Garbage pool wall structure construction scaffold structure and method thereof

Abstract

The invention provides a garbage pool wall structure construction scaffold structure and a method thereof, and belongs to the technical field of building construction.The garbage pool wall structure construction scaffold structure comprises a scaffold body, inclined supporting rods are arranged on the outer side of the scaffold body, inclined supporting rod auxiliary rods are arranged between the inclined supporting rods and the scaffold body, and moving wheels are arranged at the four corners of the bottom of the scaffold body; the anti-overturning assembly is arranged on the outer side of the scaffold body and comprises a steel wire rope, a connecting rod, a movable rod and a sleeve, the movable rod is fixed to the outer end of the connecting rod, the sleeve is arranged on the outer side of the movable rod in a sleeving mode, and the scaffold body stability monitoring system is used for monitoring the using stability of the scaffold and conducting overturning early warning. The scaffold body stability monitoring system comprises a bionics sensor network, a meteorological data acquisition module, a data fusion module, a multi-mode learning architecture module and an early warning module, and solves the problem that the safety coefficient of the scaffold is relatively low due to the fact that the anti-overturning structure of the scaffold is complicated and the overturning state of the scaffold cannot be predicted.

Description

Technical Field

[0001] The present invention belongs to the technical field of building construction, and in particular relates to a scaffolding structure for constructing a garbage pool wall structure and a method thereof. Background Art

[0002] Garbage pits are a core component of waste treatment facilities, primarily used for temporary storage of waste pending subsequent transportation and processing. Their walls must not only withstand the weight and pressure of the waste but also prevent the leakage of leachate, ensuring that the surrounding environment is not polluted. Therefore, the quality of the garbage pit wall structure is directly related to its effectiveness, safety, and environmental performance.

[0003] Due to the high walls of the garbage pit, the stability and load-bearing capacity of traditional scaffolding are significantly reduced as the height increases. External factors such as wind load and construction load can easily cause the scaffolding to shake or tilt, posing a major safety hazard. Some existing anti-overturning devices, such as anti-overturning brackets and stabilizers, can provide a certain degree of anti-overturning capability, but they are often complex in structure, difficult to install, and expensive. In addition, they only install simple sensors for horizontal monitoring, which cannot detect the overturning problem of the scaffolding in time, resulting in low safety. To this end, a more stable scaffolding structure for the construction of the garbage pit wall structure and a method thereof are designed. Summary of the Invention

[0004] The embodiment of the present invention provides a scaffolding structure and method for constructing a garbage pool wall structure, which solves the problem that the anti-overturning structure of the scaffolding is complex and the overturning state of the scaffolding cannot be predicted, resulting in a low safety factor of the scaffolding.

[0005] In view of the above problems, the technical solution proposed by the present invention is:

[0006] The present invention provides a scaffolding structure for the construction of a garbage pool wall structure, comprising a scaffolding frame body, an oblique support rod is provided on the outer side of the scaffolding frame body, an oblique support rod auxiliary rod is provided between the oblique support rod and the scaffolding frame body, and movable wheels are provided at the four corners of the bottom of the scaffolding frame body;

[0007] An anti-overturning assembly is arranged on the outside of the scaffold frame, and includes a steel wire rope, a connecting rod, a movable rod and a sleeve. The connecting rod is arranged at one end of the steel wire rope, the movable rod is fixed at the outer end of the connecting rod, and the sleeve is sleeved on the outside of the movable rod;

[0008] A frame stability monitoring system, which is used to monitor the stability of the scaffold and provide overturning warnings. The frame stability monitoring system includes a bionic sensor network, a meteorological data acquisition module, a data fusion module, a multimodal learning architecture module, and an early warning module.

[0009] The bionic sensor network collects real-time change data of the scaffold through bionic sensors;

[0010] The meteorological data acquisition module monitors the meteorological data of the garbage pit construction environment through real-time data from the weather station or installed sensors;

[0011] The data fusion module fuses the real-time data of the bionic sensor network and the meteorological data acquisition module from different aspects and performs data prediction;

[0012] The multimodal learning architecture module is used to effectively fuse data from different sources so that the model can simultaneously use information from these data to make predictions, thereby assisting the data fusion module.

[0013] The early warning module issues an early warning prompt based on the prediction result of the data fusion module.

[0014] As a preferred technical solution of the present invention, it also includes a support assembly, which is arranged at the bottom of the scaffolding frame. The support assembly includes a motor, a rotating rod, a connecting block, a support rod and a support plate. The rotating rod is arranged at the output end of the motor, and two connecting blocks are sleeved on the outside of the rotating rod. Two support rods are respectively arranged on the outside of the two connecting blocks, and support plates are provided at the ends of the two support rods.

[0015] As a preferred technical solution of the present invention, the output end of the motor is transmission-connected to the rotating rod, and threaded sections are provided near both ends of the rotating rod. The two connecting blocks are respectively threadedly engaged with the rotating rod through the threaded sections. The two support rods are respectively provided on the upper and lower sides of the connecting block, and the two ends of the support rod are respectively bolted to the connecting block and the support plate. The support rod and the rotating rod on the same side are designed in a triangle, and the support plate is screwed to the scaffolding frame body.

