Variable frequency crane speed regulation control method and system

Abstract

The invention relates to the technical field of speed regulation control, and particularly discloses a speed regulation control method and system for a variable-frequency crane, which are used for performing simulation operation on a pre-constructed crane system model based on an expected speed of a crane suspension bridge and suspended object information. The simulation hoisting object data and the simulation operation data are processed by adopting a data processing technology based on deep learning, so that time sequence mode characteristics of the simulation hoisting object data and the simulation operation data are extracted, and deep understanding of the hoisting object posture change and the operation state of the crane system is realized; and the expected speed of the crane suspension bridge is intelligently adjusted based on the interactive fusion information of the two to determine the safety speed, so that the speed control of the variable-frequency crane is realized, the computing resource consumption and the time cost in the speed regulation control process of the variable-frequency crane can be effectively reduced, and the control efficiency is improved. And meanwhile, the intelligent level and the operation efficiency of speed regulation control are improved.

Description

Technical Field

[0001] This application relates to the technical field of speed control, and more specifically, to a variable-frequency crane speed control method and system. Background Art

[0002] A crane is a mechanical device that moves in a cyclic and intermittent manner. The lifting equipment picks up an item from the picking location, then horizontally moves it to the designated address and lowers the item, and then performs a reverse movement to return the lifting equipment to its original position for the next cycle. In the modern industrial and logistics fields, as a key material handling device, the safety and efficiency of the crane's operation are directly related to the smoothness of the production line and the safety of the operators.

[0003] With the continuous development of variable-frequency technology, variable-frequency cranes have gradually become the mainstream development direction in the crane field due to their excellent speed regulation performance and energy-saving effect. A variable-frequency crane changes the rotation speed of the motor by adjusting the power supply frequency, thereby achieving precise control of the speed of the crane's suspension bridge. However, how to intelligently determine the moving speed of the crane according to the actual working environment and load conditions to improve the operation efficiency while ensuring operation safety is the main challenge faced by the current variable-frequency crane speed control.

[0004] In this regard, the invention patent with the publication number CN117657961A discloses a safe and efficient variable-frequency speed control method for cranes. First, a crane system model is constructed, and the expected speed of the crane's suspension bridge and the load information are input into the crane system model for simulation operation. According to the simulated load data and simulation operation data, the safety state of the crane and the deformation amount of the load are judged. By continuously adjusting the expected speed and re-performing the simulation judgment until the overall safety state of the crane and the deformation amount of the load are within an acceptable range, the safe speed of the crane's suspension bridge is determined. Then, according to the mapping relationship between the safe speed fitted from historical data and its corresponding control speed, the speed control of the crane's suspension bridge is realized.

[0005] Although the above solution improves the intelligent level of the crane speed control to a certain extent, it still has certain limitations. For example, when determining the safe speed of the crane's suspension bridge, it is necessary to repeatedly perform simulations and approach the safe speed through multiple simulation runs, resulting in a large consumption of computing resources and an increase in time costs. Especially in complex working conditions, the number of simulations may be very frequent, making the whole process very time-consuming.

[0006] Therefore, an optimized variable-frequency crane speed control method and system are needed to solve the above technical problems. Summary of the Invention

[0007] To solve the above technical problems, the present application is proposed. Embodiments of the present application provide a variable-frequency crane speed control method and system, which perform simulation operation on a pre-constructed crane system model based on the expected speed of the crane jib and the load information, and use data processing technology based on deep learning to process the simulated load data and simulated operation data, so as to extract the time-series pattern features of the simulated load data and simulated operation data, realize in-depth understanding of the load attitude change and the operation state of the crane system, and then intelligently adjust the expected speed of the crane jib based on the interaction and fusion information of the two to determine its safe speed, thereby realizing the speed control of the variable-frequency crane, which can effectively reduce the consumption of computing resources and time cost in the variable-frequency crane speed control process, and at the same time improve the intelligent level and operation efficiency of the speed control.

[0008] According to one aspect of the present application, a variable-frequency crane speed control method is provided, which includes: Obtain the operation data and structure data of the crane, and construct a crane system model; Obtain the expected speed of the crane jib and the load information; Input the expected speed and the load information into the pre-constructed crane system model for simulation operation, and obtain the simulated load data and simulated operation data under the simulation operation; Adjust the expected speed of the crane jib based on the simulated load data and the simulated operation data to obtain the safe speed of the crane jib; Determine the control instruction of the frequency converter according to the preset mapping relationship and the safe speed; Wherein, adjusting the expected speed of the crane jib based on the simulated load data and the simulated operation data to obtain the safe speed of the crane jib includes: Perform embedding encoding on each data item in the simulated load data and each data item in the simulated operation data respectively to obtain a time queue of simulated load data embedding encoding vectors and a time queue of simulated operation data embedding encoding vectors; Extract the time-series pattern features of the time queue of simulated load data embedding encoding vectors and the time queue of simulated operation data embedding encoding vectors respectively to obtain a simulated load data time-series pattern feature vector and a simulated operation data time-series pattern feature vector; Perform compensatory interactive fusion guided by common features on the simulated load data time-series pattern feature vector and the simulated operation data time-series pattern feature vector to obtain a simulated load data-operation data time-series feature interactive compensation fusion feature vector; Based on the interactive compensation fusion feature vector of the simulation load data - operation data time series characteristics, modulate the desired speed of the crane jib to determine the safe speed of the crane jib.

[0009] Among them, extract the time series pattern features of the time queue of the simulation load data embedded coding vector and the time queue of the simulation operation data embedded coding vector respectively to obtain the simulation load data time series pattern feature vector and the simulation operation data time series pattern feature vector, including: Input the time queue of the simulation load data embedded coding vector and the time queue of the simulation operation data embedded coding vector into the sequence encoder based on the LSTM RNN hybrid model to obtain the simulation load data time series pattern feature vector and the simulation operation data time series pattern feature vector.

[0010] Among them, perform a compensation - type interactive fusion based on the common feature guidance on the simulation load data time series pattern feature vector and the simulation operation data time series pattern feature vector to obtain the simulation load data - operation data time series feature interactive compensation fusion feature vector, including: Input the simulation load data time series pattern feature vector and the simulation operation data time series pattern feature vector into the common feature extraction network to obtain the simulation load data - operation data time series common feature representation vector; Based on the simulation load data - operation data time series common feature representation vector, perform difference measurement on the simulation load data time series pattern feature vector and the simulation operation data time series pattern feature vector respectively to obtain the simulation load data time series difference feature representation vector and the simulation operation data time series difference feature representation vector; Based on the simulation load data time series difference feature representation vector and the simulation operation data time series difference feature representation vector, perform complementary reinforcement interactive fusion on the simulation load data time series pattern feature vector and the simulation operation data time series pattern feature vector to obtain the simulation load data - operation data time series feature interactive compensation fusion feature vector.

