Method and system for monitoring uneven distribution of snow load on high and low roofs

Abstract

The invention provides a method and system for monitoring uneven distribution of snow load on high and low roofs, and relates to the technical field of load monitoring, and the method comprises the steps: arranging load, displacement and environment sensors at preset monitoring points, and measuring the snow load and structural deformation in real time. A load distribution diagram is generated through de-noising processing and an interpolation algorithm, a non-uniform load area is identified, and a monitoring result is analyzed by using a trained model, so that the safety and durability of the structure are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of load monitoring, and particularly to a monitoring method and system for uneven distribution of snow loads on high and low roofs. Background Art

[0002] With the development of modern building technology, ultra-high and super-large giant steel structure buildings have been widely used in industrial, commercial, and public buildings. These buildings usually have complex roof structures, especially the high and low roof design (i.e., there are significant differences in roof heights). Under winter snow conditions, due to factors such as wind speed, wind direction, and building height differences, the snow load often shows uneven distribution between high and low roofs. This uneven distribution may lead to local structural overloading, and then cause structural deformation or even failure.

[0003] Currently, the monitoring methods for snow loads mostly focus on single roofs or the assumption of uniform distribution, lacking real-time monitoring and analysis methods for the uneven distribution of snow loads between high and low roofs. Therefore, developing a technical method that can monitor the uneven distribution of snow loads on high and low roofs in real time has important engineering significance and application value. Summary of the Invention

[0004] In order to overcome the deficiencies of the prior art, the purpose of the present invention is to provide a monitoring method and system for uneven distribution of snow loads on high and low roofs, which can not only achieve real-time monitoring and early warning of the structure, but also provide accurate load assessment and deformation analysis, identify uneven load areas, support data-driven decision-making, and ultimately improve the safety and durability of the structure. These beneficial effects are of great significance for ensuring the safe operation of high and low roof structures.

[0005] To achieve the above purpose, the present invention provides the following solutions:

[0006] A monitoring method for uneven distribution of snow loads on high and low roofs, comprising:

[0007] Arranging load sensors, displacement sensors, and environmental sensors at each preset monitoring point of the high and low roof structure, and using the load sensors to measure the load data of local snow accumulation on the roof, and using the displacement sensors to monitor the deformation data of the roof structure;

[0008] Performing denoising processing on the load data and the deformation data to obtain load denoising information and deformation denoising information;

[0009] Determining the snow load per unit area of each region of the high and low roofs according to the load denoising information, and generating a snow load distribution map using an interpolation algorithm;

[0010] Analyzing the deformation trend of the roof structure according to the deformation denoising information;

[0011] By analyzing the load distribution diagrams of the high and low roofs, identify the unevenly distributed areas of the snow load, and use the gradient algorithm to calculate the load difference between the high and low roofs;

[0012] Input the position information of the unevenly distributed area, the load difference, and the deformation trend into the trained load monitoring model to obtain the monitoring result.

[0013] Preferably, perform denoising processing on the load data and the deformation condition data to obtain load denoising information and deformation denoising information, including:

[0014] Translate the signals where the load data and the deformation condition data are located to obtain the translated signals;

[0015] Perform wavelet decomposition on the translated signals to obtain multiple wavelet coefficients;

[0016] Determine the filtering threshold according to the decomposition scale and length of the signal;

[0017] Construct a denoising function according to the filtering threshold;

[0018] Use the denoising function to remove the noise of the translated signals to obtain the load denoising information and the deformation denoising information.

[0019] Preferably, the calculation formula of the filtering threshold is:

[0020]

[0021] where λ is the filtering threshold, j is the decomposition scale of the translated signal, d j is the wavelet coefficient with the decomposition scale of j, N represents the length of the translated signal, and median represents the median operation.

[0022] Preferably, use the denoising function to remove the noise of the translated signals to obtain the load denoising information and the deformation denoising information, including:

[0023] Use the denoising function to remove the corresponding wavelet coefficients to obtain the smoothed signal; where the denoising function is: where X represents the wavelet coefficient and n represents the adjustment coefficient;

[0024] Perform inverse translation on the smoothed signal to obtain the inversely translated signal;

[0025] Calculate the average of the inversely translated signal to obtain the load denoising information and the deformation denoising information.

