Cranes equipped with vibration fatigue failure prediction and control functions

Abstract

This invention provides a crane with vibration fatigue failure prediction and control functions, comprising: a first sample generation module for generating a first sample dataset with the lifting point and load combination as input and the fatigue life cycle count as output; a second sample generation module for generating a second sample dataset with control parameters as input and the lifting dynamic load coefficient as output; and an optimization module for performing load combination based on the output of the second prediction model and inputting it into the first prediction model to predict the remaining fatigue life of the crane, optimize the control parameters, and achieve control over the fatigue failure of the crane. This invention can predict the degree of influence of dynamic load effects on fatigue life and provide a fatigue failure control strategy. It feeds back control parameters matching the prediction results to the physical entity, guiding the adjustment, improvement, and upgrading of the physical entity, thereby delaying and controlling the fatigue failure of the crane, and has significant engineering value.

Description

Technical Field

[0001] This invention relates to the field of crane technology, specifically to a crane with vibration fatigue failure prediction and control functions. Background Technology

[0002] Cranes play a vital role in industrial manufacturing, warehousing, and logistics. The metal structure, serving as the crane's skeleton, determines its safety, performance, and economic efficiency through its fatigue life. Therefore, in-depth research into fatigue life is crucial for quickly and accurately assessing the safety of crane operation.

[0003] Vibration during the lifting process of gantry cranes reduces the stability of the lifting process, shortens the fatigue remaining life of metal structures, accelerates the fatigue failure process, and affects the inherent safety of the equipment itself. However, because the stress feedback problem of structural vibration caused by transient load impacts during complex service operations has not been systematically solved, the accuracy of predicting the vibration fatigue remaining life of cranes remains unreliable. Solving the above problems faces the following three challenges: the accuracy and complexity of crane load spectrum prediction under service conditions are contradictory; the mechanism of transient load generation that induces the structural vibration response of cranes is unclear; and the coupling between structural vibration and stress indicators during the crane's service lifting cycle is difficult.

[0004] Therefore, how to provide an accurate fatigue life prediction research scheme for the vibration fatigue failure of cranes has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] To solve the above-mentioned technical problems, the present invention provides a crane with vibration fatigue failure prediction and control functions.

[0006] The technical solution adopted in this invention is as follows:

[0007] An embodiment of the present invention proposes a crane with vibration fatigue failure prediction and control functions, comprising: a modeling module for constructing a three-dimensional virtual model based on the crane's physical model; a first sample generation module for obtaining the remaining fatigue life of the crane under different lifting points and load combinations based on the three-dimensional virtual model, and generating a first sample dataset with the lifting points and load combinations as input and the remaining fatigue life as output, wherein the load combinations include: lifting load and lifting dynamic load coefficient during lifting operation; and a second sample generation module for establishing a lifting dynamic model of the crane, solving the lifting dynamic model to obtain the lifting dynamic load coefficient under different control parameters, and generating a second sample dataset with control parameters as input and the lifting dynamic load coefficient as output, wherein the control parameters include: The system comprises: cargo mass, initial lifting speed, and wire rope stiffness; a training module, which trains the GWO-XGBoost model using a first sample dataset to generate a first prediction model, and trains the GWO-XGBoost model using a second sample dataset to generate a second prediction model; and an optimization module, which uses the trained GWO-XGBoost model, inputs different control parameters into the second prediction model, combines the loads based on the output of the second prediction model, and inputs them into the first prediction model to obtain the remaining fatigue life of the crane under different control parameters. The module then performs fatigue remaining life estimation of the crane, using the difference between the target life and the remaining fatigue life as the objective function, and minimizing the difference as the objective guideline to optimize the control parameters. Based on the found optimal control parameters, the fatigue failure of the gantry crane is controlled.

[0008] The crane with vibration fatigue failure prediction and control functions proposed in this invention also has the following additional technical features:

[0009] According to one embodiment of the present invention, the optimization module is used to optimize the control parameters using a particle swarm optimization algorithm.

[0010] According to one embodiment of the present invention, the system further includes: a virtual platform building module, which is used to: import the three-dimensional virtual model into Unity 3D software (a creation engine and development tool) in .STL (a file format of stereolithography computer-aided design software) format for configuring the crane's materials and colors, and simultaneously build a virtual scene in Unity 3D; export the built virtual scene as an HTML (hypertext markup language) file through Unity 3D's WebGL (Web Graphics Library, a 3D drawing protocol) platform, build a website using the virtual model on an IIS Web (web server) server, configure the HTML file of the virtual scene in the root directory of the built website to form a virtual space, so that users can open the virtual space by entering the URL in the webpage; embed the webpage of the virtual space into a twin system platform framework built by Visual Studio Code (a cross-platform source code editor); build a function button, data input box, and prediction result display area architecture on the twin system platform; monitor user operations and send requests through the twin system platform, call the GWO-XGBoost model according to the requests to complete the fatigue remaining life prediction of the crane, and return and display the results on the twin system platform interface.

[0011] According to an embodiment of the present invention, the first sample generation module is specifically used for: discretizing different lifting points along the length direction of the main beam according to the dimensions of the crane main beam, the trolley track, and the actual operating conditions; setting interval values ​​after the actual load combination, taking into account the influence of the lifting dynamic load coefficient, and taking a load gradient for every Δm increment; setting k input sample data corresponding to n lifting points, where Δm, n, and k are preset positive integers; importing the three-dimensional virtual model into Abaqus software to establish a global rough model of the crane gantry structure; setting classic working conditions, material properties, and constraints; analyzing the stress cloud diagram and displacement cloud diagram information of the gantry structure through finite element static simulation; determining the key parts prone to fatigue failure based on engineering practice and the most unfavorable principle; cutting out the fatigue failure parts from the global rough model and transplanting them. Boundary conditions are established, and a local fine model is constructed. Simulation analysis of this model reveals critical weld seams at key locations. Based on the lifting point and lifting load, and combined with the lifting dynamic load coefficient, a generalized load spectrum is compiled and applied to the local fine model for finite element analysis. The simulation results and the local fine model are imported into FE-SAFE software, and simulation analysis is performed using the VERITY module to obtain the equivalent structural stress at the critical weld seam. Material properties of the critical weld seam are set, and the fatigue life cycle count at the critical weld seam is analyzed based on the equivalent structural stress. Based on the fatigue life cycle count and Miner's linear cumulative damage theory, the total damage at the critical weld seam is obtained. This is then converted to obtain the remaining fatigue life in years. A first sample dataset is constructed, using the lifting point and load combination as input and the remaining fatigue life as output.

