A method and system for acquiring a real state of a portal crane structure

Abstract

The application discloses a kind of gantry crane structure real state acquisition method and system, the method includes: the working information of actual gantry crane is input to GA-FESN prediction model, obtains the stress strain value predicted by gantry crane structure;Wherein, the training method of GA-FESN prediction model includes: obtaining the working information of gantry crane and the first stress strain value of structure key position;Working information is input to the finite element model of gantry crane established in advance, obtains the stress strain simulation value of structure key position under the condition of self-weight of gantry crane;First stress strain value and stress strain simulation value are synthesized, and obtain the synthesized stress strain value;The mapping relationship between working information and synthesized stress strain value is mined using functional echo state network, and GA-FESN prediction model is obtained.The application solves the complex operation and time-consuming and labor-consuming problem of repeatedly carrying out finite element analysis when considering the influence of self-weight of gantry crane structure, and accurate prediction result can be obtained.

Description

Technical Field

[0001] This invention relates to the field of crane machinery technology, specifically to a method and system for obtaining the true structural state of a gantry crane. Background Technology

[0002] Gantry cranes are intermittently operating mechanical products, characterized by cyclical movements, frequent starting and braking, and harsh working environments. The metal structure, as the supporting framework of the crane, is a crucial foundational component, and accurate monitoring of its stress-strain changes is a prerequisite for effective health management.

[0003] Existing methods for structural health monitoring of gantry cranes mostly rely on the monitoring values ​​of various sensors to determine the stress condition of the structure. However, the sensors need to be zeroed before monitoring, so the measured values ​​can only characterize the stress and strain changes caused by the working load, while ignoring the stress and strain generated by the crane's own weight. In addition, due to the special working characteristics of gantry cranes, their slewing angle and amplitude change frequently. Repeatedly analyzing the stress and strain generated by the structure's own weight using finite element technology is extremely complex and inefficient, which is not conducive to practical engineering applications. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for obtaining the true state of a gantry crane structure, so as to solve the problem of low efficiency in the existing technology of using finite element technology to analyze the stress and strain generated by the self-weight of the structure.

[0005] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0006] In a first aspect, the present invention discloses a method for obtaining the true structural state of a gantry crane, comprising:

[0007] The actual operating information of the gantry crane is input into a pre-trained GA-FESN prediction model to obtain the predicted stress and strain values ​​of the gantry crane structure; wherein, the training method of the pre-trained GA-FESN prediction model includes:

[0008] Obtain the working information and the first stress-strain values ​​at key structural locations of the gantry crane. The working information includes lifting load, slewing angle, and luffing size.

[0009] The working information is input into the pre-established finite element model of the gantry crane to obtain the stress and strain simulation values ​​of key structural locations under the self-weight of the gantry crane.

[0010] The first stress-strain value and the simulated stress-strain value are combined to obtain the combined stress-strain value;

[0011] The mapping relationship between working information and the synthetic stress-strain values ​​is mined using a functional echo state network, and the GA-FESN prediction model is trained using the mapping relationship to obtain the pre-trained GA-FESN prediction model.

[0012] Furthermore, the simulated stress and strain values ​​at key structural locations under the self-weight of the gantry crane were obtained, including:

[0013] Based on the collected work information, the rotation angle and amplitude were set sequentially, and the working load was set to 0. The stress and strain simulation values ​​at key locations of the structure were calculated for each case.

[0014] Furthermore, the functional echo state network includes an input layer, a storage layer, and an output layer arranged sequentially; the nodes of the storage layer are processed by time aggregation operation and spatial aggregation operation respectively.

[0015] Furthermore, the mapping relationship between working information and the synthesized stress-strain values ​​is mined using a functional echo state network to obtain the GA-FESN prediction model, which includes:

[0016] The FESN prediction model is trained by mining the mapping relationship between lifting load, slewing angle, amplitude and the synthetic stress-strain value using the functional echo state network.

[0017] Using the root mean square error between the output value and the actual value of the FESN prediction model as the fitness, the hyperparameters of the FESN prediction model are optimized using the GA algorithm.

[0018] When the fitness is minimized, the hyperparameters are taken as the optimal parameters to obtain the GA-FESN prediction model.

[0019] Furthermore, the hyperparameters include storage pool sparsity, spectral radius, and output connection weights.