[0016] As a preferred technical solution of the present invention, a plurality of anti-overturning components are provided on the scaffolding frame, the other end of the steel wire rope is installed on the scaffolding frame using a rope clamp, the end of the connecting rod is provided with a connecting end, the other end of the steel wire rope passes through the connecting end and is screwed to the connecting end, and the movable rod passes through the sleeve.

[0017] As a preferred technical solution of the present invention, a spring is arranged between the movable rod and the sleeve, the spring 206 is surrounded by the outside of the movable rod, and a cover plate is provided near both ends of the movable rod. The cover plate is adapted to the sleeve and is threaded with the sleeve. The two ends of the spring are respectively screwed to the cover plate and the movable rod, a fixing rod is fixed to the outside of one of the cover plates, and the other cover plate is movably connected to the movable rod.

[0018] As a preferred technical solution of the present invention, the data fusion module includes a data alignment unit, a feature fusion unit, a model fusion unit and an output unit;

[0019] The data alignment unit aligns data of different modalities in terms of time, space and features;

[0020] The feature fusion unit extracts feature data from the data alignment unit through a generative adversarial network and fuses the feature data using a multi-core learning algorithm;

[0021] The model fusion unit is used to select a suitable model framework for the fused feature data and fuse the models using a simple average fusion or weighted average fusion method;

[0022] The output unit is used to store the fused data.

[0023] As a preferred technical solution of the present invention, the details of feature fusion performed by the feature fusion unit are as follows:

[0024] Step A: Design a generator and a discriminator. The generator is used to generate data similar to the real data, and the discriminator is used to distinguish the real data from the generated data. The generator and the discriminator are trained alternately, and the trained generator or discriminator is used to extract the features of the aligned data.

[0025] Step B: Select an appropriate predefined kernel function based on the extracted feature data, assign an initial weight to each kernel function, automatically adjust the weight of each kernel function through a multi-kernel learning algorithm, perform a weighted combination of the kernel function matrices of each modality based on the learned weights to form a new kernel function matrix, and perform standardization on the combined kernel function;

[0026] Step C: For any two data points, their similarity or distance in the new feature space is calculated by combining the kernel function matrix. The similarity or distance value represents the fused feature.

[0027] The detailed fusion steps of the model fusion unit are as follows:

[0028] Step a: Select an appropriate model architecture based on the task requirements, divide the fused data into training set, validation set, and test set, initialize the model parameters, train the model using the training set data, and adjust the model parameters through forward propagation and backpropagation;

[0029] Step b: Select an appropriate fusion method to perform model fusion on the trained models after feature fusion, and perform a simple average of the prediction results of all models, or perform a weighted average of the prediction results of each model according to the pre-calculated weights to obtain the final fusion prediction result;

[0030] Step c: Use the validation set to verify the fusion model, evaluate the final performance of the fusion model on the test set, and analyze the performance indicators of the model;

[0031] In step d, the trained fusion model is deployed in the actual application environment, the model input, output and performance indicators are monitored, and the model is updated regularly according to new data or environmental changes.

[0032] As a preferred technical solution of the present invention, the multimodal learning architecture module includes a dual-stream network establishment unit and an attention mechanism integration unit;

[0033] The dual-stream network establishment unit is used to design a stream for each modality when designing the generator and the discriminator, and to train each stream separately when training the generator and the discriminator;

[0034] The attention mechanism integration unit uses the attention mechanism to dynamically weight different modalities, enabling the model to adaptively focus on more important information according to task requirements.

[0035] On the other hand, a method for constructing a scaffolding structure for a garbage pool wall structure comprises the following steps:

[0036] S1, during the construction of the garbage pool wall, scaffolding is installed, sensors are installed around the construction site to collect real-time weather data of the construction environment, and sensors are deployed on the scaffolding to monitor the status of the scaffolding in real time;

[0037] S2, the collected data is transmitted to the frame stability monitoring system in real time, and data processing is performed to remove outliers and noise, and perform data alignment processing;

[0038] S3, fuses the gas phase data with the bionic sensor data to establish a scaffolding structure model, and applies the fused data model to the structure model for simulation;

[0039] S4, based on the simulation results, assesses the overturning risk of the frame. When a risk is detected, an early warning is issued to prompt construction workers to take measures;

[0040] S5. At the same time, the support assembly at the bottom is controlled to adjust the balance according to the simulation results, and emergency treatment is performed until the scaffolding returns to normal state. The support assembly is restored, and construction is continued to complete the work of the garbage pool wall.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] (1) The present invention realizes dynamic adjustment of the tension of the wire rope through the design of the movable rod, sleeve and spring in the anti-overturning assembly, and can adjust the anti-overturning force in real time according to the inclination of the scaffold. At the same time, the support assembly drives the rotating rod through the motor to realize the movement of the connecting block and the support rod, thereby controlling the lifting and lowering of the support plate, providing a fast and effective tilt position adjustment capability, and integrating the scaffolding stability monitoring system with the anti-overturning assembly and the support assembly to realize the functions of real-time monitoring, early warning and automatic adjustment;

[0043] (2) The present invention can monitor the status of the scaffold more comprehensively and in real time through the bionic sensor layout. Compared with the traditional monitoring method of a single or a small number of sensors, it provides a richer data source and higher monitoring accuracy, thereby improving the stability and perception ability of the system.

