AI-based anti-seismic scaffold design model optimization method
By collecting multi-source vibration data in real time and optimizing handbag design using deep neural network models, a personalized optimization solution is generated, and the problem of poor stability in traditional design is solved, and stability improvement and construction safety guarantees are achieved in complex vibration environments.
Patent Information
- Application Number
- CN202510413688.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-03
- Publication Date
- 2025-07-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional handbag design cannot fully consider complex vibration factors, resulting in poor stability in vibration environments, and risks of loose connections, structural deformation and collapse. The existing seismic design lacks intelligent processing capabilities and is difficult to personalize the use of different construction scenarios.
By collecting multi-source vibration data in real time, using deep neural network models to build a seismic handbag design optimization model, monitoring and generating targeted optimization solutions in real time, including adjusting the pole layout, wall connecting parts settings, etc., and using AI for dynamic evaluation and optimization.
It improves the stability of hand and scaffolds in complex vibration environments, reduces the number of damage and reconstructions, ensures construction safety, reduces costs, improves construction efficiency and continuity, and is suitable for various construction projects.
Smart Images
Figure CN120354488A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of construction engineering, and particularly relates to an optimization method for an anti-seismic scaffolding design model based on AI. Background Art
[0002] During the construction process, as an important support structure for construction workers to work and store materials, the scaffolding is often affected by various vibration sources. Vibrations generated by mechanical operations in the urban environment, transportation, and the operation of large equipment inside the building, etc., all pose threats to the stability of the scaffolding. Traditional scaffolding design mainly relies on experience and simple mechanical calculations, and it is difficult to comprehensively consider these complex vibration factors. Its design method cannot accurately adapt to the changes in different vibration frequencies, amplitudes, and durations, resulting in problems such as loose connections, structural deformation, and even collapse of the scaffolding in a vibrating environment during actual use, seriously endangering the lives of construction workers, causing construction delays and economic losses. The existing anti-seismic design methods lack the ability to deeply analyze and intelligently process vibration data, and it is difficult to achieve personalized adaptation to different construction scenarios and vibration conditions. There is an urgent need for an innovative design method to solve these problems. Summary of the Invention
[0003] The purpose of the present invention is to provide an optimization method for an anti-seismic scaffolding design model based on AI to solve the following technical problems.
[0004] The purpose of the present invention can be achieved through the following technical solutions: An optimization method for an anti-seismic scaffolding design model based on AI includes the following steps: S1. Multi-source vibration data collection and preprocessing: Real-time collect vibration data from different vibration sources, the structural parameters of the scaffolding itself, and the load data during the construction process, and perform preprocessing; S2. AI model construction and training: Select a deep neural network model to construct an anti-seismic scaffolding design optimization model, and use the preprocessed vibration data, scaffolding structural parameters, and load data as input vectors for training; S3. Real-time monitoring and dynamic evaluation: After the scaffolding is erected and put into use, continuously collect vibration data in real time through sensors, and transmit the data to the trained AI model in real time; S4. Generation of design optimization decisions: When the AI model evaluates that the anti-seismic performance of the scaffolding may not meet the safety requirements, the model automatically generates a targeted design optimization plan based on the optimization strategy obtained through training.
[0005] As a further solution of the present invention: The specific steps of S1 are as follows: Arrange a variety of sensors at the construction site and its surrounding environment to collect vibration data from different vibration sources in real time, including acceleration data a(t), displacement data s(t), and velocity data v(t) at a certain moment t; Collect the structural parameters of the scaffolding itself and the load data during the construction process. The structural parameters include the vertical rod spacing l, the horizontal rod step distance h, and the position and quantity of the wall connection members; Perform filtering processing on the collected vibration data, and adopt the Kalman filtering algorithm. Its prediction equation is: ; ; Wherein, is the prior state estimate at time t, F t is the state transition matrix, is the posterior state estimate at time t−1, B t is the control input matrix, u t is the control input vector, P t∣t−1 is the prior estimate error covariance, Q t is the process noise covariance; The update equation is: ; ; ; Wherein, K t is the Kalman gain, H t is the observation matrix, z t is the observation value at time t, R t is the observation noise covariance, is the posterior state estimate at time t, P t∣t is the posterior estimate error covariance; Convert the time-domain vibration data into frequency-domain data through Fourier transform, and extract the key features of the vibration by using the discrete Fourier transform formula. The key features include the characteristic frequency and the energy distribution. The formula is: ; Wherein, k = 0, 1,..., N-1, x(n) is the time-domain signal, X(k) is the frequency-domain signal, and N is the signal length.
