A method and system for optimizing and improving performance of combined bracket

Abstract

The invention relates to a method and system for optimizing and improving the performance of a combined support, which belongs to the field of intelligent combined supports, solves the defects of improving the performance of the combined support through manual inspection and empirical judgment in the prior art, and comprises the following steps: obtaining spatial parameter information of the combined support at a construction site and stress wave information released during use; obtaining an actual physical spatial model of the combined support through mapping according to the obtained spatial parameter information; obtaining spatial state information of the combined support through inversion according to the obtained stress wave information; performing three-dimensional spatial association between the obtained actual physical spatial model and spatial state information and a virtual visualization model of the combined support obtained in advance according to design requirements, and performing performance analysis of the combined support; obtaining a construction optimization plan of the combined support according to the performance analysis result of the combined support, and adopting the method of the invention to improve the intelligent and standardized application level of the combined support.

Description

Technical Field

[0001] The present invention relates to the field of combined brackets and intelligent technologies thereof, and in particular to a combined bracket performance optimization and improvement method and system. Background Art

[0002] The statements herein merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In modern bridge construction, composite supports such as supporting scaffolding are one of the important auxiliary measures. Their construction accuracy and working performance such as safety, stability, and durability during use are important guarantees for achieving project quality, progress, safety, and civilized construction. However, in the use and management of traditional composite supports, the expression of two-dimensional drawings is not clear and intuitive enough, there is a deviation between the actual erection plan and the design plan, there is a lack of effective information exchange and interaction between the design, construction, and use of composite supports, and the analysis and improvement of the working performance of composite supports rely on manual inspection and experience judgment, lacking standardized and intelligent means and methods, resulting in inaccurate analysis results and poor improvement effects. Summary of the invention

[0004] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a method and system for optimizing and improving the performance of a combined bracket, which uses intelligent means to analyze and optimize the performance of the combined bracket, thereby overcoming the defects of relying on manual judgment and experience for analysis and optimization.

[0005] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0006] In a first aspect, an embodiment of the present invention provides a method for optimizing and improving the performance of a combined bracket, characterized in that it includes the following steps:

[0007] Obtain spatial parameter information of the combined support at the construction site and stress wave information released during use;

[0008] According to the acquired spatial parameter information of the combined bracket, an actual physical space model of the combined bracket is obtained through mapping;

[0009] The spatial state information of the combined bracket is obtained by inversion according to the acquired stress wave information;

[0010] The actual physical space model and space state information of the combined bracket obtained are associated with the virtual visualization model of the combined bracket obtained in advance according to the design requirements in three-dimensional space to perform performance analysis of the combined bracket;

[0011] According to the performance analysis results of the combined support, the construction optimization plan of the combined support is obtained.

[0012] Optionally, a panoramic picture of the combined bracket acquired by drone flight photography is received, and a deep learning end-to-end semantic segmentation model is used to perform target detection, semantic segmentation and category division on the acquired panoramic picture to obtain spatial parameter information of the combined bracket, and three-dimensional spatial mapping is performed on the actual physical model obtained from the spatial parameter information of the combined bracket to obtain the actual physical space model of the combined bracket.

[0013] Optionally, stress waves transmitted by an acoustic emission sensor installed on the combined bracket are received to obtain stress wave release information of the combined bracket during use.

[0014] Optionally, a one-dimensional convolutional neural network model based on multi-scale wavelet transform performs deep picking and deep mining of the acquired stress wave information in the time-space domain, and inverts the spatial state information of the combined bracket.

[0015] Optionally, the spatial state information includes a real-time stress cloud map and a deformation cloud map of the combined bracket.

[0016] Optionally, the actual physical space model and space state information of the obtained combined bracket are associated with the virtual visualization model of the combined bracket obtained in advance according to the design requirements in three-dimensional space in a discretized manner.

[0017] Optionally, the performance analysis of the combined support includes construction accuracy analysis, working status analysis and weighted comprehensive performance analysis.