[0044] (3) The present invention can extract deep and aligned features through the design of the generator and the discriminator, thereby enhancing the feature representation capability. The invention also utilizes multi-kernel learning to automatically adjust the kernel function weights to enhance the generalization capability of the model and adapt to different data distributions. The fused feature data provides more comprehensive input information, thereby improving the accuracy of the prediction model. The combination of the model fusion strategy further optimizes the prediction results, making the prediction results of the scaffolding overturning state more accurate.

[0045] (4) The present invention effectively integrates meteorological data and scaffolding data through a dual-stream network, improving the model's ability to utilize multi-source information. The attention mechanism ensures that the model can adaptively focus on more important information sources according to task requirements.

[0046] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are specifically listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a structural schematic diagram of a scaffolding structure for construction of a garbage pool wall structure disclosed in the present invention;

[0048] Figure 2 This is a schematic diagram of the overall structure of an anti-overturning assembly of a scaffolding structure for construction of a garbage pool wall structure disclosed in the present invention;

[0049] Figure 3 This is a schematic diagram of the partially disassembled structure of an anti-overturning component of a scaffolding structure for construction of a garbage pool wall structure disclosed in the present invention;

[0050] Figure 4 This is a schematic diagram of the overall structure of a support assembly of a scaffolding structure for construction of a garbage pool wall structure disclosed in the present invention;

[0051] Figure 5 This is a block diagram of a frame stability monitoring system for a scaffolding structure for construction of a garbage pool wall structure disclosed in the present invention;

[0052] Figure 6 This is a flow chart of a method for constructing a scaffolding structure for a garbage pool wall structure disclosed in the present invention;

[0053] Description of reference numerals: 100, scaffolding frame; 101, diagonal bracing rod; 102, diagonal bracing rod auxiliary rod; 103, pad; 104, moving wheel;

[0054] 200, anti-overturning assembly; 201, wire rope; 202, connecting rod; 203, connecting end; 204, movable rod; 205, sleeve; 206, spring; 207, cover plate; 208, fixing column;

[0055] 300, support assembly; 301, motor; 302, rotating rod; 303, connecting block; 304, support rod; 305, support plate; 306, threaded segment;

[0056] 400. Frame stability monitoring system; 401. Bionic sensor network; 402. Meteorological data acquisition module; 403. Data fusion module; 4031. Data alignment unit; 4032. Feature fusion unit; 4033. Model fusion unit; 4034. Output unit; 404. Multimodal learning architecture; 4041. Dual-stream network establishment unit; 4042. Attention mechanism integration unit; 405. Early warning module. DETAILED DESCRIPTION

[0057] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0058] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention as claimed, but rather merely represents selected embodiments of the present invention. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort are also within the scope of protection of the present invention.

[0059] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings.

[0060] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0061] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature identified as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0062] Example 1

[0063] Refer to the attached Figure 1-5 As shown, the present invention provides a technical solution: a scaffolding structure for the construction of a garbage pool wall structure, comprising a scaffolding frame 100, an outer side of the scaffolding frame 100 is provided with an oblique support rod 101, an oblique support rod auxiliary rod 102 is provided between the oblique support rod 101 and the scaffolding frame 100, and movable wheels 104 are provided at the four corners of the bottom of the scaffolding frame 100, the oblique support rod 101 is installed on the four sides of the scaffolding frame 100 by screws, and the oblique support rod auxiliary rod 102 is installed at the force bearing point of the oblique support rod 101 by screws and connected to the scaffolding frame 100 by screws to form a triangular support, the bottom of the scaffolding frame 100 can be moved for use by the movable wheels 104, and a foot brake is provided for easy fixation, and a pad 103 is provided at the bottom of the oblique support rod 101 to stabilize the bottom position of the oblique support rod 101;

[0064] The anti-overturning assembly 200 is arranged on the outside of the scaffolding frame 100. The anti-overturning assembly 200 includes a steel wire rope 201, a connecting rod 202, a movable rod 204 and a sleeve 205. The diagonal brace rod 202 is arranged at one end of the steel wire rope 201, the movable rod 204 is fixed to the outer end of the connecting rod 202, and the sleeve 205 is sleeved on the outer side of the movable rod 204;

[0065] The frame stability monitoring system 400 is used to monitor the stability of the scaffold and provide overturning warnings. The frame stability monitoring system 400 includes a bionic sensor network 401, a meteorological data acquisition module 402, a data fusion module 403, a multimodal learning architecture 404, and an early warning module 405.

[0066] The bionic sensor network 401 collects real-time change data of the scaffolding through bionic sensors, such as vibration sensors, strain sensors, dynamometers, etc.;

[0067] The meteorological data acquisition module 402 monitors the meteorological data of the garbage pit construction environment through real-time data from a weather station or installed sensors;

[0068] The data fusion module 403 fuses the real-time data of the bionic sensor network 401 and the meteorological data acquisition module 402 from different aspects and performs data prediction;

[0069] The multimodal learning architecture 404 module is used to effectively fuse data from different sources (such as meteorological data and scaffolding data) so that the model can simultaneously use the information in these data to make predictions, providing assistance to the data fusion module 403;

[0070] The early warning module 405 issues an early warning prompt based on the prediction result of the data fusion module 403.

[0071] The embodiment of the present invention is also implemented through the following technical solutions.