[0026] Preferably, determine the snow load per unit area of each region of the high and low roofs according to the load denoising information, and use an interpolation algorithm to generate a snow load distribution map, including:

[0027] Determine the geometric region range of the high and low roofs according to the architectural design drawings and the sensor layout plan;

[0028] Calculate the mean value of the load data in the geometric region range, and use the mean value as the snow load per unit area;

[0029] Based on the inverse distance weighted method, perform spatial interpolation on the load data between each preset monitoring point to generate a snow load distribution map of the entire roof.

[0030] Preferably, analyze the deformation trend of the roof structure according to the deformation denoising information, including:

[0031] Construct a linear regression model;

[0032] Use the linear regression model to fit the deformation data to identify the deformation trend of the roof structure.

[0033] Preferably, by analyzing the load distribution map of the high and low roofs, identify the uneven distribution area of the snow load, and use the gradient algorithm to calculate the load difference between the high and low roofs, including:

[0034] Set a load difference threshold according to engineering experience or historical data to judge whether the load distribution is uniform;

[0035] Use an edge detection algorithm to detect the edges in the load distribution map of the high and low roofs to identify the areas where the load changes;

[0036] Start expanding outward from the edge points detected by the edge detection algorithm until the set load difference threshold is reached, form an uneven distribution area, and record the coordinates and load values of the identified uneven distribution area;

[0037] Extract the load data of the high roof and the low roof from the load distribution map of the high and low roofs, and use the gradient algorithm to calculate the load difference between the high and low roofs.

[0038] Preferably, input the position information of the uneven distribution area, the load difference, and the deformation trend into the trained load monitoring model to obtain a monitoring result, including:

[0039] Obtain a preset position information dataset, a load difference dataset, and a deformation trend dataset;

[0040] Construct a deep learning model;

[0041] Train the deep learning model according to the position information dataset, the load difference dataset, and the deformation trend dataset respectively to obtain a trained first classifier, second classifier, and third classifier;

[0042] Cascade the trained first classifier, second classifier, and third classifier to obtain a classification network;

[0043] Connect a trained LSTM neural network after the classification network to obtain the load monitoring model;

[0044] Monitor the position information, the load difference, and the deformation trend of the uneven distribution area according to the load monitoring model to obtain the monitoring result.

[0045] A monitoring method for uneven distribution of snow load on high and low roofs, including:

[0046] A sensing data acquisition unit for arranging load sensors, displacement sensors, and environmental sensors at each preset monitoring point of the high and low roof structures, and using the load sensors to measure the load data of local snow accumulation on the roof, and using the displacement sensors to monitor the deformation data of the roof structure;

[0047] A data preprocessing unit for denoising the load data and the deformation data to obtain load denoised information and deformation denoised information;

[0048] A distribution map generation unit for determining the snow load per unit area of each region of the high and low roofs according to the load denoised information, and generating a snow load distribution map using an interpolation algorithm;

[0049] A trend change unit for analyzing the deformation trend of the roof structure according to the deformation denoised information;

[0050] A load difference calculation unit for identifying the uneven distribution area of the snow load by analyzing the load distribution maps of the high and low roofs, and calculating the load difference between the high and low roofs using a gradient algorithm;

[0051] A model monitoring unit for inputting the position information of the uneven distribution area, the load difference, and the deformation trend into a trained load monitoring model to obtain a monitoring result.

