Steel material intelligent logistics management system

Abstract

The present application relates to the technical field of intelligent logistics management, and more particularly to a steel intelligent logistics management system, which comprises a collection module, a determination module, a separation module, an analysis module and an adjustment module. The present application adopts a full-process intelligent architecture with the cooperative operation of the collection module, the determination module, the separation module, the analysis module and the adjustment module, combines real-time perception of multi-source data, multi-parameter fusion early warning, accurate separation of direction and path, contribution quantification and attribution, and scene-specific active regulation, realizes closed-loop management of risk identification, path positioning, cause diagnosis and active prevention and control in steel component transportation, ensures accurate fitting of the hanging angle through quantification of contribution and determination of abnormal dominant factors, promotes the transformation of the system from passive response to active and accurate prevention and control, reduces the risk of steel damage, and effectively solves the problems of inability to respond to dynamic working conditions in real time, abnormal continuous accumulation and poor multi-scene adaptability caused by manual adjustment lag, fixed adjustment range and lack of closed-loop control.

Description

Technical Field

[0001] This invention relates to the field of intelligent logistics management technology, and in particular to an intelligent logistics management system for steel. Background Technology

[0002] With the rapid development of modern industry, steel, as a core raw material in construction, machinery, automobiles, and other fields, is experiencing explosive growth in logistics and transportation demand, with the scale and complexity of transportation continuously increasing. In road and rail transport scenarios, steel components, due to their large size and concentrated weight, are easily affected by factors such as road bumps, dynamic changes in vehicle posture, and uneven axle load distribution, resulting in dynamic anomalies such as component deflection, imbalance of force ratios across axles, and excessive force differences between left and right axles. If these anomalies are not promptly mitigated, they can lead to damage to steel structures and even transportation accidents, causing not only huge economic losses but also potentially significant safety hazards. Therefore, how to optimize the suspension system of transport vehicles to achieve proactive perception and adaptive control of dynamic anomalies has become a pressing technical challenge in the field of intelligent steel logistics.

[0003] Chinese Patent Publication No. CN219236695U discloses a multi-angle steering stabilization structure for a semi-trailer suspension system. The structure connects upper and lower fixed plates via a rotating shaft, allowing the suspension system to rotate around the shaft to adjust the angle. After adjustment, the nested components such as the first movable arm and the second movable arm are locked by the cooperation of the movable nut and the fixed nut, forming a stable structure with a fixed angle. This structure aims to solve the problem of excessive lateral tension and stress concentration in the front suspension when a semi-trailer turns.

[0004] However, this existing technology has the following significant drawbacks, failing to meet the needs of dynamic anomaly suppression in steel logistics transportation: The adjustment process relies on manual intervention and cannot respond to dynamic working conditions in real time. Angle adjustment of this structure requires manually tightening the movable nut and adjusting the relative position of the rotating shaft and movable arm, entirely dependent on operator experience. During steel transportation, road conditions and component states change in real time, requiring dynamic adaptation of the suspension angle. Manual adjustment cannot achieve real-time response, leading to large fluctuations in vehicle posture and continuous accumulation of dynamic anomalies during transportation; the fixed adjustment range cannot adapt to dynamic needs in multiple scenarios. The adjustment range of this structure is a preset fixed value, failing to consider the differentiated requirements of suspension angles in different transportation scenarios. Transporting heavy steel beams requires a larger angle adjustment to balance axle load, while transporting light steel plate coils requires a smaller angle to reduce energy consumption. The fixed range results in insufficient system versatility; and the lack of sensors and actuators prevents closed-loop control. The structural design does not integrate sensors and actuators, and only fixes the angle through mechanical locking. It cannot dynamically optimize the suspension angle based on the real-time collected component status, and it cannot form a closed-loop control of perception-decision-execution-feedback, resulting in limited anomaly suppression effect. Summary of the Invention

[0005] To address this, the present invention provides an intelligent logistics management system for steel, which overcomes the problems in the prior art caused by the inability to respond to dynamic working conditions in real time, the continuous accumulation of anomalies, and poor adaptability to multiple scenarios due to the lag in manual adjustment, fixed adjustment range, lack of closed-loop control, etc., by integrating multi-source fusion sensing, intelligent diagnosis and adaptive closed-loop adjustment technologies.

[0006] To achieve the above objectives, the present invention provides an intelligent logistics management system for steel, comprising:

[0007] The data acquisition module is used to collect in real time the road condition parameters, longitudinal attitude angle, lateral attitude angle, force ratio of each axle, force difference between left and right axles, and component deflection of the steel components of the transport vehicle operating based on a preset suspension angle during the transportation of steel components.

[0008] The determination module is used to determine whether to enter an abnormal mode based on the threshold comparison result of the structural health index. The structural health index is determined based on the structural offset index and the duration of the warning signal. The structural offset index is determined based on the change of the component deflection and the change of the force ratio of each axis. The warning signal is generated based on the threshold comparison result of the change trend of the component deflection and the distribution stability of the force ratio of each axis.

[0009] The separation module is used to determine the directional path based on the determination result of entering the abnormal mode, according to the longitudinal attitude angle, the lateral attitude angle, the force difference between the left and right axes, and the road condition parameters, wherein the directional path includes the longitudinal path, the lateral path, and the coupling path.

[0010] The analysis module is used to determine the dominant abnormal factor based on the coupling path and the comparison result of the longitudinal contribution and the lateral contribution. The longitudinal contribution is determined based on the change sequence of the component deflection and the longitudinal attitude angle within a preset judgment time, and the lateral contribution is determined based on the change sequence of the left and right axis force difference and the lateral attitude angle within a preset judgment time.

[0011] An adjustment module is used to adjust the preset suspension angle based on the aforementioned abnormal dominant factors.

[0012] Furthermore, the determination module includes:

[0013] The risk warning submodule is used to generate the warning signal when the trend parameter or stability parameter exceeds its respective preset warning threshold. The trend parameter is determined based on the linear regression slope of the component deflection within a preset trend period, and the stability parameter is determined based on the information entropy of the force ratio of each axis. The warning signal includes risk type and risk level.

[0014] The structural offset index calculation submodule is used to calculate the structural offset index based on the change sequence of the component deflection and the force ratio of each axis.

[0015] The status persistence factor calculation submodule is used to calculate the status persistence factor based on the severity weight and the duration of the warning signal, wherein the severity weight is determined based on the risk type and the risk level;

[0016] The structural health index calculation submodule is used to determine the structural health index based on the structural offset index, the state persistence factor, and the amplification factor, wherein the amplification factor is determined based on the road condition parameters.

[0017] The trigger decision submodule is used to determine whether to enter an abnormal mode when the structural health index exceeds a preset structural trigger threshold.

[0018] Furthermore, the risk warning submodule includes:

[0019] The trend parameter early warning unit is used to obtain the trend parameter by calculating the linear regression slope of the component deflection within a preset trend period. When the trend parameter exceeds the preset early warning threshold of the trend parameter within a preset number of consecutive sampling periods, a trend warning is triggered.

[0020] A stability warning unit is used to determine the stability parameter based on the information entropy of the force ratio of each axis, and to trigger a stability warning when the stability parameter is less than the corresponding preset warning threshold.

[0021] A risk classification unit is used to classify the risk level and the risk type based on the trend parameter, the stability parameter, the preset medium risk threshold, the preset high risk threshold and the corresponding preset warning threshold;

[0022] A signal generation unit is used to generate the early warning signal based on the trend parameter, the stability parameter, the risk level, and the risk type.

[0023] Furthermore, the structural offset index calculation submodule includes:

[0024] The variation feature extraction unit is used to extract the mean, variance, and peak value as quantitative feature values ​​based on the variation sequence of the component deflection and the force ratio of each axis.