[0012] According to an embodiment of the present invention, the second sample generation module is specifically used for: equating the crane to a lifting dynamic model composed of mass / moment of inertia, stiffness, and damping based on the crane's structural characteristics, load characteristics, and service characteristics; establishing the crane's vibration differential equation based on the lifting dynamic model; calculating and solving the parameters of the vibration differential equation based on the crane's physical dimensions, cross-sectional dimensions, connection methods, and material properties, wherein the parameters include the mass, equivalent stiffness, equivalent damping, and moment of inertia of each component; establishing a gantry crane lifting dynamic simulation model using MATLAB software; adjusting the control parameters by adjusting the cargo mass, initial lifting speed, wire rope stiffness, and response time; and simulating and solving the lifting dynamic simulation model using an oscilloscope to obtain the change of wire rope tension F within a set time, combining the wire rope tension F and the lifting dynamic load coefficient. Relationship formula Obtain the lifting dynamic load coefficient under different control parameters n is the crane pulley block ratio, and m2 is the cargo mass.

[0013] According to one embodiment of the present invention, the vibration differential equation of the crane is obtained according to the following formula:

[0014]

[0015] Where M is the output torque of the reducer, F is the tension of the wire rope, R is the drum radius, i is the commutator transmission ratio, n is the pulley block ratio, J1 is the moment of inertia of the reducer motor and coupling, J2 is the moment of inertia of the commutator, drum, and coupling, θ1 is the angular displacement corresponding to J1, θ2 is the angular displacement corresponding to J2, c0 and k0 are the damping and stiffness coefficients of the coupling, respectively, c1, k1, and m1 are the damping, stiffness coefficients, and mass of the main beam, respectively, c2, k2, and m2 are the damping, stiffness coefficients, and mass of the wire rope of the pulley block, c3, k3, and m3 are the damping, stiffness coefficients, and mass of the left outrigger, c4, k4, and m4 are the damping, stiffness coefficients, and mass of the right outrigger, respectively, and x1, x2, x3, and x4 are the displacements of the main beam, cargo, left outrigger, and right outrigger, respectively. This represents the first derivative of θ1 with respect to time. Let θ1 be the second derivative with respect to time. This represents the first derivative of θ² with respect to time. Let θ² represent the second derivative of θ² with respect to time. Table x1 is the first derivative with respect to time. Let x1 be the second derivative with respect to time. This represents the first derivative of x² with respect to time. Let x² represent the second derivative of x² with respect to time. Table x3 shows the first derivative with respect to time. Let x3 be the second derivative with respect to time. Table x4 shows the first derivative with respect to time. This represents the second derivative of x4 with respect to time.

[0016] According to one embodiment of the present invention, the IGWO-XGBoost model specifically includes: using the XGBoost regression algorithm as the prediction model, and using the improved Grey Wolf Optimization Algorithm (IGWO) to optimize the hyperparameters of the model.

[0017] According to one embodiment of the present invention, the following steps are specifically adopted to train the GWO-XGBoost model: Data processing is performed on the sample dataset, and the processed dataset is divided into a training set and a test set according to a certain ratio; the parameters of the GWO-XGBoost model are set, including the gray wolf population size, the maximum number of iterations, the dimension of the gray wolf individual position vector, and the parameter range of each hyperparameter; the positions of the gray wolf population are initialized, and fitness is calculated to classify the wolf pack levels and determine the positions of α wolves, β wolves, and δ wolves, wherein the fitness is set as the mean square error (MSE) of the XGBoost model. Error (mean squared error); During the iteration process, each iteration traverses the position of each wolf, determines the fitness, and divides the wolf pack into α, β, and δ wolves, completing the wolf pack position update and outputting the optimal solution α wolf's position vector and fitness value. It compares whether the fitness meets the preset threshold or whether the number of iterations has reached the maximum value. If it meets the threshold, it outputs the model evaluation index. At the same time, after the number of iterations reaches the set threshold, a certain probability mutation is performed on the individual positions of the wolf pack to avoid getting trapped in local optima. After the iteration terminates, the obtained optimal hyperparameters are used to train the XGBoost model. After training, the model is tested using test set samples, and the model evaluation index is output.

[0018] The present invention has the following beneficial effects:

[0019] This invention establishes a hoisting dynamics model, studies the influencing factors of hoisting dynamic load effects, deduces the degree of influence of dynamic load effects on fatigue life, and provides a control strategy for fatigue failure. The control parameters that match the deduction results are fed back to the physical entity to guide the regulation, improvement and upgrading of the physical entity, thereby delaying and controlling the fatigue failure of the crane and maximizing the service life of the crane to approach the design life, which has important engineering value.

[0020] The fatigue remaining life was extrapolated using the IGWO-XGBoost model. The hyperparameters of XGBoost were optimized by leveraging the global fast search capability of IGWO, ensuring the model's prediction accuracy while controlling its complexity.

[0021] Based on the digital twin framework for the simulation and control of crane vibration fatigue failure, and supported by the simulation and control of crane vibration fatigue failure, a twin system platform for the simulation and control of crane vibration fatigue failure is developed. This platform enables real-time behavior mapping between physical and virtual entities, completes the simulation of crane fatigue life, and traces the impact of dynamic load effects, thereby achieving control of operating parameters for fatigue failure. Attached Figure Description

[0022] Figure 1This is a block diagram of a crane with vibration fatigue failure prediction and control functions according to an embodiment of the present invention;

[0023] Figure 2 This is a schematic diagram illustrating the generation principle of a first sample dataset according to an embodiment of the present invention. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] The following is a description, in conjunction with the accompanying drawings, of a crane with vibration fatigue failure prediction and control functions proposed in an embodiment of the present invention.