[0018] Optionally, the method for establishing the virtual visualization model of the combined bracket is:

[0019] Identify the various parts of the combined bracket and their spatial position relationship according to the design drawings of the combined bracket;

[0020] According to the identified parts of the combined bracket and their spatial position relationship, combined with the set unrecognizable parameter information, a virtual visualization model of the combined bracket is established by adopting a parametric driving method.

[0021] Optionally, a deep learning method is used to identify the design drawings of the combined bracket, so as to identify the various parts of the combined bracket in the design drawings and their spatial positional relationships.

[0022] In a second aspect, an embodiment of the present invention provides a system for optimizing and improving the performance of a combined support, comprising:

[0023] Physical space information acquisition module: configured to acquire spatial parameter information of the combined support at the construction site and stress wave information released during use;

[0024] Physical space model mapping module: configured to obtain the actual physical space model of the combined bracket by mapping according to the acquired spatial parameter information of the combined bracket;

[0025] Physical space state inversion module: configured to invert the space state information of the combined bracket according to the acquired stress wave information;

[0026] Combined bracket performance analysis module: configured to associate the acquired actual physical space model and space state information of the combined bracket with the virtual visualization model of the combined bracket obtained in advance according to the design requirements in three-dimensional space, and perform performance analysis of the combined bracket;

[0027] Optimization module: configured to obtain a construction optimization plan for the combined support according to the performance analysis results of the combined support.

[0028] The beneficial effects of the present invention are as follows:

[0029] The combined bracket performance optimization and improvement method and system of the present invention obtain the actual physical space model of the combined bracket at the construction site and the spatial state information of the combined bracket, and then perform three-dimensional spatial association between the obtained actual physical space model and spatial state information of the combined bracket and the virtual visualization model of the combined bracket established in advance according to the design requirements, so that the combined bracket at the construction site can be compared with the design drawings, and the performance of the combined bracket at the construction site can be analyzed, so that optimization plans can be proposed in a targeted manner according to the analysis results, which promotes information interaction between the design, construction and use of the combined bracket, and provides standardized and intelligent means, methods and systems for physical perception, performance analysis and function improvement of the combined bracket, which makes up for the shortcomings of low efficiency, large errors, detection blind spots and subjective judgment of manual inspection, improves the intelligent and standardized application level of the combined bracket, promotes intelligent decision-making and precise control of the combined bracket, and is suitable for cast-in-place reinforced concrete structure construction in different scenarios such as bridge engineering, industrial and civil buildings. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0031] Figure 1 is a flow chart of the method of embodiment 1 of the present invention;

[0032] Figure 2 It is a schematic diagram of a cloud diagram obtained by inversion of the physical space state in Example 1 of the present invention;

[0033] Figure 3 is a schematic diagram of a virtual visualization model of a combined bracket according to Example 1 of the present invention;

[0034] Figure 4 is a schematic diagram of a system in Example 2 of the present invention; DETAILED DESCRIPTION

[0035] Example 1

[0036] This embodiment provides a method for optimizing and improving the performance of a combined bracket, such as Figure 1 As shown, the following steps are included:

[0037] Step 1: Obtain physical space information, obtain the spatial parameter information of the combined bracket at the construction site and the stress wave information during use.

[0038] The spatial parameter information of the combined bracket during its erection and use is collected in a multi-directional and active panoramic manner using drone flight photography.

[0039] Specifically, drone flight photography has high pixels and clarity, and can perform surround shooting according to pre-set positioning targets, angles and light requirements based on test technology and image processing requirements. The panoramic pictures obtained by drone flight photography are matched with positioning targets, angles, distances, time and other information and stored in the central database.

[0040] Acoustic emission sensors arranged at key positions of the combined bracket are used to collect real-time and passive information on stress waves released by various components during the use of the combined bracket.

[0041] In this embodiment, the key parts of the combined bracket are near the nodes where the local stress of the combined bracket is large or the force is complex. By utilizing the large propagation range of stress waves, only a small number of acoustic emission sensors are needed to fully perceive the stress wave release of the combined bracket.

[0042] The arrangement location and quantity of the acoustic emission sensors can be set according to the actual working conditions and will not be described in detail here.