[0025] The offset index generation unit is used to generate the structural offset index based on the quantized feature value through a preset calculation model.

[0026] Furthermore, the state persistence factor calculation submodule includes:

[0027] The severity weight determination unit is used to determine the corresponding severity weight based on the risk type and risk level of the warning signal through a preset risk weight rule base.

[0028] The state persistence factor calculation unit is used to perform a fusion calculation based on the severity weight and the duration of the warning signal to obtain the state persistence factor.

[0029] Furthermore, the separation module includes:

[0030] The initial probability calculation submodule is used to determine the longitudinal path probability based on the longitudinal attitude angle change rate and the road condition parameters, and to determine the lateral path probability based on the lateral attitude angle change rate and the left and right axis force difference, and to determine the coupled path probability based on the curve coupling factor and pitch and roll synchronization, wherein the curve coupling factor is determined based on the proportion of the left and right axis force difference and the road condition parameters, and the pitch and roll synchronization is determined based on the longitudinal attitude angle and the lateral attitude angle within a preset judgment time.

[0031] The path allocation submodule is used to determine the directional path as the longitudinal path when the longitudinal path probability is greater than the product of the lateral path probability and the preset dominant determination coefficient, and the lateral path probability is less than the preset probability baseline; and to determine the directional path as the lateral path when the lateral path probability is greater than the product of the longitudinal path probability and the preset dominant determination coefficient, and the longitudinal path probability is less than the preset probability baseline; and to determine the coupling path based on the longitudinal path and the lateral path.

[0032] Furthermore, the analysis module includes:

[0033] The contribution calculation submodule is used to determine the longitudinal contribution based on the change sequence of the component deflection and the longitudinal attitude angle within a preset determination time, and to determine the lateral contribution based on the change sequence of the left and right axis force difference and the lateral attitude angle within a preset determination time.

[0034] The abnormal dominant factor determination submodule is used to calculate the relative proportion and absolute value of the difference between the horizontal contribution and the vertical contribution, and to determine the abnormal dominant factor based on preset determination rules.

[0035] Furthermore, the contribution calculation submodule includes:

[0036] The feature extraction unit is used to extract the mean of the component deflection, the variance of the longitudinal attitude angle, and the covariance of the component deflection and the longitudinal attitude angle based on the respective change sequences of the component deflection and the longitudinal attitude angle to form a longitudinal feature set; and to extract the mean of the change in the left and right axis force difference, the peak value of the change in the lateral attitude angle, and the correlation coefficient between the left and right axis force difference and the change in the lateral attitude angle based on the respective change sequences of the left and right axis force difference and the lateral attitude angle within a preset determination time period to form a lateral feature set.

[0037] The feature fusion subunit calculates the vertical contribution and the horizontal contribution based on the vertical feature set and the horizontal feature set.

[0038] Furthermore, the abnormal dominant factor determination submodule is used to determine that the abnormal dominant factor is a vertical dominant factor when the absolute value of the difference is greater than a preset difference threshold and the relative proportion is less than a preset low proportion threshold; and to determine that the abnormal dominant factor is a horizontal dominant factor when the absolute value of the difference is greater than a preset difference threshold and the relative proportion is greater than a preset high proportion threshold; and to determine that the abnormal dominant factor is a composite dominant factor when the absolute value of the difference is less than a preset difference threshold and the relative proportion is between a preset low proportion threshold and a preset high proportion threshold.

[0039] Furthermore, the adjustment module includes:

[0040] The first adjustment submodule is used to adjust the preset suspension angle based on the longitudinal suspension angle adjustment amount when the abnormal dominant factor is the longitudinal dominant factor, wherein the longitudinal suspension angle adjustment amount is calculated based on the longitudinal contribution, the preset longitudinal contribution benchmark value and the longitudinal adjustment coefficient.

[0041] The second adjustment submodule is used to adjust the preset suspension angle based on the lateral suspension angle adjustment amount when the abnormal dominant factor is the lateral dominant factor. The lateral suspension angle adjustment amount is calculated based on the ratio of the lateral contribution to the preset lateral contribution benchmark value.

[0042] The third adjustment submodule is used to adjust the preset suspension angle based on the composite suspension angle adjustment amount when the abnormal dominant factor is a composite dominant factor. The composite suspension angle adjustment amount is calculated based on the longitudinal contribution, the lateral contribution, the longitudinal suspension angle adjustment amount, and the lateral suspension angle adjustment amount.

[0043] Compared with existing technologies, the beneficial effects of this invention are as follows: by adopting a fully intelligent architecture that coordinates the operation of acquisition, judgment, separation, analysis, and adjustment modules, combined with real-time perception of multi-source data, multi-parameter fusion early warning, precise separation of direction and path, contribution quantification and attribution, and proactive control in different scenarios, it achieves closed-loop management of risk identification, path positioning, cause diagnosis, and proactive prevention and control in the transportation of steel components. Multi-parameter fusion early warning improves the comprehensiveness of anomaly identification, dynamic feature quantification enhances the depth of risk judgment, dynamic adaptation of working conditions improves the accuracy of judgment, and threshold triggering ensures timely response. Through contribution quantification and determination of the dominant factors of anomalies, it achieves an upgrade from phenomenon identification to cause attribution. Then, through scenario-based adjustment strategies and closed-loop feedback, it ensures precise adaptation of suspension angle. Ultimately, it promotes the transformation of the system from passive response to proactive and precise prevention and control, significantly improving the depth of risk judgment, adjustment efficiency, and transportation stability, reducing the risk of steel damage, and effectively solving the problems of inability to respond to dynamic working conditions in real time, continuous accumulation of anomalies, and poor adaptability to multiple scenarios caused by manual adjustment lag, fixed adjustment range, and lack of closed-loop control. Attached Figure Description

[0044] Figure 1 This is a structural diagram of the intelligent steel logistics management system in this embodiment;

[0045] Figure 2 This is a flowchart illustrating the process of the determination module in this embodiment determining whether to enter an abnormal mode;

[0046] Figure 3 A flowchart for generating early warning signals in this embodiment;

[0047] Figure 4 This is a flowchart for adjusting the preset suspension angle in this embodiment. Detailed Implementation

[0048] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0049] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0050] Please see Figure 1 As shown, it is a structural diagram of the intelligent steel logistics management system of this embodiment;

[0051] This embodiment provides a smart logistics management system for steel, including:

[0052] The data acquisition module is used to collect in real time the road condition parameters, longitudinal attitude angle, lateral attitude angle, force ratio of each axle, force difference between left and right axles, and component deflection of the steel components of the transport vehicle operating based on a preset suspension angle during the transportation of steel components.

[0053] The determination module is used to determine whether to enter an abnormal mode based on the threshold comparison result of the structural health index. The structural health index is determined based on the structural offset index and the duration of the warning signal. The structural offset index is determined based on the change of the component deflection and the change of the force ratio of each axis. The warning signal is generated based on the threshold comparison result of the change trend of the component deflection and the distribution stability of the force ratio of each axis.

[0054] The separation module is used to determine the directional path based on the determination result of entering the abnormal mode, according to the longitudinal attitude angle, the lateral attitude angle, the force difference between the left and right axes, and the road condition parameters, wherein the directional path includes the longitudinal path, the lateral path, and the coupling path.

[0055] The analysis module is used to determine the dominant abnormal factor based on the coupling path and the comparison result of the longitudinal contribution and the lateral contribution. The longitudinal contribution is determined based on the change sequence of the component deflection and the longitudinal attitude angle within a preset judgment time, and the lateral contribution is determined based on the change sequence of the left and right axis force difference and the lateral attitude angle within a preset judgment time.

[0056] An adjustment module is used to adjust the preset suspension angle based on the aforementioned abnormal dominant factors.