[0006] As a further solution of the present invention: The specific steps of S2 are as follows: Select a deep neural network model to construct an optimized model for seismic scaffolding design. Use the preprocessed vibration data, scaffolding structure parameters, and load data as the input vector X = (x1, x2,..., x m ), and use the stability evaluation results of the scaffolding under different vibration conditions as the output label y; For the long short-term memory network, the input gate i of its core unit t , forget gate f t , output gate o t and cell state C t are updated according to the following formulas: ; ; ; ; ; ; where σ represents the activation function of the neural network, tanh is the hyperbolic tangent function, W is the weight matrix, b is the bias vector, ⊙ is element-wise multiplication, h t−1 is the hidden state at the previous moment, and x t is the input at the current moment.
[0007] As a further solution of the present invention: In the said S2: For the convolutional neural network, including a convolutional layer, a pooling layer, and a fully connected layer, the convolutional layer performs a convolution operation on the input data through a convolution kernel W. The convolution operation formula is: ; where x is the input data, w is the convolution kernel element, b is the bias, y is the convolution output, i and j are used to locate the element position in the output feature map y, and y i,j represents the element in the i-th row and j-th column of the output feature map; The pooling layer performs downsampling on the output of the convolutional layer. The commonly used max pooling formula is: ; where S is the pooling window.
[0008] As a further solution of the present invention: It further includes: According to historical data and real-time collected data, continuously adjust the weights and parameters of the model through the backpropagation algorithm, where the loss function to be minimized is: ; where i is used to represent the index of the sample, N is the number of samples, j is used to represent the category index, K is the number of categories, and yij represents the true label of the \(i\)-th sample in the \(j\)-th category, represents the predicted probability that the model assigns the \(i\)-th sample to the \(j\)-th category.
[0009] As a further solution of the present invention: The specific steps of S3 are as follows: After the scaffolding is built and put into use, continuously collect vibration data in real time through sensors and transmit the data to the trained AI model in real time; The AI model, based on the input real-time data \(X\) new , combines the learned knowledge to dynamically evaluate the seismic performance of the current scaffolding and predict the stability change trend of the scaffolding during subsequent vibrations; Let the output of the model be , representing the probability distribution of the scaffolding in different stability states.
[0010] As a further solution of the present invention: In S5: When the AI model evaluates and finds that the seismic performance of the scaffolding does not meet the safety requirements, the model automatically generates a targeted design optimization plan based on the trained optimization strategy; The optimization plan includes adjusting the layout of vertical poles and horizontal bars and optimizing the settings of connecting wall members.
[0011] As a further solution of the present invention: Adjusting the layout of vertical poles includes increasing the density of vertical poles, and the specific steps are as follows: Let the original spacing between vertical poles be \(l_0\) and the adjustment coefficient be \(\alpha\); The adjusted spacing between vertical poles is \(l_1 = \alpha l_0\), where \(0 < \alpha < 1\).
[0012] Advantages of the present invention: By comprehensively collecting multi-source vibration data and leveraging the powerful learning ability of AI, the scaffolding design can accurately consider the influence of different vibration characteristics, effectively improve the stability in complex vibration environments, and ensure construction safety; With real-time monitoring and dynamic evaluation of the AI model, potential safety hazards during the vibration process of the scaffolding can be promptly detected, and an optimization plan can be quickly generated to achieve real-time adjustment of the scaffolding, greatly improving the timeliness and effectiveness in dealing with vibration risks; Reduce the number of damages and reconstructions of the scaffolding caused by vibrations, avoid construction delays, reduce construction costs, and at the same time improve the continuity and efficiency of the construction process, with significant economic benefits; The present invention can generate personalized optimization plans according to different construction scenarios, vibration conditions, and scaffolding structure characteristics, and is applicable to various types of building construction projects, with strong versatility and scalability. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] The present invention will be further described below with reference to the accompanying drawings.
