Optimal design method for seismic performance of reinforced concrete structure of hotel building

Abstract

The application relates to the technical field of building seismic design, and discloses a hotel building reinforced concrete structure seismic performance optimization design method. The method first acquires original structure data and dynamic response characteristics of a hotel building structure, constructs a structure system graph to identify key structure elements and seismic weak links; extracts structure characteristics of the key structure elements and calculates complexity characteristics, according to which optimal seismic design parameters of a user are determined; a structure response characteristic sequence is reconstructed and modal decomposition is carried out to obtain slow-varying components and fast-varying components; based on the two components, a seismic optimization design category is acquired, and an optimization design scheme is generated; finally, the scheme is dynamically corrected according to a user input signal, and a structure system graph is updated. The method can accurately identify structure weak links, meet personalized seismic requirements, and improve the seismic performance of a hotel building structure.

Description

Technical Field

[0001] This invention relates to the field of seismic design technology for buildings, specifically a method for optimizing the seismic performance of reinforced concrete structures in hotel buildings. Background Technology

[0002] As public buildings with high population density and complex functions, the seismic performance of hotel structures is directly related to the safety of life and property. With the rapid development of the construction industry, hotel buildings are trending towards higher heights, larger spans, and more multifunctional designs, which presents numerous challenges to the seismic design of reinforced concrete structures. Traditional seismic design methods are mostly based on empirical formulas and code requirements, using static analysis to determine structural parameters, making it difficult to fully consider the dynamic response characteristics of buildings under seismic loading.

[0003] In practical applications, the original structural data of hotel buildings are often scattered, including multiple aspects such as building dimensions, material properties, and component layout. The integration and analysis of this data has always been a challenge in seismic design. At the same time, seismic loads are random and uncertain, and traditional design methods cannot accurately capture the dynamic response of structures under seismic loading, which may result in seismic weaknesses in the designed structures, making them prone to damage in the event of a strong earthquake.

[0004] Different hotel buildings have varying seismic performance requirements due to differences in functional needs, geographical location, and architectural style. Traditional design methods often employ uniform design standards, which are insufficient to meet the individual needs of different buildings, potentially leading to over-design or under-design. With increasing emphasis on building safety and the continuous development of seismic design concepts, there is a need for a seismic performance optimization design method that can comprehensively consider structural dynamic response, identify weak points, and meet individualized requirements. Summary of the Invention

[0005] The purpose of this invention is to provide a method for optimizing the seismic performance of reinforced concrete structures in hotel buildings, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a method for optimizing the seismic performance of reinforced concrete structures in hotel buildings, the method comprising:

[0007] Obtain the original structural data and corresponding dynamic response characteristics of the hotel building structure;

[0008] Based on the original structural data and dynamic response characteristics, a structural system map is constructed;

[0009] Key structural elements and weak points in seismic resistance are identified through the structural system map.

[0010] Extract the structural features corresponding to the key structural elements;

[0011] Calculate the complexity features of the structural features;

[0012] Based on the key structural elements and complexity characteristics, determine the optimal seismic design parameters personalized for each user.

[0013] Based on the key structural elements and optimal seismic design parameters, a structural response characteristic sequence is reconstructed and generated.

[0014] Modal decomposition is performed on the structural response feature sequence to obtain the slow-varying components reflecting the steady-state characteristics of the structure and the fast-varying components reflecting the dynamic changes of the structure.

[0015] Based on the slow-changing component and the fast-changing component, the seismic optimization design category is obtained to generate a seismic optimization design scheme;

[0016] The seismic optimization design scheme is dynamically corrected based on user input signals, and the structural system map is updated according to the correction results.

[0017] Preferably, the acquisition of the original structural data and corresponding dynamic response characteristics of the hotel building structure includes:

[0018] Obtain the property characteristics of candidate building materials; the property characteristics include at least the elastic modulus and yield strength;

[0019] Based on the structural analysis algorithm, the attribute features, the dynamic response features, and the preset decomposition layer range are input to determine the optimal combination of material attributes and the corresponding decomposition layer from the candidate building material attributes.

[0020] Preferably, the construction of the structural system map includes:

[0021] A structural modeling algorithm is used to construct a graph of the original structural data based on the optimal combination of material properties and the corresponding number of decomposition layers.

[0022] Symmetrical extension processing is applied to the boundaries of the original structural data to reduce the impact of boundary distortion.

[0023] Obtain multiple structural node parameters, and construct the structural system map based on the multiple structural node parameters.

[0024] Preferably, the identification of key structural elements and seismic weak points includes:

[0025] The structural modal characteristics under different load states are obtained through a multi-stage acquisition process corresponding to the dynamic response characteristics; the multi-stage process includes at least static load state, dynamic load state, and seismic simulation state.

[0026] The key structural elements and weak points in seismic resistance are obtained based on the structural modal characteristics.

[0027] Preferably, determining the user-specific optimal seismic design parameters includes:

[0028] Based on the frequency band energy distribution characteristics, load state, and complexity characteristics of the dynamic response features, the adaptation score of the preset candidate design parameters is calculated through multi-layer weighted fusion; the adaptation score includes the matching degree score with the dominant structural mode, the load state adaptation score, and the complexity-related score.

[0029] An optimized search algorithm is used to dynamically weight the adaptation score, and the optimal seismic design parameters personalized to the user are determined from the candidate design parameters.