[0011] Among them, after the common feature extraction network performs dot - plus fusion on the simulation load data time series pattern feature vector and the simulation operation data time series pattern feature vector, use the neural network layer based on the tanh function to perform common feature extraction on it to obtain the simulation load data - operation data time series common feature representation vector.

[0012] Among them, taking the common feature representation vector of the simulation suspended load data - operation data time series as a benchmark, the difference metrics are respectively performed on the time series pattern feature vector of the simulation suspended load data and the time series pattern feature vector of the simulation operation data to obtain the time series difference feature representation vector of the simulation suspended load data and the time series difference feature representation vector of the simulation operation data, including: Input the time series pattern feature vector of the simulation suspended load data, the time series pattern feature vector of the simulation operation data, and the common feature representation vector of the simulation suspended load data - operation data time series into a probabilistic unit based on the Sigmoid function to obtain a probabilistic time series pattern feature vector of the simulation suspended load data, a probabilistic time series pattern feature vector of the simulation operation data, and a probabilistic common feature representation vector of the simulation suspended load data - operation data time series; Calculate the difference features of the probabilistic time series pattern feature vector of the simulation suspended load data with respect to the probabilistic common feature representation vector of the simulation suspended load data - operation data time series to obtain the time series difference feature representation vector of the simulation suspended load data; Calculate the difference features of the probabilistic time series pattern feature vector of the simulation operation data with respect to the probabilistic common feature representation vector of the simulation suspended load data - operation data time series to obtain the time series difference feature representation vector of the simulation operation data.

[0013] Among them, calculating the difference features of the probabilistic time series pattern feature vector of the simulation suspended load data with respect to the probabilistic common feature representation vector of the simulation suspended load data - operation data time series to obtain the time series difference feature representation vector of the simulation suspended load data includes: Calculate the point-division vector between the probabilistic time series pattern feature vector of the simulation suspended load data and the probabilistic common feature representation vector of the simulation suspended load data - operation data time series, and calculate the base-2 logarithm of the absolute value of each eigenvalue in the point-division vector to obtain the normalized representation vector of the time series difference features of the simulation suspended load data; Calculate the dot-product vector between the normalized representation vector of the time series difference features of the simulation suspended load data and the probabilistic time series pattern feature vector of the simulation suspended load data, and calculate the exponential function values with base e and the eigenvalues of the dot-product vector as exponents to obtain the exponentialized time series difference feature representation vector of the simulation suspended load data; Input the exponentialized time series difference feature representation vector of the simulation suspended load data into the softmax function for normalization processing to obtain the time series difference feature representation vector of the simulation suspended load data.

[0014] Among them, based on the time series difference feature representation vectors of the simulated suspended load data and the time series difference feature representation vectors of the simulated operation data, complementary reinforcement interaction fusion is performed on the time series pattern feature vectors of the simulated suspended load data and the time series pattern feature vectors of the simulated operation data to obtain the time series feature interaction compensation fusion feature vectors of the simulated suspended load data - operation data, including: Calculate the element-wise multiplication between the time series difference feature representation vectors of the simulated suspended load data and the time series pattern feature vectors of the simulated suspended load data to obtain the time series feature reinforcement compensation vectors of the simulated suspended load data; Calculate the element-wise multiplication between the time series difference feature representation vectors of the simulated operation data and the time series pattern feature vectors of the simulated operation data to obtain the time series feature reinforcement compensation vectors of the simulated operation data; Concatenate the time series feature reinforcement compensation vectors of the simulated suspended load data, the time series feature reinforcement compensation vectors of the simulated operation data, and the time series common feature representation vectors of the simulated suspended load data - operation data to obtain the time series feature interaction compensation fusion feature vectors of the simulated suspended load data - operation data.

[0015] Among them, based on the time series feature interaction compensation fusion feature vectors of the simulated suspended load data - operation data, modulating the desired speed of the crane jib to determine the safe speed of the crane jib, including: Input the time series feature interaction compensation fusion feature vectors of the simulated suspended load data - operation data into a speed control module based on a decoder to obtain a speed modulation safety factor; Multiply the speed modulation safety factor by the desired speed of the crane jib to obtain the safe speed of the crane jib.

[0016] According to another aspect of the present application, a variable frequency crane speed control system is provided, which includes: A data embedding and encoding module for respectively performing embedding encoding on each data item in the simulated suspended load data and each data item in the simulated operation data to obtain a time queue of the simulated suspended load data embedding encoding vectors and a time queue of the simulated operation data embedding encoding vectors; A time series pattern feature extraction module for respectively extracting the time series pattern features of the time queue of the simulated suspended load data embedding encoding vectors and the time queue of the simulated operation data embedding encoding vectors to obtain the time series pattern feature vectors of the simulated suspended load data and the time series pattern feature vectors of the simulated operation data; A feature interaction fusion module for performing compensation - type interaction fusion based on common feature guidance on the time series pattern feature vectors of the simulated suspended load data and the time series pattern feature vectors of the simulated operation data to obtain the time series feature interaction compensation fusion feature vectors of the simulated suspended load data - operation data; It further includes a safety speed decision module for determining the safety speed of the crane's suspension bridge based on the interaction compensation fusion feature vector of the simulation load data - operation data timing characteristics.

[0017] The present application has at least the following technical effects: Compared with the prior art, the variable frequency crane speed control method and system provided by the present application simulate the operation of a pre - constructed crane system model based on the desired speed of the crane's suspension bridge and the load information, and use data - processing techniques based on deep learning to process the simulation load data and simulation operation data to extract the timing pattern characteristics of the simulation load data and simulation operation data, so as to achieve an in - depth understanding of the load attitude change and the operation state of the crane system. Furthermore, based on the interaction fusion information of the two, the desired speed of the crane's suspension bridge is intelligently adjusted to determine its safety speed, thereby realizing the speed control of the variable frequency crane, which can effectively reduce the consumption of computing resources and time cost in the variable frequency crane speed control process, and at the same time improve the intelligent level and operation efficiency of the speed control. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By describing the embodiments of the present application in more detail in conjunction with the drawings, the above - mentioned and other objects, features, and advantages of the present application will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification, and are used together with the embodiments of the present application to explain the present application, and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0019] Figure 1 It is a flowchart of the variable frequency crane speed control method according to the embodiment of the present application.

[0020] Figure 2 It is a schematic diagram of data flow of the variable frequency crane speed control method according to the embodiment of the present application.

[0021] Figure 3 It is a flowchart of sub - step S3 of the variable frequency crane speed control method according to the embodiment of the present application.

[0022] Figure 4 It is a flowchart of sub - step S32 of the variable frequency crane speed control method according to the embodiment of the present application.

[0023] Figure 5 It is a flowchart of sub - step S4 of the variable frequency crane speed control method according to the embodiment of the present application.