[0057] The data acquisition module is deployed on the chassis of the transport vehicle and the fixed brackets of the steel components, collecting parameters in real time through multiple types of sensors. Specifically, a lidar mounted on the front of the vehicle scans the road ahead to identify road surface smoothness, radius of curvature, and slope to collect road condition parameters; a dual-axis tilt sensor is installed at the vehicle's center of gravity to measure longitudinal and lateral attitude angles; piezoelectric pressure sensors are installed on the front and rear axle suspension systems to measure the vertical force on each axle, calculate the force ratio of each axle and the force difference between the left and right axles; and a fiber optic strain sensor is deployed at the mid-span of the steel components to sense strain through wavelength drift and then convert the strain into component deflection.

[0058] The preset suspension angle is based on the neutral angle of the suspension system when the transport vehicle is unloaded, and is initially set in conjunction with the stable bearing angle of the steel components. In this embodiment, based on simulation experiments, the preset suspension angle is set to 1.5°, which can keep the mid-span deflection of the steel beam in a low deformation range, thereby adapting to different types of steel components and transportation conditions.

[0059] By collecting multi-dimensional dynamic parameters in real time, it avoids the anomaly omissions caused by single parameters in existing technologies. The structural offset index integrates the changes in component deflection and the force ratio of each axis, combines the time effect of the accumulated warning signal of the state persistence factor, and introduces a road condition parameter amplification coefficient to reflect the impact of severe working conditions. Anomaly patterns are determined by comparing structural health index thresholds, achieving an upgrade from single over-limit alarms to multi-parameter fusion diagnosis, significantly improving the anomaly identification accuracy in complex scenarios. Based on initial probability, it distinguishes between longitudinal paths, lateral paths, and coupled paths to clarify the direction of anomaly propagation. Furthermore, it compares the longitudinal contribution and the lateral contribution. Identifying the dominant factors of anomalies provides a precise basis for differentiated adjustments. Based on these factors, the suspension angle adjustment is calculated using differentiated strategies. Closed-loop control is achieved by combining safety limits and execution feedback submodules, avoiding adjustment lag or overshoot, effectively reducing component deflection and stress differences, and suppressing the continuous accumulation of anomalies. The entire system forms a complete closed loop of perception, diagnosis, decision-making, execution, and feedback, enabling steel logistics transportation to shift from passive response to proactive prevention and control. This effectively solves the problems of inability to respond to dynamic working conditions in real time, continuous accumulation of anomalies, and poor adaptability to multiple scenarios caused by manual adjustment lag, fixed adjustment range, lack of closed-loop control, etc.

[0060] Please continue reading. Figure 2 As shown, it is a flowchart of the determination module in this embodiment determining whether to enter the abnormal mode.

[0061] The determination module includes:

[0062] The risk warning submodule is used to generate the warning signal when the trend parameter or stability parameter exceeds its respective preset warning threshold. The trend parameter is determined based on the linear regression slope of the component deflection within a preset trend period, and the stability parameter is determined based on the information entropy of the force ratio of each axis. The warning signal includes risk type and risk level.

[0063] The structural offset index calculation submodule is used to calculate the structural offset index based on the change sequence of the component deflection and the force ratio of each axis.

[0064] The status persistence factor calculation submodule is used to calculate the status persistence factor based on the severity weight and the duration of the warning signal, wherein the severity weight is determined based on the risk type and the risk level;

[0065] The structural health index calculation submodule is used to determine the structural health index based on the structural offset index, the state persistence factor, and the amplification factor, wherein the amplification factor is determined based on the road condition parameters.

[0066] The trigger decision submodule is used to determine whether to enter an abnormal mode when the structural health index exceeds a preset structural trigger threshold.

[0067] The setting of the preset trend period depends on real-time requirements, trend representativeness, and statistical validity. In this embodiment, heavy trucks transporting steel beam components are used as the object, and the preset trend period is set to 10 minutes to avoid short-term noise interference, cover typical scene units, and balance real-time performance and robustness.

[0068] The preset warning thresholds include trend parameter preset warning thresholds and stability parameter preset warning thresholds. The setting of the trend parameter preset warning threshold depends on the deformation characteristics of the steel component, the critical value of safety risk, and the dynamic requirements of the transportation scenario. In this embodiment, the maximum slope of undamaged cases and the minimum slope of damaged cases in past transportation are combined, and the median value of 0.015 mm / s is taken as the preset warning threshold to balance safety and economy.

[0069] The setting of the preset early warning threshold for the stability parameter depends on the critical state of stress balance, the safe stress characteristics of the steel component, historical damage data, and industry standards. In this embodiment, based on historical transportation cases, the preset early warning threshold for the stability parameter is set to 1.0, which can provide early warning of cumulative axle load damage.

[0070] The amplification factor is set based on road condition parameters, industry standards, and risk control requirements. Road condition parameters include three types of road condition factors: pavement smoothness, pavement curvature, and pavement slope. In this embodiment, the amplification factor calculation model uses a weighted summation model to integrate the three types of road condition factors. Specifically, pavement smoothness is represented by the International Roughness Index, pavement curvature by the reciprocal of curvature, and pavement slope by the slope angle. First, the three types of road condition parameters are normalized, and then the amplification factor is calculated by weighted summation according to the following formula:

[0071]

[0072] in, , , These are the weighting coefficients. The normalized road surface smoothness, The normalized road surface curvature, The normalized road slope is represented by a base coefficient of 1, indicating that there is no amplification effect on flat road conditions.

[0073] The weighting coefficients are set based on the degree of influence of different road conditions on the stress of steel components. In this embodiment, the actual vehicle transportation experiment and multibody dynamics simulation are used. =0.3, =0.4, =0.3, prioritizing the impact of curvature on attitude, can accurately quantify the differential amplification effect of different road conditions on structural risks.

[0074] The setting of the preset structural trigger threshold depends on the combination of experimental and simulation damage threshold calibration, historical case statistical correction, and industry standards. In this embodiment, heavy trucks transporting steel beam components are used as the object. Based on experimental calibration and historical data correction, the preset structural trigger threshold is set to 1.2, which can accurately distinguish between normal fluctuations and abnormal accumulation, avoiding false alarms and missed judgments. The preset structural trigger threshold is updated quarterly using new transportation data: if a certain type of component is frequently damaged at the preset structural trigger threshold, the threshold is lowered; if there is no damage for a long period of time, the preset structural trigger threshold is raised.

[0075] By integrating the linear regression slope of the deflection of trend parameters and the information entropy of the force ratio of each axis of stability parameters, early warning signals containing risk types and levels are generated. Based on the change sequence of these two parameters, the synergistic effect of dynamic deformation and force imbalance is quantified. Combining the severity weight and the cumulative abnormal time effect of the duration of the early warning signal, a road condition parameter amplification coefficient is introduced to achieve dynamic adaptation of working conditions. Anomaly patterns are accurately determined by comparison with preset thresholds. A full-chain logic of risk early warning, feature fusion, cumulative assessment, working condition adaptation, and precise triggering is constructed. This breaks through the shortcomings of existing technologies such as single parameter monitoring, lack of cumulative effect, and lack of dynamic adaptation. It achieves multi-parameter fusion early warning to improve the comprehensiveness of anomaly identification, dynamic feature quantification to enhance the depth of risk judgment, dynamic adaptation of working conditions to improve the accuracy of judgment results, and threshold-based precise triggering to ensure timely anomaly response, thus promoting the transformation of steel logistics from passive response to proactive prevention and control.

[0076] Please continue reading. Figure 3 As shown, this is a flowchart of the process for generating the warning signal in this embodiment.