[0030] Preferably, the reconstructed structural response feature sequence includes:

[0031] Construct a time-domain mapping matrix; the matrix elements of the time-domain mapping matrix are generated by nonlinear combination of the key structural elements, the frequency matching degree corresponding to the optimal seismic design parameters, and the preset stress modulation coefficient;

[0032] Multi-scale time analysis is performed on the time-domain mapping matrix; the multi-scale time analysis includes short-scale analysis to capture structural response transformation characteristics, medium-scale analysis to track stress state, and long-scale analysis to monitor deformation trends.

[0033] By adaptively weighting and fusing the analysis results of multi-scale time analysis, a structural response feature sequence containing features of multiple structural response modes is generated.

[0034] Preferably, the step of performing mode decomposition on the structural response feature sequence to obtain the slowly varying components reflecting the steady-state characteristics of the structure and the rapidly varying components reflecting the dynamic changes of the structure includes:

[0035] The components of the structural response feature sequence are subjected to extreme point detection and mirror symmetric extension processing, and multi-order intrinsic structural components that satisfy the intrinsic mode conditions are obtained through iterative screening.

[0036] Based on the average period and energy distribution characteristics of the multi-order intrinsic structural components, the preceding short-period components corresponding to the multi-order intrinsic structural components are classified as fast-changing components reflecting structural response transformation, and the subsequent long-period components and residual terms corresponding to the multi-order intrinsic structural components are classified as slow-changing components reflecting structural maintenance.

[0037] Preferably, obtaining the seismic optimization design category includes:

[0038] The steady-state response feature vector is constructed by extracting time-domain statistical features, frequency band energy distribution features, and inter-component correlation features from the slowly varying components.

[0039] Extract the instantaneous deformation rate, peak value, and waveform morphology features from the rapidly changing components to construct a dynamic response feature vector;

[0040] A multi-stage classification strategy is adopted to classify the steady-state response feature vector, and an adaptive threshold detector is used to identify the response transition events corresponding to the dynamic response feature vector.

[0041] The seismic optimization design category is output by fusing the classification results and response transition events corresponding to the state classification through a probabilistic network, and combining the spatiotemporal continuity characteristics of the response transition to evaluate its credibility.

[0042] Preferably, the step of dynamically correcting the seismic optimization design scheme based on user input signals includes:

[0043] Receive user design constraint signals and determine the optimization operation type based on the user design constraint signals;

[0044] The seismic optimization design scheme is dynamically modified based on the optimization operation type; the optimization operation type includes at least structural node adjustment, connection relationship modification and weight dynamic update.

[0045] Preferably, updating the structural system map based on the correction result includes:

[0046] Implement the dynamically revised seismic optimization design scheme;

[0047] Real-time acquisition of feedback data during the execution process; the feedback data includes at least the stress distribution completion rate, the actual change time of the deformation path, and the change value of the nodal material utilization rate;

[0048] The structural node parameters and connection relationships in the structural system map are updated based on the feedback data.

[0049] Compared with the prior art, the beneficial effects of the present invention are:

[0050] Starting with acquiring raw structural data and dynamic response characteristics, a comprehensive and accurate foundation is laid for subsequent structural analysis and optimization. By constructing a structural system map, scattered structural information can be integrated and visualized, making the identification of key structural elements and seismic weak points more intuitive and accurate, which helps designers carry out targeted optimization work.

[0051] Extracting the structural features corresponding to key structural elements and calculating their complexity characteristics allows for a deeper understanding of the structure's intrinsic properties, providing a detailed basis for determining optimal seismic design parameters. This parameter determination method based on structural features and complexity avoids the blind reliance on experience in traditional design, making the design parameters more closely aligned with the actual conditions of the structure.

[0052] Reconstructing structural response characteristic sequences based on key structural elements and optimal seismic design parameters can simulate the structural response under different working conditions, providing reliable data support for subsequent modal decomposition. Modal decomposition of the structural response characteristic sequences separates the slowly varying components reflecting the steady-state characteristics of the structure from the rapidly varying components reflecting its dynamic changes. This helps to more clearly understand the stress mechanism and deformation patterns of the structure under seismic loading, thus providing a scientific reference for determining the seismic optimization design category.

[0053] Obtaining seismic optimization design categories and generating optimized design schemes based on slow-varying and fast-varying components ensures the relevance and effectiveness of the schemes, enabling them to better address potential structural risks during earthquakes. Dynamically correcting the optimized design schemes based on user input signals and updating the structural system map allows for dynamic adjustment and continuous optimization of the design process. This adapts to the personalized needs of different users and potential structural changes throughout its lifecycle, enhancing the flexibility and adaptability of the design methodology. Attached Figure Description

[0054] Figure 1 This is a schematic diagram illustrating the working principle of the seismic performance optimization design method for reinforced concrete structures in hotel buildings according to the present invention.