[0026] Figure 1 This is a block diagram of a crane with vibration fatigue failure prediction and control functions according to an embodiment of the present invention, as shown below. Figure 1 As shown, the crane with vibration fatigue failure prediction and control functions includes: modeling module 1, first sample generation module 2, second sample generation module 3, training module 4, and optimization module 5.

[0027] The system comprises the following modules: Modeling Module 1 constructs a 3D virtual model based on the crane's physical model; First Sample Generation Module 2 obtains the remaining fatigue life of the crane under different lifting points and load combinations based on the 3D virtual model, generating a first sample dataset with the lifting points and load combinations as input and the remaining fatigue life as output. The load combinations include lifting load and lifting dynamic load coefficient during lifting operation; Second Sample Generation Module 3 establishes a lifting dynamic model of the crane, solves the lifting dynamic model to obtain the lifting dynamic load coefficient under different control parameters, and generates a second sample dataset with control parameters as input and the lifting dynamic load coefficient as output. The control parameters include cargo mass, initial lifting speed, and wire rope stiffness; Training Module 4 utilizes... The first sample dataset is used to train the GWO-XGBoost model to generate the first prediction model. The second sample dataset is used to train the GWO-XGBoost model to generate the second prediction model. The optimization module 5 uses the trained GWO-XGBoost model to input different control parameters into the second prediction model. Based on the output of the second prediction model, the load combination is set and then input into the first prediction model to obtain the remaining fatigue life of the crane under different control parameters. The remaining fatigue life of the crane is then extrapolated. The difference between the target life and the remaining fatigue life is used as the objective function. With minimizing the difference as the objective, the control parameters are optimized. The optimal control parameters are used to control the fatigue failure of the crane.

[0028] Specifically, the crane in this invention can be a gantry crane, which mainly includes: a hoisting mechanism, a trolley traveling mechanism, a crane traveling mechanism, a gantry structure, a lifting device, a sensing system, an electrical control system, a communication system, and safety devices. The gantry structure consists of a main beam (with tracks), end beams, outriggers, and a lower crossbeam, serving to support and connect other mechanisms, and determining the overall service life of the machine. The trolley traveling mechanism is connected to the lower crossbeam of the gantry structure, enabling the entire machine to move longitudinally along the tracks laid on the ground. The crane traveling mechanism is placed on the main beam tracks of the gantry structure, allowing the lifting trolley to move laterally along the main beam tracks. The cooperation between the trolley and crane traveling mechanisms achieves full coverage of lifting / unloading points within the work area. The hoisting mechanism is fixed to the hoisting trolley and mainly includes: wire rope, drum, coupling, stepper motor, and stepper reducer. One end of the wire rope is fixed to the trolley frame, and the other end is connected to the movable pulley of the lifting device through winding on the drum. Driven by the motor, it achieves the lifting and lowering of the goods. Safety devices prevent the trolley and crane from derailing due to accidents, and prevent goods from colliding with the machine body after exceeding the lifting limit. The electrical control system controls the movement of all mechanisms of the entire machine. The sensing system monitors the operating status and performance characteristics of the entire gantry crane through various sensor devices, including surface strain gauges, tension sensors, lifting stroke encoders, and lidar rangefinders. The communication system serves to transmit signals between various mechanisms and systems, including but not limited to Bluetooth signal transmission during the control system's operation of each mechanism, and data stream transmission after sensor data collection.

[0029] Based on the material, geometry, shape, and location attributes of the crane and its service environment, a 1:1 model can be created using SolidWorks software to construct a three-dimensional virtual model.

[0030] A 3D virtual model of the crane is used to obtain the number of fatigue life cycles under different lifting points and load combinations. A first sample dataset is generated, taking the lifting points and load combinations as input and the number of fatigue life cycles as output. The load combinations include lifting load and lifting dynamic load coefficient during lifting operation. Simultaneously, the lifting dynamic model of the crane is solved to obtain the lifting dynamic load coefficient under different control parameters (cargo mass, initial lifting speed, and wire rope stiffness), generating a second sample dataset with the control parameters as input and the lifting dynamic load coefficient as output. The GWO-XGBoost model is trained using the first sample dataset to generate a first prediction model, and the GWO-XGBoost model is trained using the second sample dataset to generate a second prediction model. The optimization module utilizes a trained GWO-XGBoost model. Different control parameters are input into a second prediction model, which outputs the corresponding lifting dynamic load coefficient. Based on the output of the second prediction model and the lifting load, a load combination is set and input into the first prediction model. The first prediction model then outputs the number of fatigue life cycles for the crane under different control parameters. The remaining fatigue life of the crane can be obtained from the number of fatigue life cycles, allowing for the extra fatigue life to be extrapolated. The optimization module continuously adjusts the control parameters, using the difference between the target life and the remaining fatigue life as the objective function, aiming to minimize this difference. It optimizes the control parameters to control fatigue failure of the gantry crane. The target life is set based on the crane's structural design and service conditions. Therefore, by establishing a hoisting dynamics model, we can study the influencing factors of the hoisting dynamic load effect, deduce the degree of influence of the dynamic load effect on fatigue life, and provide a control strategy for fatigue failure. The control parameters that match the deduction results are fed back to the physical entity to guide the regulation, improvement and upgrading of the physical entity, so as to delay and control the fatigue failure of the crane and make the fatigue life of the crane as close as possible to the expected life, which has important engineering value.