[0077] The risk warning submodule includes:

[0078] The trend parameter early warning unit is used to obtain the trend parameter by calculating the linear regression slope of the component deflection within a preset trend period. When the trend parameter exceeds the preset early warning threshold of the trend parameter within a preset number of consecutive sampling periods, a trend warning is triggered.

[0079] A stability warning unit is used to determine the stability parameter based on the information entropy of the force ratio of each axis, and to trigger a stability warning when the stability parameter is less than the corresponding preset warning threshold.

[0080] A risk classification unit is used to classify the risk level and the risk type based on the trend parameter, the stability parameter, the preset medium risk threshold, the preset high risk threshold and the corresponding preset warning threshold;

[0081] A signal generation unit is used to generate the early warning signal based on the trend parameter, the stability parameter, the risk level, and the risk type.

[0082] Specifically, in this embodiment, the trend parameter is the linear regression slope. The least squares method is used to fit the component deflection time series within a preset trend period, and the linear regression slope is calculated. Before the calculation, the component deflection time series is preprocessed, specifically including filtering out instantaneous deflection fluctuations caused by bumpy road surfaces through wavelet transform, filling in missing sampling points using linear interpolation, and then normalizing the component deflection.

[0083] The preset trend warning quantity refers to the critical number of consecutive sampling periods during which a trend parameter exceeds a preset warning threshold before a trend warning is triggered. Its value depends on the quantitative correlation verified by experiments, the adaptability calibration of the sampling frequency, and the precise balance between false alarm rate and response speed. Specifically, firstly, based on historical transportation experiments and multivariate statistical analysis, the false alarm rate and response effect under different number of periods are tested, using the number of consecutive periods where the trend parameter exceeds the threshold as a variable: when the number of consecutive periods is 3, the false alarm rate is reduced by approximately 70% compared to a single period, and the accuracy in distinguishing between single dominant and compound dominant anomalies reaches 92%, representing the optimal balance between false alarm rate and response speed. Secondly, considering the sensor sampling frequency adaptation for transportation vehicles: the sampling frequency determines the sampling period (e.g., 1Hz sampling corresponds to a 1-second period), and the number of consecutive periods needs to match the period length to control the total confirmation time, i.e., the number of consecutive periods × sampling period should be within a reasonable range of 2-5 seconds. With 1Hz sampling, 3 periods correspond to a total confirmation time of 3 seconds, which can filter instantaneous noise and respond promptly to the true trend. Finally, through engineering practice, it was verified that three cycles perform stably in most transportation scenarios, avoiding both noise residue from too few cycles and response delay from too many cycles. In this embodiment, the sampling frequency of the transport vehicle sensor is set to 1Hz. Based on the experimentally verified optimal value, the adaptability of 1Hz sampling, and considerations of noise suppression and response balance, the preset trend warning number is set to 3, which can significantly reduce the false alarm rate, ensure timely response to anomalies, and achieve a precise balance between noise suppression and trend confirmation.

[0084] Risk types include trend anomalies, stability anomalies, or combined anomalies. Trend anomalies are only those where the linear regression slope of component deflection exceeds the threshold, reflecting an accelerating deformation trend. Stability anomalies are only those where the information entropy of the force ratio of each axis is lower than the threshold, reflecting an imbalance in the distribution of axis loads. Combined anomalies are those where both trend parameters and stability parameters exceed the threshold, reflecting deformation-force coupling risks.

[0085] The preset medium-risk threshold and preset high-risk threshold are critical values ​​used to determine the medium-risk and high-risk levels. Their values ​​are determined by a comprehensive analysis of the plastic deformation characteristics of steel, experimental data, and engineering safety standards. Specifically: First, based on the plastic deformation characteristics of steel, the stress-strain curves of the component under axial tension and bending moment are clarified. Combined with the component's geometric dimensions, such as the span of the steel beam and the moment of inertia of the cross section, the theoretical deformation threshold is calculated to define the starting point of plastic deformation and the critical point of failure. Second, based on experimental data calibration, through historical transportation measurements, multibody dynamics simulations of stress distribution under curve tilt and longitudinal bump conditions, and the cumulative law of fatigue deformation under cyclic loading in material mechanics tests, the upper limit of parameters in undamaged cases and the lower limit of parameters in damaged cases are statistically analyzed to determine the critical range of deformation and stress imbalance in actual scenarios. Finally, combined with engineering safety standards as a safety net, mandatory requirements for the allowable deformation value of the component and the balance of axle load distribution are introduced, balancing economy and safety. In this embodiment, the preset high-risk threshold is set to 0.015 mm / s and the preset high-risk threshold is set to 0.03 mm / s, which can provide early warning of elastic deformation and plastic deformation risks.

[0086] The preset stability parameter warning threshold is a critical value used to determine whether the force distribution of each axle is unbalanced. It depends on a comprehensive analysis of historical experimental data and engineering safety standards. Specifically, through statistical analysis of the axle force ratio sequence under normal / abnormal working conditions in historical transportation, simulation of curve tilt in multibody dynamics, axle load transfer during emergency braking, and the strain accumulation law of components under different axle load ratios in material mechanics tests, it was found that the information entropy of the force ratio in undamaged cases is generally >1.0, while the entropy value in damaged cases is <1.0. Therefore, 1.0 is used as the imbalance threshold. Referring to the requirement of axle load distribution imbalance ≤30% in the "Technical Specification for Safety of Road Freight Transport", and combining the mapping relationship between information entropy and imbalance (entropy value 1.0 corresponds to imbalance of 30%), a safety margin coefficient of 1.2 is introduced to ensure that the threshold is higher than the actual risk lower limit. In this embodiment, the preset stability parameter warning threshold is set to 1.0, which can improve the quantitative accuracy of the structural health index on the cumulative effect of force imbalance.

[0087] In this embodiment, the correspondence between threshold ranges and risk levels and types is quantified by a table to achieve intuitive determination of risk classification. The preset medium risk threshold is K1, the preset high risk threshold is K2, and the preset stability parameter warning threshold is H.

[0088]

[0089] The warning signal includes the following fields: timestamp, trend parameter, stability parameter, risk type, risk level, warning duration, and force ratio of each axis. It uses both JSON and CAN bus DBC protocols. For transmission, it is directed to the structural offset index calculation submodule and the state duration factor calculation submodule of the judgment module via the system's internal bus.

[0090] By using the component deflection time series within a preset trend period, the linear regression slope is calculated using the least squares method. After wavelet transform denoising and linear interpolation for gap filling, the true deformation trend is confirmed by exceeding the standard for multiple consecutive sampling periods, thus suppressing false alarms due to instantaneous noise. Stability parameters are defined based on the entropy of the force ratio information of each axis, and latent force imbalances are identified by the early warning threshold when the entropy value is lower than the preset stability parameter. Risk levels and risk types are classified through table mapping. Early warning signals are output to achieve noise filtering, latent imbalance capture, composite risk classification, and closed-loop control, promoting the transformation of steel logistics from passive response to proactive prevention and control, and significantly improving transportation safety and economy.

[0091] The structural offset index calculation submodule includes:

[0092] The variation feature extraction unit is used to extract the mean, variance, and peak value as quantitative feature values ​​based on the variation sequence of the component deflection and the force ratio of each axis.

[0093] The offset index generation unit is used to generate the structural offset index based on the quantized feature value through a preset calculation model.

[0094] Specifically, the variation sequences of component deflection and the force ratio of each axis are preprocessed. This includes using wavelet transform to filter out transportation noise, using linear interpolation to fill in missing values, and using Z-score normalization for normalization. The number of decomposition levels depends on the noise characteristics of the transportation scenario, signal fidelity requirements, and computational efficiency. In this embodiment, when using wavelet transform to filter out transportation noise, a db4 wavelet basis is selected, with a decomposition level of 3. Low-frequency approximate components are retained to reflect the true deformation trend, while high-frequency detail components are removed, which can accurately filter out transportation noise and avoid distortion of the effective signal.