[0055] Figure 2 A flowchart for constructing a structural system diagram;

[0056] Figure 3 Flowchart for determining optimal seismic design parameters;

[0057] Figure 4 A flowchart for reconstructing the structural response feature sequence;

[0058] Figure 5 This is a flowchart for mode decomposition and component classification. Detailed Implementation

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] Please see Figure 1 This invention provides a method for optimizing the seismic performance of reinforced concrete structures in hotel buildings, the method comprising:

[0061] The process involves acquiring raw structural data and corresponding dynamic response characteristics of the hotel building structure. Raw structural data includes information such as building geometry, material properties, and component layout. Dynamic response characteristics are obtained through sensor measurements or simulation models, covering indicators such as displacement, acceleration, and stress. Based on the raw structural data and dynamic response characteristics, a structural system map is constructed. This map integrates structural connections, load transfer paths, and inherent dynamic characteristics to form a visualized topological network. The structural system map identifies key structural elements and seismic weak points. Key structural elements include beam-column joints and core tubes, while seismic weak points include stress concentration areas or modally unstable sections. Structural features corresponding to key structural elements are extracted, including geometric dimensions, material types, and boundary conditions. The complexity of these structural features is calculated; algorithms are used to analyze feature variability and correlation, quantifying the uncertainty of component behavior. Based on the key structural elements and complexity characteristics, personalized optimal seismic design parameters are determined. These parameters include reinforcement ratio, concrete strength grade, and joint reinforcement methods, considering user needs such as cost constraints or safety redundancy. Based on key structural elements and optimal seismic design parameters, a structural response characteristic sequence is reconstructed and generated; the sequence simulates the displacement and stress evolution process under seismic loading. Modal decomposition is performed on the structural response characteristic sequence to obtain slow-varying components reflecting the steady-state characteristics of the structure and fast-varying components reflecting the dynamic changes of the structure; the slow-varying components characterize the overall deformation trend, while the fast-varying components capture instantaneous fluctuations. Based on the slow-varying and fast-varying components, seismic optimization design categories are obtained; these categories are defined as different strengthening or optimization strategies, such as adding dampers or adjusting component stiffness, to generate seismic optimization design schemes. The seismic optimization design schemes are dynamically corrected based on user input signals, and the structural system map is updated according to the correction results; corrections include parameter adjustments or scheme iterations to ensure that the design schemes adapt to real-time requirements.

[0062] Example 1: See Figure 2In the initial stage of acquiring basic structural information for the hotel building, the raw structural data encompasses the building's geometric dimensions, component layout, material specifications, and node connection details. This data is extracted through building information modeling (BIM) or engineering drawings. Dynamic response characteristics are acquired through a sensor network installed at key structural nodes, including accelerometers, strain gauges, and displacement monitoring devices, recording the structural response behavior in real time under simulated seismic conditions. Sensor data includes time-series displacement changes, acceleration vibration waveforms, and stress distribution values, with sampling frequencies satisfying Nyquist's law to capture effective dynamic components. Attribute feature acquisition is performed using a candidate building material library, which pre-stores various concrete grades and steel reinforcement levels. The elastic modulus reflects the material's resistance to elastic deformation, while the yield strength defines the critical point at which the material transitions from the elastic to the plastic stage. Material selection considers regional supply capacity, avoiding the use of difficult-to-procure special alloys. The elastic modulus range of candidate materials covers 30 GPa to 50 GPa, and the yield strength range is set between 335 MPa and 500 MPa to ensure the engineering practicality of the material properties.

[0063] The structural analysis algorithm uses a finite element analysis engine as its core calculation module. This module takes into account the attribute feature set, dynamic response feature set, and a preset range of decomposition layers. The default range of decomposition layers is an integer between 3 and 7 layers, corresponding to different levels of structural performance analysis at different granularities. Level 3 focuses on the macroscopic overall structural stiffness assessment, level 5 analyzes the synergistic effect of component groups, and level 7 focuses on the microscopic nodal stress state. The algorithm execution process includes a three-layer loop mechanism: the outer loop iterates through candidate material combination schemes, the middle loop iterates the decomposition layer parameters, and the inner loop calculates the dynamic response matching degree. After normalization, the dynamic response features are compared with the simulation results using waveform correlation coefficients, with a correlation coefficient threshold set above 0.85. After nine iterations, the algorithm outputs the optimal material property combination, such as C40 concrete with HRB400 steel reinforcement, and determines the optimal analysis depth of 5 decomposition layers. This number of layers balances computational efficiency and accuracy, avoiding the neglect of local performance due to too few layers or data redundancy due to too many layers. After the number of layers is fixed, the layer division principle is as follows: the first layer corresponds to the foundation slab system, the second layer covers the vertical load-bearing components, the third layer deals with the horizontal connection system, the fourth layer focuses on the core area of ​​the nodes, and the fifth layer details the reinforcement anchorage.

[0064] The structural system atlas construction phase employs a graph theory-based topological mapping method. The structural modeling algorithm receives the optimal material property combination and decomposition layer parameters to establish a 3D digital structural model. The model node coordinate system uses the building's geometric center as its origin, and component connections are converted into adjacency matrices for storage. Beam elements are modeled using Euler beam theory, and wall elements are simulated using shell element theory. Material property assignments are performed based on the optimal combination parameters. The boundary processing employs symmetric extension technology, mirroring and extending the building's outer contour. Specifically, the original structural data boundary point set is identified, and data is mirrored using each boundary plane as a symmetry plane. The extension distance is set to 1.5 times the maximum component size. This processing eliminates boundary stress anomalies in finite element analysis, making the stress gradient in the model's edge regions more consistent with actual distribution.