[0024] Figure 6 It is a block diagram of the variable frequency crane speed control system according to the embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] As shown in this application and the claims, unless the context clearly indicates otherwise, words such as "a", "an", "one", and / or "the" are not specifically singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0026] Although this application makes various references to certain modules in the system according to the embodiments of this application, however, any number of different modules can be used and run on the user terminal and / or server. The modules are merely illustrative, and different aspects of the system and method can use different modules.

[0027] Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the operations before or below do not necessarily need to be executed precisely in sequence. On the contrary, according to the need, various steps can be processed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0028] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments of this application. It should be understood that this application is not limited by the exemplary embodiments described herein.

[0029] The crane variable frequency speed regulation control method is an advanced control technology that can dynamically adjust the motor speed according to the actual operating state and external conditions of the crane, thereby achieving a more stable, efficient, and safe operation. In the crane variable frequency speed regulation control method, constructing a crane system model based on the operating data and structural data of the crane is a key step, which can be roughly divided into several stages: data collection, data processing, model construction, and computer simulation. The specific implementation of this process will be introduced in detail below: In the data collection stage, it is necessary to start from two main aspects: one is the collection of operation data, and the other is the acquisition of structural data. For operation data, various types of sensors usually need to be installed at key parts of the crane, such as motors, hydraulic systems, hooks, etc., including but not limited to: position sensors (such as encoders) for monitoring the position changes of hooks or jibs; speed sensors (such as optical encoders) that can provide real-time feedback on the motor speed; accelerometers to help detect abnormal vibrations in the mechanical parts; torque sensors are crucial for identifying the load weight and distribution; load cells can monitor the mass and position of the lifted object to understand the crane load; torque sensors are used to measure the torque borne by the hook or jib during operation, and obtain information on the center of gravity position of each component of the crane and the lifted object and the distance from the center of gravity to the tipping edge. By monitoring various parameters such as pressure, speed, torque, position, and vibration, the operating state of the crane can be comprehensively grasped.

[0030] Structural data refers to a series of data that describes the physical structure and mechanical characteristics of the crane, mainly used to establish a mathematical model of the crane system. Specifically, it includes but is not limited to the following categories: Mechanical structure parameters: including geometric parameters such as the boom length, wheelbase, and axle distance of the crane. These parameters determine the spatial layout and movement range of the crane. Material properties: including physical properties such as the material strength and elastic modulus of each component of the crane. Material properties affect the durability and reliability of the crane. Dynamic parameters: including dynamic parameters such as the rated power, maximum speed, and torque coefficient of the motor. Dynamic parameters determine the dynamic performance of the crane. Control parameters: including control algorithm parameters of the frequency converter, PID controller parameters, etc. Control parameters affect the speed regulation accuracy and response speed of the crane.

[0031] After obtaining the above data, it is necessary to preprocess the data to ensure the quality and consistency of the data. The specific steps include: Data cleaning: removing outliers and noise to ensure the accuracy of the data. Outliers and noise may lead to error accumulation during model training, affecting the final simulation results. Data alignment: aligning data from different sources to the same time axis to ensure the time synchronization of the data. Time synchronization is crucial for analyzing the temporal pattern characteristics of the data. Data standardization: normalizing or standardizing the data so that data with different units and dimensions can be used in the same model. Standardization helps improve the training efficiency and generalization ability of the model.

[0032] Next, building a crane system model is a key step in achieving intelligent speed control. This model not only needs to accurately reflect the actual operating state of the crane but also needs to have good scalability and robustness. The specific steps are as follows: First, conduct physical modeling. Physical modeling is to establish a mathematical model of the crane system based on physical principles and mechanical theories. The specific steps include: Dynamics modeling: Using Newton's second law and Lagrange's equations, establish the dynamic equations of the crane system. These equations describe the motion states of the various components of the crane and their interactions. Dynamics modeling helps to understand the dynamic behavior of the crane under different working conditions. Static modeling: Considering the force conditions of the crane in a static state, establish static equilibrium equations. Static modeling helps to evaluate the stability of the crane under different load conditions. Rigid body modeling: Treat the various components of the crane as rigid bodies and establish rigid body motion equations. Rigid body modeling simplifies the complexity of the model and improves the calculation efficiency. Elastic modeling: Considering the elastic deformation of the various components of the crane, establish elastic equations. Elastic modeling helps to evaluate the structural integrity of the crane under high dynamic response conditions.

[0033] Next, conduct mathematical modeling. Mathematical modeling is to transform physical equations into mathematical expressions on the basis of physical modeling, in order to facilitate computer simulation and calculation. The specific steps include: Establishment of equation systems: Transform physical equations into a set of differential equations or algebraic equations. These equation systems form the mathematical basis of the crane system model. Parameterization processing: Transform physical parameters into parameters in the mathematical model to ensure the generality and flexibility of the model. Parameterization processing enables the model to adapt to different working conditions and load requirements. Model verification: Verify the model through experimental data to ensure the accuracy and reliability of the model. Model verification is an important step to ensure the credibility of simulation results.

[0034] Secondly, conduct computer simulation. Computer simulation is to use the established mathematical model to simulate the operating state of the crane through a computer program. The specific steps include: Selection of simulation software: Select appropriate simulation software, such as MATLAB / Simulink, ANSYS, etc. The selection of simulation software needs to consider the complexity of the model and the requirements of computing resources. Setting of simulation parameters: According to the actual working conditions, set simulation parameters, such as initial conditions, boundary conditions, etc. The selection of simulation parameters directly affects the accuracy of simulation results. Simulation run: Run the simulation program to obtain simulation results, including simulation load data and simulation operation data. The simulation results are the basis for subsequent data processing and intelligent speed control. Result analysis: Analyze the simulation results and extract useful information, such as the attitude change of the load and the operating state of the crane system. Result analysis helps to optimize control strategies and improve operation efficiency.

[0035] After constructing the crane system model, input the desired speed and load information into the crane system model for simulation operation to obtain simulation load data and simulation operation data under simulation operation. The following is a specific description of this embodiment: First, define the desired speed and the information of the lifted object. The desired speed refers to the target speed value that the operator hopes to achieve, which can be a fixed value. The information of the lifted object mainly includes, but is not limited to, attributes such as the mass, size, and center-of-gravity position of the load. These factors directly affect the working state and stability of the crane.

[0036] Then, import the above two sets of information as input variables into the established crane system model. This step requires the developer to have good programming skills and be able to proficiently use the API interfaces or scripting languages provided by the selected simulation platform to write code for data reading and transmission. For example, in the Simulink environment, different input conditions can be simulated by setting signal source modules; if a general programming language such as Python is used, third-party libraries (such as SciPy) need to be used to complete the corresponding functions.

[0037] Subsequently, start the simulation run. During this process, the crane system model will advance step by step according to the preset time step, and at each step, the state change at the next moment will be calculated based on the current state. For the crane, this means continuously solving a series of complex differential equations covering multiple aspects such as dynamics and kinematics. In this way, the change trends of various parameters in the entire operation process can be observed in detail, providing a basis for subsequent analysis.