[0043] In this embodiment, the temperature, humidity and wind speed of the combined bracket in various usage scenarios are collected and recorded by temperature sensors, humidity sensors and wind speed sensors, and the stress fluctuation signal of the acoustic emission sensor is corrected accordingly to improve the stress and deformation inversion accuracy, providing a reference for the analysis and optimization of the performance of the combined bracket.

[0044] The stress waves collected by the acoustic emission sensor, the temperature information collected by the temperature sensor, the humidity information collected by the humidity sensor, and the wind speed information collected by the wind speed sensor are transmitted to the remote terminal device by wireless transmission. The information sequence is matched with the sensor model, location, temperature, humidity, wind speed, etc. and stored in the central database.

[0045] Step 2: Physical space model mapping: obtaining the actual physical space model of the combined bracket through mapping according to the acquired spatial parameter information of the combined bracket.

[0046] Specifically, the following steps are included:

[0047] Step 2.1: Use deep learning methods to perform target detection, semantic segmentation, and category division on the panoramic photos of the combined bracket obtained in step 1 to obtain the spatial parameter information of each component, and then obtain the actual physical model of the combined bracket.

[0048] Step 2.2: Fuse the dot matrix model (target positioning), shooting angle, and distance information, perform three-dimensional spatial mapping on the actual physical model of the combined bracket, and obtain the actual physical spatial model of the combined bracket.

[0049] Furthermore, in the method of using deep learning methods to perform target detection, semantic segmentation and category division on panoramic photos of combined brackets, the input information is a two-dimensional panoramic picture taken by a drone, and the output information is spatial parameter information, wherein the spatial parameter information includes the physical parameters, layout forms and their corresponding spatial position relationships, connection methods of each component of the combined bracket, the actual loading, mechanical equipment, types of potential hazards, corner point position coordinates, construction personnel, driving routes, etc. during the use of the combined bracket. Technical personnel in this field can obtain the required spatial parameter information according to the actual working conditions, and obtain the actual physical model of the combined bracket according to the spatial parameter information of the combined bracket.

[0050] Preferably, an existing deep learning end-to-end semantic segmentation model is used to identify spatial parameter information of the combined bracket in the two-dimensional panoramic image.

[0051] The deep learning end-to-end semantic segmentation model picks up pixel information through several convolution and pooling operations, reorganizes each pixel through a focusing layer and several deconvolution operations, and finally connects a classification layer based on dynamic particle swarm optimization (DPSO) to achieve global optimization and accurate identification and classification of each component module.

[0052] Furthermore, before performing target detection, semantic segmentation and category classification on the input two-dimensional panoramic images taken by the drone, the two-dimensional panoramic images taken by the drone are cleaned and preprocessed, specifically including de-distortion, stitching, noise reduction, interpolation, etc. Technical personnel in this field can select corresponding cleaning and preprocessing steps according to actual needs.

[0053] The actual physical model of the combined bracket is mapped in three-dimensional space to obtain the actual physical space model of the combined bracket. The three-dimensional space mapping process requires matching the components of the combined bracket identified in the deep learning model with the corresponding components in the component cluster library, and using a parametric drive method to establish the actual physical space model of the combined bracket in the conversion coordinate system.

[0054] Step 3: Physical space state inversion: the space state information of the combined bracket is obtained by inversion based on the acquired stress wave information.

[0055] Specifically, Figure 2 As shown, a deep learning method is used to deeply pick up and mine the time-space characteristics of the stress waves released from various parts of the combined bracket, and the real-time spatial state information of the combined bracket is inverted. The spatial state information includes the stress cloud map and deformation cloud map of the combined bracket.

[0056] When using the existing deep learning methods to invert and predict the stress and deformation of each part of the combined bracket, the input information is the time and space domain stress waves obtained by the acoustic emission sensor, and the output information is the stress and deformation values ​​of each discrete point of the combined bracket in the whole space, thus obtaining the overall stress cloud map and deformation cloud map of the combined bracket.