[0031] In a specific embodiment of the present invention, the optimization module 5 can be used to optimize the control parameters using a particle swarm optimization algorithm. The specific optimization process is as follows: determine the optimized control parameters (initial lifting speed, lifting mass, wire rope stiffness) and set their respective value ranges according to actual needs and service conditions; determine the fitness as the difference between the predicted life and the target life under these parameters; initialize the particle swarm, determine the number of particles in the swarm, and randomly determine the initial position of each particle; set the number of iterations and the iteration termination condition; determine the running speed and direction of each particle based on the fitness value, perform iteration, and calculate a new fitness; complete the iteration, and output the optimal control parameters that meet the target life. Adjust the actual operation of the crane according to the obtained optimal control parameters to make the crane's fatigue life close to the expected life.

[0032] The following describes how to generate the first sample dataset in detail with reference to specific embodiments.

[0033] According to one embodiment of the present invention, the first sample generation module is specifically used to: discretize different lifting points along the length direction of the main beam according to the dimensions of the crane main beam, the trolley track, and the actual operating conditions; combine the rated lifting load and consider the influence of the lifting dynamic load coefficient to set the interval value of the actual load combination; take a load gradient for every Δm increment; and set k input sample data corresponding to n lifting points, where Δm, n, and k are preset positive integers. This completes the setting of the lifting point and load combination. The three-dimensional virtual model is imported into Abaqus software to establish a global rough model of the crane gantry structure. Classical working conditions, material properties, and constraints are set. The stress and displacement cloud diagrams of the gantry structure are analyzed through finite element static simulation. Based on engineering practice and according to the most unfavorable principle, the key parts prone to fatigue failure are determined. Fatigue failure parts are cut out from the global rough model, and boundary conditions are transferred to establish a local fine model. The dangerous welds on the key parts are obtained through simulation analysis of the local fine model. Based on the lifting point and lifting load, and combined with the lifting dynamic load coefficient, a generalized load spectrum is compiled and applied to a local fine model for finite element analysis. The simulation results and the local fine model are imported into FE-SAFE software, and simulation analysis is performed using the VERITY module to obtain the equivalent structural stress at the critical weld. The material properties of the critical weld are set, and the fatigue life cycle number at the critical weld is analyzed based on the equivalent structural stress. Based on the fatigue life cycle number and combined with Miner's linear cumulative damage theory, the total damage at the critical weld is obtained. After conversion, the remaining fatigue life in years is obtained. A first sample dataset is constructed with the lifting point and load combination as input and the remaining fatigue life as output.

[0034] Specifically, the vibration during the lifting process of a gantry crane reduces the stability of the lifting process, shortens the fatigue remaining life of the metal structure, accelerates the fatigue failure process, and affects the inherent safety of the equipment. Therefore, it is necessary to consider the impact of lifting dynamic load vibration on the fatigue remaining life, and introduce a lifting dynamic load factor, which is set by technicians based on experience. Figure 2 As shown, a 3D virtual model of the crane was constructed using SolidWorks software and then imported into Abaqus software to establish a global rough model of the gantry structure. After setting the classic working conditions, material properties, and constraints, finite element static simulation analysis was conducted to obtain stress and displacement cloud diagrams of the gantry structure. Based on engineering practice and following the most unfavorable principle, key areas prone to fatigue failure were identified. Fatigue failure areas were cut from the global rough model, and boundary conditions were transferred to establish a local fine model. Simulation analysis revealed the dangerous welds at these key areas.

[0035] Based on the lifting point and lifting load, combined with the lifting dynamic load coefficient, a generalized load spectrum is compiled and applied to a local fine model for finite element analysis. The simulation results and the fine model are then imported into the FE-SAFE software. Critical welds are defined, and simulation analysis is performed using the VERITY module to obtain the equivalent structural stress at the critical welds. The material properties of the critical welds are set, and the fatigue life cycle number at the critical welds is analyzed. After conversion, the remaining fatigue life in years is obtained. Thus, a first sample dataset is constructed with the lifting point and load combination as input and the remaining fatigue life as output.

[0036] The following describes, with reference to specific embodiments, how to analyze the fatigue life cycle number of dangerous welds based on equivalent structural stress.

[0037] As a typical welded structural component, the metal structure of a crane is susceptible to cracks at its welds. Once a crack appears, it will slowly propagate under working loads until fracture occurs. Under a constant stress amplitude σ... i Under the influence of [various factors], the fatigue life cycle number at the weld joint can be calculated using the Paris formula for fracture mechanics. Based on the ratio of weld length to plate thickness a / t, 0 [various factors]...

[0038] da / dN=C[f1(ΔK) a / t≤0.1 f2(ΔK) a / t＞0.1 (1);

[0039] Where C is a constant; ΔK represents the range of stress intensity factor variation within the stress range at the far end, f1(ΔK). a / t≤0.1 Let f2(ΔK) be the short crack propagation rate function. a / t＞0.1 This is the rate function for the propagation of long cracks.

[0040] Considering the stress concentration effect of the notch in the self-balancing part under actual stress conditions, assume f1(ΔK) a / t≤0.1 and f2(ΔK) a / t＞0.1 In accordance with the law of conservation of energy, equation (1) can be rewritten as:

[0041] da / dN=C(M kn ) n (ΔK) m (2);

[0042] In the formula, M kn is the notch stress intensity amplification factor; n and m are exponents, determined based on experimental data of typical short and long cracks.

[0043] ​Integrating equation (2) above, we obtain the fatigue life from the initiation of the weld crack to the penetration of the plate thickness t:

[0044]

[0045] In the formula, I(r) is a dimensionless function of the load bending ratio r; Δσ s This represents the range of structural stress variation.

[0046] Based on formula (3), and taking into account engineering practice, fatigue test data is used to correct the two constants C and m in the above formula, resulting in:

[0047] N=(ΔS s / C d ) -1 / h (4);

[0048] In the formula, ΔS s The range of equivalent structural stress variation can be calculated using the following formula; C d h and N are experimental parameters; N is the number of fatigue life cycles.