[0014] Figure 1 It is a schematic flow chart of an optimization method for an AI-based anti-seismic scaffolding design model of the present invention. Specific implementation manners
[0015] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.
[0016] Please refer to Figure 1 As shown, the present invention is an optimization method for an AI-based anti-seismic scaffolding design model, including the following steps: S1. Multi-source vibration data collection and preprocessing: At some large-scale construction sites, sensors are installed at key nodes of the scaffolding, at different height positions, and near equipment that may generate vibrations around. Vibration data and relevant construction data are continuously collected for one week. The Kalman filtering algorithm is used to denoise the collected vibration data, and the Fourier transform is used to convert the time-domain data into frequency-domain data to extract features such as the main frequency components and energy ratios of the vibrations.
[0017] Vibration data collection: A variety of sensors are arranged at the construction site and its surrounding environment, including acceleration sensors, displacement sensors, velocity sensors, etc., to collect vibration data from different vibration sources in real time. Let the acceleration data collected at a certain moment t be a(t), the displacement data be s(t), and the velocity data be v(t); Data filtering: Filter the collected vibration data to remove noise interference and outliers. The Kalman filtering algorithm is adopted, and its prediction equation is: ; ; Among them, is the prior state estimate at time t, F t is the state transition matrix, is the posterior state estimate at time t−1, B t is the control input matrix, u t is the control input vector, P t∣t−1 is the prior estimate error covariance, Q t is the process noise covariance; The update equation is: ; ; ; Among them, K t is the Kalman gain, H t is the observation matrix, z t is the observed value at time t, R t is the observation noise covariance, is the posterior state estimate at time t, P t∣t is the posterior estimation error covariance; Feature extraction: Convert the vibration data in the time domain to the frequency domain data through Fourier transform, and extract key features such as the characteristic frequency and energy distribution of the vibration. The discrete Fourier transform formula is: ; Among them, k = 0, 1,..., N - 1, x(n) is the time-domain signal, X(k) is the frequency-domain signal, and N is the signal length.
[0018] At the same time, collect the structural parameters of the scaffolding itself, such as the vertical pole spacing l, the horizontal bar step distance h, the position and quantity of the wall connecting members, etc., as well as the load data during the construction process, such as the weight distribution of personnel and materials, etc.
[0019] S2. AI model construction and training: Select the LSTM network to construct the seismic scaffolding design optimization model, and set parameters such as the number of network layers and the number of nodes. Divide the preprocessed data into a training set and a test set according to a ratio of 7:3, and use the training set to train the model for three days. During the training process, continuously adjust hyperparameters such as the learning rate until the accuracy of the model on the test set reaches more than 90%.
[0020] Select a deep neural network model, such as a convolutional neural network (CNN) or a long short-term memory network (LSTM), to construct the seismic scaffolding design optimization model. Use the preprocessed vibration data (including characteristic frequency, energy distribution, etc.), scaffolding structure parameters (l, h, etc.) and load data as the input vector X=(x1, x2,..., x m ), and use the stability evaluation results (stable, slightly shaking, significantly deformed, etc.) of the scaffolding under different vibration conditions as the output label y.
[0021] For the long short-term memory network, the update formulas for the input gate i t , forget gate f t , output gate o t and cell state C t are: ; ; ; ; ; ; Among them, σ represents the activation function of the neural network, tanh is the hyperbolic tangent function, W is the weight matrix, b is the bias vector, ⊙ is element-wise multiplication, h t−1 is the hidden state at the previous moment, and x t is the input at the current moment.