[0065] The parameterized definition of the nodes in the digital twin map covers ten key physical quantities: including three-dimensional spatial coordinates, node mass parameters, moment of inertia values, local stiffness coefficients, damping characteristic parameters, current stress state values, strain rate of change, upper limit of displacement threshold, material fatigue characteristic index, and failure probability parameters. The total number of nodes fluctuates between 2000 and 5000 depending on the structural complexity, and each node generates a 128-dimensional feature vector. The weighting of connection relationships adopts a dual criterion: physical connection weights are calculated based on the cross-sectional characteristics of the components, with values ​​ranging from 0 to 1; dynamic transmission weights are determined based on modal analysis results, and the weight values ​​are correlated with the natural frequency matching degree. The final structural system map is stored in a sparse matrix format, with the proportion of non-zero elements in the matrix controlled below 15% to improve computational efficiency. The map is dynamically refreshed every 60 seconds to ensure data timeliness. The map visualization supports a three-dimensional topology rendering mode, with different color gradients mapping stress state distribution, and red lines marking dangerous node areas exceeding the yield strength. The entire process requires no manual intervention and automatically generates a digital twin model of the structure.

[0066] Example 2: See Figure 3The multi-stage dynamic response characteristic acquisition process sets up three load state analysis scenarios. The static load state is achieved through a hydraulic loading system, applying 1.2 times the design gravity load to the building foundation and holding it for 600 seconds to measure the static deformation characteristics of the structure. Displacement gauge arrays are deployed at key nodes on each floor to record vertical settlement with an accuracy of 0.5 mm. The dynamic load state uses an electromagnetic excitation device, inputting 0.1g to 0.3g of white noise excitation at the top of the structure, with a frequency scanning range covering 0.5Hz to 30Hz. Accelerometers capture the structural oscillation response at a sampling rate of 200Hz, with each frequency step lasting 120 seconds, recording the modal damping ratio parameters for each order. The seismic simulation state uses a shaking table test platform, inputting three seismic motion time histories: El-Centro wave, Kobe wave, and synthetic wave, with the peak ground acceleration adjusted to 0.35g and a time step of 0.02 seconds. A three-dimensional accelerometer network synchronously acquires the structural dynamic response, focusing on monitoring the inter-story drift angle changes. The load condition switching interval is set to a stabilization period of 900 seconds to eliminate the influence of residual vibration.

[0067] The structural modal features acquired during the data acquisition process are expressed through a five-dimensional feature matrix. The row vectors of the matrix correspond to different load states, while the column vectors contain five parameters: natural frequency, modal participation factor, equivalent damping ratio, modal mass ratio, and strain energy distribution coefficient. Under static conditions, the first six buckling modes are extracted; under dynamic conditions, twelve vibration modes are identified; and under seismic simulation conditions, time-varying nonlinear modal parameters are captured. Principal component analysis is used for data fusion of structural modal features, projecting the three types of state data onto a unified feature space to generate a modal feature map containing 36 core indicators. The operational procedure for identifying key structural elements based on this map is as follows: scan the modal participation factors of all nodes, and define nodes with participation factors exceeding 0.75 as key structural elements; analyze the strain energy distribution gradient of each component, and identify areas with strain energy concentrations exceeding three times the average value as seismic weak points. An outlier detection mechanism is introduced into the identification process to eliminate misjudged data caused by sensor noise.

[0068] The frequency band energy analysis of dynamic response characteristics employs an improved Welch power spectrum estimation method. Acceleration response data under seismic simulation conditions are divided into 128 subsequences of 5 seconds each, weighted using a Hanning window, and then subjected to a 1024-point Fast Fourier Transform. The frequency bands are divided into five intervals: 0-2Hz characterizes the foundation frequency response, 2-5Hz corresponds to the main structural vibration, 5-10Hz reflects the local component response, 10-20Hz captures high-frequency nodal oscillations, and frequencies above 20Hz filter out instrument noise. The energy proportion and standard deviation of each frequency band are calculated as indicators of frequency band energy distribution characteristics. Load state parameterization is performed: static state quantifies load values ​​and duration, dynamic state records the excitation power spectral density function, and seismic state extracts triaxial peak acceleration and duration. Complexity feature calculation is based on a sample entropy algorithm using structural feature parameter sequences. The time series includes 30 structural feature parameters such as section moment of inertia, reinforcement ratio, and constraint conditions. The sampling interval is set to 0.5 seconds, the embedding dimension is set to 3, the similarity tolerance is 0.2 times the standard deviation, and the final complexity index range is 0-1.

[0069] The multi-layer weighted fusion architecture for matching scores includes parallel computing pathways. The matching score calculation pathway employs a three-layer processing: the fundamental modal matching layer calculates the correlation between frequency band energy distribution and the fundamental frequency; the higher-order modal response layer analyzes the uniformity of energy distribution in higher-order modes across each frequency band; and the frequency band leakage assessment layer detects the energy dispersion of abnormal frequency bands. The load state adaptation scoring pathway consists of static adaptation units, dynamic adaptation units, and seismic adaptation units: the static adaptation unit assesses the deformation control capability of design parameters under dead load; the dynamic adaptation unit verifies the influence of parameter offset on resonant frequency; and the seismic adaptation unit analyzes the energy consumption efficiency of parameters in the nonlinear stage. The complexity-related scoring pathway is executed by an entropy mapping algorithm, establishing a regression model between the complexity index and the sensitivity of structural parameters. The output values ​​of the three scoring pathways are normalized to the [0,1] interval, with initial weights allocated in proportions of 0.4, 0.35, and 0.25.