[0095] The preset calculation model includes at least one of the following: a weighted summation model, a fuzzy comprehensive evaluation model, a neural network model, a support vector machine model, or a rule-based reasoning model. Since the weighted summation model is computationally efficient and highly interpretable, it is adopted in this embodiment. Based on historical transportation experiments, the contribution of each feature to the structural offset risk is determined through multiple regression analysis. In this embodiment, the weights are allocated as mean 0.25, variance 0.4, and peak 0.35, with a sum of weights of 1.0, which improves the quantitative accuracy of the structural offset index in assessing the synergistic effect of deformation and stress.

[0096] By employing wavelet transform denoising, linear interpolation for missing data, and Z-score normalization as data preprocessing techniques, the quality and comparability of input data are significantly improved, suppressing the impact of noise interference and missing value bias on subsequent analysis. Multi-dimensional feature quantification of mean, variance, and peak values ​​can comprehensively quantify the dynamic coupling effect of deformation and stress imbalance, avoiding the loss of information from single features and ensuring the complete representation of dynamic characteristics. Based on the above quantified feature values, a structural offset index is generated according to the weights calibrated by historical experiments through a preset weighted summation model, transforming the deformation-stress synergy effect into a single quantitative index that intuitively reflects the degree of risk. This achieves accurate quantification of the synergy effect of dynamic deformation and stress imbalance in the transportation of steel components, providing accurate input for subsequent calculation of state persistence factors and anomaly mode determination.

[0097] The state duration factor calculation submodule includes:

[0098] The severity weight determination unit is used to determine the corresponding severity weight based on the risk type and risk level of the warning signal through a preset risk weight rule base.

[0099] The state persistence factor calculation unit is used to perform a fusion calculation based on the severity weight and the duration of the warning signal to obtain the state persistence factor.

[0100] The preset risk weight rule base is built based on historical transportation damage data and industry standards. It stores the severity weights corresponding to different risk types and risk level combinations in tabular form. The severity weights range from 0 to 1, with larger values ​​indicating higher risk severity.

[0101] The exponential decay model can reflect the cumulative effect as the duration increases, while avoiding the excessive amplification of long-term anomalies by the linear model. It fits the engineering reality that risk accumulation is approaching saturation. In this embodiment, based on the severity weight and the duration of the warning signal, the state duration factor is generated by the exponential decay fusion model, where the duration of the warning signal accumulates from the first anomaly trigger.

[0102]

[0103] in, For state persistence factor; λ is the severity weight; λ is the attenuation coefficient, in units of s⁻¹, used to control the rate at which duration affects the cumulative effect; e is the natural constant; and T is the duration of the warning signal.

[0104] The value of the attenuation coefficient is determined based on experimental data fitting, model adaptability, and actual engineering needs. The core objective is to ensure that the state duration factor can reasonably quantify the cumulative effect of risk severity versus duration. In this embodiment, the attenuation coefficient is calibrated based on experimental data and set to 0.01 / s, which ensures that the cumulative effect approaches saturation when the duration is sufficiently long, preventing it from increasing indefinitely.

[0105] By invoking a pre-defined risk weight rule base based on the risk type and level of the early warning signal, abstract risks are transformed into quantifiable severity weights, eliminating subjective judgment bias and ensuring that the severity of different risk combinations is comparable and traceable. Based on the severity weights and the duration of the early warning signal, an exponential decay fusion model is used to calculate the state persistence factor, quantifying the pattern that the longer the duration, the more significant the cumulative effect, but asymptotically saturates. This distinguishes between short-term fluctuations and long-term anomalies, avoiding the limitations of existing technologies that only identify single anomalies. It achieves an upgrade from single risk identification to cumulative effect assessment, providing dynamic input of risk severity and duration for anomaly pattern determination, supporting precise interventions such as subsequent path separation and suspension angle adjustment, capturing progressive damage, quantifying cumulative risk, and avoiding over-response.

[0106] The separation module includes:

[0107] The initial probability calculation submodule is used to determine the longitudinal path probability based on the longitudinal attitude angle change rate and the road condition parameters, and to determine the lateral path probability based on the lateral attitude angle change rate and the left and right axis force difference, and to determine the coupled path probability based on the curve coupling factor and pitch and roll synchronization, wherein the curve coupling factor is determined based on the proportion of the left and right axis force difference and the road condition parameters, and the pitch and roll synchronization is determined based on the longitudinal attitude angle and the lateral attitude angle within a preset judgment time.

[0108] The path allocation submodule is used to determine the directional path as the longitudinal path when the longitudinal path probability is greater than the product of the lateral path probability and the preset dominant determination coefficient, and the lateral path probability is less than the preset probability baseline; and to determine the directional path as the lateral path when the lateral path probability is greater than the product of the longitudinal path probability and the preset dominant determination coefficient, and the longitudinal path probability is less than the preset probability baseline; and to determine the coupling path based on the longitudinal path and the lateral path.

[0109] Specifically, it is used to normalize the longitudinal attitude angle, lateral attitude angle, left-right axle force difference, and road condition parameters using Min-Max, mapping each parameter to the [0,1] interval. Then, it calculates the rate of change for the longitudinal and lateral attitude angles to reflect dynamic attitude changes, and calculates the proportion of the left-right axle force difference to quantify the degree of axle load imbalance. Based on the normalized rate of change, road condition parameters, and their respective weights, a linear weighted model is used to calculate the longitudinal path probability, where each weight is calibrated based on the contribution of the corresponding variable to path anomalies in historical data. Based on the lateral attitude angle rate of change and the proportion of the left-right axle force difference, a product-saturation model is used to calculate the lateral path probability, with the following formula:

[0110]

[0111] in Here, k represents the lateral path probability, and k is the gain coefficient. This represents the normalized rate of change of the lateral attitude angle. This represents the normalized percentage of the force difference between the left and right axes.

[0112] The coupling path probability is based on a weighted sum of the curve coupling factor and pitch and roll synchronization. In this embodiment, through historical curve transportation experiments, the accuracy of coupling path identification with different weight allocations was compared. It was determined that the accuracy was the highest and the false positive rate was the lowest when the weights were equal. Therefore, the curve coupling factor and pitch and roll synchronization weights are each 0.5.

[0113] The gain coefficient is set based on the amplification requirement of the coupling effect between the rate of change of the lateral attitude angle and the proportion of the force difference between the left and right axes, as well as the probability boundary constraints. In this embodiment, through abnormal case experiments, the gain coefficient is set to 2.0, which can achieve the highest recognition accuracy and the lowest false positive rate.

[0114] The curve coupling factor is determined based on the product of the normalized ratio of the left and right axis force differences and road condition parameters. The pitch and roll synchronization is determined based on the cross-correlation coefficient between the longitudinal attitude angle and the lateral attitude angle within a preset judgment period. The closer the synchronization is to 1, the stronger the pitch and roll synchronization. The specific calculation formula is as follows:

[0115]

[0116] in, For pitch and roll synchronization, t is a time variable in seconds, and N is the preset judgment duration in seconds. For the longitudinal attitude angle time series within the preset judgment time period, This is the time series of lateral attitude angles within a preset judgment period. and This represents the mean of the corresponding sequence.

[0117] The preset judgment duration is determined based on scenario requirements, experimental data, and computational efficiency. The core objective is to ensure that the pitch and roll synchronization accurately reflects the synchronous changes in longitudinal and lateral attitude angles under curve conditions. In this embodiment, the preset judgment duration is set to 5 seconds to ensure coverage of most curve scenarios, adapt to system real-time performance and computational efficiency, and improve the accuracy of coupled path recognition.