[0038] Finally, collect and analyze the simulation results. After the simulation ends, the system will output a large amount of data on the movement trajectory of the lifted object, the fluctuation of the motor speed / current, etc. By statistically processing these data, the response performance of the system under different control strategies can be evaluated, such as indicators like the magnitude of the steady-state error and the percentage of overshoot.

[0039] In summary, through simulation, different control strategies can be tested in a virtual environment, such as adjusting the parameters of the PID controller, fuzzy logic control, etc., so as to find the optimal control scheme. This helps to reduce the trial-and-error cost in the actual debugging process and ensure that the selected control method can achieve the expected effect in practical applications.

[0040] Furthermore, Figure 1 is a flowchart of the variable-frequency crane speed control method according to an embodiment of the present application. Figure 2 is a schematic diagram of data flow of the variable-frequency crane speed control method according to an embodiment of the present application. As Figure 1 and Figure 2As shown, the variable-frequency crane speed control method includes the steps of: S1, respectively performing embedding encoding on each data item in the simulated load data and each data item in the simulated operation data to obtain a time queue of the simulated load data embedding encoding vectors and a time queue of the simulated operation data embedding encoding vectors; S2, respectively extracting the time series pattern features of the time queue of the simulated load data embedding encoding vectors and the time queue of the simulated operation data embedding encoding vectors to obtain a simulated load data time series pattern feature vector and a simulated operation data time series pattern feature vector; S3, performing compensation-based interactive fusion guided by common features on the simulated load data time series pattern feature vector and the simulated operation data time series pattern feature vector to obtain a simulated load data-operation data time series feature interactive compensation fusion feature vector; S4, based on the simulated load data-operation data time series feature interactive compensation fusion feature vector, modulating the expected speed of the crane bridge to determine the safe speed of the crane bridge.

[0041] Among them, the simulated load data and the simulated operation data are simulation data generated by the crane system model under specific conditions. Specifically, the simulated load data includes, but is not limited to, the mass of the load, the position coordinates and attitude angles of the load in space, the speed and acceleration of the load, the vibration frequency and amplitude generated during the movement of the load, etc. The simulated operation data includes, but is not limited to, the running speed and acceleration of the crane, the center of gravity positions of each component of the crane and the load and the distance from the center of gravity to the tipping edge, the gravity moments of each component of the crane and the load, etc. Among them, the simulated load data and the simulated operation data reflect the running state of the crane system model and the attitude change of the load under specific load information and expected speed. Therefore, in the technical solution of this application, it is expected to analyze the dynamic change patterns of the load and the crane system during the movement process to evaluate the impact of the current expected speed on the safety of the crane system, and accordingly adjust the expected speed to ensure the safe operation of the crane.

[0042] In the above-mentioned variable frequency crane speed control method, the step S1, each data item in the simulation hanging object data and each data item in the simulation operation data are respectively embedded and coded to obtain the time queue of the simulation hanging object data embedded coding vector and the time queue of the simulation operation data embedded coding vector. Wherein, in the technical solution of the present application, each data item in the simulation hanging object data and the simulation operation data represents a plurality of related parameter measurement values ​​obtained at the same time node, considering that various types of related parameters have different data units and dimensions, making it difficult for the machine learning model to directly process data. To this end, the present application uses embedded coding technology to process each data item in the simulation hanging object data and each data item in the simulation operation data respectively, so as to unify the data format, compress the multidimensional data in each data item into a vector representation of a low-dimensional space, and obtain the time queue of the simulation hanging object data embedded coding vector and the time queue of the simulation operation data embedded coding vector, so as to facilitate subsequent data analysis. In a specific example of the present application, the Word2Vec model is used to train and generate the hanging object data embedding coding matrix and the operation data embedding coding matrix, so as to realize the embedded coding of the simulation hanging object data and the simulation operation data.

[0043] In the above-mentioned variable frequency crane speed control method, the step S2, respectively extracts the time sequence pattern characteristics of the time queue of the simulated hanging object data embedded in the coding vector and the time queue of the simulated operation data embedded in the coding vector to obtain the simulated hanging object data time sequence pattern feature vector and the simulated operation data time sequence pattern feature vector. In a specific example of the present application, the step S2 includes: inputting the time queue of the simulated hanging object data embedded in the coding vector and the time queue of the simulated operation data embedded in the coding vector into the sequence encoder based on the LSTM RNN hybrid model to obtain the simulated hanging object data time sequence pattern feature vector and the simulated operation data time sequence pattern feature vector. Among them, considering that during the operation of the crane system, the operating state of the hanging object and each component of the crane changes continuously over time, therefore, in order to fully understand the posture change of the hanging object and the operating state change of the crane system, the present application further uses the time series analysis method to analyze the time queue of the simulated hanging object data embedded in the coding vector and the time queue of the simulated operation data embedded in the coding vector, so as to capture the time sequence trend change of the data, and dig out the key information such as the position change of the hanging object in space, the posture adjustment, the gravity moment change of the crane system, and the overturning risk. In a specific example of the present application, the time queue of the simulated hanging object data embedded in the coding vector and the time queue of the simulated operation data embedded in the coding vector are input into a sequence encoder based on the LSTM RNN hybrid model to obtain the simulated hanging object data timing pattern feature vector and the simulation operation data timing pattern feature vector.

[0044] Specifically, the LSTM-RNN (Long Short-Term Memory - Recurrent Neural Network) hybrid model is a deep learning architecture that combines the advantages of LSTM (Long Short-Term Memory) and RNN (Recurrent Neural Network), and is particularly suitable for processing sequential data, such as in tasks like time series analysis.

[0045] At the core of LSTM is a chain called the "cell state" that runs through the entire network. Information can be passed along this chain simply without fundamental changes. LSTM controls the flow of information by introducing the concept of gates: input gate, forget gate, and output gate. The input gate determines which new information will be stored in the cell state. The forget gate determines which old information will be discarded from the cell state. The output gate generates an output based on the current cell state and determines which parts will affect subsequent nodes. Additionally, LSTM typically uses tanh as the activation function to ensure numerical stability and the ability of nonlinear transformation.

[0046] RNN remembers previous information through the state of the hidden layer, which enables it to process sequential data of variable lengths. However, traditional RNNs have problems of vanishing or exploding gradients, especially performing poorly in the case of long-term dependencies.

[0047] The effect of the LSTM-RNN hybrid model: LSTM solves the problem that traditional RNNs have difficulty effectively capturing long-term dependencies through its unique structure. In many tasks, especially when considering the relationships between data over a long time span, LSTM performs better than the standard RNN. Therefore, in the technical solution of this application, the combined use of the LSTM model and the RNN model helps to capture both the long-term dependencies and short-term dynamic changes of the simulation load data and the simulation operation data in the time dimension, thus more accurately understanding the behavioral changes of the crane system during operation.