[0072] In an embodiment of the present invention, a support assembly 300 is also included. The support assembly 300 is arranged at the bottom of the scaffolding frame 100. The support assembly 300 includes a motor 301, a rotating rod 302, a connecting block 303, a support rod 304 and a support plate 305. The rotating rod 302 is arranged at the output end of the motor 301, and two connecting blocks 303 are sleeved on the outer side of the rotating rod 302. Two support rods 304 are respectively arranged on the outer sides of the two connecting blocks 303, and support plates 305 are provided at the ends of the two support rods 304.

[0073] In an embodiment of the present invention, the output end of the motor 301 is connected to the rotating rod 302 for transmission, and the rotating rod 302 is controlled to rotate by the motor 301. A threaded section 306 is provided near both ends of the rotating rod 302. The two connecting blocks 303 are respectively threadedly engaged with the rotating rod 302 through the threaded sections 306. The rotation of the rotating rod 302 can control the movement of the connecting block 303 on the rotating rod 302 through the threaded sections 306. The two support rods 304 are respectively provided on the upper and lower sides of the connecting block 303. The two ends of the support rod 304 are respectively engaged with the connecting rod 302. The block 303 and the support plate 305 are bolted together, and the support rod 304 on the same side is designed in a triangle with the rotating rod 302. The support plate 305 is screwed to the scaffold frame 100. The support rod 304 can be driven to move together through the connecting block 303, so that when the two connecting blocks 303 move inward or outward at the same time, the angle between the support rod 304 and the connecting block 303 becomes larger or smaller, so that the angle between the support plate 305 at the end of the support rod 304 on the same side and the support rod 304 becomes smaller or larger, and the support rod 304 controls the support plate 305 to rise and fall.

[0074] Specifically, when the frame stability monitoring system 400 detects that the scaffold frame 100 is at risk of overturning, the system controls the motor 301 to start according to the tilt direction of the scaffold frame 100. The motor 301 directly drives the rotating rod 302 to rotate. Due to the spiral action of the thread, the connecting block 303 moves axially on the rotating rod 302 (i.e., moves left and right). When the two connecting blocks 303 move inward at the same time, the angle between the connecting block 303 and the support rod 304 becomes smaller. Since the support rod 304 and the support plate 305 are connected by bolts, the movement of the connecting block 303 drives the support rod 304 The support rod 304 tilts inward, causing the support plate 305 at its end to rise, thereby achieving the raising of the scaffolding frame 100. The motor 301 is controlled in reverse to rotate the rotating rod 302, causing the two connecting blocks 303 to move outward at the same time. The outward movement of the connecting block 303 increases the angle between the support rod 304 and the connecting block 303, and the tilt angle of the support rod 304 decreases, causing the support plate 305 at its end to descend, thereby achieving the lowering of the scaffolding frame 100. The tilted position of the scaffolding frame 100 is temporarily handled, providing maintenance time for the staff and preventing the scaffolding from tipping over.

[0075] In an embodiment of the present invention, a plurality of anti-overturning assemblies 200 are provided on the scaffolding frame 100. The anti-overturning assemblies 200 are installed at the installation nodes of the scaffolding frame 100, or installed at specific nodes as required. The other end of the wire rope 201 is installed on the scaffolding frame 100 by means of a rope clamp, and the wire rope 201 is installed at the node by means of the rope clamp. The end of the connecting rod 202 is provided with a connecting end 203. The other end of the wire rope 201 passes through the connecting end 203 and is screwed to the connecting end 203. The other end of the wire rope 201 is fixed to the connecting rod 202 by screws. The movable rod 204 passes through the sleeve 205 and is slidably connected to the cover plate 207. A guide rod is provided in the cover plate 207, and a guide groove adapted to the guide rod is provided inside the movable rod 204, which serves to guide the movable rod 204 to move in the sleeve 205.

[0076] In an embodiment of the present invention, a spring 206 is provided between the movable rod 204 and the sleeve 205. The spring 206 surrounds the outer side of the movable rod 204. Cover plates 207 are provided near both ends of the movable rod 204. The cover plates 207 are adapted to the sleeve 205 and are threadedly matched with the sleeve 205. The two ends of the spring 206 are screwed to the cover plates 207 and the movable rod 204 respectively. The spring 206 assists the steel wire rope 201 to prevent the scaffolding from overturning. A fixed rod is fixed to the outer side of one of the cover plates 207, and the anti-overturning assembly 200 is installed on the building through the fixed rod. The other cover plate 207 is movably connected to the movable rod 204, so that the cover plate 207 can move on the outside of the movable rod 204 as the spring 206 expands and contracts.

[0077] Specifically, when the scaffold is in a normal state, the wire rope 201 maintains a certain tension, the anti-overturning assembly 200 is in a stationary state, and the spring 206 is in a compressed state, providing a certain pre-tightening force to ensure that the tension of the wire rope 201 is stable. When the scaffold is at risk of overturning, when the frame tilts or displaces, the tilt data is sent to the frame stability monitoring system 400 for processing. At the same time, the tension of the wire rope 201 changes and is transmitted to the connecting rod 202 and the movable rod 204. The movable rod 204 moves in the sleeve 205, and the spring 206 expands and contracts accordingly to provide additional tension adjustment. The expansion and contraction of the spring 206 assists the wire rope 201 in anti-overturning work on the scaffold, thereby enhancing the anti-overturning effect. The movement of the cover plate 207 further adjusts the tension of the wire rope 201 to ensure the effectiveness of the anti-overturning assembly 200.