[0022] For a convolutional neural network, including a convolutional layer, a pooling layer, and a fully connected layer, the convolutional layer performs a convolution operation on the input data through a convolution kernel W. The convolution operation formula is: ; where x is the input data, w is the convolution kernel element, b is the bias, y is the convolution output, i and j are used to locate the element position in the output feature map y, and y i,j represents the element in the i-th row and j-th column of the output feature map; The pooling layer downsamples the output of the convolutional layer. The commonly used max pooling formula is: ; where S is the pooling window.
[0023] Using a large amount of historical data and real-time collected data, continuously adjust the weights and parameters of the model through the backpropagation algorithm, so that the model learns the complex relationship between the vibration characteristics and the stability of the scaffolding. The goal of training is to minimize the loss function L. Commonly used loss functions such as the cross-entropy loss function: ; where i is used to represent the index of the sample, N is the number of samples, j is used to represent the class index, K is the number of classes, y ij represents the true label of the i-th sample in the j-th class, represents the predicted probability that the model assigns the i-th sample to the j-th class.
[0024] S3. Real-time Monitoring and Dynamic Evaluation: After the scaffolding is built and put into use, continuously collect vibration data in real time through sensors and transmit the data to the trained AI model in real time. The AI model dynamically evaluates the seismic performance of the current scaffolding based on the input real-time data X new , combined with the learned knowledge, and predicts the stability change trend of the scaffolding during subsequent vibrations. Let the output of the model be , representing the probability distribution of the scaffolding in different stability states.
[0025] S4. Design Optimization Decision Generation: When the AI model evaluates and finds that the seismic performance of the scaffolding may not meet the safety requirements, the model automatically generates a targeted design optimization plan based on the optimized strategy obtained through training. The optimization plan may involve adjusting the layout of the vertical poles and horizontal bars. For example, increasing the density of the vertical poles. Suppose the original spacing between vertical poles is \(l_0\), the adjusted spacing between vertical poles is \(l_1\), and the adjustment coefficient is \(\alpha\), then \(l_1 = \alpha l_0\), where \(0 < \alpha < 1\); changing the connection method of the horizontal bars; optimizing the settings of the wall connecting members. For example, increasing the number of wall connecting members. Suppose the original number of wall connecting members is \(n_0\), the increased number of wall connecting members is \(n_1\), and the increase coefficient is \(\beta\), then \(n_1=(1 + \beta)n_0\), where \(\beta>0\); or even replacing the material with higher strength and greater toughness. The construction workers make corresponding adjustments and reinforcements to the scaffolding according to the optimization plan generated by the AI model to ensure its safety in a vibrating environment.
[0026] The above has described an embodiment of the present invention in detail, but the described content is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the present invention application should still fall within the scope covered by the patent of the present invention.
Claims
1. An optimization method for an AI-based seismic scaffolding design model, characterized in that, It includes the following steps: S1. Multi-source vibration data acquisition and preprocessing: Vibration data from different vibration sources, the structural parameters of the scaffolding itself, and the load data during the construction process are collected in real time and preprocessed; S2. AI model construction and training: Select a deep neural network model to construct an anti-seismic scaffolding design optimization model, and use the preprocessed vibration data, scaffolding structural parameters, and load data as input vectors for training; S3. Real-time monitoring and dynamic evaluation: After the scaffolding is erected and put into use, vibration data is continuously collected in real time through sensors, and the data is transmitted to the trained AI model in real time; S4. Generation of design optimization decisions: When the AI model evaluates and finds that the anti-seismic performance of the scaffolding does not meet the safety requirements, the model automatically generates a targeted design optimization plan based on the optimized strategy obtained from training.