[0070] The optimized search algorithm employs an improved particle swarm optimization mechanism, setting 200 particles to form the parameter search space. Each particle position vector contains 15 design parameter components, including reinforcement ratio, concrete strength grade, damper arrangement density, and node connection plate thickness, with parameter value ranges set according to national seismic design codes. The adaptation score serves as the optimization objective function, and the particle position update frequency is 50Hz. The dynamic weight optimization module includes a gradient estimator and a weight corrector: the gradient estimator calculates the partial derivative of the score with respect to the weights by perturbing the particle positions; the weight corrector adjusts the three-layer weight values ​​according to the gradient direction, with each adjustment not exceeding 15% of the original value. The optimization converges after 150 iterations, recording the design parameter combination corresponding to the globally optimal particle position. The parameter customization module receives user-required parameters, such as a cost coefficient of 0.65 and a construction period constraint of 180 days, and achieves parameter customization by adding a penalty term to the objective function. The final output list of optimal seismic design parameters includes specific values ​​such as: a reinforcement ratio of 2.85% in the core area of ​​the beam-column joint, a shear wall distributed reinforcement diameter of 10mm, a concrete strength grade of C45, and an additional damping ratio of 0.12. The parameter values ​​are chosen with consideration for the continuity of material market supply, and special specifications of materials are avoided.

[0071] Example 3: See Figure 4 The construction of the time-domain mapping matrix is ​​based on the interaction between the dynamic response characteristics of key structural elements and the optimal seismic design parameters. The matrix dimension is determined by the number of structural nodes N and the seismic motion duration T, where N ranges from 2000 to 5000, and T is set to a 40-second standard seismic motion time history. Matrix elements are generated using a nonlinear transformation function. :

[0072]

[0073] in: This represents the material damping ratio of the i-th node. The frequency matching degree corresponding to the j-th design parameter, It is the stiffness adjustment factor (values ​​1.2 and 1.8). The attenuation coefficient (values ​​range from 0.05 to 0.15) This represents the time step index. This formula ensures that the matrix element values ​​are controlled within the (-1, 1) range, while also reflecting the attenuation characteristics of stress wave propagation.

[0074] The matrix filling process is implemented in three stages. The first stage scans all key structural nodes, extracting node coordinates, section properties, and current stress state. Node coordinates are converted into normalized position vectors relative to the building center, section properties include moment of inertia and torsional coefficient, and the current stress state is recorded as triaxial stress components. The second stage processes optimal seismic design parameters, converting parameters such as reinforcement ratio and concrete strength into equivalent frequency parameters. Frequency matching is calculated using the ratio of the parameter's natural frequency to the structure's fundamental frequency, with the ratio controlled within the reasonable engineering range of 0.8 to 1.2. The third stage performs matrix element calculations, with each element corresponding to the node-parameter coupling effect at a specific time step. The calculation process employs a parallel processing architecture, with the maximum computation time for a single time step limited to within 50 milliseconds.

[0075] Multi-scale time analysis employs a triple sliding window mechanism. The short-scale analysis window has a fixed width of 2 seconds, with a displacement increment threshold of 5 mm; exceeding this threshold triggers a response conversion flag. The medium-scale analysis window has a width of 10 seconds, calculating the stress state index every 0.5 seconds; the index is determined by both the direction and magnitude of the principal stresses. The long-scale analysis window covers the entire 40-second time history; deformation trends are evaluated using the cumulative displacement vector magnitude, with a vector sampling interval of 10 seconds. The window sliding step size is set to 1 / 10 of the width of each window to ensure analytical continuity.

[0076] An adaptive weighted fusion algorithm establishes a three-layer decision-making mechanism. The first layer calculates the signal-to-noise ratio (SNR) of the analysis results at each scale: short-term SNR is based on the energy proportion of high-frequency components, medium-term SNR on the stress fluctuation standard deviation, and long-term SNR on the trend line fit. The second layer performs dynamic weight allocation, with initial weights set to 0.4, 0.3, and 0.3, adjusted every 5 seconds based on the SNR, with a single adjustment not exceeding 0.1. The third layer implements weighted fusion, storing the fusion result as a 128-dimensional feature vector containing key indicators such as peak time, maximum stress location, and residual deformation.

[0077] The generation of structural response feature sequences follows the time-space coupling principle. In the time dimension, the sequence sampling interval is 0.1 seconds, with a total of 400 time points. In the spatial dimension, the sequence contains three types of data channels: a displacement channel recording the three-dimensional displacement of nodes, a stress channel storing principal stress values, and a damage channel marking the degree of plastic development. The sequence encoding uses a binary floating-point format, with each sequence data size approximately 25MB. The feature extraction process identifies six typical response modes: elastic vibration, local yielding, energy dissipation, modal transition, stiffness degradation, and residual deformation. Each mode corresponds to a combination of waveform features; for example, the local yielding mode is characterized by high-frequency vibration superimposed with steady-state drift.

[0078] In the modal decomposition preprocessing stage, data normalization is performed. Zero-mean normalization is applied to the original sequence to eliminate the influence of DC offset. The data frame length is set to 1024 points, with a frame shift of 256 points, and Hamming window weighting is used to reduce spectral leakage. Boundary processing employs a mirror extension method, extending the sequence by 512 points at both ends. The extreme point detection algorithm is based on cubic spline interpolation to identify local maxima and minima, with the extreme point interval controlled within the range of 10 to 100 sampling points. Three criteria are set for intrinsic modal condition determination: the difference between the number of zero-crossing points and the number of extreme points does not exceed 1, the envelope symmetry is greater than 0.7, and the instantaneous frequency monotonicity retention rate is higher than 80%.