[0048] In the above variable-frequency crane speed control method, in step S3, a compensation-based interactive fusion guided by common features is performed on the time-series pattern feature vector of the simulated load data and the time-series pattern feature vector of the simulated operation data to obtain a time-series feature interactive compensation fusion feature vector of the simulated load data-operation data. Among them, considering that the simulated load data and the simulated operation data respectively describe different aspects of the load and the crane, and there is mutual influence and dependence between the two. For example, the swing of the load may be affected by the speed and acceleration of the crane, and the stability of the crane system may also be affected by the weight and position of the load. Therefore, in order to achieve a comprehensive understanding of the overall operating safety state of the crane system, the present application further performs information fusion on the time-series pattern feature vector of the simulated load data and the time-series pattern feature vector of the simulated operation data, so as to comprehensively utilize the complementary information of the simulated load data and the simulated operation data and more accurately evaluate the safety performance of the crane under specific operating conditions. Among them, Figure 3 is a flowchart of sub-step S3 of the variable-frequency crane speed control method according to an embodiment of the present application. As Figure 3 shown, step S3 includes the steps of: S31, inputting the time-series pattern feature vector of the simulated load data and the time-series pattern feature vector of the simulated operation data into a common feature extraction network to obtain a time-series common feature representation vector of the simulated load data-operation data; S32, taking the time-series common feature representation vector of the simulated load data-operation data as a reference, respectively performing difference measurement on the time-series pattern feature vector of the simulated load data and the time-series pattern feature vector of the simulated operation data to obtain a time-series difference feature representation vector of the simulated load data and a time-series difference feature representation vector of the simulated operation data; S33, based on the time-series difference feature representation vector of the simulated load data and the time-series difference feature representation vector of the simulated operation data, performing complementary reinforcement interactive fusion on the time-series pattern feature vector of the simulated load data and the time-series pattern feature vector of the simulated operation data to obtain the time-series feature interactive compensation fusion feature vector of the simulated load data-operation data.

[0049] Specifically, step S31 further includes: after the common feature extraction network performs dot-plus fusion on the time-series pattern feature vector of the simulated load data and the time-series pattern feature vector of the simulated operation data, using a neural network layer based on the tanh function to perform common feature extraction on it to obtain the time-series common feature representation vector of the simulated load data-operation data, which is represented by the formula: Among them, represents the time-series pattern feature vector of the simulated load data, represents the time-series pattern feature vector of the simulated operation data, Denote the time - series common - feature representation vector of the simulation load data - operation data, and respectively denote the weight - parameter matrix and the bias term, denotes the hyperbolic tangent function.

[0050] Specifically, by constructing a common - feature extraction network to process the time - series pattern feature vector of the simulation load data and the time - series pattern feature vector of the simulation operation data, the time - series common pattern between the two is learned, the common information of the two is extracted, and the time - series common - feature representation vector of the simulation load data - operation data is generated.

[0051] Figure 4 is a flowchart of sub - step S32 of the variable - frequency crane speed - regulation control method according to an embodiment of the present application. As Figure 4 shown, the step S32 includes steps: S321, inputting the time - series pattern feature vector of the simulation load data, the time - series pattern feature vector of the simulation operation data, and the time - series common - feature representation vector of the simulation load data - operation data into a probability unit based on the Sigmoid function to obtain a probability - based simulation load data time - series pattern feature vector, a probability - based simulation operation data time - series pattern feature vector, and a probability - based simulation load data - operation data time - series common - feature representation vector; S322, calculating the difference feature of the probability - based simulation load data time - series pattern feature vector relative to the probability - based simulation load data - operation data time - series common - feature representation vector to obtain the simulation load data time - series difference - feature representation vector; S323, calculating the difference feature of the probability - based simulation operation data time - series pattern feature vector relative to the probability - based simulation load data - operation data time - series common - feature representation vector to obtain the simulation operation data time - series difference - feature representation vector.

[0052] More specifically, the step S321 is expressed by the formula: where, denotes the sigmoid function, denotes the probability - based simulation load data time - series pattern feature vector, denotes the probability - based simulation operation data time - series pattern feature vector, denotes the probability - based simulation load data - operation data time - series common - feature representation vector.

[0053] That is, the sigmoid function is used to perform probabilistic processing on the time-series pattern feature vector of the simulated suspended load data, the time-series pattern feature vector of the simulated operation data, and the generated time-series common feature representation vector of the simulated suspended load data - operation data, so as to convert the feature value range to between 0 and 1, thus facilitating subsequent data processing.

[0054] More specifically, step S322 further includes: calculating the dot-division vector between the probabilistic time-series pattern feature vector of the simulated suspended load data and the probabilistic time-series common feature representation vector of the simulated suspended load data - operation data, and calculating the base-2 logarithm of the absolute value of each eigenvalue in the dot-division vector to obtain the standardized representation vector of the time-series difference features of the simulated suspended load data; calculating the dot-multiplication vector between the standardized representation vector of the time-series difference features of the simulated suspended load data and the probabilistic time-series pattern feature vector of the simulated suspended load data, and calculating the exponential function value with e as the base and each eigenvalue of the dot-multiplication vector as the exponent to obtain the exponential representation vector of the time-series difference features of the simulated suspended load data; inputting the exponential representation vector of the time-series difference features of the simulated suspended load data into the softmax function for normalization processing to obtain the representation vector of the time-series difference features of the simulated suspended load data.

[0055] Correspondingly, step S323 also adopts the above processing flow to calculate the difference features of the probabilistic time-series pattern feature vector of the simulated operation data relative to the probabilistic time-series common feature representation vector of the simulated suspended load data - operation data to obtain the representation vector of the time-series difference features of the simulated operation data, which is expressed by the formula: where 、 and respectively represent the -th eigenvalue of the probabilistic time-series pattern feature vector of the simulated suspended load data, the -th eigenvalue of the probabilistic time-series pattern feature vector of the simulated operation data, and the -th eigenvalue of the probabilistic time-series common feature representation vector of the simulated suspended load data - operation data, represents the logarithm function with base 2, represents the natural exponential function, represents the normalization exponential function, and respectively represent the representation vector of the time-series difference features of the simulated suspended load data and the representation vector of the time-series difference features of the simulated operation data.

[0056] That is, by respectively measuring the feature differences between the probability-based simulation hoisting data time-series pattern feature vectors and the probability-based simulation operation data time-series pattern feature vectors and the probability-based simulation hoisting data-operation data time-series common feature representation vectors, the unique feature information of the simulation hoisting data and the simulation operation data outside the common features is mined, so as to generate the simulation hoisting data time-series difference feature representation vector and the simulation operation data time-series difference feature representation vector.

