divided into training and validation sets according to proportions; the network weights are adjusted using the gradient descent method, with the initial learning rate set according to the model convergence speed, and adjusted according to the validation set error after a certain number of iterations, iteratively training until the validation set error reaches the preset condition; the trained neural network is embedded into the computational model of the digital twin model through an interface to correct the stress and vibration data output by the physical model in real time.

[0035] Specifically, the real-time data parameters in the real-time adjustment model are used to maintain the consistency between the digital twin model of the loading and unloading arm and the actual equipment. The specific process is as follows: select the real-time parameters to be adjusted; collect the current wind speed, actual flow rate of the medium, and hydraulic pipeline oil temperature at the dock in real time through edge nodes; dynamically adjust the corresponding parameter values ​​according to the collected data; complete the parameter update according to a fixed cycle set according to the frequency of changes in the actual operating environment of the equipment; and compare the deviation between the model output and the actual monitoring data after each update.

[0036] Specifically, the parameter adjustment optimization model accuracy is used to improve the prediction accuracy of the parameter-performance mapping model. The specific process is as follows: determine the model parameters that need to be adjusted; adjust the parameters within a reasonable range preset in combination with the algorithm characteristics; after each adjustment, input the test set data into the model to calculate the prediction error. If the error exceeds the preset threshold, continue to adjust; if the error is within the preset threshold, stop adjusting; record the adjustment trajectory of each parameter during the optimization process.

[0037] In this embodiment, for the boom (model RC / E, diameter DN200) of a 50,000-ton crude oil terminal, the operation is combined with the actual working scenario when adjusting the parameters of the digital twin model in real time. The specific process is as follows: First, determine the real-time parameters to be adjusted—wind load coefficient and hydraulic pressure transmission efficiency. These two parameters are the core indicators affecting the matching degree between the model and reality. Second, collect real-time data—obtain the current wind speed from the weather station through the terminal edge node, read the oil temperature from the temperature sensor on the outer wall of the hydraulic pipeline of the boom, and calculate the actual flow velocity of the medium from the pipeline flow sensor. Third, dynamically adjust the parameters; when the oil temperature changes, the hydraulic oil viscosity decreases, and the hydraulic pressure transmission efficiency is adjusted accordingly. Fourth, verify the matching degree—compare the "boom vibration amplitude" output by the model with the actual sensor value every 30 seconds to ensure that the model is consistent with the actual operating state of the boom.

[0038] Specifically, the process of identifying common failure modes of loading and unloading booms is used to confirm the influencing factors and judgment criteria for each failure mode. The specific process is as follows: Identifying common failure modes of loading and unloading booms; for hydraulic system leakage, influencing factors include pressure decay rate and seal wear degree, with the judgment criterion being that the pressure decay rate exceeds the normal operating range or the seal wear reaches a set level that affects the sealing effect; for boom structure fatigue, influencing factors include stress cycle count and vibration frequency, with the judgment criterion being that the stress cycle count is close to the material fatigue limit or the vibration frequency is continuously within the material fatigue characteristic frequency range; for rotary joint jamming, influencing factors include angle change rate and lubricating grease viscosity, with the judgment criterion being that the angle change rate suddenly drops below the normal operating range or the lubricating grease viscosity exceeds the equipment's compatibility range.

[0039] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for intelligently optimized adaptive setting and control of marine loading and unloading boom parameters, characterized in that, include: S1: By establishing a unified spatiotemporal data platform, collect multi-source heterogeneous data types; A clock synchronization mechanism is deployed, designating the acquisition node closest to the dock's central control as the reference node, with other nodes receiving clock calibration signals from the reference node at fixed intervals; outliers are handled by using interpolation to fill in missing values ​​in data transmission gaps; and spatiotemporal and spatial tags are added to each data entry. S2: Construct a digital twin model of the loading and unloading boom, calculate the force on the boom at different angles; simulate the boom vibration characteristics under different medium flow velocities; record the pressure transmission process from the hydraulic station to the actuator; Historical operation data of the loading and unloading boom is collected and divided into training and validation sets. The output of the digital twin model of the loading and unloading boom is corrected using a backpropagation neural network. Real-time data parameters in the model are adjusted in real time. S3: Establish a parameter-performance mapping model, using input parameters as model features and output performance as model labels. Substitute these into the training set to train the model, adjust the parameters to optimize model accuracy, and verify the prediction results of the method through the test set. If the prediction error exceeds the preset range, supplement the data under different work scenarios and retrain the model. S4: Through failure mode and effect analysis, identify common failure modes of loading and unloading arms and confirm the influencing factors and judgment criteria for each failure mode; Based on the failure mode and effects analysis results, record the key monitoring data corresponding to each failure mode; collect stress and vibration data of key parts in real time, and calculate the stress cycle number and current fatigue damage degree by combining the fatigue characteristic curve of the loading arm material; calculate the structural safety factor according to the material mechanics formula. The fatigue damage level and safety factor are updated synchronously.