2. The optimization method for the anti-seismic scaffolding design model based on AI according to claim 1, wherein The specific steps of S1 are as follows: A variety of sensors are arranged at the construction site and its surrounding environment to collect vibration data from different vibration sources in real time, including acceleration data a(t), displacement data s(t), and velocity data v(t) at a certain moment t; Collect the structural parameters of the scaffolding itself and the load data during the construction process. The structural parameters include the vertical rod spacing l, the horizontal rod step distance h, and the position and quantity of the connecting wall members; Filter the collected vibration data. The Kalman filter algorithm is used, and its prediction equation is: ； ； Among them, is the prior state estimate at time t, and F t is the state transition matrix, is the posterior state estimate at time t−1, and B t is the control input matrix, and u t is the control input vector, and P t∣t−1 is the prior estimate error covariance, and Q t is the process noise covariance; The update equation is: ； ； ； Among them, K t is the Kalman gain, H t is the observation matrix, z t is the observation value at time t, R t is the observation noise covariance, is the posterior state estimate at time t, P t∣t is the posterior estimation error covariance; The time-domain vibration data is converted into frequency-domain data through Fourier transform, and the discrete Fourier transform formula is used to extract the key features of the vibration. The key features include the characteristic frequency and energy distribution. The formula is: ； where k = 0, 1,..., N - 1, x(n) is the time-domain signal, X(k) is the frequency-domain signal, and N is the signal length.
3. The optimization method of an AI-based seismic scaffolding design model according to claim 1, characterized in that The specific steps of S2 are as follows: Select a deep neural network model and construct an optimization model for the design of seismic scaffolding. Use the preprocessed vibration data, scaffolding structure parameters, and load data as the input vector X=(x1,x2,...,x m ), and use the stability evaluation results of the scaffolding under different vibration conditions as the output label y; For the long short-term memory network, the update formulas for the input gate i t , forget gate f t , output gate o t and cell state C t are as follows: ； ； ； ； ； ； Among them, σ represents the activation function of the neural network, tanh is the hyperbolic tangent function, W is the weight matrix, b is the bias vector, ⊙ is element-wise multiplication, and h t−1 is the hidden state at the previous moment, and x t is the input at the current moment.
4. A method for optimizing the design model of an earthquake-resistant scaffolding based on AI according to claim 1, characterized in that, In S2: For the convolutional neural network, including the convolutional layer, pooling layer, and fully connected layer, the convolutional layer performs a convolution operation on the input data through the convolution kernel W. The convolution operation formula is: ； where x is the input data, w is the convolutional kernel element, b is the bias, y is the convolutional output, and i and j are used to locate the element position in the output feature map y, and y i,j represents the element in the i-th row and j-th column of the output feature map; The pooling layer performs downsampling on the output of the convolutional layer. The commonly used maximum pooling formula is: ； where S is the pooling window.
5. A method for optimizing an AI-based anti-seismic scaffolding design model according to claim 3 or 4, characterized in that It also includes: According to historical data and real-time collected data, the weights and parameters of the model are continuously adjusted through the backpropagation algorithm. The minimized loss function is: ； where i is used to represent the index of the sample, N is the number of samples, j is used to represent the class index, K is the number of classes, and y ij represents the true label of the i-th sample in the j-th class, represents the predicted probability that the model assigns to the i-th sample belonging to the j-th class.
6. The optimization method of an AI-based seismic scaffolding design model according to claim 1, characterized in that, The specific steps of S3 are as follows: After the scaffolding is erected and put into use, vibration data is continuously collected in real time through sensors, and the data is transmitted to the trained AI model in real time; The AI model, based on the input real-time data X new , combines the learned knowledge to dynamically evaluate the seismic performance of the current scaffolding and predict the trend of stability change of the scaffolding during subsequent vibrations; Let the output of the model be , representing the probability distribution of the scaffolding in different stability states.
7. A method for optimizing the design model of an earthquake-resistant scaffolding based on AI according to claim 1, characterized in that, In S5: When the AI model evaluates and finds that the anti-seismic performance of the scaffolding does not meet the safety requirements, the model automatically generates a targeted design optimization plan based on the optimized strategy obtained from training; The optimization plan includes adjusting the layout of the vertical rods and horizontal rods and optimizing the setting of the connecting wall members.
8. An optimization method for an earthquake-resistant scaffolding design model based on AI according to claim 5, characterized in that, Adjusting the layout of the vertical rods includes increasing the density of the vertical rods. The specific steps are as follows: Let the original vertical rod spacing be l0 and the adjustment coefficient be α; The adjusted vertical rod spacing is l1 = αl0, where 0 < α < 1.