[0049]

[0050] In the formula, r is the bending ratio; Δσ s The range of structural stress variation; Δσ b The range of bending stress variation; Δσ m This represents the range of membrane stress variation.

[0051] The following describes how to construct the second sample dataset using specific examples:

[0052] According to one embodiment of the present invention, the second sample generation module is specifically used for: equating the crane to a lifting dynamic model composed of mass / moment of inertia, stiffness, and damping, based on the crane's structural characteristics, load characteristics, and service characteristics; establishing the crane's vibration differential equation based on the lifting dynamic model; calculating and solving the parameters of the vibration differential equation based on the crane's physical dimensions, cross-sectional dimensions, connection methods, and material properties, including the mass, equivalent stiffness, equivalent damping, and moment of inertia of each component; establishing a gantry crane lifting dynamic simulation model using MATLAB software; adjusting the control parameters by adjusting the cargo mass, initial lifting speed, wire rope stiffness, and response time, and simulating the lifting dynamic simulation model; and using an oscilloscope to acquire the change of wire rope tension F within a set time, combining the wire rope tension F and the lifting dynamic load coefficient. Relationship formula Obtain the lifting dynamic load coefficient under different control parameters n is the crane pulley block ratio, and m2 is the cargo mass.

[0053] Specifically, the structural fatigue failure mechanism caused by transient impacts during the dynamic operation of cranes remains unresolved, leading to an inability to guarantee the accuracy of structural fatigue remaining life prediction. Therefore, it is necessary to study the key factors inducing transient impacts during crane operation and to explore the influence of transient load impacts on the stress cycle indicators of crane structures through dynamic theory, which can provide a theoretical basis for the accurate prediction of crane structural fatigue remaining life. To this end, this invention establishes a hoisting dynamics model to study the influencing factors of hoisting dynamic load effects and deduce the degree of influence of the hoisting dynamic load coefficient on fatigue life.

[0054] Based on the crane's lifting power model and considering the crane's complex mechanisms and structures, a vibration differential equation can be established that reflects the crane's vibration characteristics while facilitating modeling and calculation, while retaining key factors. In a specific embodiment of this invention, the crane's vibration differential equation can be obtained using the following formula:

[0055]

[0056] Where M is the output torque of the reducer, F is the tension of the wire rope, R is the drum radius, i is the commutator transmission ratio, n is the pulley block ratio, J1 is the moment of inertia of the reducer motor and coupling, J2 is the moment of inertia of the commutator, drum, and coupling, θ1 is the angular displacement corresponding to J1, θ2 is the angular displacement corresponding to J2, c0 and k0 are the damping and stiffness coefficients of the coupling, respectively, c1, k1, and m1 are the damping, stiffness coefficients, and mass of the main beam, respectively, c2, k2, and m2 are the damping, stiffness coefficients, and mass of the wire rope of the pulley block, c3, k3, and m3 are the damping, stiffness coefficients, and mass of the left outrigger, c4, k4, and m4 are the damping, stiffness coefficients, and mass of the right outrigger, respectively, and x1, x2, x3, and x4 are the displacements of the main beam, cargo, left outrigger, and right outrigger, respectively. This represents the first derivative of θ1 with respect to time. Let θ1 be the second derivative with respect to time. This represents the first derivative of θ² with respect to time. Let θ² represent the second derivative of θ² with respect to time. Table x1 is the first derivative with respect to time. Let x1 be the second derivative with respect to time. This represents the first derivative of x² with respect to time. Let x² represent the second derivative of x² with respect to time. Table x3 shows the first derivative with respect to time. Let x3 be the second derivative with respect to time. Table x4 shows the first derivative with respect to time. This represents the second derivative of x4 with respect to time.

[0057] Based on the structural dimensions, cross-sectional dimensions, connection methods, and material properties of the physical entity, and combined with theoretical calculations, the parameters in the vibration differential equation, including the mass of each component, equivalent stiffness, equivalent damping, and moment of inertia, are solved. Using the Simulink tools and functional modules built into MATLAB software, such as the integration module, multiplication module, addition / subtraction module, constant module, and integration module, a hoisting dynamic simulation model of the crane is established. On this basis, by adjusting the cargo mass m2 and the initial hoisting speed (i.e., ... The model was simulated using the wire rope stiffness k2 and response time. An oscilloscope in Simulink was used to obtain the change of the target variable (i.e., the wire rope tension F) within a set time period, and the relationship between the wire rope tension and the lifting dynamic load coefficient was analyzed. Obtain the lifting dynamic load coefficient under different working conditions This lifting dynamic load factor This leads to the construction of a second sample dataset with control parameters as input and lifting dynamic load coefficient as output.

[0058] In one embodiment of the present invention, the IGWO-XGBoost model specifically includes: using the XGBoost regression algorithm as the prediction model, and using the improved Grey Wolf Optimization Algorithm (IGWO) to optimize the hyperparameters of the model.

[0059] Specifically, XGBoost (Extreme Gradient Boosting) is an ensemble model that improves upon Boosting by using Gradient Boosting Decision Tree (GBDT) as the foundation and Classification and Regression Tree (CART) as the base classifier. During model training, the results of multiple CART trees are integrated to compensate for the insufficient accuracy of a single CART prediction. If the total space of CART subtrees is M, the model output is as follows:

[0060]

[0061] In the formula, The prediction result for the i-th sample; m is the number of trees; x i Let f represent the i-th sample. m (x i ) represents the prediction result for the m-th round sample.

[0062] The objective function of XGBoost consists of two parts: a regularization term and a loss function, and the minimum objective function Obj can be expressed as:

[0063]

[0064] In the formula, l is the loss function; C is a constant term; Ω is a regularization term; γ and λ are regularization parameters that control the complexity of the model, and y i Ω(f) represents the true label value of the training samples. t ) represents the regularization term, T represents the number of leaf nodes, and ω represents the regularization term. j The weights are those of the leaf nodes.