2. The method according to claim 1, characterized in that, The multi-source heterogeneous data type is used as the basic data source for constructing a unified spatiotemporal data platform. The specific process is as follows: Time-series data is collected, including real-time pressure of the hydraulic pipeline of the loading and unloading boom, vibration amplitude of various parts of the boom, and temperature of the medium being transported, continuously collected at a fixed frequency; Spatial data includes the berth identification of the loading and unloading boom, distance from the center of the column to the front edge of the wharf, and boom coordinates, with basic values ​​determined based on wharf survey drawings; Physical quantity data includes the actual extension length of the boom, rotation angle of each joint, hydraulic drive speed, and viscosity of the transported medium, obtained through real-time reading from sensors and extraction from equipment parameter documents. The three types of data are then summarized to obtain a multi-source heterogeneous dataset.

3. The method according to claim 1, characterized in that, The deployed clock synchronization mechanism is used to ensure the time consistency of multi-source heterogeneous data acquisition. The specific process is as follows: the acquisition node close to the dock control center is selected as the reference node, which must meet the conditions of stable power supply and network signal; the interval of the reference node sending calibration signals is set in combination with the data acquisition frequency; after receiving the calibration signal, the other nodes compare the time difference with the reference node through the local clock chip, record the time breakpoints during the correction process, and subsequently ensure the continuity of timestamps through data supplementation; after all nodes have completed the correction, they feed back the synchronization result to the reference node.

4. The method according to claim 1, characterized in that, The outliers are used to process multi-source heterogeneous data. The specific process is as follows: When setting the equipment parameter thresholds, the thresholds are determined by referring to the factory design standards of the loading and unloading arm, industry safety specifications, and the statistical range of historical normal operation data. The collected data is compared with the thresholds one by one through the data inspection program. If the hydraulic pressure exceeds the design range or the vibration amplitude exceeds the structural tolerance of the equipment, it is directly marked as abnormal data and removed. At the same time, the time of occurrence of the abnormality and the corresponding sensor location are recorded. For missing values ​​caused by data transmission gaps, linear interpolation is used to calculate them. The value at the time of the missing data is calculated based on the ratio of the values ​​of the valid data before and after and the time interval. After filling, the filled value is checked a second time. If it exceeds the limit, it is recalculated.

5. The method according to claim 1, characterized in that, The process of adding spatiotemporal and spatial tags to each data point for data association is as follows: After data cleaning, tag addition is automatically triggered; the time tag collection time information is extracted from the clock signal synchronized by the reference node, and the current tidal period information is obtained by calling the real-time data interface of the wharf hydrological monitoring system; the spatial tag berth number and the unique code of the loading and unloading arm equipment are matched and retrieved from the equipment basic information database, and the distance from the center of the column to the wharf front edge is queried from the wharf surveying database; the data processing program converts the extracted time and spatial information into tags of a unified format, binds them to each valid data point according to the data collection order, and finally forms a structured dataset with corresponding tags.

6. The method according to claim 1, characterized in that, The calculation of the force on the boom at different angles is used to construct a digital twin model of the boom. The specific process is as follows: Based on the moment balance formula in rigid body dynamics, the boom material parameters and structural dimensions are input; when considering three types of loads, the wind load is calculated based on the wind pressure according to the long-term meteorological data provided by the meteorological department of the wharf location, the medium weight is calculated based on the product of the density of the transported medium and the pipeline volume, and the boom self-weight is calculated based on the product of the material density and the boom volume; the boom model is constructed, the mesh is densified at the joints, and the mesh is divided according to the conventional size for the other parts. The three types of loads are applied to the corresponding positions of the model to simulate the rotation process of the boom from horizontal to vertical. At different rotation angles, the axial force and radial force of each joint, as well as the maximum stress value and stress distribution of the middle section of the boom are output respectively.