[0118] The preset dominant judgment coefficient is set based on historical transportation data statistics, experimental verification, and scenario adaptability. In this embodiment, the preset dominant judgment coefficient adopts the optimal ratio of 1.2 from historical data statistics to filter out instantaneous probability fluctuations caused by a single turbulence and improve the stability of the judgment.

[0119] The preset probability baseline is set based on the background noise level of the path probability in normal transportation. A probability below this value is considered as no significant anomaly. In this embodiment, it is set to 0.3, which is statistically calibrated based on the background noise of normal transportation. This can eliminate interference from dual low-probability coupled paths, identify scenarios without significant anomalies, and improve the specificity of path allocation determination.

[0120] When neither the longitudinal path condition nor the lateral path condition is met, the directional path is determined to be a coupled path. Specifically, a coupled path is determined when the longitudinal path probability is not greater than the product of the lateral path probability and the preset dominant determination coefficient, or when the lateral path probability is not lower than the preset probability baseline and the lateral path probability is not greater than the product of the longitudinal path probability and the preset dominant determination coefficient, or when the longitudinal path probability is not lower than the preset probability baseline.

[0121] By combining longitudinal attitude angle change rate with road condition parameters through longitudinal path probability fusion, lateral attitude angle change rate with left and right axis force difference through lateral path probability fusion, and curve coupling factor with pitch and roll synchronization through coupled path probability fusion, the core characteristics of different paths are comprehensively captured. The dynamic coupling relationship of deformation-force-road condition-attitude synchronization is quantified to avoid omissions due to single parameter monitoring. Based on the judgment logic of preset dominant judgment coefficient and preset probability baseline, the boundary between single path dominance and coupled path is clearly distinguished, reducing short-term fluctuation interference and improving judgment specificity.

[0122] The analysis module includes:

[0123] The contribution calculation submodule is used to determine the longitudinal contribution based on the change sequence of the component deflection and the longitudinal attitude angle within a preset judgment time period by feature extraction and fusion calculation, and to determine the lateral contribution based on the change sequence of the left and right axis force difference and the lateral attitude angle within a preset judgment time period by feature extraction and fusion calculation.

[0124] The abnormal dominant factor determination submodule is used to calculate the relative proportion and absolute value of the difference between the horizontal contribution and the vertical contribution, and to determine the abnormal dominant factor based on preset determination rules.

[0125] The pre-defined judgment rules are constructed based on a dual-indicator framework of absolute difference and relative proportion. The absolute difference reflects the absolute difference between horizontal and vertical contributions, used to distinguish between single-dominant and multi-dominant scenarios; the relative proportion reflects the ratio of horizontal contribution to the total contribution, used to clarify the direction of single-dominant factors. The judgment logic achieves classification through threshold conditions: when the absolute difference is greater than a preset difference threshold and the relative proportion is less than a preset low proportion threshold, it is judged as a vertical dominant factor, indicating that the vertical abnormal contribution is significantly dominant; when the absolute difference is greater than a preset difference threshold and the relative proportion is greater than a preset high proportion threshold, it is judged as a horizontal dominant factor, indicating that the horizontal abnormal contribution is significantly dominant; when the absolute difference is less than or equal to a preset difference threshold and the relative proportion is between the preset low proportion threshold and the preset high proportion threshold, it is judged as a multi-dominant factor, indicating a synergistic effect between vertical and horizontal abnormalities. This rule, constructed through historical transportation experiment statistics, multivariate regression model validation, and scenario adaptation optimization, achieves accurate identification and clear direction of the root cause of anomalies, providing a scientific basis for subsequent targeted adjustments to the suspension angle.

[0126] By calculating the relative proportions of horizontal and vertical contributions, as well as the absolute values ​​of the differences, and based on preset judgment rules, three types of abnormal scenarios—vertical dominance, horizontal dominance, and composite dominance—are clearly distinguished. The accuracy of the distinction is high, effectively filtering out short-term fluctuation interference and significantly reducing the misjudgment rate. This provides scientific input for proactive prevention and control strategies and significantly reduces the risk of damage to steel logistics caused by misjudgment of abnormal dominant factors.

[0127] The contribution calculation submodule includes:

[0128] The feature extraction unit is used to extract the mean of the component deflection, the variance of the longitudinal attitude angle, and the covariance of the component deflection and the longitudinal attitude angle based on the respective change sequences of the component deflection and the longitudinal attitude angle to form a longitudinal feature set; and to extract the mean of the change in the left and right axis force difference, the peak value of the change in the lateral attitude angle, and the correlation coefficient between the left and right axis force difference and the change in the lateral attitude angle based on the respective change sequences of the left and right axis force difference and the lateral attitude angle within a preset determination time period to form a lateral feature set.

[0129] The feature fusion subunit calculates the vertical contribution and the horizontal contribution based on the vertical feature set and the horizontal feature set.

[0130] The correlation coefficient between the change in the force difference between the left and right axes and the change in the lateral attitude angle is the Pearson correlation coefficient between the two change sequences. A weighted summation model is used, normalizing each feature value in the longitudinal feature set and then merging them according to the weights calibrated from historical transportation experiments. The mean and covariance of component deflection are given higher weights, while the variance of the longitudinal attitude angle is given an auxiliary weight, ensuring that the fusion result accurately represents the dominance of longitudinal factors on anomalies. The lateral contribution uses a product saturation model. After normalizing the features in the lateral feature set, the coupling effect is amplified through product operations, and then saturation processing is used to limit the result to a reasonable range, ensuring that the lateral contribution can sensitively reflect the comprehensive influence of lateral factors. All feature values ​​before fusion and the fusion result are linearly normalized and mapped to the zero-to-one interval, ensuring that the longitudinal and lateral contributions are comparable under the same dimensions, providing standardized input for subsequent determination of the dominant anomaly factors.

[0131] By constructing feature sets in different directions, the shortcomings of single-parameter attribution in existing technologies are overcome, ensuring that the contributions of different directions to anomalies are fully represented. Through differentiated models, the transformation from features to contribution is realized, completing the upgrade from coarse attribution based on a single parameter to precise quantification through multi-feature collaboration. This improves the depth and accuracy of anomaly contribution assessment, provides reliable evidence for the anomaly dominant factor determination module, provides targeted input for proactive prevention and control strategies, and significantly reduces the risk of damage caused by attribution bias.

[0132] The abnormal dominant factor determination submodule is used to determine that the abnormal dominant factor is a vertical dominant factor when the absolute value of the difference is greater than a preset difference threshold and the relative proportion is less than a preset low proportion threshold; and to determine that the abnormal dominant factor is a horizontal dominant factor when the absolute value of the difference is greater than a preset difference threshold and the relative proportion is greater than a preset high proportion threshold; and to determine that the abnormal dominant factor is a composite dominant factor when the absolute value of the difference is less than a preset difference threshold and the relative proportion is between a preset low proportion threshold and a preset high proportion threshold.

[0133] The module for determining abnormal dominant factors outputs the type of abnormal dominant factor and the contribution quantification value. The abnormal dominant factor type includes vertical dominance, horizontal dominance, and composite dominance, and the contribution quantification value includes vertical contribution degree and horizontal contribution degree.

[0134] The relative proportion is determined by the ratio of the horizontal contribution to the sum of the horizontal and vertical contributions.