[0052] According to the specific embodiments provided by the present invention, the following technical effects are disclosed by the present invention:

[0053] The present invention provides a monitoring method and system for non-uniform distribution of snow loads on high-low roofs, including: arranging load sensors, displacement sensors and environmental sensors at each preset monitoring point of the high-low roof structure, and using the load sensors to measure the load data of local snow accumulation on the roof, and using the displacement sensors to monitor the deformation data of the roof structure; denoising the load data and the deformation data to obtain denoised load information and denoised deformation information; determining the snow load per unit area of each region of the high-low roof according to the denoised load information, and generating a snow load distribution map using an interpolation algorithm; analyzing the deformation trend of the roof structure according to the denoised deformation information; by analyzing the load distribution map of the high-low roof, identifying the non-uniform distribution area of the snow load, and calculating the load difference between the high-low roofs using a gradient algorithm; inputting the position information of the non-uniform distribution area, the load difference and the deformation trend into a trained load monitoring model to obtain a monitoring result. The present invention can not only realize real-time monitoring and early warning of the structure, but also provide accurate load assessment and deformation analysis, identify non-uniform load areas, support data-driven decision-making, and ultimately improve the safety and durability of the structure. These beneficial effects are of great significance for ensuring the safe operation of high-low roof structures. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0055] Figure 1 It is a flowchart of the method provided by the embodiment of the present invention;

[0056] Figure 2 It is a schematic structural diagram of the system provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all 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.

[0058] The object of the present invention is to provide a monitoring method and system for the uneven distribution of snow load on high and low roofs, which can not only achieve real-time monitoring and early warning of the structure, but also provide accurate load assessment and deformation analysis, identify uneven load areas, support data-driven decision-making, and ultimately improve the safety and durability of the structure. These beneficial effects are of great significance for ensuring the safe operation of high and low roof structures.

[0059] To make the above objects, features, and advantages of the present invention more obvious and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] Figure 1 The flowchart of the method provided by the embodiment of the present invention is shown as Figure 1 As shown, the present invention provides a monitoring method for the uneven distribution of snow load on high and low roofs, including:

[0061] Step 100: Arrange load sensors, displacement sensors, and environmental sensors at each preset monitoring point of the high and low roof structure, and use the load sensors to measure the load data of local snow accumulation on the roof, and use the displacement sensors to monitor the deformation data of the roof structure;

[0062] Step 200: Denoise the load data and deformation data to obtain load denoised information and deformation denoised information;

[0063] Step 300: Determine the snow load per unit area of each region of the high and low roofs according to the load denoised information, and use the interpolation algorithm to generate a snow load distribution map;

[0064] Step 400: Analyze the deformation trend of the roof structure according to the deformation denoised information;

[0065] Step 500: Identify the uneven distribution area of the snow load by analyzing the load distribution map of the high and low roofs, and use the gradient algorithm to calculate the load difference between the high and low roofs;

[0066] Step 600: Input the position information, load difference, and deformation trend of the uneven distribution area into the trained load monitoring model to obtain the monitoring result.

[0067] Specifically, step 100 of this embodiment involves the selection and layout design of sensors. The load sensor selects a model suitable for measuring snow load and can work normally in low temperature and humid environments. The displacement sensor needs to have high precision and stability to ensure that it can accurately monitor the small deformation of the roof structure. The environmental sensor is used to record meteorological data such as temperature, humidity, and wind speed, providing the necessary environmental background information for subsequent analysis.

[0068] When arranging sensors, the positions of monitoring points should be reasonably selected according to the geometric shape and load distribution characteristics of the roof. Generally, dense load sensors are arranged at the junctions of high and low roofs, in areas with large load changes, and near structural support points to capture local load changes. At the same time, displacement sensors should be evenly distributed on the roof, especially in areas where large deformations may occur, to ensure comprehensive monitoring of the roof deformation. Environmental sensors should be arranged above or near the roof to accurately obtain the meteorological conditions affecting the snow load.

[0069] After the sensors are arranged, the system in this embodiment is calibrated and tested to ensure the accuracy and reliability of data acquisition. The load sensors measure the local snow load on the roof, record the load data in real time, and transmit it to the data acquisition system. The displacement sensors monitor the deformation of the roof structure and record the deformation data. These data will provide a basis for subsequent denoising processing, load distribution analysis, and structural safety assessment, ensuring that the monitoring system can operate effectively and provide accurate monitoring results.