[0078] In the embodiment of the present invention, the data fusion module 403 includes a data alignment unit 4031, a feature fusion unit 4032, a model fusion unit 4033 and an output unit 4034;

[0079] The data alignment unit 4031 aligns data of different modalities in terms of time, space and features;

[0080] Specifically: for time series data, ensure that data of different modalities are aligned in the time dimension;

[0081] For image or video data, ensure that data from different modalities are aligned in spatial dimensions;

[0082] Through mapping or transformation, the features of different modalities are represented in the same scale or space;

[0083] The feature fusion unit 4032 extracts feature data from the data alignment unit 4031 through a generative adversarial network and fuses the feature data using a multi-core learning algorithm;

[0084] The model fusion unit 4033 is used to select a suitable model framework for the fused feature data and fuse the models using a simple average fusion or weighted average fusion method;

[0085] The output unit 4034 is used to store the fused data.

[0086] In the embodiment of the present invention, the details of feature fusion performed by the feature fusion unit 4032 are as follows:

[0087] Step A: Design a generator and a discriminator. The network structures of the generator and discriminator can be designed using CNN, RNN, or Transformer. The generator is used to generate data similar to real data, and the discriminator is used to distinguish real data from generated data. By alternately training the generator and discriminator, the generator generates data close to the real data, and the trained generator or discriminator is used to extract features of the aligned data.

[0088] The detailed steps of alternating training of the generator and discriminator in step A are as follows:

[0089] Step A1: Initialize the parameters of the generator and discriminator, set the loss function (such as binary cross entropy loss) and optimizer (such as Adam or SGD), sample real data from the real data distribution, generate fake data through the generator, calculate the loss of the discriminator for the real data, usually the cross entropy loss between the discriminator output and the real label, calculate the loss of the discriminator for the fake data, usually the cross entropy loss between the discriminator output and the fake label, add the loss of the real data and the loss of the fake data to get the combined loss, and use the optimizer to update the parameters of the discriminator;

[0090] In step A2, fake data is generated through the generator, and the loss of the generator is calculated, which is the cross entropy loss between the discriminator's output of the fake data and the true label. The parameters of the generator are updated using the optimizer. The discriminator and generator are alternately trained in step A1, and the number of training times is set to control the training process. During or after the training, features are extracted from specific layers of the generator or discriminator.

[0091] Step A3: Use dimensionality reduction techniques such as principal component analysis to reduce the extracted features to 2D or 3D and visualize them. Use different colors or markers to distinguish different categories of data. Calculate the similarity between the extracted features, such as cosine similarity and Euclidean distance, to verify whether similar data has similar feature representations and whether different data has different feature representations, so as to determine whether the features extracted by the generator or discriminator are required for processing.

[0092] Step B: Select a suitable predefined kernel function based on the extracted feature data, such as Gaussian kernel, polynomial kernel, linear kernel, etc. The selection criteria include the distribution, dimension and intrinsic structure of the data, assign an initial weight to each kernel function, which can usually be set to equal weight or random weight, and automatically adjust the weight of each kernel function through a multi-kernel learning algorithm, such as gradient descent, convex optimization, etc. Regularization terms, such as L1 or L2 regularization, are added during the optimization process to prevent overfitting and improve the generalization ability of the model. According to the learned weights, the kernel function matrix of each modality is weighted and combined to form a new kernel function matrix. The combined kernel function is standardized to ensure that it meets the properties of the kernel function, such as positive definiteness, and a unified kernel function that can effectively integrate multiple modal features is constructed to provide a powerful feature representation for subsequent classification or regression tasks;

[0093] The detailed steps for automatically adjusting the kernel function weights in step B are as follows:

[0094] Step B1, defining an objective function based on a weighted combination of kernel functions, for example, minimizing the classification error rate, ensuring that the sum of the kernel function weights is 1;

[0095] Step B2: Select an optimization algorithm to optimize, such as using the convex property of the kernel function to find the optimal weight by solving a convex optimization problem;

[0096] Step B3: Initially, the weights of all kernel functions are equal. The weights of the kernel functions are updated according to the selected optimization algorithm until the constraints are met. The kernel matrices of each modality are weighted combined using the learned weights to form a new kernel matrix.

[0097] Step C: The combined kernel function matrix defines a new feature space, which is a weighted combination of all modal feature spaces. In this new space, the features of different modalities are effectively fused together. For any two data points, their similarity or distance in the new feature space is calculated by combining the kernel function matrix. The similarity or distance value represents the fused feature. The similarity can be measured using cosine similarity, Pearson correlation coefficient, and other metrics to measure the similarity between feature vectors. The distance can be used to measure the difference between feature vectors using Euclidean distance, Manhattan distance, or Mahalanobis distance.

[0098] The detailed fusion steps of the model fusion unit 4033 are as follows:

[0099] Step a: Select an appropriate model architecture based on the task requirements, such as support vector machine, random forest, neural network, etc., divide the fused data into training set, validation set, and test set, initialize the model parameters, set the random seed to ensure repeatability, train the model using the training set data, adjust the model parameters through forward propagation and backpropagation, use the validation set to tune hyperparameters such as learning rate, regularization coefficient, batch size, etc., use the validation set to evaluate the model performance, and record indicators such as loss function value and accuracy;

[0100] Step b: Select an appropriate fusion method to perform model fusion on the trained models after feature fusion. After the fusion model, use the validation set for preliminary verification to ensure the fusion effect, and perform a simple average of the prediction results of all models, or perform a weighted average of the prediction results of each model according to the pre-calculated weights to obtain the final fusion prediction result.