[0079] An adaptive stopping criterion is introduced in the component reconstruction process. The maximum number of screening iterations is set to 10, and the residual energy ratio threshold is set to 5%. Two evaluation metrics are calculated in each iteration: component orthogonality index and energy concentration. The orthogonality index requires the correlation coefficient between different components to be less than 0.3, and the energy concentration requires a single-order component to contain more than 15% of the total energy. For fast-varying components, the first three orders are extracted and focused, and the instantaneous frequency is calculated using a 5-point difference operator, with the frequency range limited to above 3Hz. Slow-varying components include the last two orders and residual terms; a moving average filter is applied to eliminate high-frequency fluctuations, with a window width of 1 second. After component separation, energy normalization is performed to ensure that each component has a comparable energy scale.

[0080] Example 4: See Figure 5 In the modal decomposition processing of the structural response feature sequence, the input data came from the measured seismic response records of a standard floor of a hotel building. This building is a 15-story reinforced concrete frame-shear wall structure. The monitoring system deployed 42 accelerometers in the east-west direction, 38 in the north-south direction, and 24 vertically. The seismic input used the El-Centro wave NS component, with the peak ground acceleration adjusted to 0.28g and a sampling frequency of 200Hz. After preprocessing, the raw data formed a structural response feature sequence, containing three channels of data: displacement, velocity, and acceleration, with a duration of 40 seconds and a total of 8000 data points.

[0081] The slow-varying component feature extraction process targets the last three modal components and residual terms. The time-domain statistical feature calculation window width is set to 5 seconds, with a sliding step size of 0.5 seconds, resulting in 80 sets of statistics. Each set of statistics includes six parameters: the mean reflects the overall offset, the standard deviation represents the fluctuation amplitude, the skewness coefficient describes the distribution symmetry, the kurtosis coefficient measures the probability of extreme values, the autocorrelation delay value characterizes periodicity, and the zero-crossing rate indicates the oscillation frequency. Frequency band energy distribution characteristics are analyzed using an eight-band approach, with band divisions determined based on the structure's inherent frequencies: 0-1Hz, 1-2Hz, 2-3.5Hz, 3.5-5Hz, 5-8Hz, 8-12Hz, 12-18Hz, and 18-25Hz. Energy calculation uses a modified periodogram method with a Hanning window width of 1024 points and an overlap rate of 50%. The inter-component correlation characteristics are obtained by calculating the cross-correlation matrix between fourth-order components, with a matrix dimension of 4×4 and element values ​​ranging from 0 to 1.

[0082] Fast-variant component processing focuses on the first three high-frequency components. The instantaneous deformation rate is calculated using the five-point center difference method, with a time resolution of 0.01 seconds. Peak feature detection employs a dynamic threshold mechanism, with an initial threshold set to three times the root mean square of the signal, adjusted every 5 seconds based on signal characteristics. Waveform morphology feature extraction includes seven indicators: rise time, fall time, pulse width, duty cycle, waveform asymmetry, envelope smoothness, and oscillation persistence. These features are obtained through waveform extreme point identification and envelope fitting, with the mean of each feature calculated every 2 seconds.

[0083] Table 1: Feature extraction results of a typical node (N-7F-12) in the 15-20 second time period.

[0084]

[0085] The multi-stage classification strategy is implemented in three levels. The first level uses a support vector machine (SVM) to classify the steady-state response feature vectors, with a radial basis function (RBF) kernel and a penalty coefficient of 1.0. The classification target distinguishes four states: elastic state, micro-crack state, significant damage state, and extreme state. The training dataset contains feature vectors from 120 historical earthquake records, and the test set uses 30 independent datasets. The second level processes dynamic response feature vectors. The adaptive threshold detector includes three threshold levels: warning threshold, alert threshold, and danger threshold, corresponding to normal, attention, and emergency response levels, respectively. The initial threshold values ​​are set according to structural design specifications, and are dynamically updated every 10 seconds based on the latest 30 seconds of data. The third level implements probabilistic network fusion. The Bayesian network node setup includes six parent nodes and three child nodes, and the conditional probability table is obtained through 200 Monte Carlo simulations.

[0086] The spatiotemporal continuity verification of response transition events employs a four-dimensional verification mechanism. The spatial dimension examines the node distribution pattern of the event, requiring at least three nodes within three adjacent layers to simultaneously trigger events of the same level. The temporal dimension analyzes the event duration; short-duration events (<0.5s) must occur within three consecutive time steps, while long-duration events (≥0.5s) must maintain state consistency. The frequency dimension verifies the matching degree between the event's characteristic frequency and the structure's local intrinsic frequency, with a difference not exceeding 15%. The energy dimension requires that the energy change rate synchronization of the event-related nodes reach at least 80%. Only events that pass the verification are confirmed as valid response transitions.

[0087] The decision logic for seismic optimization design categories is based on a combination of state classification and response transformation. When the structure is in an elastic state without response transformation, category A (maintain the status quo) is output; when the elastic state is accompanied by local response transformation, category B (add local constraints) is output; when the microcrack state is without response transformation, category C (repair microcracks) is output; when the microcrack state is accompanied by multi-node response transformation, category D (add dampers) is output; and when the structure is significantly damaged, category E (structural strengthening) is output regardless of whether there is response transformation. Each category corresponds to a specific list of technical measures; for example, category D includes the quantity, location, and parameter specifications of viscous dampers to be installed. The design scheme generation module automatically matches the category with the technical measure library and outputs a complete scheme document containing construction drawings, a material list, and process requirements. The document format adopts the IndustryFoundationClasses standard and supports direct import into the Building Information Modeling (BIM) system.