[0070] Preferably, the load data and the deformation data are denoised to obtain load denoising information and deformation denoising information, including:

[0071] The signals where the load data and the deformation data are located are translated to obtain the translated signals;

[0072] The translated signals are decomposed by wavelet to obtain multiple wavelet coefficients;

[0073] The filtering threshold is determined according to the decomposition scale and length of the signal;

[0074] A denoising function is constructed according to the filtering threshold;

[0075] The noise of the translated signals is removed by using the denoising function to obtain the load denoising information and the deformation denoising information.

[0076] Preferably, the calculation formula of the filtering threshold is:

[0077]

[0078] where λ is the filtering threshold, j is the decomposition scale of the translated signal, d j is the wavelet coefficient with the decomposition scale of j, N represents the length of the translated signal, and median represents the median operation.

[0079] Preferably, removing the noise of the translated signals by using the denoising function to obtain the load denoising information and the deformation denoising information includes:

[0080] Remove the corresponding wavelet coefficients using the denoising function to obtain a smoothed signal; wherein, the denoising function is: wherein, X represents the wavelet coefficient and n represents the adjustment coefficient;

[0081] Perform inverse translation on the smoothed signal to obtain an inversely translated signal;

[0082] Obtain the average of the inversely translated signal to obtain the load denoising information and the deformation denoising information.

[0083] Specifically, in this embodiment, by using wavelet coefficients to measure the attenuation deviation, the attenuation of wavelet coefficients with larger absolute values can be made to decrease as their absolute values increase, effectively avoiding the loss of high-frequency information and improving the signal-to-noise ratio of the signal.

[0084] Preferably, determine the snow load per unit area of each region of the high and low roofs according to the load denoising information, and use an interpolation algorithm to generate a snow load distribution map, including:

[0085] Determine the geometric region range of the high and low roofs according to the building design drawings and the sensor layout scheme;

[0086] Calculate the mean value of the load data in the geometric region range and use the mean value as the snow load per unit area;

[0087] Based on the inverse distance weighted method, perform spatial interpolation on the load data between each preset monitoring point to generate a snow load distribution map of the entire roof.

[0088] Specifically, in this embodiment, based on the building design drawings and the sensor layout scheme, first extract the boundary coordinates of the high roof and the low roof. By analyzing the layout positions of the sensors, determine the areas covered by each sensor, and divide these areas into high roof areas and low roof areas. Ensure that the geometric characteristics of each area and the distribution of the sensors can reflect the actual load distribution characteristics.

[0089] Next, calculate the mean value of the load data in the geometric region range and use it as the snow load per unit area. For each region, collect the load denoising information of all relevant sensors and calculate the mean value of these data. The mean value can be calculated through a simple arithmetic mean formula, that is, add up the load values of all sensors and divide by the number of sensors. This mean value will represent the snow load per unit area of this region and provide basic data for subsequent interpolation calculations.

[0090] Finally, based on the Inverse Distance Weighting (IDW) method, spatial interpolation is performed on the load data between each preset monitoring point to generate a snow load distribution map for the entire roof. The core idea of the Inverse Distance Weighting method is that the data of sensors closer to the interpolation point have a greater impact on the interpolation result. In the implementation process, first, the distance from each monitoring point to the interpolation point is calculated, and then the load value is calculated by weighting according to the reciprocal of the distance. The specific formula is:

[0091]

[0092] where P i is the load value of the i-th sensor, d i is the distance from this sensor to the interpolation point, and k is the weight exponent (usually taken as 2). By performing interpolation calculations on the entire roof, the finally generated snow load distribution map will clearly show the load distribution of each area of the high and low roofs, providing an important basis for subsequent structural analysis and safety assessment.

[0093] Preferably, the deformation trend of the roof structure is analyzed according to the deformation denoising information, including:

[0094] Construct a linear regression model;

[0095] Use the linear regression model to fit the deformation data to identify the deformation trend of the roof structure.

[0096] Specifically, the construction of the linear regression model involves selecting appropriate independent and dependent variables. In this embodiment, the independent variable can be time series data (such as monitoring time points), and the dependent variable is the deformation data obtained by displacement sensors. By collecting deformation data at multiple time points, a linear equation model is established in the form of D(t) = a×t + b, where D(t) is the deformation value at time t, a is the slope (representing the change rate of deformation), and b is the intercept (representing the initial deformation value).