[0101] Details on which fusion method to choose:

[0102] Check the distribution of the data after feature fusion to see if it is uniform. Use feature selection techniques, such as tree-based methods and linear model coefficients, to assess the importance of each feature. Train multiple basic models on the data after feature fusion. Use cross-validation to evaluate the performance of each model, recording indicators such as accuracy, recall, and F1 score. Compare the performance indicators of each model to see if there are significant differences. Analyze the stability of model performance to see if it is significantly affected by data fluctuations. If the model performance is similar and the data distribution is uniform, choose simple average fusion. If there are significant differences in model performance, choose weighted average fusion and assign weights based on the performance differences.

[0103] Step c: Use the validation set to verify the fusion model to ensure that there is no overfitting. Evaluate the final performance of the fusion model on the test set and analyze the model's performance indicators, such as accuracy, recall, and F1 score.

[0104] In step d, the trained fusion model is deployed in the actual application environment to predict the overturning state of the scaffold frame 100, monitor the model input, output and performance indicators, ensure the stable operation of the model, and regularly update the model according to new data or environmental changes.

[0105] In an embodiment of the present invention, the multimodal learning architecture 404 module includes a dual-stream network establishment unit 4041 and an attention mechanism integration unit 4042;

[0106] The dual-stream network establishment unit 4041 is used to design a stream for each modality when designing the generator and discriminator. For example, one stream is dedicated to processing meteorological data, and the other stream is dedicated to processing scaffolding data. When training the generator and discriminator, each stream is trained separately to ensure that they can effectively extract the features of their respective modalities. In the feature fusion stage, the output features of the dual-stream network are combined, and splicing, summing, or other fusion strategies can be used. In the model fusion stage, the features extracted by the dual-stream network can be used as input for further model training and optimization.

[0107] The attention mechanism integration unit 4042 uses the attention mechanism to dynamically weight different modalities, so that the model can adaptively focus on more important information according to task requirements, improve the interpretability of the model, understand the decision basis of the model through the attention weight, integrate the attention mechanism in the network structure of the generator and the discriminator, and add the Self-Attention layer after the feature extraction layer. During the training process, the attention mechanism will learn how to weight the features of different modalities. When fusing features, the weights output by the attention mechanism can be used to weightedly fuse the features of different modalities. As part of feature fusion, the attention mechanism dynamically adjusts the feature weights in the fusion process. In the model fusion stage, the attention mechanism can be used to adjust the weights of different model outputs to achieve weighted average fusion. When selecting a model architecture, a neural network with an attention mechanism can be integrated, such as Attention-based During RNN or CNN training, the attention mechanism will help the model focus on key features. During model fusion, the attention mechanism can be used to perform a weighted average of the prediction results of different models and distribute the contributions of different models according to the attention weights. During the verification and evaluation phases, the impact of the attention mechanism on model performance is analyzed to ensure that the attention mechanism effectively improves model performance. After the model is deployed, the weight distribution of the attention mechanism is continuously monitored to ensure the stability and adaptability of the model in practical applications.

[0108] Example 2

[0109] Refer to the attached Figure 6 As shown, an embodiment of the present invention further provides a method for constructing a scaffolding structure of a garbage pool wall structure, comprising the following steps:

[0110] S1. During the construction of the garbage pool wall, scaffolding is installed, and sensors are installed around the construction site to collect meteorological data of the construction environment in real time, such as wind speed sensors, wind direction sensors, temperature sensors, humidity sensors, etc. Sensors are arranged on the scaffolding to monitor the status of the frame in real time, such as vibration sensors, strain sensors, dynamometers, etc., wherein the dynamometer is used to monitor the tension changes of the steel wire rope 201, and an anti-overturning component 200 is installed at the installation node or specific node of the scaffolding frame 100. One end of the steel wire rope 201 is fixed to the scaffolding frame 100 with a rope clamp, and the other end passes through the connecting end of the connecting rod 202 and is screwed to ensure firmness, and a support component is installed at the bottom of the scaffolding frame to ensure that the output end of the motor is connected to the rotating rod in transmission, and the support rod, connecting block, and support plate form an adjustable support structure;

[0111] S2, the collected data is transmitted to the frame stability monitoring system 400 in real time, and data processing is performed to remove outliers and noise, and perform data alignment processing;

[0112] S3, fuses the gas phase data with the bionic sensor data to establish a scaffolding structure model, applies the fused data model to the structure model, uses the fused feature data to input the prediction model to predict the scaffolding overturning state, and optimizes the prediction results by combining the model fusion strategy;

[0113] S4, based on the simulation results, assesses the overturning risk of the frame. When a risk is detected, an early warning is issued to prompt construction workers to take measures;

[0114] S5. At the same time, the bottom support assembly 300 is controlled to adjust the balance according to the simulation results, and the rotating rod 302 is driven by the motor 301 to control the movement of the connecting block 303 and the support rod 304, so as to adjust the lifting and lowering of the support plate 305, correct the tilt position of the scaffolding, and perform emergency treatment until the scaffolding returns to normal state, restore the support assembly 300, continue construction, and complete the work of the garbage pool wall.