[0065] Regularization term Ω, number of leaf nodes T, and leaf node weight ω j The relationship is as shown in equation (8), then the objective function Obj(t) in the t-th iteration is:

[0066]

[0067] Expanding the objective function (Equation (9)) using the second Taylor formula and removing the constant term yields:

[0068]

[0069] In the formula, g i and h i These are the first and second gradients of the loss function, respectively; v j For the sample at the j-th node; ω j This represents the node weight.

[0070] make For ω j Taking the partial derivative and setting it to 0, we obtain the optimal solution. Substituting the optimal solution into the objective function, we get:

[0071]

[0072] ω j * Minimize the node weights.

[0073] The Grey Wolf Optimizer (GWO) is a swarm intelligence optimization algorithm that simulates the hunting behavior of grey wolves. Its principle is primarily based on the social hierarchy and hunting strategies of grey wolves. The optimal solution alpha (α) simulates the position of the alpha wolf, the second-best solutions beta (β) and delta (δ) simulate the positions of second- and third-order wolves, and the remaining candidate solutions omega (ω) simulate the positions of subordinate wolves (i.e., fourth-order wolves). In the GWO algorithm, the hunt (solution) is initiated and guided by alpha, β, and δ wolves, with ω wolves surrounding the prey. The mathematical model of this behavior can be expressed as:

[0074]

[0075] In the formula, D is the distance between the individual gray wolf and its prey; t is the current iteration number; X p X represents the prey's position; X represents the current position of the gray wolf; A and C are the coordination coefficient vectors. Adjusting A and C can change the influence of the prey's position on the gray wolf's next position. To simulate the process of approaching the prey, A can be set to a random number in [-a, a], and a decreases from 2 to 0 during the iteration process; r1 and r2 are random numbers in the interval [0, 1].

[0076] As the wolf pack begins to surround its prey, the positions of α wolf, β wolf, and δ wolf also iteratively change as the prey's position shifts, and the position update formula is:

[0077]

[0078] X(t+1)=(X1+X2+X3) / 3 (16)

[0079] In the formula, X α (t), X β (t) and X δ (t) represent the position vectors of wolf α, wolf β, and wolf δ after iteration t, respectively; D α D β and D δ ω represents the distance between the current candidate gray wolf ω and the three best wolves after t iterations; X(t+1) is the position of the gray wolf after t+1 iterations.

[0080] To overcome the problem of getting trapped in local optima during iteration, a combination of the GWO optimization algorithm and a later-stage perturbation mutation strategy is used. When the number of iterations reaches a set threshold (0.7T, where T is the maximum number of iterations), mutation judgments are initiated on the wolf pack positions, causing them to undergo random mutations with a certain probability. Simultaneously, the length of the mutation position vector is controlled, and the mutated position vector is pruned to ensure that the mutation result remains within the pre-defined wolf pack position range. The mutated individual positions are obtained, and the fitness values ​​of the corresponding individuals are calculated. After all individual positions in the wolf pack have been updated, the wolf pack levels are reclassified based on fitness, and the wolf pack positions are updated and mutated again. This process is iterated and updated until prey is successfully captured, thus obtaining the optimal solution. XGBoost exhibits good performance on highly complex, nonlinear, and small-scale datasets, and reduces the risk of overfitting through pruning and column sampling. Therefore, in one embodiment of this invention, by using the XGBoost regression algorithm as the prediction model and optimizing the hyperparameters of the model using the improved gray wolf optimization algorithm IGWO, the model's prediction accuracy can be ensured while controlling its complexity.

[0081] In one embodiment of the present invention, the training module 4 is specifically used for: processing the sample dataset, dividing the processed dataset into a training set and a test set according to a certain ratio; setting the parameters of the IGWO-XGBoost model, including the number of gray wolves, the maximum number of iterations, the dimension of the gray wolf individual position vector, and the parameter range of each hyperparameter; initializing the position of the gray wolf population and calculating the fitness to classify the wolf pack into different levels and determine the positions of α wolves, β wolves, and δ wolves, wherein the fitness is set as the mean squared error of the XGBoost model. During the iteration process, each iteration traverses the position of each wolf, determines its fitness, and classifies wolves into α, β, and δ groups, updating the wolf pack's position and outputting the optimal α wolf's position vector and fitness value. The fitness is compared to whether it meets a preset threshold or the maximum number of iterations has been reached. If so, a model evaluation metric is output. Simultaneously, after the number of iterations reaches the set threshold, a certain probability mutation is performed on the individual wolf positions to avoid getting trapped in local optima. After the iteration terminates, the obtained optimal hyperparameters are used to train the XGBoost model. After training, the model is tested using a test set, and the model's evaluation metrics are output. Evaluation metrics include: fitting accuracy, root mean square percentage error, and mean absolute percentage error.

[0082] In one embodiment of the present invention, the crane further includes a virtual platform building module, which is used to: import the 3D virtual model in .STL format into Unity 3D software for configuring the crane's materials and colors, and simultaneously build a virtual scene in Unity 3D. The built virtual scene is exported as an HTML file through Unity 3D's WebGL platform, and a website is built using the IIS Web server virtual model. The HTML file of the virtual scene is configured in the root directory of the built website to form a virtual space, allowing users to open the virtual space by entering a URL in a webpage. The virtual scene is exported as an HTML file through Unity 3D's WebGL platform, and a website is built using the IIS Web server virtual model. The HTML file of the virtual scene is configured in the root directory of the built website to form a virtual space, allowing users to open the virtual space by entering a URL in a webpage. Function buttons, data input boxes, and a prediction result display area architecture are built on the twin system platform. User operations are monitored and requests are sent through the twin system platform. Based on the requests, the GWO-XGBoost model is invoked to complete the fatigue remaining life estimation of the crane, and the results are returned and displayed on the twin system platform interface.