[0135] The preset difference threshold is a critical value used to distinguish between single-dominant and compound-dominant anomaly types. It depends on a comprehensive analysis of historical transportation experiment statistics, multiple regression model validation, and scenario adaptability. First, by collecting a large number of steel transportation anomaly cases, the correspondence between the absolute difference of horizontal and vertical contributions (i.e., the absolute value of the difference) and the anomaly dominance type was statistically analyzed. It was found that when the absolute value of the difference is greater than 0.2, the accuracy in distinguishing between single-dominant and compound-dominant anomalies is highest. Next, a logistic regression model was fitted with the absolute value of the difference as the independent variable and the anomaly dominance type as the dependent variable. The results showed that 0.2 is the optimal critical value for distinguishing between the two types of dominant factors, with the lowest false positive rate of approximately 8%. Finally, combined with the characteristics of transportation scenarios, multibody dynamics simulation and actual testing verified that the threshold of 0.2 is stable in most scenarios such as highways and construction sites. It avoids noise residue caused by an excessively low threshold and response delay caused by an excessively high threshold, ensuring its effectiveness and adaptability under different working conditions, ultimately achieving accurate identification of anomaly dominance factors. In this embodiment, the preset difference threshold is set to 0.2, which can achieve the highest accuracy in distinguishing between single-dominant and compound-dominant anomalies and the lowest false positive rate.

[0136] The preset high percentage and preset low percentage are threshold parameters used to quantify the relative percentage, together forming the judgment boundary of the relative percentage index. The preset low percentage threshold is the upper limit of the vertical dominant factor; when the relative percentage is less than this value, the horizontal contribution percentage is low, and the vertical anomaly is dominant. The preset high percentage threshold is the lower limit of the horizontal dominant factor; when the relative percentage is greater than this value, the horizontal contribution percentage is high, and the horizontal anomaly is dominant. The interval between the two is the composite dominant factor. The setting of these values ​​depends on the dual support of historical transportation experiment statistics and multiple regression model verification. By collecting abnormal cases in steel transportation, including scenarios such as vertical deformation, lateral tilting, and composite disturbances, the relative percentage distribution of different anomaly dominant types is statistically analyzed. Furthermore, the differentiation effect of relative percentage versus anomaly dominant type is fitted using a multiple regression model to determine the preset low percentage and preset high percentage thresholds. In this embodiment, the preset low percentage threshold is set to 0.4 and the preset high percentage threshold is set to 0.6, which ensures the highest differentiation accuracy and the lowest misjudgment rate.

[0137] By using a dual-indicator logic of absolute difference and relative proportion, the system achieves accurate classification of abnormal dominant factors. This avoids the shortcomings of existing technologies, such as vague attribution and blind adjustment, and provides a scientific basis for subsequent targeted adjustment of suspension angle. Ultimately, it improves the accuracy of judgment, reduces the misjudgment rate, and supports the optimization of adjustment efficiency and structural safety.

[0138] Please continue reading. Figure 4 The diagram shown is a flowchart illustrating the adjustment of the preset suspension angle in this embodiment. It is used to adjust the preset suspension angle based on the aforementioned dominant abnormal factor.

[0139] Furthermore, the adjustment module includes:

[0140] The first adjustment submodule is used to adjust the preset suspension angle based on the longitudinal suspension angle adjustment amount when the abnormal dominant factor is the longitudinal dominant factor, wherein the longitudinal suspension angle adjustment amount is calculated based on the longitudinal contribution, the preset longitudinal contribution benchmark value and the longitudinal adjustment coefficient.

[0141] The second adjustment submodule is used to adjust the preset suspension angle based on the lateral suspension angle adjustment amount when the abnormal dominant factor is the lateral dominant factor. The lateral suspension angle adjustment amount is calculated based on the ratio of the lateral contribution to the preset lateral contribution benchmark value.

[0142] The third adjustment submodule is used to adjust the preset suspension angle based on the composite suspension angle adjustment amount when the abnormal dominant factor is a composite dominant factor. The composite suspension angle adjustment amount is calculated based on the longitudinal contribution, the lateral contribution, the longitudinal suspension angle adjustment amount, and the lateral suspension angle adjustment amount.

[0143] When the abnormal dominant factor type is longitudinal, the longitudinal contribution is used as input, and the longitudinal suspension angle adjustment is calculated through the following linear mapping model. The target longitudinal suspension angle is equal to the sum of the current longitudinal suspension angle and the longitudinal suspension angle adjustment.

[0144]

[0145] in, This refers to the longitudinal suspension angle adjustment amount. This is the vertical adjustment coefficient. As the baseline value for longitudinal contribution, The vertical contribution of the input.

[0146] When the input longitudinal contribution is greater than the longitudinal contribution benchmark value, the angle needs to be adjusted upward to alleviate deformation; otherwise, there is no need to change the suspension angle.

[0147] Both the longitudinal adjustment coefficient and the longitudinal contribution benchmark value are based on the quantitative calibration of steel transportation experiments. In this embodiment, based on the statistics of real transportation cases, the steel transportation case data is fitted by the least squares method, so that the longitudinal adjustment coefficient is set to 5° and the longitudinal contribution benchmark value is set to 0.3, which can ensure that the adjustment amount matches the contribution and accurately alleviate the anomaly.

[0148] When the dominant abnormal factor is lateral, the suspension angle adjustment is calculated using the following nonlinear mapping model. The target lateral suspension angle is equal to the sum of the current lateral suspension angle and the lateral suspension angle adjustment.

[0149]

[0150] in, This refers to the lateral suspension angle adjustment amount. This is the horizontal adjustment coefficient. As the baseline value for horizontal contribution, The horizontal contribution of the input.

[0151] Both the lateral adjustment coefficient and the lateral contribution baseline value are based on the quantitative calibration of steel transportation experiments. In this embodiment, the steel transportation case data is fitted using the least squares method, the lateral adjustment coefficient is set to 3°, and the lateral contribution baseline value is set to 0.2. This can adapt to the lateral tilt nonlinear response, accurately amplify the coupling effect of force difference and attitude change, avoid over-adjustment, and improve the efficiency of lateral anomaly mitigation.

[0152] When the dominant abnormal factor is a composite dominant factor, the lateral suspension angle adjustment and longitudinal suspension angle adjustment are calculated based on the aforementioned logic. The suspension angle adjustment is calculated using the following formula, and the target suspension angle is equal to the sum of the current suspension angle and the suspension angle adjustment.

[0153]

[0154]

[0155]

[0156] in, This is the amount of suspension angle adjustment. and These are the vertical contribution weight and the horizontal contribution weight, respectively. and These are the longitudinal suspension angle adjustment amount and the lateral suspension angle adjustment amount, respectively. and These are the vertical contribution weight and the horizontal contribution weight, respectively.

[0157] By combining three differentiated calculation models—linear mapping, nonlinear mapping, and weighted fusion—when the dominant anomaly factor is longitudinal, the linear mapping model accurately correlates the contribution and adjustment amount, adapting to the linear response characteristics of longitudinal deformation and attitude-related anomalies, ensuring a strict match between the adjustment direction and the root cause of the longitudinal anomaly. The adjustment amount is calculated using the ratio of the lateral contribution to the baseline value and the adjustment coefficient, adapting to the nonlinear coupling effect of lateral force imbalance and attitude fluctuations, such as the nonlinear characteristics of the tilt response, accurately amplifying key influencing factors. The longitudinal and lateral adjustment amounts are fused and weighted based on their contributions, quantifying the synergistic effect of longitudinal and lateral anomalies and avoiding the limitations of single-direction adjustments. This enables dynamic quantitative generation of the suspension angle adjustment amount and the target angle, improving the matching degree between the adjustment amount and the severity of the anomaly, significantly reducing the risk of structural damage due to mismatched adjustment strategies, and enhancing the stability and reliability of steel logistics transportation.