[0057] Preferably, the deep learning model is a one-dimensional convolutional neural network architecture of multi-scale wavelet transform. A one-dimensional convolutional neural network model based on multi-scale wavelet transform is used to deeply pick up and mine the waveform, amplitude, arrival time, number of impacts, accumulated energy, etc. of the stress wave obtained by the acoustic emission sensor. The real-time stress cloud map and deformation cloud map of the combined bracket are obtained through high-dimensional space fitting and time-space domain comparative inversion.

[0058] The one-dimensional convolutional neural network model of multi-scale wavelet transform fully displays the time and space domain characteristics of the signal through multi-scale wavelet transform, and then adaptively extracts the multi-channel time and space characteristics through the extended convolutional long short-term memory network. Finally, the stress and deformation values ​​of each discrete point of the combined bracket are inverted and predicted through global connection, regression fitting and time and space domain comparison.

[0059] The wavelet transform basis functions specifically include: Haar series wavelets, Daubeshies series wavelets, Coif series wavelets, Sym series wavelets, Hat wavelets, Morlet wavelets, etc. Wavelet basis functions are selected to perform multi-scale "isomorphic" transformation on stress wave signals, and error reconstruction is performed based on their decomposition. Formula (1) enables the temporal and spatial characteristics of stress wave signals to be fully reflected.

[0060] (1)

[0061] In the formula, srec , i is the reconstructed signal obtained after wavelet decomposition, s is the original signal, i is the signal sequence.

[0062] The extended convolutional LSTM network uses convolution operations to replace matrix multiplication in the LSTM network, and sets all inputs, outputs, hidden layer states, and LSTM network gates as three-dimensional tensors, of which two dimensions are time and space information, and the third dimension is the state information that changes over time in the LSTM network. Through a series of convolution operations, the frequency of the multi-scale wavelet reconstruction signal is extracted, and the LSTM network is used for adaptive association learning of time and space domain features.

[0063] Step 4: Combined bracket performance analysis: the actual physical space model and spatial state information of the combined bracket obtained are associated with the virtual visualization model of the combined bracket obtained in advance according to the design requirements in three-dimensional space to perform performance analysis of the combined bracket.

[0064] The method for establishing the virtual visualization model of the combined bracket comprises the following steps:

[0065] Step a: Automatically identify the components of the combined bracket in the design drawing and their spatial position relationship, and convert them to a unified transformation coordinate system, and match them with the corresponding components in the component cluster library. The transformation coordinate system is consistent with the transformation coordinate system of the actual physical space model of the combined bracket in step 3.

[0066] Specifically, automatic recognition of design drawings is achieved using existing deep learning methods. By establishing an end-to-end semantic segmentation model for pre-training and transfer learning, the components of the combined bracket in the design drawings can be accurately identified. Based on the pre-trained end-to-end semantic segmentation model, the components of the combined bracket, such as columns, Bailey beams, I-beams, scaffolding rods and their connection nodes, and their spatial position relationships can be automatically identified.

[0067] The end-to-end semantic segmentation model picks up pixel information through several convolution and pooling operations, reorganizes each pixel through a focusing layer and several deconvolution operations, and finally connects a classification layer based on dynamic particle swarm optimization (DPSO) to achieve global optimization and accurate identification and classification of each component module.

[0068] The input information of the end-to-end semantic segmentation model is the two-dimensional design drawing of the combined bracket, and the output information is the physical parameters, layout form and corresponding spatial position relationship, connection method, etc. of each component of the combined bracket.

[0069] Step b: receiving the user's instruction and correcting the automatic recognition result.

[0070] Step c: receiving the user's instruction, and setting the parameter information that the combined bracket cannot automatically identify, the parameter information that the combined bracket cannot automatically identify includes the service life of the scaffolding, material performance parameters, and constitutive model.

[0071] Step d: Figure 3 As shown, according to the automatically recognized information corrected in step b and the parameter information of the combined bracket set in step c that cannot be automatically recognized, a virtual visualization model of the combined bracket is established by a parametric driving method.

[0072] The obtained virtual visualization model of the combined bracket is three-dimensionally associated with the actual physical space model of the combined bracket obtained in step 2 and step 3 and its spatial state information, and then the performance analysis of the combined bracket is performed.