[0057] Specifically, step S33 further includes: calculating the point-by-point multiplication between the simulation hoisting data time-series difference feature representation vector and the simulation hoisting data time-series pattern feature vector to obtain the simulation hoisting data time-series feature enhancement compensation vector; calculating the point-by-point multiplication between the simulation operation data time-series difference feature representation vector and the simulation operation data time-series pattern feature vector to obtain the simulation operation data time-series feature enhancement compensation vector; cascading the simulation hoisting data time-series feature enhancement compensation vector, the simulation operation data time-series feature enhancement compensation vector and the simulation hoisting data-operation data time-series common feature representation vector to obtain the simulation hoisting data-operation data time-series feature interaction compensation fusion feature vector, which is expressed by the formula: Wherein, represents point multiplication, and respectively represent the simulation hoisting data time-series feature enhancement compensation vector and the simulation operation data time-series feature enhancement compensation vector, represents the cascading operation, represents the simulation hoisting data-operation data time-series feature interaction compensation fusion feature vector.

[0058] That is, taking the simulation hoisting data time-series difference feature representation vector and the simulation operation data time-series difference feature representation vector as weights, the original simulation hoisting data time-series pattern feature vector and the simulation operation data time-series pattern feature vector are compensated and enhanced by point-by-point multiplication to ensure that as much unique information of the two is retained as possible during the fusion process. Then, the compensated and enhanced simulation hoisting data time-series pattern feature vector, the simulation operation data time-series pattern feature vector and the simulation hoisting data-operation data time-series common feature representation vector are fused to comprehensively consider the common information and their respective unique information between the two, and generate the simulation hoisting data-operation data time-series feature interaction compensation fusion feature vector, thereby effectively improving the quality of feature interaction fusion and realizing the accurate description of the overall operating safety state of the crane system.

[0059] In the above variable-frequency crane speed control method, in step S4, based on the simulation load data - operating data time-series feature interaction compensation fusion feature vector, the desired speed of the crane's jib is modulated to determine the safe speed of the crane's jib. Among them, Figure 5 is a flowchart of sub-step S4 of the variable-frequency crane speed control method according to an embodiment of the present application. As Figure 5 shown, step S4 includes steps: S41, inputting the simulation load data - operating data time-series feature interaction compensation fusion feature vector into a speed control module based on a decoder to obtain a speed modulation safety factor; S42, multiplying the speed modulation safety factor by the desired speed of the crane's jib to obtain the safe speed of the crane's jib.

[0060] Specifically, in step S41, the simulation load data - operating data time-series feature interaction compensation fusion feature vector is input into a speed control module based on a decoder to obtain a speed modulation safety factor. Specifically, through multi-layer non-linear transformation, the decoder gradually extracts the key information in the simulation load data - operating data time-series feature interaction compensation fusion feature vector, learns the operating state characteristics and potential safety risks of the crane system under the current working conditions, and generates a speed modulation safety factor for the crane's jib accordingly, so as to represent the distance between the desired speed and the safe operating boundary of the crane.

[0061] The decoder is a neural network model designed to recover valuable information from the simulation load data - operating data time-series feature interaction compensation fusion feature vector and generate a speed modulation safety factor. Specifically, the speed modulation safety factor actually reflects the safety level of the crane under the current working state and is a key parameter determining whether the crane can operate at a higher or lower speed. In practical applications, the decoder contains multiple neural network layers and uses the method of multi-layer non-linear transformation to conduct in-depth analysis layer by layer on the input simulation load data - operating data time-series feature interaction compensation fusion feature vector, gradually refining the key operating modes and rules of the crane. As the number of training times increases, the decoder can gradually learn how to identify the potential safety risks of the crane system from the given input data and generate the corresponding speed modulation safety factor to guide the subsequent speed adjustment strategy. Simply put, the smaller the speed modulation safety factor, the lower the allowable safe speed under the current conditions; vice versa.

[0062] Specifically, in step S42, multiply the speed modulation safety factor by the expected speed of the crane jib to obtain the safe speed of the crane jib. That is, using the speed modulation safety factor as a weight, the expected speed of the crane jib is weighted and adjusted to ensure that the operating speed of the crane is within the safety boundary, thereby minimizing the safety risk while ensuring the operating efficiency of the crane.

[0063] Particularly, when the simulated load data time-series pattern feature vector and the simulated operation data time-series pattern feature vector respectively represent the multi-scale pattern features of the simulated load data time-series and the multi-scale pattern features of the simulated operation data time-series, during the feature-intercompensation interaction between them, the simulated load data-operation data time-series feature interaction compensation fusion feature vector will also exhibit diversity in the cross-modal interaction compensation distribution of the feature-intercompensation interaction due to the difference in the source time-domain distribution of the simulated load data and the simulated operation data. Thus, when the simulated load data-operation data time-series feature interaction compensation fusion feature vector is decoded and regressed through a decoder-based speed control module, it will affect the accuracy of the decoding result.

[0064] Preferably, during the process of obtaining the speed modulation safety factor by passing the simulated load data-operation data time-series feature interaction compensation fusion feature vector through a decoder-based speed control module, feature optimization is performed on the simulated load data-operation data time-series feature interaction compensation fusion feature vector. The specific process includes: Using a non-linear mapping function to perform probability distribution processing on the feature information at each position of the simulated load data-operation data time-series feature interaction compensation fusion feature vector to obtain the simulated load data-operation data time-series feature interaction compensation fusion coding probability characterization vector, denoted as: where, represents the simulated load data-operation data time-series feature interaction compensation fusion feature vector, represents the S-shaped activation function, represents the simulated load data-operation data time-series feature interaction compensation fusion coding probability characterization vector; Using the statistical central quantity and dispersion index of the simulated load data-operation data time-series feature interaction compensation fusion coding probability characterization vector itself, respectively perform feature manifold boundary constraints on it to obtain the first simulated load data-operation data time-series feature interaction compensation fusion coding probability constraint boundary vector and the second simulated load data-operation data time-series feature interaction compensation fusion coding probability constraint boundary vector, denoted as: Among them, represents the statistical central quantity of the simulation hoisting object data - the time - series feature interaction compensation fusion coding probability characterization vector of the operation data itself, represents the dispersion index of the simulation hoisting object data - the time - series feature interaction compensation fusion coding probability characterization vector of the operation data itself, represents dot - multiplication by position, represents a preset hyper - parameter, represents the first simulation hoisting object data - the time - series feature interaction compensation fusion coding probability constraint boundary vector of the operation data, represents the second simulation hoisting object data - the time - series feature interaction compensation fusion coding probability constraint boundary vector of the operation data; Based on the first simulation hoisting object data - the time - series feature interaction compensation fusion coding probability constraint boundary vector and the second simulation hoisting object data - the time - series feature interaction compensation fusion coding probability constraint boundary vector of the operation data, construct an information propagation chain and apply the information propagation chain to the to obtain the simulation hoisting object data - the time - series feature interaction compensation fusion coding probability characterization vector to obtain the simulation hoisting object data - the time - series feature interaction compensation fusion coding information propagation vector, denoted as: Among them, represents subtraction by position, represents the natural exponential function, represents the simulation hoisting object data - the time - series feature interaction compensation fusion coding information propagation vector; Based on the first simulation hoisting object data - the time - series feature interaction compensation fusion coding probability constraint boundary vector and the second simulation hoisting object data - the time - series feature interaction compensation fusion coding probability constraint boundary vector of the operation data, as well as the statistical central quantity and dispersion index of the simulation hoisting object data - the time - series feature interaction compensation fusion coding probability characterization vector of the operation data itself, construct the simulation hoisting object data - the time - series feature interaction compensation fusion coding cross - granularity balanced collaborative vector, denoted as: Among them, represents addition by position, represents the simulation hoisting object data - the time - series feature interaction compensation fusion coding cross - granularity balanced collaborative vector; Perform element - wise stacking on the simulation hoisting object data - the time - series feature interaction compensation fusion coding information propagation vector and the simulation hoisting object data - the time - series feature interaction compensation fusion coding cross - granularity balanced collaborative vector to obtain the optimized simulation hoisting object data - the time - series feature interaction compensation fusion feature vector, denoted as: Among them, represents the optimized simulation load data - time series feature interaction compensation fusion feature vector of operation data.