[0097] Next, use the constructed linear regression model to fit the deformation data. Calculate the regression coefficients a and b by the Least Squares Method to minimize the error between the actual deformation data and the model prediction values. The specific steps include inputting the deformation data into the model, calculating the predicted deformation values at each time point, and comparing them with the actual measured values. In this way, the goodness of fit of the model can be evaluated. Usually, the coefficient of determination R2 is used to measure the explanatory ability of the model to ensure that the model can effectively capture the change trend of the deformation data.

[0098] Finally, by analyzing the results of the linear regression model, the deformation trend of the roof structure is identified. According to the slope a of the regression model, it can be judged whether the deformation of the roof structure is increasing, decreasing or remaining stable. If the slope is positive, it indicates that the roof structure is continuously deforming; if the slope is negative, it means that the structure is gradually recovering; if the slope is close to zero, it shows that the deformation trend is relatively stable. Combining with the time series of deformation data, a deformation trend graph can be drawn to visually display the deformation changes of the roof structure, providing a basis for subsequent structural safety assessment and maintenance decision-making.

[0099] Preferably, by analyzing the load distribution diagrams of the high and low roofs, the uneven distribution areas of the snow load are identified, and the gradient algorithm is used to calculate the load difference between the high and low roofs, including:

[0100] Set a load difference threshold according to engineering experience or historical data to judge whether the load distribution is uniform;

[0101] Use the edge detection algorithm to detect the edges in the load distribution diagrams of the high and low roofs to identify the areas where the load changes;

[0102] Start expanding outward from the edge points detected by the edge detection algorithm until the set load difference threshold is reached, form an uneven distribution area, and record the coordinates and load values of the identified uneven distribution area;

[0103] Extract the load data of the high roof and the low roof from the load distribution diagrams of the high and low roofs, and use the gradient algorithm to calculate the load difference between the high and low roofs.

[0104] Specifically, in this embodiment, a load difference threshold is set according to engineering experience or historical data. This threshold is used to judge whether the load distribution is uniform, and in this embodiment, it is determined by analyzing the standard deviation of the historical load data or the difference between the maximum value and the minimum value. The set threshold should be able to effectively distinguish between normal load distribution and abnormal load distribution for subsequent edge detection and area identification.

[0105] Next, in this embodiment, the edge detection algorithm is used to process the load distribution diagrams of the high and low roofs to identify the areas where the load changes. Commonly used edge detection algorithms include Sobel operator, Canny operator, etc. These algorithms can effectively identify the edge points with large load changes in the load distribution diagrams. By performing a convolution operation on the load distribution diagrams, the gradient value of each pixel point is calculated to determine the edge position. The results of the edge detection will provide basic data for subsequent identification of uneven distribution areas.

[0106] After identifying the edge points, expand outward from these edge points to form an uneven distribution area. The expansion process can adopt the region growing algorithm. Starting from the edge points, gradually check the load values of the surrounding pixels to determine whether they exceed the set load difference threshold. If the load value of the surrounding pixels exceeds the threshold, include them in the current uneven distribution area. This process continues until no more eligible adjacent pixels can be found. Finally, record the coordinates and corresponding load values of the identified uneven distribution area for subsequent analysis.

[0107] Finally, extract the load data of the high roof and the low roof from the high and low roof load distribution map, and use the gradient algorithm to calculate the load difference between the two. Specifically, the load difference map can be obtained by calculating the load value difference between the high roof and the low roof at the same position. The formula is:

[0108] ΔP(x,y)=P high (x,y)-P low (x,y)ΔP(x,y)=P high (x,y)-P low (x,y)

[0109] where P high (x,y) and P low (x,y) are the load values of the high roof and the low roof respectively. In this way, the load difference between the high and low roofs can be clearly shown, providing an important basis for subsequent structural safety assessment and maintenance decision-making.