[0158] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A smart logistics management system for steel, characterized in that, include: The data acquisition module is used to collect in real time the road condition parameters, longitudinal attitude angle, lateral attitude angle, force ratio of each axle, force difference between left and right axles, and component deflection of the steel components of the transport vehicle operating based on a preset suspension angle during the transportation of steel components. The determination module is used to determine whether to enter an abnormal mode based on the threshold comparison result of the structural health index. The structural health index is determined based on the structural offset index and the duration of the warning signal. The structural offset index is determined based on the change of the component deflection and the change of the force ratio of each axis. The warning signal is generated based on the threshold comparison result of the change trend of the component deflection and the distribution stability of the force ratio of each axis. The separation module is used to determine the directional path based on the determination result of entering the abnormal mode, according to the longitudinal attitude angle, the lateral attitude angle, the force difference between the left and right axes, and the road condition parameters, wherein the directional path includes the longitudinal path, the lateral path, and the coupling path. The analysis module is used to determine the dominant abnormal factor based on the coupling path and the comparison result of the longitudinal contribution and the lateral contribution. The longitudinal contribution is determined based on the change sequence of the component deflection and the longitudinal attitude angle within a preset judgment time, and the lateral contribution is determined based on the change sequence of the left and right axis force difference and the lateral attitude angle within a preset judgment time. An adjustment module is used to adjust the preset suspension angle based on the aforementioned abnormal dominant factors.

2. The intelligent logistics management system for steel according to claim 1, characterized in that, The determination module includes: The risk warning submodule is used to generate the warning signal when the trend parameter or stability parameter exceeds its respective preset warning threshold. The trend parameter is determined based on the linear regression slope of the component deflection within a preset trend period, and the stability parameter is determined based on the information entropy of the force ratio of each axis. The warning signal includes risk type and risk level. The structural offset index calculation submodule is used to calculate the structural offset index based on the change sequence of the component deflection and the force ratio of each axis. The status persistence factor calculation submodule is used to calculate the status persistence factor based on the severity weight and the duration of the warning signal, wherein the severity weight is determined based on the risk type and the risk level; The structural health index calculation submodule is used to determine the structural health index based on the structural offset index, the state persistence factor, and the amplification factor, wherein the amplification factor is determined based on the road condition parameters. The trigger decision submodule is used to determine whether to enter an abnormal mode when the structural health index exceeds a preset structural trigger threshold.

3. The intelligent steel logistics management system according to claim 2, characterized in that, The risk warning submodule includes: The trend parameter early warning unit is used to obtain the trend parameter by calculating the linear regression slope of the component deflection within a preset trend period. When the trend parameter exceeds the preset early warning threshold of the trend parameter within a preset number of consecutive sampling periods, a trend warning is triggered. A stability warning unit is used to determine the stability parameter based on the information entropy of the force ratio of each axis, and to trigger a stability warning when the stability parameter is less than the corresponding preset warning threshold. A risk classification unit is used to classify the risk level and the risk type based on the trend parameter, the stability parameter, the preset medium risk threshold, the preset high risk threshold and the corresponding preset warning threshold; A signal generation unit is used to generate the early warning signal based on the trend parameter, the stability parameter, the risk level, and the risk type.

4. The intelligent logistics management system for steel according to claim 2, characterized in that, The structural offset index calculation submodule includes: The variation feature extraction unit is used to extract the mean, variance, and peak value as quantitative feature values ​​based on the variation sequence of the component deflection and the force ratio of each axis. The offset index generation unit is used to generate the structural offset index based on the quantized feature value through a preset calculation model.

5. The intelligent logistics management system for steel according to claim 2, characterized in that, The state duration factor calculation submodule includes: The severity weight determination unit is used to determine the corresponding severity weight based on the risk type and risk level of the warning signal through a preset risk weight rule base. The state persistence factor calculation unit is used to perform a fusion calculation based on the severity weight and the duration of the warning signal to obtain the state persistence factor.

6. The intelligent logistics management system for steel according to claim 1, characterized in that, The separation module includes: The initial probability calculation submodule is used to determine the longitudinal path probability based on the longitudinal attitude angle change rate and the road condition parameters, and to determine the lateral path probability based on the lateral attitude angle change rate and the left and right axis force difference, and to determine the coupled path probability based on the curve coupling factor and pitch and roll synchronization, wherein the curve coupling factor is determined based on the proportion of the left and right axis force difference and the road condition parameters, and the pitch and roll synchronization is determined based on the longitudinal attitude angle and the lateral attitude angle within a preset judgment time. The path allocation submodule is used to determine the directional path as the longitudinal path when the longitudinal path probability is greater than the product of the lateral path probability and the preset dominant determination coefficient, and the lateral path probability is less than the preset probability baseline; and to determine the directional path as the lateral path when the lateral path probability is greater than the product of the longitudinal path probability and the preset dominant determination coefficient, and the longitudinal path probability is less than the preset probability baseline; and to determine the coupling path based on the longitudinal path and the lateral path.

7. The intelligent steel logistics management system according to claim 1, characterized in that, The analysis module includes: The contribution calculation submodule is used to determine the longitudinal contribution based on the change sequence of the component deflection and the longitudinal attitude angle within a preset determination time, and to determine the lateral contribution based on the change sequence of the left and right axis force difference and the lateral attitude angle within a preset determination time. The abnormal dominant factor determination submodule is used to calculate the relative proportion and absolute value of the difference between the horizontal contribution and the vertical contribution, and to determine the abnormal dominant factor based on preset determination rules.

8. The intelligent logistics management system for steel according to claim 7, characterized in that, The contribution calculation submodule includes: The feature extraction unit is used to extract the mean of the component deflection, the variance of the longitudinal attitude angle, and the covariance of the component deflection and the longitudinal attitude angle based on the respective change sequences of the component deflection and the longitudinal attitude angle to form a longitudinal feature set; and to extract the mean of the change in the left and right axis force difference, the peak value of the change in the lateral attitude angle, and the correlation coefficient between the left and right axis force difference and the change in the lateral attitude angle based on the respective change sequences of the left and right axis force difference and the lateral attitude angle within a preset determination time period to form a lateral feature set. The feature fusion subunit calculates the vertical contribution and the horizontal contribution based on the vertical feature set and the horizontal feature set.

9. The intelligent logistics management system for steel according to claim 1, characterized in that, The abnormal dominant factor determination submodule is used to determine that the abnormal dominant factor is a vertical dominant factor when the absolute value of the difference is greater than a preset difference threshold and the relative proportion is less than a preset low proportion threshold; and to determine that the abnormal dominant factor is a horizontal dominant factor when the absolute value of the difference is greater than a preset difference threshold and the relative proportion is greater than a preset high proportion threshold; and to determine that the abnormal dominant factor is a composite dominant factor when the absolute value of the difference is less than a preset difference threshold and the relative proportion is between a preset low proportion threshold and a preset high proportion threshold.

10. The intelligent logistics management system for steel according to claim 9, characterized in that, The adjustment module includes: The first adjustment submodule is used to adjust the preset suspension angle based on the longitudinal suspension angle adjustment amount when the abnormal dominant factor is the longitudinal dominant factor, wherein the longitudinal suspension angle adjustment amount is calculated based on the longitudinal contribution, the preset longitudinal contribution benchmark value and the longitudinal adjustment coefficient. The second adjustment submodule is used to adjust the preset suspension angle based on the lateral suspension angle adjustment amount when the abnormal dominant factor is the lateral dominant factor. The lateral suspension angle adjustment amount is calculated based on the ratio of the lateral contribution to the preset lateral contribution benchmark value. The third adjustment submodule is used to adjust the preset suspension angle based on the composite suspension angle adjustment amount when the abnormal dominant factor is a composite dominant factor. The composite suspension angle adjustment amount is calculated based on the longitudinal contribution, the lateral contribution, the longitudinal suspension angle adjustment amount, and the lateral suspension angle adjustment amount.

Images