[0020] Furthermore, the formula for calculating the root mean square error is as follows:

[0021]

[0022] Where, n t y represents the number of test samples. i (t) represents the actual value of the i-th test sample at time t. Let be the predicted value of the i-th test sample at time t.

[0023] Furthermore, the following conditions must be met when obtaining the working information of the gantry crane and the first stress-strain values ​​at key structural locations:

[0024] The lifting load, slewing angle, and amplitude in the working information correspond one-to-one with the first stress-strain values ​​at the key locations of the structure in that state.

[0025] Furthermore, the synthesis of the first stress-strain value and the simulated stress-strain value specifically involves:

[0026] Analyze the force characteristics at key locations; the force characteristics include any one or more combinations of tension, compression, and torsion.

[0027] Based on the stress characteristics, the first stress-strain value and the simulated stress-strain value are synthesized using an appropriate synthesis method.

[0028] Secondly, the present invention discloses a system for acquiring the true state of a gantry crane structure, including a processor and a storage medium;

[0029] The storage medium is used to store instructions;

[0030] The processor is configured to operate according to the instructions to perform the steps of the method described in any of the first aspects.

[0031] Thirdly, the present invention discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, comprises the steps of the method described in any of the first aspects.

[0032] According to the above technical solution, the beneficial effects of the present invention are as follows:

[0033] This invention mines the mapping relationship between working information and synthetic stress-strain values ​​through functional echo state network (FESN), and optimizes the selection of hyperparameters in the FESN using a genetic algorithm, effectively improving the performance of the FESN model in nonlinear multivariate time series prediction. It allows direct input of various sensor monitoring values ​​into the model without the need for data processing methods such as feature extraction. The GA-FESN prediction model obtained by this invention enables rapid and efficient conversion of gantry crane sensor monitoring data into the actual structural state response, solving the complex and time-consuming problem of repeatedly performing finite element analysis when considering the influence of the gantry crane's structural self-weight. Accurate prediction results can be obtained through the GA-FESN prediction model. Attached Figure Description

[0034] Figure 1 This is a flowchart illustrating the overall technical route of the present invention;

[0035] Figure 2 This is a flowchart of the finite element analysis considering the influence of the crane's own weight in this invention;

[0036] Figure 3 This is a schematic diagram of the functional echo state network of the present invention;

[0037] Figure 4 This is a schematic diagram of the genetic algorithm optimization function echo state network of the present invention;

[0038] Figure 5 This is a flowchart of the GA-FESN prediction model of the present invention. Detailed Implementation

[0039] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0040] Example 1

[0041] like Figures 1 to 5 As shown, a method for obtaining the true state of a gantry crane structure includes: inputting the actual working information of the gantry crane into a pre-trained GA-FESN prediction model to obtain the predicted stress and strain values ​​of the gantry crane structure; wherein, the training method of the pre-trained GA-FESN prediction model includes: obtaining the working information of the gantry crane and the first stress and strain values ​​at key structural locations, the working information including lifting load, slewing angle, and luffing magnitude; inputting the working information into a pre-established finite element model of the gantry crane to obtain simulated stress and strain values ​​at key structural locations under the gantry crane's own weight; synthesizing the first stress and strain values ​​and the simulated stress and strain values ​​to obtain synthesized stress and strain values; using a functional echo state network to mine the mapping relationship between the working information and the synthesized stress and strain values, and using the mapping relationship to train the GA-FESN prediction model to obtain the pre-trained GA-FESN prediction model.

[0042] This invention mines the mapping relationship between working information and synthetic stress-strain values ​​through functional echo state network (FESN), and optimizes the selection of hyperparameters in the FESN using a genetic algorithm, effectively improving the performance of the FESN model in nonlinear multivariate time series prediction. It allows direct input of various sensor monitoring values ​​into the model without the need for data processing methods such as feature extraction. The GA-FESN prediction model obtained by this invention can quickly and efficiently convert sensor monitoring data of gantry cranes into the actual structural response (predicted stress-strain values), solving the complex and time-consuming problem of repeatedly performing finite element analysis when considering the influence of the gantry crane's structural self-weight. Accurate prediction results can be obtained through the GA-FESN prediction model.

[0043] The present invention will now be described in further detail with reference to the accompanying drawings.