[0083] Specifically, based on the digital twin framework for the simulation and control of crane vibration fatigue failure, and supported by the simulation and control of crane vibration fatigue failure, this project utilizes VS Code as the platform, Node.js as the runtime environment, Python and C# as the development languages, and integrates software such as Unity 3D, MySQL, and Matlab Simulink with corresponding communication protocols to build a data transmission and reception method. Focusing on system functions, development processes, and system platform, a twin system platform for the simulation and control of crane vibration fatigue failure is developed. This platform enables real-time behavior mapping between physical and virtual entities, completes the simulation of crane fatigue life, traces the impact of dynamic load effects, and thus achieves control of operating parameters for fatigue failure.

[0084] The twin system platform interface can be divided into six main areas: the overall system performance parameter area (Area 1), the virtual and physical space window area (Area 2), the overall system operation information area (Area 3), the fatigue remaining life simulation area (Area 4), the simulated fatigue failure control information area (Area 5), ​​and the function button area (Area 6). The function button area consists of seven function buttons: virtual-physical interaction start, transmission communication start, operation information monitoring, fatigue life simulation, fatigue failure control, information data management, and system exit. The different function services and implementation processes are shown below:

[0085] ① The overall performance parameter area displays the relevant design parameters of the gantry crane, thus representing some of the inherent attribute information of the equipment. ② Clicking the "Virtual-Real Interaction Start → Transmission Communication Start → Operation Information Monitoring" function buttons sequentially allows observation of the synchronous operation of the virtual and real models in Zone 2, and displays real-time service status information such as actual lifting capacity, cargo lifting position, trolley / gantry running position, and measuring point stress in Zone 3. ③ Clicking the "Fatigue Life Prediction" function button uses the collected information and some derived data (historical and real-time) to complete the prediction of historical / current / future life, displaying the prediction results such as design life, crack length, service life, and remaining life in Zone 4. ④ In Zone 5, select the parameters to be changed and enter the corresponding values. After clicking the "Fatigue Failure Control" button, the dynamic load coefficient and related life information after failure control will be displayed in Zone 4. ⑤ Clicking the "Information Data Management → System Exit" buttons sequentially allows you to disconnect all connections, close all channels and functions, and exit the system after saving and updating data information.

[0086] Therefore, taking the integrated hoisting power model of the crane structure as the center and the fatigue remaining life simulation results as the basis, and focusing on the fatigue failure control strategy, the dynamic load effect coefficient is obtained by dynamically adjusting control parameters such as cargo mass, initial hoisting speed and wire rope stiffness. Then, the fatigue life can be re-simulated on the virtual entity (twin system platform). The control parameters that match the simulation results are fed back to the physical entity to guide the regulation, improvement and upgrading of the physical entity, thus realizing virtual control of the real.

[0087] In summary, the crane with vibration fatigue failure prediction and control functions according to embodiments of the present invention establishes a lifting dynamics model, studies the influencing factors of lifting dynamic load effects, predicts the degree of influence of dynamic load effects on fatigue life, and provides a fatigue failure control strategy. Control parameters matching the prediction results are fed back to the physical entity to guide the regulation, improvement, and upgrading of the physical entity, thereby delaying and controlling the fatigue failure of the crane and maximizing the crane's service life to approach its design life, which has significant engineering value. The IGWO-XGBoost model is used for fatigue remaining life prediction, and the global fast search capability of IGWO is used to optimize the hyperparameters of XGBoost, ensuring model prediction accuracy while controlling model complexity. Based on the digital twin framework for crane vibration fatigue failure prediction and control, and supported by crane vibration fatigue failure prediction and control, a twin system platform for crane vibration fatigue failure prediction and control is developed. This platform enables real-time behavior mapping between physical and virtual entities, completes the prediction of crane fatigue life, traces the influence of dynamic load effects, and thus achieves control of operating parameters for fatigue failure.

[0088] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. "A plurality of" means two or more, unless otherwise explicitly specified.

[0089] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0090] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of the invention pertain.

[0091] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any of the following techniques known in the art, or a combination thereof: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0092] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0093] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0094] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.

[0095] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A crane with a function of deducing and controlling vibration fatigue failure, characterized in that, Comprise: A modeling module for constructing a three-dimensional virtual model according to a physical model of the crane; A first sample generation module for obtaining the fatigue residual life of the crane under different lifting points and load combinations according to the three-dimensional virtual model, generating a first sample dataset with the lifting point and load combination as input and the fatigue residual life as output, and the load combination comprising: hoisting load and hoisting dynamic load coefficient during operation; A second sample generation module for establishing a hoisting dynamic model of the crane, solving the hoisting dynamic model to obtain the hoisting dynamic load coefficient under different control parameters, generating a second sample dataset with the control parameters as input and the hoisting dynamic load coefficient as output, and the control parameters including: cargo mass, hoisting initial speed and wire rope stiffness; A training module for training the GWO-XGBoost model using the first sample dataset to generate a first prediction model, and training the GWO-XGBoost model using the second sample dataset to generate a second prediction model; An optimization module for inputting different control parameters into the second prediction model using the trained GWO-XGBoost model, inputting the load combination set according to the output of the second prediction model into the first prediction model, obtaining the fatigue residual life of the crane under different control parameters, and performing fatigue residual life deduction of the crane, taking the difference between the target life and the fatigue residual life as the objective function, and taking the minimum difference as the target orientation, optimizing the control parameters, and controlling the fatigue damage of the crane according to the optimal control parameters found.

2. The crane with vibration fatigue failure deduction and control function according to claim 1, characterized in that, The optimization module is used to optimize the control parameters by using a particle swarm optimization algorithm.

3. The crane with vibration fatigue failure deduction and control function according to claim 1, characterized in that, Also include: A virtual platform building module for: Importing the three-dimensional virtual model into the Unity 3D software in the.STL format to configure the material and color of the crane, and building a virtual scene in Unity 3D; Exporting the built virtual scene to an HTML file through the WebGL platform of Unity 3D, using an IIS Web server virtual model to build a website, configuring the HTML file of the virtual scene in the root directory of the built website to form a virtual space, so that the user can input the website address in the webpage to open the virtual space; Nesting the webpage of the virtual space into the twin system platform framework built by Visual Studio Code; Building function buttons, data input boxes and prediction result display areas on the twin system platform; Monitoring user operations and sending requests through the twin system platform, calling the GWO-XGBoost model according to the request to complete the fatigue residual life deduction of the crane and returning and displaying the results on the twin system platform interface.