[0115] The above is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

[0116] It should be understood that the specific order or hierarchy of steps in the disclosed processes is an example of an exemplary method. Based on design preferences, it should be understood that the specific order or hierarchy of steps in the process can be rearranged without departing from the scope of the present disclosure. The accompanying method claims present elements of the various steps in an exemplary order and are not intended to be limited to a specific order or hierarchy.

[0117] In the foregoing detailed description, various features are grouped together in a single embodiment to simplify the disclosure. This method of disclosure should not be interpreted as reflecting an intention that embodiments of the claimed subject matter require more features than are expressly recited in each claim. On the contrary, as reflected in the appended claims, the invention comprises less than all the features of any individual disclosed embodiment. The appended claims are therefore hereby expressly incorporated into the detailed description, with each claim standing on its own as a separate preferred embodiment of the invention.

[0118] Those skilled in the art will also appreciate that the various illustrative logic blocks, modules, circuits, and algorithmic steps described in conjunction with the embodiments herein may be implemented as electronic hardware, computer software, or a combination thereof. In order to clearly illustrate the interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps described above are generally described around their functions. Whether such functions are implemented as hardware or software depends on the specific application and the design constraints imposed on the entire system. A skilled person may implement the described functions in an adaptable manner for each specific application, but such implementation decisions should not be interpreted as departing from the scope of protection of this disclosure.

[0119] The steps of the methods or algorithms described in conjunction with the embodiments herein may be directly embodied as hardware, software modules executed by a processor, or a combination thereof. The software module may be located in a RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, register, hard disk, removable disk, CD-ROM, or any other form of storage medium well known in the art. An exemplary storage medium is connected to the processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium may also be an integral part of the processor. The processor and storage medium may be located in an ASIC. The ASIC may be located in a user terminal. Of course, the processor and storage medium may also be present in a user terminal as discrete components.

[0120] For software implementation, the techniques described in this application can be implemented using modules (e.g., procedures, functions, etc.) that perform the functions described in this application. These software codes can be stored in a memory unit and executed by a processor. The memory unit can be implemented within the processor or external to the processor. In the latter case, it is communicatively coupled to the processor via various means, which are well known in the art.

[0121] The foregoing description includes examples of one or more embodiments. Of course, it is not possible to describe all possible combinations of components or methods for the purposes of describing the above embodiments, but one of ordinary skill in the art will recognize that the various embodiments may be further combined and arranged. Therefore, the embodiments described herein are intended to encompass all such changes, modifications and variations that fall within the scope of the appended claims. Furthermore, to the extent the term "comprising" is used in the specification or claims, the term is intended to be encompassed in a manner similar to the term "including," as explained in terms of "including," used as a transitional word in the claims. Furthermore, any use of the term "or" in the specification of the claims is intended to mean a "non-exclusive or."

Claims

1. A scaffolding structure for the construction of a garbage pool wall structure, characterized in that: The scaffolding frame (100) comprises a scaffolding frame (100), wherein an oblique support rod (101) is provided on the outside of the scaffolding frame (100), an oblique support rod auxiliary rod (102) is provided between the oblique support rod (101) and the scaffolding frame (100), and movable wheels (104) are provided at the four corners of the bottom of the scaffolding frame (100); An anti-overturning assembly (200), the anti-overturning assembly (200) is arranged on the outside of the scaffold frame (100), the anti-overturning assembly (200) comprises a steel wire rope (201), a connecting rod (202), a movable rod (204) and a sleeve (205), the connecting rod (202) is arranged at one end of the steel wire rope (201), the movable rod (204) is fixed to the outer end of the connecting rod (202), and the sleeve (205) is sleeved on the outside of the movable rod (204); A frame stability monitoring system (400) is used to monitor the stability of a scaffold and provide an early warning of a scaffold overturning. The frame stability monitoring system (400) comprises a bionic sensor network (401), a meteorological data acquisition module (402), a data fusion module (403), a multimodal learning architecture (404) module, and an early warning module (405); The bionic sensor network (401) collects real-time change data of the scaffold through bionic sensors; The meteorological data acquisition module (402) monitors the meteorological data of the garbage pit construction environment through real-time data from a weather station or installed sensors; The data fusion module (403) performs data fusion on the real-time data of the bionic sensor network (401) and the meteorological data acquisition module (402) through different aspects, and performs data prediction; The multimodal learning architecture (404) module is used to effectively fuse data from different sources so that the model can simultaneously use information from these data to make predictions, thereby assisting the data fusion module (403); The early warning module (405) issues an early warning prompt based on the prediction result of the data fusion module (403).

2. A scaffolding structure for construction of a garbage pool wall structure according to claim 1, characterized in that: The scaffolding frame (100) further comprises a support assembly (300), wherein the support assembly (300) is arranged at the bottom of the scaffolding frame (100), and the support assembly (300) comprises a motor (301), a rotating rod (302), a connecting block (303), a supporting rod (304) and a supporting plate (305). The rotating rod (302) is arranged at the output end of the motor (301), two connecting blocks (303) are sleeved on the outer side of the rotating rod (302), two supporting rods (304) are respectively arranged on the outer sides of the two connecting blocks (303), and support plates (305) are arranged at the ends of the two supporting rods (304).