[0044] Step 1: Obtain the working information of the gantry crane and the first stress and strain values ​​at key structural locations. The working information includes the lifting load, slewing angle, and luffing amplitude.

[0045] Specifically, various sensors installed on the gantry crane are used to collect and record sensor monitoring information such as lifting load, slewing angle, amplitude change, and stress and strain values ​​at key structural locations during the operation.

[0046] Step 2: Input the work information into the pre-established finite element model of the gantry crane to obtain the stress and strain simulation values ​​of key structural locations under the self-weight of the gantry crane.

[0047] like Figure 2 As shown, the finite element analysis process considering the influence of the crane's own weight is as follows:

[0048] a. Based on the design parameters, establish a finite element model of the gantry crane, and set material parameters, mesh size, boundary constraints, etc.

[0049] b. Based on the working data recorded in step 1, set the corresponding slewing angle and amplitude, and set the working load to 0.

[0050] c. Calculate and analyze the stress and strain values ​​at key locations.

[0051] d. Record the corresponding stress and strain values ​​generated by self-weight under the given rotation angle and amplitude.

[0052] e. Based on the working data recorded in step 1, update the next set of rotation angles and amplitudes, and perform finite element analysis until the self-weight influence analysis under all rotation angles and amplitudes is completed.

[0053] f. Store the stress and strain simulation information of the gantry crane at all recorded slewing angles, luffing sizes, and corresponding key locations.

[0054] Step 3: Combine the first stress-strain value and the simulated stress-strain value to obtain the combined stress-strain value.

[0055] Specifically, based on mechanics principles, the stress characteristics at key locations are analyzed, such as tension, compression, torsion, or other combinations. Combining the fourth strength theory in materials mechanics, a reasonable synthesis method is set according to different stress characteristics. The stress-strain values ​​monitored by sensors and the stress-strain simulation information are synthesized to obtain the synthesized stress-strain values, i.e., the actual structural response.

[0056] Step 4: Use the functional echo state network to mine the mapping relationship between the working information and the synthesized stress and strain values ​​to obtain the GA-FESN prediction model.

[0057] like Figure 3 As shown, the functional echo state network consists of an input layer with M neurons, a pooling layer with N internal neurons, and an output layer with K neurons, where: the input matrix W in ∈R N×M The weights from the input layer to the storage layer are defined, W∈R N×N The weights between nodes within the storage layer are defined, and the output matrix W is... out ∈R K×N The weights from the storage layer to the output layer are defined.

[0058] The workflow of a functional echo state network is as follows:

[0059] a. Initialize the model parameters and input multivariate time series data u(t). At this time, the state of x(t) in the storage layer is continuously updated, and the update method is x(t+1)=f(W x (t)+W in u(t+1)), where the node state x(t) at time t is x(t) = [x1(t), x2(t), ..., x N (t)] T f() is the activation function.

[0060] b. All internal nodes in the storage layer need to undergo time aggregation and spatial aggregation operations sequentially. The result after time aggregation is: The result of spatial aggregation is Where w ji The weights are the weights from the i-th internal unit to the j-th output unit.

[0061] c. After continuous updating of neurons in the storage layer, combined with spatiotemporal aggregation operations, the final output layer outputs the following result: Where f out This is the output activation function.

[0062] like Figure 4 As shown, based on the data of lifting load, slewing angle and amplitude obtained in step 3) and the corresponding actual structural state response, the predictive performance of the functional echo state network is optimized using a genetic algorithm. The specific steps are as follows.

[0063] a. Establish the matching relationship between lifting load, rotation angle, and amplitude magnitude and the actual state response of the structure, and use it as training data for the model.

[0064] b. Utilize the Functional Echo State Network (FESN) to explore the potential mapping relationship between monitoring data such as lifting load, slewing angle, and amplitude variation and the actual state response of the structure, and carry out model training.

[0065] c. During model training, the root mean square error between the output value and the actual value of the FESN prediction model is used as the fitness. A genetic algorithm (GA) is used to optimize the selection of hyperparameters such as pool sparsity, spectral radius, and output connection weights.

[0066] d. When the fitness is minimized, i.e. the performance of the FESN prediction model reaches its optimal state, the values ​​of the storage pool sparsity, spectral radius and output connection weight at this point are taken as the optimal parameters to obtain the GA-FESN prediction model.