[0110] Preferably, input the position information of the uneven distribution area, the load difference, and the deformation trend into the trained load monitoring model to obtain the monitoring results, including:

[0111] Obtain the preset position information dataset, load difference dataset, and deformation trend dataset;

[0112] Construct a deep learning model;

[0113] Train the deep learning model according to the position information dataset, the load difference dataset, and the deformation trend dataset respectively to obtain the trained first classifier, second classifier, and third classifier;

[0114] Cascade the trained first classifier, second classifier, and third classifier to obtain a classification network;

[0115] Connect the trained LSTM neural network after the classification network to obtain the load monitoring model;

[0116] Monitoring the location information, load difference, and deformation trend of the uneven distribution area according to the load monitoring model to obtain the monitoring results.

[0117] Specifically, in this embodiment, it is first necessary to obtain a preset location information dataset, a load difference dataset, and a deformation trend dataset. These datasets should contain a sufficient sample size to ensure the training effect of the model. The location information dataset should include the coordinates and related features of the uneven distribution area, the load difference dataset should contain the load difference values between the high and low roofs, and the deformation trend dataset should record the deformation changes of the roof structure. The preparation of these datasets is the basis for subsequent model training.

[0118] Next, a deep learning model is constructed to process the above data. A deep learning architecture suitable for classification tasks can be selected, such as a convolutional neural network (CNN) or a fully connected neural network (DNN). The input layer of the model should match the feature dimensions of the dataset, and the number of hidden layers and nodes can be adjusted according to the complexity of the dataset. The output layer of the model should be set to multi-classification output to predict different load states or deformation trends based on the input data. After construction, a loss function and an optimization algorithm need to be defined to continuously adjust the model parameters during training.

[0119] After the model is constructed, the deep learning model is trained according to the location information dataset, the load difference dataset, and the deformation trend dataset respectively to obtain the trained first classifier, second classifier, and third classifier. Each classifier should focus on specific dataset features, optimize the model parameters through the backpropagation algorithm, and minimize the error between the prediction result and the actual label. During the training process, the cross-validation method can be used to evaluate the performance of the model to ensure the accuracy and generalization ability of each classifier.

[0120] Finally, the trained first classifier, second classifier, and third classifier are cascaded to form a comprehensive classification network, and a trained long short-term memory (LSTM) neural network is connected after this classification network to obtain the final load monitoring model. The LSTM network can effectively process time series data and capture the dynamic changes of the deformation trend. According to this load monitoring model, the location information, load difference, and deformation trend of the uneven distribution area are monitored, and finally the monitoring results are obtained. These results will provide an important basis for structural safety assessment and maintenance decision-making, helping engineers to identify potential risks in a timely manner and take corresponding measures.

[0121] Exemplarily, the monitoring results of this embodiment include the location information of the uneven distribution area, the load difference values of each area, and the deformation trend of the roof structure. These results are generated by a trained load monitoring model, which can provide a comprehensive assessment of the safety status of the high-low roof structure, help engineers identify potential risk areas, and take maintenance measures in a timely manner, thus ensuring the safety and stability of the building. Specifically, the monitoring results will reveal which areas are at risk of overload, the specific values of load changes, and the dynamic changes in structural deformation, providing a scientific basis for subsequent decision-making.

[0122] Corresponding to the above method, as Figure 2 shown, this embodiment also provides a method for monitoring the uneven distribution of snow load on a high-low roof, including:

[0123] A sensing data acquisition unit, configured to arrange load sensors, displacement sensors, and environmental sensors at each preset monitoring point of the high-low roof structure, and use the load sensors to measure the load data of local snow accumulation on the roof, and use the displacement sensors to monitor the deformation data of the roof structure;

[0124] A data preprocessing unit, configured to perform denoising processing on the load data and the deformation data to obtain load denoising information and deformation denoising information;

[0125] A distribution map generation unit, configured to determine the snow load per unit area of each area of the high-low roof according to the load denoising information, and generate a snow load distribution map using an interpolation algorithm;

[0126] A trend change unit, configured to analyze the deformation trend of the roof structure according to the deformation denoising information;

[0127] A load difference calculation unit, configured to identify the uneven distribution area of the snow load by analyzing the load distribution maps of the high-low roof, and calculate the load difference between the high-low roof using a gradient algorithm;

[0128] A model monitoring unit, configured to input the location information of the uneven distribution area, the load difference, and the deformation trend into a trained load monitoring model to obtain monitoring results.