[0073] The three-dimensional space association refers to the one-to-one correspondence and comparison between the actual physical space model of the combined bracket and the corresponding parts of the virtual visualization model.

[0074] Preferably, the obtained virtual visualization model of the combined bracket is associated with the actual physical space model of the combined bracket obtained in step 2 and step 3 and the spatial state information thereof in three-dimensional space in a discretized manner.

[0075] The performance analysis of the combined bracket includes:

[0076] (1) Calculate the position deviation between the actual physical space model of the combined bracket and the design scheme, so as to conduct construction accuracy analysis and rating.

[0077] The construction accuracy analysis includes aspects such as the positioning accuracy of each component of the combined bracket, the inclination angle of rods and the frame, and the safety level is divided according to the deviation between the actual error and the maximum allowable value.

[0078] (2) Calculate the deviation between the actual stress state, deformation, stability coefficient and the maximum allowable value of the combined bracket to conduct working status analysis and rating.

[0079] The working status analysis includes aspects such as local stress, deformation, and stability of rods and frames, and the safety level is divided according to the deviation between the actual inversion value and the maximum allowable value (or design value).

[0080] (3) Weight the construction accuracy analysis results of the combined support and the working status analysis results to conduct a comprehensive performance analysis and rating of the combined support.

[0081] The specific weighting method can be based on the existing technology according to the actual working conditions and will not be described in detail here.

[0082] In this embodiment, when performing actual physical space model mapping, spatial state information inversion and performance analysis in steps 2, 3 and 4, the combined bracket is discretized to match the inversion and analysis results to the three-dimensional view. The user selects each discrete point in the three-dimensional view to simultaneously display the construction accuracy rating, working status rating and comprehensive performance rating, etc., and the parts with lower safety ratings are marked with dark colors for emphasis.

[0083] Step 5: Improve and optimize the function of the combined bracket, and obtain the construction optimization plan of the combined bracket based on the performance analysis results of the combined bracket.

[0084] According to the construction accuracy analysis results, working status analysis results, comprehensive performance analysis results and safety ratings in step 4, corresponding construction improvement plans, optimized usage suggestions and reinforcement measures are given respectively.

[0085] For situations with higher safety ratings, follow-up observation can be carried out and no measures will be taken for the time being. For situations with lower safety ratings, optimized lifting strategies are automatically given, including: correcting the installation position, local reinforcement, balanced loading, limiting driving routes, demarcating safe restricted areas, immediately ceasing use, dismantling and re-erecting, etc. Specifically, for example: for frames with large deviations in construction accuracy analysis, they should be dismantled and re-erected; for situations where local stress is close to the critical value, a local reinforcement plan or a balanced loading plan should be given. Technical personnel in this field can pre-set the optimized lifting strategies corresponding to each safety level, which will not be described in detail here.

[0086] In this embodiment, based on the performance analysis and function improvement analysis in steps 4 and 5, a combined bracket working performance analysis report and a corresponding function improvement strategy report are generated, including aspects such as the positioning accuracy of each component, the inclination angle of rods and frames, local stress, deformation, stability of rods and frames, and comprehensive performance of the combined bracket.

[0087] In this embodiment, according to the needs of construction and inspection, after each panoramic photo acquired by the drone flight photography and the stress wave information of the acoustic emission sensor are extracted, the physical space model mapping and physical space state inversion are automatically performed, and the performance analysis report and function improvement strategy of the combined bracket are updated and generated.

[0088] The combined bracket performance optimization and improvement method of this embodiment deeply integrates flight photography, wave testing and other technologies and applies them to the field of combined bracket performance analysis and function improvement. It adopts a combination of active and passive methods to obtain the actual physical space model and real-time space status information of the combined bracket, and digitally maps the real physical space model with the virtual visualization space, which promotes the information interaction between the design, construction and use of the combined bracket. It integrates digital modeling, image analysis, data cleaning, and deep learning methods to provide standardized and intelligent means, methods and systems for physical perception, performance analysis and function improvement of the combined bracket, making up for the shortcomings of low efficiency, large errors, detection blind spots and subjective judgment of manual inspections, improving the level of intelligent and standardized application of combined brackets, and promoting intelligent decision-making and precise control of combined brackets. It is suitable for the construction of cast-in-place reinforced concrete structures in different scenarios such as bridge engineering, industrial and civil buildings.