7. The method according to claim 1, characterized in that, The simulated boom vibration characteristics under different medium flow velocities are used to reflect the dynamic characteristics of the digital twin model of the loading and unloading boom. The specific process is as follows: Based on fluid-structure interaction theory, a loading and unloading boom pipeline model is constructed, and the roughness of the pipeline inner wall is set according to the actual machining accuracy; the model is imported into the fluid-structure interaction simulation software, and a medium flow velocity variable covering the design flow range of the loading and unloading boom is set, and the viscosity parameters of different media are input; boundary conditions are set, and the impact pressure and impact force of the medium on the pipeline inner wall under different flow velocities are calculated; the impact force is applied as a load to the boom structure model, modal analysis is performed, and the natural vibration frequency of the boom, the vibration frequency and vibration amplitude under different flow velocities are output; the simulation results are sorted and organized according to medium type and flow velocity.

8. The method according to claim 1, characterized in that, The recorded pressure transmission process from the hydraulic station to the actuator is used to supplement the hydraulic system characteristics of the loading and unloading arm digital twin model. The specific process is as follows: pressure sensors are installed at the main outlet pipe of the hydraulic station, the middle section of the pipeline, and the inlet of the actuator; the pressure value and pressure transmission time at each point are recorded synchronously. Taking the moment when the pressure rises at the outlet of the hydraulic station as the starting point, the time difference between that moment and the moment when the pressure rises at the middle section and the inlet of the actuator is recorded; when calculating the pressure transmission delay, extreme values ​​are removed after multiple measurements and the average value is taken; the pressure loss value of each section is calculated according to the pipeline friction resistance formula; at the same time, the transmission delay and the pressure loss data of each section are recorded, and the correspondence between oil temperature and pressure loss is correlated.

9. The method according to claim 1, characterized in that, The backpropagation neural network is used to correct the output of the loading arm digital twin model. The specific process is as follows: a multi-layer backpropagation neural network is constructed, with input layer neurons corresponding to the stress values ​​and vibration amplitudes output by the physical model, and output layer neurons corresponding to the corrected stress values ​​and vibration amplitudes; when filtering historical operation data, data from equipment failure periods are removed, covering operation scenarios of different media and different tidal periods, and divided into training and validation sets according to proportions; the network weights are adjusted using the gradient descent method, with the initial learning rate set according to the model convergence speed, and adjusted according to the validation set error after a certain number of iterations, iteratively training until the validation set error reaches the preset condition; the trained neural network is embedded into the computational model of the digital twin model through an interface to correct the stress and vibration data output by the physical model in real time.

10. The method according to claim 1, characterized in that, The real-time data parameters in the real-time adjustment model are used to maintain the consistency between the digital twin model of the loading and unloading arm and the actual equipment. The specific process is as follows: select the real-time parameters to be adjusted; collect the current wind speed, actual flow rate of the medium, and hydraulic pipeline oil temperature at the dock in real time through edge nodes; dynamically adjust the corresponding parameter values ​​according to the collected data; complete the parameter update according to a fixed cycle set according to the frequency of changes in the actual operating environment of the equipment; and compare the deviation between the model output and the actual monitoring data after each update.

11. The method according to claim 1, characterized in that, The parameter adjustment optimization model accuracy is used to improve the prediction accuracy of the parameter-performance mapping model. The specific process is as follows: determine the model parameters that need to be adjusted; adjust the parameters within a reasonable range preset in combination with the algorithm characteristics; after each adjustment, input the test set data into the model to calculate the prediction error. If the error exceeds the preset threshold, continue to adjust; if the error is within the preset threshold, stop adjusting; record the adjustment trajectory of each parameter during the optimization process.

12. The method according to claim 1, characterized in that, The common failure modes of the loading and unloading arm are identified to determine the influencing factors and judgment criteria for each failure mode. The specific process is as follows: identify common failure modes of the loading and unloading arm; for hydraulic system leakage, the influencing factors include pressure decay rate and seal wear degree, and the judgment criteria are that the pressure decay rate exceeds the normal operating range or the seal wear reaches the set degree that affects the sealing effect. For boom structure fatigue, influencing factors include stress cycle count and vibration frequency. The judgment criteria are that the stress cycle count is close to the material fatigue limit or the vibration frequency is continuously within the material fatigue characteristic frequency range. For swivel joint jamming, influencing factors include angle change rate and lubricating grease viscosity. The judgment criteria are that the angle change rate suddenly drops below the normal operating range or the lubricating grease viscosity exceeds the equipment's compatibility range.

Images