[0088] Example 5: User design constraint signals are received via a graphical user interface or application programming interface (API). The interface includes a structural schematic layer and a parameter input panel. Users can select the area to be adjusted on the 3D model and input constraint parameters. Signal transmission uses JSON format data packets, containing six key fields: design target identifier, constraint type code, target parameter value, allowable fluctuation range, priority weight, and effective time identifier. The constraint type code mapping table defines value 1 as a structural node adjustment instruction, value 2 as a connection relationship correction instruction, and value 3 as a weight dynamic update instruction.

[0089] The execution mechanism for optimization operations is based on specific algorithm rules. The structural node adjustment operation activates the coordinate optimizer module, which reads historical displacement data of the target node and establishes a correlation model between the node position and maximum stress. Distance adjustments are implemented incrementally using a linear interpolation algorithm, with each adjustment step not exceeding 10% of the initial distance. The connection relationship correction operation initiates the topology reconstruction engine, first disconnecting the original connection definitions and reconstructing connection weights based on the mode shape correlation coefficients between nodes. The reconstruction process retains 75% of the original connection topology, optimizing the remaining portion according to the principle of minimum energy transfer path. The dynamic weight update operation triggers the parameter regulator, which constructs a sensitivity matrix between design parameters and constraints, where matrix elements represent the influence coefficient of a single parameter change on the constraint target. Each update iteration calculates the gradient direction, adjusting the weight values ​​along the negative gradient direction; the magnitude of a single adjustment is limited by the constraint boundary conditions.

[0090] The dynamic correction process employs a closed-loop control strategy. The initial design scheme is loaded into the finite element simulation environment with a time step set to 0.05 seconds. Correction commands are injected into the control points, and the correction deviation value is checked every 3 seconds. When a structural node adjustment command is executed, the actual coordinate offset of the target node is monitored, forming a deviation feedback loop with the coordinates required by the command. If the deviation exceeds the allowable range, a compensation algorithm is activated, with the compensation amount implemented in four incremental increments according to the deviation ratio. During the connection relationship correction process, the change curve of the connection weights is tracked in real time, and a smooth transition function is inserted when abrupt transitions occur between the old and new weights. The dynamic weight update operation is equipped with a self-checking mechanism, performing a sensitivity consistency check after every two weight adjustments to check whether the parameter's direction of action remains stable. All correction operations are completed within 40 seconds of the simulation time, maintaining time synchronization with the structural response sequence.

[0091] The revised scheme deploys a virtual sensor network during execution. Monitoring points are established at 253 key locations on the structural model, each equipped with a stress calculation unit, displacement recording unit, and material state analysis unit. A stress distribution completion rate calculator scans all monitoring points, calculating the percentage of areas reaching over 90% of the design stress, with a calculation cycle of once every 5 seconds. A deformation path actual change time tracker records the displacement trajectory of typical nodes, focusing on the time span of key monitoring nodes from their initial position to their maximum displacement point, with an accuracy of 0.01 seconds. A node material utilization rate evaluator obtains the ratio of the effective stress area to the theoretical area through cross-sectional strain integration, outputting 20 sample values ​​per second per node. Feedback data storage adopts a time-series database structure, recording timestamps, monitoring point IDs, and three-dimensional components of physical quantities.

[0092] The structural system map update operation is implemented through a data fusion engine. This engine contains three parallel processing channels: The first channel processes stress distribution completion rate data, creates a node state thermal map, and converts the completion rate into state coefficients in the [0,1] interval. The new coefficients are fused with the original map coefficients with a weight of 0.3, and the fusion formula is α×new value + (1-α)×original value, where α is 0.7. The second channel analyzes the actual change time of the deformation path, extracts abnormal nodes with a time deviation exceeding 20%, and reconstructs their connection relationship weight factors. The reconstruction principle adds a time-varying adjustment coefficient, so that the connection weights change adaptively with time. The third channel receives the node material utilization rate change value, establishes a material utilization efficiency matrix, and triggers node attribute recalculation when the node value change exceeds 0.15. Attribute updates include strength reduction coefficient adjustment, stiffness decay curve correction, and failure threshold recalibration. The overall map update cycle is controlled to be completed within 60 seconds of simulation time. During the update process, the connectivity of the map structure is maintained to avoid system splitting due to local modifications. The updated map automatically generates a version identifier and establishes an association index with the correction scheme.