[0089] Example 2

[0090] This embodiment provides a system for optimizing and improving the performance of a combined support. Figure 4 As shown, including:

[0091] Physical space information acquisition module: configured to obtain the spatial parameter information of the combined support at the construction site and the stress wave release information during use.

[0092] Specifically, the physical space information acquisition module is divided into two sub-modules: active acquisition and passive acquisition. The active acquisition module is configured as follows: using drone flight photography to perform multi-directional and active panoramic acquisition of the spatial parameter information of each component during the erection and use of the combined bracket, and all the collected panoramic pictures are matched with the positioning target, angle, distance, time and other information and stored in the central database C1; the passive acquisition module is configured as follows: acoustic emission sensors are arranged at key stress-bearing parts of the combined bracket, and the stress wave release information, temperature, humidity and wind speed of each component during the use of the combined bracket are collected in real time and passively, and the stress wave signal emitted by the acoustic emission sensor is corrected using the temperature, humidity and wind speed information, and the stress wave release information, temperature, humidity and wind speed of the combined bracket in each use scenario are recorded at the same time, and the information sequence is matched with the sensor model, location, temperature, humidity, wind speed, etc. and stored in the central database C2.

[0093] Physical space model mapping module: configured to obtain the actual physical space model of the combined bracket through mapping according to the spatial parameter information of the combined bracket obtained by the physical space information obtaining module.

[0094] Specifically, the physical space model mapping module is configured as follows: the actual physical model of the combined bracket is obtained by identifying the components of the combined bracket and the load, construction personnel, mechanical equipment, potential hazards, etc. in multi-directional and panoramic pictures through a deep learning end-to-end semantic segmentation model, and the positioning target, angle, distance and other information are integrated to perform three-dimensional spatial mapping of the actual physical model of the combined bracket to obtain the actual physical space model of the combined bracket.

[0095] Physical space state inversion module: configured to obtain the space state information of the combined bracket by inverting the stress wave release information obtained by the physical space information acquisition module.

[0096] Specifically, the physical space state inversion module is configured as follows: a one-dimensional convolutional neural network model based on multi-scale wavelet transform is used to deeply pick up and mine the time-space domain features of the stress waves obtained by the acoustic emission sensor, and the real-time stress cloud map and deformation cloud map of the combined bracket are inverted.

[0097] Combined bracket performance analysis module: It is configured to perform three-dimensional spatial association between the actual physical space model and space state information of the combined bracket obtained according to the physical space model mapping module and the physical space state inversion module and the virtual visualization model of the combined bracket obtained in advance according to the design requirements, so as to perform performance analysis of the combined bracket.

[0098] Specifically, the performance analysis module of the combined bracket is configured as follows: to associate the virtual visualization model of the combined bracket obtained in advance with the actual physical space model of the combined bracket and its spatial state information obtained by the physical space mapping module and the physical space state inversion module in three-dimensional space in a discretized manner, calculate the position deviation between the actual physical space model and the design scheme for construction accuracy analysis and rating; calculate the deviation between the actual stress state, deformation, stability coefficient and the maximum allowable value for working state analysis and rating; and weight the construction accuracy analysis results and the working state analysis results for comprehensive performance analysis and rating.

[0099] Optimization module: configured to obtain a construction optimization plan for the combined support according to the analysis results obtained by the performance analysis module of the combined support.

[0100] Specifically, the optimization module is configured as follows: according to the combined bracket performance analysis module, the construction accuracy analysis results, working status analysis results, comprehensive performance analysis results and safety ratings are obtained, and corresponding function improvement strategies are given to guide the reinforcement and improvement of the combined bracket, optimized use, etc., and support the generation of the combined bracket working performance analysis report and the corresponding function improvement strategy report.