[0129] The beneficial effects of the present invention are as follows:

[0130] (1) By arranging load sensors, displacement sensors, and environmental sensors at the preset monitoring points of the high-low roof structure, the present invention can obtain the local snow load and structural deformation data of the roof in real time. This real-time monitoring can timely detect potential safety hazards; through the analysis of the load and deformation data, an early warning can be issued in a timely manner when the snow load reaches a dangerous level or the structural deformation exceeds the safety range, helping relevant personnel take necessary emergency measures.

[0131] (2) By denoising the load data and generating a load distribution map through an interpolation algorithm, the present invention can accurately evaluate the snow load per unit area in each region of the high and low roofs. This precise load evaluation helps to judge the safety of the structure; by analyzing the deformation trend of the roof structure in combination with the denoised deformation information, the response of the structure under the action of the load can be better understood, and its bearing capacity and stability can be evaluated.

[0132] (3) By analyzing the load distribution map, the present invention can identify the uneven distribution regions of the snow load. This is of great significance for the maintenance and reinforcement of the high and low roof structures, and can effectively prevent structural damage caused by local overloading; calculating the load difference between the high and low roofs using the gradient algorithm can provide data support for subsequent structural optimization and design.

[0133] (4) By inputting the position information, load difference, and deformation trend of the uneven distribution region into the trained load monitoring model, the present invention can obtain more accurate monitoring results. This data-driven method can provide a scientific decision-making basis for engineers and optimize the maintenance and management strategies; this method can be long-term applied to the monitoring of high and low roof structures, accumulate data and conduct trend analysis, providing a reference for future design and improvement.

[0134] (5) By timely identifying and analyzing the uneven load distribution and its impact on the structure, the present invention can carry out maintenance and reinforcement in advance, extending the service life of the structure; through comprehensive monitoring and analysis, the risk of structural failure caused by snow load can be effectively reduced, improving the safety of the building.

[0135] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the system disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0136] Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention; at the same time, for those of ordinary skill in the art, there will be changes in the specific implementation manner and application scope according to the idea of the present invention. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A monitoring method for uneven distribution of snow load on high and low roofs, characterized in that Including: Load sensors, displacement sensors, and environmental sensors are arranged at each preset monitoring point of the high-low roof structure. The load sensors are used to measure the load data of local snow accumulation on the roof, and the displacement sensors are used to monitor the deformation data of the roof structure; The load data and the deformation data are denoised to obtain load denoised information and deformation denoised information; According to the load denoised information, the snow load per unit area of each region of the high-low roof is determined, and an interpolation algorithm is used to generate a snow load distribution map; According to the deformation denoised information, the deformation trend of the roof structure is analyzed; By analyzing the high-low roof load distribution map, the uneven distribution areas of the snow load are identified, and a gradient algorithm is used to calculate the load difference between the high and low roofs; The position information of the uneven distribution area, the load difference, and the deformation trend are input into the trained load monitoring model to obtain the monitoring result.

2. The monitoring method for uneven distribution of snow load on high-low houses according to claim 1, characterized in that Denoising the load data and the deformation data to obtain load denoised information and deformation denoised information, including: Translating the signal where the load data and the deformation data are located to obtain a translated signal; Performing wavelet decomposition on the translated signal to obtain multiple wavelet coefficients; Determining a filtering threshold according to the decomposition scale and length of the signal; Constructing a denoising function according to the filtering threshold; Using the denoising function to remove the noise of the translated signal to obtain the load denoised information and the deformation denoised information.

3. The monitoring method for uneven distribution of snow load on high-low houses according to claim 2, characterized in that The calculation formula of the filtering threshold is: where λ is the filtering threshold, j is the decomposition scale of the translated signal, d j is the wavelet coefficient at the decomposition scale j, N represents the length of the translated signal, and median represents the median operation.