3. A scaffolding structure for construction of a garbage pool wall structure according to claim 2, characterized in that: The output end of the motor (301) is connected to the rotating rod (302) in a transmission manner. Threaded sections (306) are provided near the two ends of the rotating rod (302). The two connecting blocks (303) are respectively threadedly engaged with the rotating rod (302) through the threaded sections (306). The two support rods (304) are respectively provided on the upper and lower sides of the connecting block (303). The two ends of the support rod (304) are respectively bolted to the connecting block (303) and the support plate (305). The support rod (304) and the rotating rod (302) on the same side are designed in a triangular shape. The support plate (305) is screwed to the scaffold frame (100).

4. A scaffolding structure for construction of a garbage pool wall structure according to claim 3, characterized in that: A plurality of anti-overturning components (200) are provided on the scaffolding frame (100), the other end of the steel wire rope (201) is mounted on the scaffolding frame (100) by means of a rope clamp, a connecting end (203) is provided at the end of the connecting rod (202), the other end of the steel wire rope (201) passes through the connecting end (203) and is screwed to the connecting end (203), and the movable rod (204) passes through the sleeve (205).

5. A scaffolding structure for construction of a garbage pool wall structure according to claim 4, characterized in that: A spring (206) is provided between the movable rod (204) and the sleeve (205), and the spring (206) surrounds the outer side of the movable rod (204). Cover plates (207) are provided near both ends of the movable rod (204), and the cover plates (207) are adapted to the sleeve (205) and threadedly engaged with the sleeve (205). The two ends of the spring (206) are respectively screwed to the cover plates (207) and the movable rod (204), and a fixing rod is fixed to the outer side of one of the cover plates (207), and the other cover plate (207) is movably connected to the movable rod (204).

6. A scaffolding structure for construction of a garbage pool wall structure according to claim 5, characterized in that: The data fusion module (403) includes a data alignment unit (4031), a feature fusion unit (4032), a model fusion unit (4033) and an output unit (4034); The data alignment unit (4031) performs alignment processing on data of different modalities in terms of time, space and features; The feature fusion unit (4032) extracts feature data from the data alignment unit (4031) through a generative adversarial network and fuses the feature data using a multi-core learning algorithm; The model fusion unit (4033) is used to select a suitable model framework for the fused feature data and fuse the models using a simple average fusion or weighted average fusion method; The output unit (4034) is used to store the fused data.

7. A scaffolding structure for construction of a garbage pool wall structure according to claim 6, characterized in that: The details of the feature fusion performed by the feature fusion unit (4032) are as follows: Step A: Design a generator and a discriminator. The generator is used to generate data similar to the real data, and the discriminator is used to distinguish the real data from the generated data. The generator and the discriminator are trained alternately, and the trained generator or discriminator is used to extract the features of the aligned data. Step B: Select an appropriate predefined kernel function based on the extracted feature data, assign an initial weight to each kernel function, automatically adjust the weight of each kernel function through a multi-kernel learning algorithm, perform a weighted combination of the kernel function matrices of each modality based on the learned weights to form a new kernel function matrix, and perform standardization on the combined kernel function; Step C: For any two data points, their similarity or distance in the new feature space is calculated by combining the kernel function matrix. The similarity or distance value represents the fused feature. The detailed fusion steps of the model fusion unit (4033) are as follows: Step a: Select an appropriate model architecture based on the task requirements, divide the fused data into training set, validation set, and test set, initialize the model parameters, train the model using the training set data, and adjust the model parameters through forward propagation and backpropagation; Step b: Select an appropriate fusion method to perform model fusion on the trained models after feature fusion, and perform a simple average of the prediction results of all models, or perform a weighted average of the prediction results of each model according to the pre-calculated weights to obtain the final fusion prediction result; Step c: Use the validation set to verify the fusion model, evaluate the final performance of the fusion model on the test set, and analyze the performance indicators of the model; In step d, the trained fusion model is deployed in the actual application environment, the model input, output and performance indicators are monitored, and the model is updated regularly according to new data or environmental changes.

8. A scaffolding structure for construction of a garbage pool wall structure according to claim 7, characterized in that: The multimodal learning architecture (404) module includes a dual-stream network establishment unit (4041) and an attention mechanism integration unit (4042); The dual-stream network establishment unit (4041) is used to design a stream for each modality when designing the generator and the discriminator, and to train each stream separately when training the generator and the discriminator; The attention mechanism integration unit (4042) uses the attention mechanism to dynamically weight different modalities, so that the model can adaptively focus on more important information according to task requirements.

9. A method for constructing a scaffolding structure for a garbage pit wall structure, applied to a scaffolding structure for constructing a garbage pit wall structure according to any one of claims 1 to 8, characterized in that: The following steps are involved: S1, during the construction of the garbage pool wall, scaffolding is installed, sensors are installed around the construction site to collect real-time weather data of the construction environment, and sensors are deployed on the scaffolding to monitor the status of the frame in real time; S2, the collected data is transmitted to the frame stability monitoring system (400) in real time, and data processing is performed to remove abnormal values and noise, and data alignment is performed; S3, fuses the gas phase data with the bionic sensor data to establish a scaffolding structure model, and applies the fused data model to the structure model for simulation; S4, based on the simulation results, assesses the overturning risk of the frame. When a risk is detected, an early warning is issued to prompt construction workers to take measures; S5, at the same time, the support assembly (300) at the bottom is controlled to adjust the balance according to the simulation results, and emergency treatment is performed until the scaffolding returns to a normal state, the support assembly (300) is restored, and construction is continued to complete the work of the garbage pool wall.

Images