[0101] It also includes a virtual space visualization modeling module, which is configured to establish an end-to-end semantic segmentation model through a deep learning method to automatically identify the components of the combined bracket in the design drawing and their spatial position relationships, and receive user instructions to correct the recognition results, receive user instructions to set relevant information that cannot be directly obtained from the drawings, and then use a parametric drive method to establish a virtual visualization model of the combined bracket.

[0102] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for optimizing and improving the performance of a combined support, characterized in that: The following steps are involved: Acquire spatial parameter information of the combined bracket at the construction site and stress wave information released during use; obtain the actual physical spatial model of the combined bracket through mapping based on the acquired spatial parameter information of the combined bracket; The spatial state information of the combined bracket is obtained by inversion according to the acquired stress wave information; specifically: a one-dimensional convolutional neural network model based on multi-scale wavelet transform performs deep picking and deep mining of the acquired stress wave information in the time-space domain, and inverts to obtain the spatial state information of the combined bracket; the spatial state information includes a real-time stress cloud map and a deformation cloud map of the combined bracket; The actual physical space model and space state information of the combined bracket obtained are three-dimensionally associated with the virtual visualization model of the combined bracket obtained in advance according to the design requirements, and the performance analysis of the combined bracket is carried out; the construction optimization plan of the combined bracket is obtained according to the performance analysis results of the combined bracket.

2. A method for optimizing and improving the performance of a combined support as claimed in claim 1, characterized in that: A panoramic picture of the combined bracket obtained by drone flight photography is received, and a deep learning end-to-end semantic segmentation model is used to perform target detection, semantic segmentation and category division on the obtained panoramic picture to obtain spatial parameter information of the combined bracket. The actual physical model obtained from the spatial parameter information of the combined bracket is mapped into three-dimensional space to obtain the actual physical space model of the combined bracket.

3. A method for optimizing and improving the performance of a combined support as claimed in claim 1, characterized in that: The stress wave transmitted by the acoustic emission sensor installed on the combined bracket is received to obtain the stress wave release information of the combined bracket during use.

4. A method for optimizing and improving the performance of a combined support as claimed in claim 1, characterized in that: The actual physical space model and space state information of the obtained combined bracket are associated with the virtual visualization model of the combined bracket obtained in advance according to the design requirements in three-dimensional space in a discretized manner.

5. The method for optimizing and improving the performance of a combined support according to claim 1, characterized in that: The performance analysis of the combined support includes construction accuracy analysis, working status analysis and weighted comprehensive performance analysis.

6. A method for optimizing and improving the performance of a combined support as claimed in claim 1, characterized in that: The method for establishing the virtual visualization model of the combined bracket is as follows: Identify the various parts of the combined bracket and their spatial position relationship according to the design drawings of the combined bracket; According to the identified parts of the combined bracket and their spatial position relationship, combined with the set unrecognizable parameter information, a virtual visualization model of the combined bracket is established by adopting a parametric driving method.

7. A method for optimizing and improving the performance of a combined support as claimed in claim 6, characterized in that: The deep learning method is used to identify the design drawings of the combined bracket, and the various parts of the combined bracket in the design drawings and their spatial position relationships are identified.

8. A system for optimizing and improving the performance of a combined support, characterized in that: include: Physical space information acquisition module: configured to acquire spatial parameter information of the combined support at the construction site and stress wave information released during use; Physical space model mapping module: configured to obtain the actual physical space model of the combined bracket by mapping according to the acquired spatial parameter information of the combined bracket; Physical space state inversion module: configured to obtain the space state information of the combined bracket by inverting the acquired stress wave information; specifically: a one-dimensional convolutional neural network model based on multi-scale wavelet transform performs deep time-space domain feature picking and deep mining on the acquired stress wave information, and inverts the space state information of the combined bracket; the space state information includes a real-time stress cloud map and a deformation cloud map of the combined bracket; Combined bracket performance analysis module: configured to associate the acquired actual physical space model and space state information of the combined bracket with the virtual visualization model of the combined bracket obtained in advance according to the design requirements in three-dimensional space, and perform performance analysis of the combined bracket; Optimization module: configured to obtain a construction optimization plan for the combined support according to the performance analysis results of the combined support.

Images