4. The monitoring method for uneven distribution of snow load on high-low houses according to claim 3, characterized in that Using the denoising function to remove the noise of the translated signal to obtain the load denoised information and the deformation denoised information, including: Using the denoising function to remove the corresponding wavelet coefficients to obtain a smoothed signal; where the denoising function is: Wherein, X represents a wavelet coefficient, and n represents an adjustment coefficient; Performing inverse translation on the smoothed signal to obtain an inversely translated signal; Calculating the average of the inversely translated signal to obtain the load denoised information and the deformation denoised information.

5. The monitoring method for uneven distribution of snow load on high-low houses according to claim 1, characterized in that, Determining the snow load per unit area of each region of the high-low roof according to the load denoised information and using an interpolation algorithm to generate a snow load distribution map, including: Determining the geometric area range of the high-low roof according to the architectural design drawings and the sensor layout plan; Calculating the mean value of the load data in the geometric area range and taking the mean value as the snow load per unit area; Based on the inverse distance weighted method, spatial interpolation is performed on the load data between each preset monitoring point to generate a snow load distribution map of the entire roof.

6. The monitoring method for uneven distribution of snow load on high-low houses according to claim 1, characterized in that Analyzing the deformation trend of the roof structure according to the deformation denoised information, including: Constructing a linear regression model; Using the linear regression model to fit the deformation data to identify the deformation trend of the roof structure.

7. The monitoring method for uneven distribution of snow load on high-low houses according to claim 1, characterized in that, By analyzing the high-low roof load distribution map, identifying the uneven distribution areas of the snow load, and using a gradient algorithm to calculate the load difference between the high and low roofs, including: Setting a load difference threshold according to engineering experience or historical data to judge whether the load distribution is uniform; Using an edge detection algorithm to detect the edges in the high-low roof load distribution map to identify the areas where the load changes; Start expanding outward from the edge points detected by the edge detection algorithm until the set load difference threshold is reached, forming an uneven distribution area, and record the coordinates and load values of the identified uneven distribution area; Extract the load data of the high roof and the low roof from the high-low roof load distribution map, and use the gradient algorithm to calculate the load difference between the high and low roofs.

8. The monitoring method for uneven distribution of snow load on high-low houses according to claim 1, characterized in that, Input the position information of the uneven distribution area, the load difference, and the deformation trend into the trained load monitoring model to obtain the monitoring results, including: Obtain the preset position information dataset, load difference dataset, and deformation trend dataset; Construct a deep learning model; Train the deep learning model according to the position information dataset, the load difference dataset, and the deformation trend dataset respectively to obtain the trained first classifier, second classifier, and third classifier; Cascade the trained first classifier, second classifier, and third classifier to obtain a classification network; Connect the trained LSTM neural network after the classification network to obtain the load monitoring model; Monitor the position information of the uneven distribution area, the load difference, and the deformation trend according to the load monitoring model to obtain the monitoring results.

9. A monitoring method for the uneven distribution of snow load on high and low roofs, characterized in that, Including: A sensing data acquisition unit, which is used to arrange load sensors, displacement sensors, and environmental sensors at each preset monitoring point of the high-low roof structure, and use the load sensors to measure the load data of local snow accumulation on the roof, and use the displacement sensors to monitor the deformation data of the roof structure; A data preprocessing unit, which is used to denoise the load data and the deformation data to obtain load denoised information and deformation denoised information; A distribution map generation unit, which is used to determine the snow load per unit area of each area of the high and low roofs according to the load denoised information, and use the interpolation algorithm to generate a snow load distribution map; A trend change unit, which is used to analyze the deformation trend of the roof structure according to the deformation denoised information; A load difference calculation unit, which is used to identify the uneven distribution area of the snow load by analyzing the high-low roof load distribution map, and use the gradient algorithm to calculate the load difference between the high and low roofs; A model monitoring unit, which is used to input the position information of the uneven distribution area, the load difference, and the deformation trend into the trained load monitoring model to obtain the monitoring results.

Images