[0093] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0094] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A hotel building reinforced concrete structure seismic performance optimization design method, characterized in that, The method comprises the following steps: Obtain the original structure data and corresponding dynamic response characteristics of the hotel building structure; Construct a structure system graph based on the original structure data and dynamic response characteristics; Identify key structural elements and weak links in seismic resistance through the structure system graph; Extract the structure characteristics corresponding to the key structural elements; Calculate the complexity characteristics of the structure characteristics; Determine the optimal seismic design parameters personalized for the user according to the key structural elements and complexity characteristics; Reconstruct and generate a structure response characteristic sequence based on the key structural elements and optimal seismic design parameters; Perform modal decomposition on the structure response characteristic sequence to obtain slow-varying components reflecting the steady-state characteristics of the structure and fast-varying components reflecting the dynamic changes of the structure; According to the slow-varying components and fast-varying components, obtain the seismic optimization design category, which is defined as different reinforcement or optimization strategies to generate a seismic optimization design scheme. Based on the user input signal, dynamically correct the seismic optimization design scheme, and update the structure system graph according to the correction result; Correction Including parameter adjustment or scheme iteration to ensure that the design scheme adapts to real-time needs; The determination of the optimal seismic design parameters personalized for the user comprises: According to the frequency band energy distribution characteristics, load state, and complexity characteristics of the dynamic response characteristics, calculate the adaptation score of the pre-set candidate design parameters through multi-layer weighted fusion. The adaptation score includes matching degree score, load state adaptation score, and complexity related score. Use an optimization search algorithm to dynamically optimize the adaptation score to determine the optimal seismic design parameters personalized for the user among the candidate design parameters. The reconstruction and generation of the structure response characteristic sequence comprises: Construct a time domain mapping matrix; the matrix elements of the time domain mapping matrix are generated by nonlinear combination of the frequency matching degree corresponding to the key structural elements, optimal seismic design parameters, and pre-set stress modulation coefficient; Perform multi-scale time analysis on the time domain mapping matrix; the multi-scale time analysis includes short-time scale analysis for capturing structure response conversion characteristics, medium-scale analysis for tracking stress state, and long-time scale analysis for monitoring deformation trend; Generate a structure response characteristic sequence containing multiple structure response mode characteristics by adaptively fusing the analysis results of multi-scale time analysis.

2. The hotel building reinforced concrete structure seismic performance optimization design method according to claim 1, characterized in that, The original structure data and corresponding dynamic response characteristics of the hotel building structure are obtained, which comprises: Obtain the attribute characteristics of the candidate building material properties; the attribute characteristics at least include elastic modulus and yield strength; According to the structure analysis algorithm, input the attribute characteristics, the dynamic response characteristics, and the pre-set decomposition layer number range to determine the optimal material property combination and the corresponding decomposition layer number in the candidate building material properties.

3. The hotel building reinforced concrete structure seismic performance optimization design method according to claim 2, characterized in that, The construction of the structure system graph comprises: Use a structure modeling algorithm to construct a graph based on the optimal material property combination and the corresponding decomposition layer number for the original structure data; Use symmetric continuation processing on the boundary of the original structure data to reduce the influence of boundary distortion; Obtain a plurality of structure node parameters, and construct the structure system graph according to a plurality of structure node parameters.

4. The hotel building reinforced concrete structure seismic performance optimization design method according to claim 1, characterized in that, The key structural elements and aseismic weak links are identified, including: Obtaining structural modal characteristics under different load states through a multi-stage dynamic response characteristic acquisition process; the multi-stage includes at least a static load state, a dynamic load state, and a seismic simulation state; Obtaining the key structural elements and aseismic weak links according to the structural modal characteristics.

5. The hotel building reinforced concrete structure seismic performance optimization design method according to claim 1, characterized in that, The structural response characteristic sequence is subjected to modal decomposition to obtain slow-varying components reflecting structural steady-state characteristics and fast-varying components reflecting structural dynamic changes, including: Performing extreme point detection and mirror symmetry extension processing on components of the structural response characteristic sequence, and obtaining multi-order inherent structural components satisfying intrinsic modal conditions through iterative screening; Based on the average period and energy distribution characteristics of the multi-order inherent structural components, the front-order short-period components corresponding to the multi-order inherent structural components are classified as fast-varying components reflecting structural response conversion, and the back-order long-period components and residual terms corresponding to the multi-order inherent structural components are classified as slow-varying components reflecting structural maintenance.

6. The hotel building reinforced concrete structure seismic performance optimization design method according to claim 1, characterized in that, The aseismic optimization design category is obtained, including: Extracting time-domain statistical characteristics, frequency band energy distribution characteristics, and component correlation characteristics from the slow-varying components to construct a steady-state response feature vector; Extracting instantaneous deformation change rate, peak value characteristics, and waveform morphology characteristics from the fast-varying components to construct a dynamic response feature vector; Using a multi-stage classification strategy, the steady-state response feature vector is subjected to state classification, and the dynamic response feature vector is identified through an adaptive threshold detector to identify response conversion events; Through probability network fusion of classification results corresponding to state classification and response conversion events, and combining the spatiotemporal continuity characteristics of response conversion, a credibility evaluation is performed to output the aseismic optimization design category.

7. The hotel building reinforced concrete structure seismic performance optimization design method according to claim 1, characterized in that, The aseismic optimization design scheme is dynamically corrected based on a user input signal, including: Receiving a user design constraint signal, and determining an optimization operation type according to the user design constraint signal; Based on the optimization operation type, the aseismic optimization design scheme is dynamically corrected; the optimization operation type includes at least structural node adjustment, connection relationship correction, and weight dynamic update.

8. The hotel building reinforced concrete structure seismic performance optimization design method according to claim 7, characterized in that, The structural system atlas is updated according to the correction result, including: Executing the aseismic optimization design scheme after dynamic correction; Real-time acquisition of feedback data during execution; the feedback data includes at least stress distribution completion rate, deformation path actual change time, and node material utilization rate change value; Based on the feedback data, the structural node parameters and connection relationships in the structural system atlas are updated.

Images