[0065] Correspondingly, in this preferred embodiment, the two-way balance metric of the composite probability constraint boundary is used as the interaction constraint benchmark to dynamically optimize the population feature information transfer mechanism of the simulation load data - time series feature interaction compensation fusion feature vector of operation data. At the same time, the hierarchical balance correction coefficient driven by global statistical feedback is fused to establish a robust distribution consistency expression framework under the cross-granularity fair cooperation paradigm of the simulation load data - time series feature interaction compensation fusion feature vector, thereby constructing a fair cooperation system with adaptive regulation characteristics, precisely coordinating the multi-dimensional interaction mapping of the simulation load data - time series feature interaction compensation fusion feature vector in the feature space, and thus improving the accuracy of generating the speed modulation safety factor by the speed control module based on the decoder.

[0066] In summary, the variable-frequency crane speed control method based on the embodiments of the present application is elucidated. It performs simulation operation on a pre-constructed crane system model based on the expected speed of the crane bridge and load information, and uses data processing techniques based on deep learning to process the simulation load data and simulation operation data to extract the time series pattern features of the simulation load data and simulation operation data, realizing an in-depth understanding of the load attitude change and the operating state of the crane system. Furthermore, based on the interaction fusion information of the two, the expected speed of the crane bridge is intelligently adjusted to determine its safe speed, thereby achieving the speed control of the variable-frequency crane. In this way, the computational resource consumption and time cost in the variable-frequency crane speed control process can be effectively reduced, and at the same time, the intelligent level and operation efficiency of the speed control can be improved.

[0067] Furthermore, a variable-frequency crane speed control system is also provided.

[0068] Figure 6 is a block diagram of the variable-frequency crane speed control system according to the embodiments of the present application. As Figure 6As shown in the figure, the variable-frequency crane speed control system 100 according to an embodiment of the present application includes: a data embedding and encoding module 110, configured to perform embedding encoding on each data item in the simulated hoisting object data and each data item in the simulated operation data respectively to obtain a time queue of the simulated hoisting object data embedding encoding vectors and a time queue of the simulated operation data embedding encoding vectors; a time-series pattern feature extraction module 120, configured to extract the time-series pattern features of the time queue of the simulated hoisting object data embedding encoding vectors and the time queue of the simulated operation data embedding encoding vectors respectively to obtain a simulated hoisting object data time-series pattern feature vector and a simulated operation data time-series pattern feature vector; a feature interaction and fusion module 130, configured to perform a compensation-based interactive fusion guided by common features on the simulated hoisting object data time-series pattern feature vector and the simulated operation data time-series pattern feature vector to obtain a simulated hoisting object data-operation data time-series feature interactive compensation fusion feature vector; a safe speed decision module 140, configured to determine the safe speed of the crane bridge based on the simulated hoisting object data-operation data time-series feature interactive compensation fusion feature vector.

[0069] The specific operations of the various modules in the above variable-frequency crane speed control system have been described in detail above with reference to Figures 1 to 5 the description of the variable-frequency crane speed control method, and therefore, the repeated description thereof will be omitted.

[0070] The basic principles of the present invention have been described above in combination with specific embodiments. However, it should be noted that the advantages, advantages, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, advantages, effects, etc. are essential for each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and easy understanding, rather than limitations, and the above details do not limit the present invention to necessarily adopt the above specific details to implement.

[0071] In the above embodiments, the descriptions of the various embodiments have their own emphases. For the parts not detailed or recorded in a certain embodiment, reference may be made to the relevant descriptions of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiments described above are only illustrative. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0072] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.

[0073] In addition, it is obvious that the word "comprising" does not exclude other elements or steps, and the singular does not exclude the plural. The multiple elements stated in the system claims can also be implemented by one element through software or hardware.

[0074] Finally, it should be noted that the above description has been given for purposes of illustration and description. In addition, the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A speed control method for a variable-frequency crane, comprising: Obtain the operating data and structural data of the crane to construct a crane system model; obtain the desired speed of the crane's jib and the information of the lifted load. Input the desired speed and the lifted load information into the pre-constructed crane system model for simulation operation, and obtain the simulated lifted load data and simulation operation data under the simulation operation; adjust the desired speed of the crane's jib based on the simulated lifted load data and the simulation operation data to obtain the safe speed of the crane's jib. Determine the control instruction of the frequency converter according to the preset mapping relationship and the safe speed. It is characterized in that adjusting the desired speed of the crane's jib based on the simulated lifted load data and the simulation operation data to obtain the safe speed of the crane's jib includes: Perform embedding encoding on each data item in the simulated lifted load data and each data item in the simulation operation data respectively to obtain a time queue of simulated lifted load data embedding encoding vectors and a time queue of simulation operation data embedding encoding vectors. Extract the temporal pattern features of the time queue of simulated lifted load data embedding encoding vectors and the time queue of simulation operation data embedding encoding vectors respectively to obtain a simulated lifted load data temporal pattern feature vector and a simulation operation data temporal pattern feature vector. Perform a compensatory interactive fusion guided by common features on the simulated lifted load data temporal pattern feature vector and the simulation operation data temporal pattern feature vector to obtain a simulated lifted load data - operation data temporal feature interactive compensatory fusion feature vector. Based on the simulated lifted load data - operation data temporal feature interactive compensatory fusion feature vector, modulate the desired speed of the crane's jib to determine the safe speed of the crane's jib.

2. The variable-frequency crane speed control method according to claim 1, characterized in that Extracting the temporal pattern features of the time queue of simulated lifted load data embedding encoding vectors and the time queue of simulation operation data embedding encoding vectors respectively to obtain a simulated lifted load data temporal pattern feature vector and a simulation operation data temporal pattern feature vector includes: Input the time queue of simulated lifted load data embedding encoding vectors and the time queue of simulation operation data embedding encoding vectors into a sequence encoder based on an LSTM RNN hybrid model to obtain the simulated lifted load data temporal pattern feature vector and the simulation operation data temporal pattern feature vector.