[0067] Step 5: Input the actual working information of the gantry crane into the GA-FESN prediction model to obtain the actual structural state response of the gantry crane.

[0068] like Figure 5 As shown, according to the workflow of the GA-FESN prediction model, new sensor monitoring data is collected on-site and transmitted to the analysis platform. New data such as lifting load, slewing angle, and amplitude are directly input into the input layer of the trained GA-FESN prediction model. After implicit processing by neurons within the dynamic storage layer, the predicted values ​​of the structural real-state response (i.e., predicted stress and strain values) are output. In practical engineering, workers can directly utilize the monitoring data collected by on-site sensors to quickly and efficiently obtain the structural real-state response at key locations of the gantry crane, avoiding the inefficiency caused by tedious and repetitive finite element analysis.

[0069] Example 2

[0070] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0071] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0072] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0073] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0074] As is known from common technical knowledge, this invention can be implemented through other embodiments that do not depart from its spirit or essential characteristics. Therefore, the disclosed embodiments described above are merely illustrative in all respects and are not the only ones. All modifications within the scope of this invention or its equivalents are included in this invention.

Claims

1. A method for obtaining the actual structural state of a gantry crane, characterized in that, include: The actual operating information of the gantry crane is input into a pre-trained GA-FESN prediction model to obtain the predicted stress and strain values ​​of the gantry crane structure; wherein, the training method of the pre-trained GA-FESN prediction model includes: Obtain the working information and the first stress-strain values ​​at key structural locations of the gantry crane. The working information includes lifting load, slewing angle, and luffing size. The working information is input into the pre-established finite element model of the gantry crane to obtain the stress and strain simulation values ​​of key structural locations under the self-weight of the gantry crane. The first stress-strain value and the simulated stress-strain value are combined to obtain the combined stress-strain value; A functional echo state network is used to mine the mapping relationship between working information and the synthesized stress-strain values, and the mapping relationship is used to train a GA-FESN prediction model to obtain the pre-trained GA-FESN prediction model; wherein, the construction process of the GA-FESN prediction model includes: The FESN prediction model is trained by mining the mapping relationship between lifting load, slewing angle, amplitude and the synthetic stress-strain value using the functional echo state network. Using the root mean square error between the output value and the actual value of the FESN prediction model as the fitness, the hyperparameters of the FESN prediction model are optimized using the GA algorithm. When the fitness is minimized, the hyperparameters are taken as the optimal parameters to obtain the GA-FESN prediction model.

2. The method for obtaining the true structural state of a gantry crane according to claim 1, characterized in that, The simulated stress and strain values ​​at key locations of the gantry crane structure under its own weight were obtained, including: Based on the collected work information, the rotation angle and amplitude were set sequentially, and the working load was set to 0. The stress and strain simulation values ​​at key locations of the structure were calculated for each case.

3. The method for obtaining the true structural state of a gantry crane according to claim 1, characterized in that, The functional echo state network includes an input layer, a storage layer, and an output layer arranged sequentially. The nodes in the storage layer are processed sequentially using time aggregation operations and then spatial aggregation operations.

4. The method for obtaining the true structural state of a gantry crane according to claim 1, characterized in that, The hyperparameters include storage pool sparsity, spectral radius, and output connection weights.

5. The method for obtaining the true structural state of a gantry crane according to claim 1, characterized in that, The formula for calculating the root mean square error is as follows: ； in, For the number of test samples, For the first One test sample in The actual value at time, For the first One test sample in The predicted value at any given time.

6. The method for obtaining the true structural state of a gantry crane according to claim 1, characterized in that, The following conditions must be met when obtaining the working information and the first stress-strain values ​​at key structural locations of the gantry crane: The lifting load, slewing angle, and amplitude in the working information correspond one-to-one with the first stress-strain values ​​at the key locations of the structure in that state.

7. The method for obtaining the true structural state of a gantry crane according to claim 1, characterized in that, The synthesis of the first stress-strain value and the simulated stress-strain value is specifically as follows: Analyze the force characteristics at key locations; the force characteristics include any one or more combinations of tension, compression, and torsion. Based on the stress characteristics, the first stress-strain value and the simulated stress-strain value are synthesized using an appropriate synthesis method.

8. A system for acquiring the actual structural state of a gantry crane, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1 to 7.

Images