4. The crane with vibration fatigue failure deduction and control function according to claim 1, characterized in that, The first sample generation module is specifically used for: According to the crane girder, trolley track size and actual operation, the main beam length direction is discretized into different lifting points, combined with the rated lifting load, considering the influence of the lifting dynamic load coefficient, the interval value of the actual load combination is set, each increment takes a load gradient, sets k input sample data corresponding to n lifting points, and Δm, n and k are preset positive integers; The three-dimensional virtual model is imported into Abaqus software to establish a global rough model of the crane gantry structure; The typical working conditions, material properties and constraint conditions are set, the stress nephogram and displacement nephogram information of the gantry structure are analyzed through finite element static simulation, the key parts prone to fatigue failure are determined according to the engineering practice and the most unfavorable principle; The key parts prone to fatigue failure are cut from the global rough model and the boundary conditions are transplanted to establish a local fine model, and the dangerous welds on the key parts are obtained through simulation analysis of the local fine model; According to the lifting point and the lifting load, combined with the lifting dynamic load coefficient, a generalized load spectrum is compiled and applied to the local fine model for finite element analysis, the simulation results and the local fine model are imported into the FE-SAFE software, and the equivalent structural stress at the dangerous weld is obtained through simulation analysis by using the VERITY module; The material properties of the dangerous weld are set, the fatigue life cycle number of the dangerous weld is analyzed according to the equivalent structural stress at the dangerous weld, the total damage of the dangerous weld is obtained according to the fatigue life cycle number and combined with the Miner linear cumulative damage theory, the fatigue residual life in years is obtained through conversion, and a first sample data set is constructed, which takes the lifting point and the load combination as input and takes the fatigue residual life as output.

5. The crane with vibration fatigue failure deduction and control function according to claim 1, characterized in that, The second sample generation module is specifically used for: According to the structural characteristics, load characteristics and service characteristics of the crane, the crane is equivalent to a lifting dynamic model composed of mass / rotational inertia, stiffness and damping; According to the lifting dynamic model, a vibration differential equation of the crane is established; According to the structural size, sectional size, connection mode and material properties of the physical entity of the crane, the parameters of the vibration differential equation are calculated and solved, the parameters include the mass, equivalent stiffness, equivalent damping and rotational inertia of each component, and a lifting dynamic simulation model of the gantry crane is established by using MATLAB software; The lifting dynamic simulation model is simulated and solved by adjusting the cargo mass, lifting initial speed, wire rope stiffness and response time to adjust the control parameters; The oscilloscope is used to obtain the change of the steel wire rope tension F in the set time, and the lifting dynamic load coefficient under different control parameters is obtained by combining the relationship formula between the steel wire rope tension F and the lifting dynamic load coefficient is the hoist pulley block ratio, is the cargo mass.​​​ 6. The crane with vibration fatigue failure deduction and control function according to claim 5, characterized in that, The vibration differential equation of the crane is obtained according to the following formula: ； wherein, is the output torque of the speed reducer, is the steel wire rope tension, is the drum radius, is the gear ratio of the reversing gear, is the block ratio of the pulley block, is the moment of inertia of the speed reducer motor and the coupling, is the moment of inertia of the reversing gear, the drum and the coupling, is is the corresponding angular displacement, is is the corresponding angular displacement, , are the damping and stiffness coefficients of the coupling, respectively, , , are the damping, stiffness coefficients and mass of the main beam, respectively, , , are the damping, stiffness coefficients and mass of the pulley block steel wire rope, respectively, , , are the damping, stiffness coefficients and mass of the left side leg, , , are the damping, stiffness coefficients and mass of the right side leg, , , , are the displacements of the main beam, the cargo, the left side leg and the right side leg, respectively, represents the first order differential with respect to time, represents the second order differential with respect to time, represents the first order differential with respect to time, represents the second order differential with respect to time, represents the first order differential with respect to time, represents the second order differential with respect to time, represents the first order differential with respect to time, represents the second order differential with respect to time, represents the first order differential with respect to time, represents the second order differential with respect to time, represents the first order differential with respect to time, representative second derivative with respect to time.

7. The crane with vibration fatigue failure deduction and control function according to claim 1, characterized in that, The GWO-XGBoost model specifically includes: The XGBoost regression algorithm is used as the prediction model, and the improved grey wolf optimization algorithm IGWO is used to optimize the hyperparameters of the model.

8. The crane with vibration fatigue failure deduction and control function according to claim 7, characterized in that, The training module is specifically used for: Data processing is performed on the sample data set, and the processed data set is divided into a training set and a test set according to a certain proportion; The parameters of the GWO-XGBoost model are set, including the number of grey wolf populations, the maximum number of iterations, the dimension of the grey wolf individual position vector, and the parameter range of each hyperparameter. Initialize the positions of the gray wolf population, and perform fitness calculation to divide the wolf population into ranks and determine wolf, wolf, the position of the wolf, wherein the fitness is set as the mean square error of the XGBoost model; In each iteration, the position of each wolf is traversed to determine the fitness and to perform wolves, wolves and wolf division, complete wolf pack position update, and output the optimal solution the position vector of the wolf and the fitness value, compare whether the fitness meets the preset threshold or the number of iterations has reached the maximum value, if yes, output the model evaluation index, and meanwhile, after the number of iterations reaches the set threshold, mutate the individual position of the wolf pack with a certain probability; After the iteration is terminated, the XGBoost model is trained using the obtained optimal hyperparameters, and after the training is completed, the model is tested using the test set samples, and the evaluation indicators of the model are output.

Images