3. The variable-frequency crane speed control method according to claim 2, wherein, Performing a compensatory interactive fusion guided by common features on the simulated lifted load data temporal pattern feature vector and the simulation operation data temporal pattern feature vector to obtain a simulated lifted load data - operation data temporal feature interactive compensatory fusion feature vector includes: Input the simulated lifted load data temporal pattern feature vector and the simulation operation data temporal pattern feature vector into a common feature extraction network to obtain a simulated lifted load data - operation data temporal common feature representation vector. Taking the simulated lifted load data - operation data temporal common feature representation vector as a reference, perform difference measurement on the simulated lifted load data temporal pattern feature vector and the simulation operation data temporal pattern feature vector respectively to obtain a simulated lifted load data temporal difference feature representation vector and a simulation operation data temporal difference feature representation vector. Based on the time - series difference feature representation vectors of the simulated hoisted object data and the time - series difference feature representation vectors of the simulated operation data, the time - series pattern feature vectors of the simulated hoisted object data and the time - series pattern feature vectors of the simulated operation data are subjected to complementary reinforcement interactive fusion to obtain the time - series feature interactive compensation fusion feature vectors of the simulated hoisted object data - operation data.

4. The variable-frequency crane speed control method according to claim 3, characterized in that, After the common feature extraction network performs dot - addition fusion on the time - series pattern feature vectors of the simulated hoisted object data and the time - series pattern feature vectors of the simulated operation data, a neural network layer based on the tanh function is used to extract common features therefrom to obtain the time - series common feature representation vectors of the simulated hoisted object data - operation data.

5. The variable-frequency crane speed regulation control method according to claim 4, characterized in that, Taking the time - series common feature representation vectors of the simulated hoisted object data - operation data as a reference, the time - series pattern feature vectors of the simulated hoisted object data and the time - series pattern feature vectors of the simulated operation data are respectively subjected to difference measurement to obtain the time - series difference feature representation vectors of the simulated hoisted object data and the time - series difference feature representation vectors of the simulated operation data, including: Inputting the time - series pattern feature vectors of the simulated hoisted object data, the time - series pattern feature vectors of the simulated operation data, and the time - series common feature representation vectors of the simulated hoisted object data - operation data into a probability unit based on the Sigmoid function to obtain the probability - based time - series pattern feature vectors of the simulated hoisted object data, the probability - based time - series pattern feature vectors of the simulated operation data, and the probability - based time - series common feature representation vectors of the simulated hoisted object data - operation data; Calculating the difference features of the probability - based time - series pattern feature vectors of the simulated hoisted object data with respect to the probability - based time - series common feature representation vectors of the simulated hoisted object data - operation data to obtain the time - series difference feature representation vectors of the simulated hoisted object data; Calculating the difference features of the probability - based time - series pattern feature vectors of the simulated operation data with respect to the probability - based time - series common feature representation vectors of the simulated hoisted object data - operation data to obtain the time - series difference feature representation vectors of the simulated operation data.

6. The variable-frequency crane speed control method according to claim 5, wherein Calculating the difference features of the probability - based time - series pattern feature vectors of the simulated hoisted object data with respect to the probability - based time - series common feature representation vectors of the simulated hoisted object data - operation data to obtain the time - series difference feature representation vectors of the simulated hoisted object data, including: Calculating the dot - division vector between the probability - based time - series pattern feature vectors of the simulated hoisted object data and the probability - based time - series common feature representation vectors of the simulated hoisted object data - operation data, and calculating the base - 2 logarithm of the absolute value of each eigenvalue in the dot - division vector to obtain the normalized representation vectors of the time - series difference features of the simulated hoisted object data; Calculating the dot - multiplication vector between the normalized representation vectors of the time - series difference features of the simulated hoisted object data and the probability - based time - series pattern feature vectors of the simulated hoisted object data, and calculating the exponential function values with base e and the eigenvalues of the dot - multiplication vector as exponents to obtain the exponential - based time - series difference feature representation vectors of the simulated hoisted object data; Inputting the exponential - based time - series difference feature representation vectors of the simulated hoisted object data into the softmax function for normalization processing to obtain the time - series difference feature representation vectors of the simulated hoisted object data.

7. The variable-frequency crane speed control method according to claim 6, characterized in that, Based on the time - series difference feature representation vectors of the simulated suspended load data and the time - series difference feature representation vectors of the simulated operation data, perform complementary reinforcement interactive fusion on the time - series pattern feature vectors of the simulated suspended load data and the time - series pattern feature vectors of the simulated operation data to obtain the time - series feature interactive compensation fusion feature vector of the simulated suspended load data - operation data, including: Calculate the element - wise multiplication between the time - series difference feature representation vector of the simulated suspended load data and the time - series pattern feature vector of the simulated suspended load data to obtain the time - series feature reinforcement compensation vector of the simulated suspended load data; Calculate the element - wise multiplication between the time - series difference feature representation vector of the simulated operation data and the time - series pattern feature vector of the simulated operation data to obtain the time - series feature reinforcement compensation vector of the simulated operation data; Cascade the time - series feature reinforcement compensation vector of the simulated suspended load data, the time - series feature reinforcement compensation vector of the simulated operation data, and the time - series common feature representation vector of the simulated suspended load data - operation data to obtain the time - series feature interactive compensation fusion feature vector of the simulated suspended load data - operation data.

8. The variable-frequency crane speed control method according to claim 7, wherein Based on the time - series feature interactive compensation fusion feature vector of the simulated suspended load data - operation data, modulate the desired speed of the crane jib to determine the safe speed of the crane jib, including: Input the time - series feature interactive compensation fusion feature vector of the simulated suspended load data - operation data into a speed - control module based on a decoder to obtain a speed modulation safety factor; Multiply the speed modulation safety factor by the desired speed of the crane jib to obtain the safe speed of the crane jib.

9. A variable-frequency crane speed control system, characterized in that, Including: A data embedding and encoding module for respectively performing embedding encoding on each data item in the simulated suspended load data and each data item in the simulated operation data to obtain a time queue of the simulated suspended load data embedding encoding vectors and a time queue of the simulated operation data embedding encoding vectors; A time - series pattern feature extraction module for respectively extracting the time - series pattern features of the time queue of the simulated suspended load data embedding encoding vectors and the time queue of the simulated operation data embedding encoding vectors to obtain the time - series pattern feature vector of the simulated suspended load data and the time - series pattern feature vector of the simulated operation data; A feature interactive fusion module for performing compensation - type interactive fusion on the time - series pattern feature vector of the simulated suspended load data and the time - series pattern feature vector of the simulated operation data guided by common features to obtain the time - series feature interactive compensation fusion feature vector of the simulated suspended load data - operation data.

10. The variable-frequency crane speed control system according to claim 9, characterized in that, Also including: A safe speed decision - making module for determining the safe speed of the crane jib based on the time - series feature interactive compensation fusion feature vector of the simulated suspended load data - operation data.

Images