Dam crack and reinforcing steel bar stress combined prediction method, equipment and medium

Abstract

The invention discloses a dam crack and steel bar stress combined prediction method, which comprises the following steps: S1, collecting dam monitoring data, and weighting temperature data and reservoir water level data after preprocessing operation; s2, constructing multi-scale features for the data by using a rolling window technology; s3, inputting multi-scale characteristics of a plurality of continuous moments, constructing a time sequence sample by adopting a sliding window method, and taking crack data and steel bar stress data of a window at the last moment as target variables of joint prediction; s4, constructing a multi-task time sequence model and training a time sequence sample for independently outputting predicted values of crack and reinforcement stress through the time sequence sample; and S5, performing evaluation and parameter optimization on the multi-task time sequence model by adopting a specified performance evaluation index, applying the trained model, and providing early warning information of dam crack and reinforcement stress through a combined predicted target variable. The method can improve the prediction accuracy and reliability through a weighting mechanism and a multi-task learning model.

Description

Technical Field

[0001] The present invention relates to the technical field of building safety monitoring, and particularly to a method, device, and storage medium for jointly predicting dam cracks and steel bar stress based on multi-scale feature fusion. Background Art

[0002] As an important water conservancy project facility, the structural safety of a reservoir dam is directly related to the flood control safety of the downstream area and the protection of people's lives and property. During the operation of the dam, the generation of cracks and the change of steel bar stress are key factors affecting the structural safety of the dam. Therefore, real-time and accurate monitoring and prediction of dam cracks and steel bar stress are of great significance for preventing structural failure and formulating maintenance strategies.

[0003] Traditional dam monitoring methods mainly rely on single-type sensor data, such as piezometers, strain gauges, etc. However, these methods often have problems such as single data dimension and insufficient information content, making it difficult to reflect the true state of the dam structure in a timely and comprehensive manner. In addition, traditional methods usually analyze cracks and steel bar stress independently, failing to fully consider the coupling relationship between the two, which may lead to insufficient accuracy of the prediction results.

[0004] Therefore, time series models such as long short-term memory networks (LSTM) or gated recurrent units (GRU) are used in the prior art to analyze dam monitoring data. However, these methods usually only target a single target variable and fail to consider the joint prediction requirements of cracks and steel bar stress at the same time. At the same time, a multi-task learning model with a shared encoder and a dual-branch decoder is also used in the process of analyzing monitoring data. Therefore, in specific applications, the design of the model structure still needs to be adjusted according to actual needs, and there are still deficiencies in generality and generalization.

[0005] For this reason, the present application specifically proposes a method for jointly predicting dam cracks and steel bar stress to solve the above technical problems. Summary of the Invention

[0006] The main object of the present invention is to provide a method for jointly predicting dam cracks and steel bar stress to solve the technical problems proposed in the background art.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] A method for jointly predicting dam cracks and steel bar stress, comprising:

[0009] S1. Collect dam monitoring data, and perform weighted processing on temperature data and reservoir water level data according to preset weights after preprocessing operations;

[0010] S2. Use the rolling window technique to construct multi-scale features for the weighted temperature and water level data;

[0011] S3. Input multi-scale features at consecutive moments, use the sliding window method to construct time series samples, and take the crack data and steel bar stress data at the last moment of the window as the target variables for joint prediction;

[0012] S4. Construct a multi-task time series model and train it with the time series samples. The multi-task time series model is used to independently output the predicted values of cracks and steel bar stresses through the time series samples;

[0013] S5. Evaluate and optimize the parameters of the multi-task time series model using specified performance evaluation indicators, and apply the trained model to the prediction of real-time dam safety monitoring data to provide early warning information on dam cracks and steel bar stresses through the target variables of joint prediction.

[0014] Preferably, the specific operation process of step S1 includes:

[0015] S11. Collect crack data and steel bar stress data, with the data volume N of crack data c , the data volume N of steel bar stress data s and the data volume N of water level data H ;

[0016] S12. Inner-join and merge the crack data and steel bar stress data based on time t to form joint monitoring data:

[0017] JointData(t) = {(t, C(t), S(t))|t ∈ T C ∩ T S}

[0018] where t is the data collection time, T C and T S represent the collection time sets of crack data and steel bar stress data respectively, C(t) is the data set of crack sizes, and S(t) is the data set of steel bar stress values;

[0019] S13. For each moment t in the joint monitoring data, align the water level data using the approximate time-based merging method;

[0020] S14. Preset the temperature weight as α and the reservoir water level weight as β = 1 - α, and weight the temperature and reservoir water level data respectively to obtain the weighted temperature data M'(t) and water level data H'(t). The calculation formulas are:

[0021] M'(t) = α · M(t)

[0022] H'(t) = (1 - α)H(t)

[0023] Among them, M(t) is the temperature data at the acquisition time t, and H(t) is the reservoir water level data at the acquisition time t.

[0024] Preferably, the specific operation process of aligning the water level data in step S13 includes:

[0025] For each joint monitoring time t, find the time t in the water level data that is closest to t * , there is:

[0026]

[0027] Among them, T H is the time set of the water level data;

[0028] Let the water level value H(t * ) at this moment be used as the water level of the joint monitoring data at time t, and there is reservoir water level data H(t) = H(t * ).

[0029] Preferably, the multi-scale features in step S2 include, but are not limited to, the cumulative values, means, and standard deviations of short-term features, medium-term features, and long-term features as the input feature set. For any weighted data X′(t), the statistical features within the corresponding window at time t are defined as:

[0030]

[0031]

[0032]

[0033] Among them, Sum n (t) is the cumulative value of the multi-scale features, n is the rolling window size, μ n (t) is the mean of the multi-scale features, and σ n (t) is the standard deviation of the multi-scale features;

[0034] Apply the above formulas to the weighted temperature data M′(t) and water level data H′(t) at the acquisition time t respectively to construct a multi-scale temperature and water level statistical feature set for subsequent joint prediction model training.

[0035] Preferably, the specific operation process of constructing time series samples by using the sliding window method in step S3 includes:

[0036] S31. Set the multi-scale input feature vector obtained at each time point t in the joint monitoring data as F(t) ∈ R d , where F(t) includes temperature, reservoir water level, and various scale statistical features calculated by its rolling window, and R dis a d-dimensional vector feature map;

[0037] S32. At time t, set the target variables as the crack size C(t) and the steel bar stress S(t);

[0038] S33. Use the sliding window method to construct time series samples. Let the window size be n. Then the input X of the i-th time series sample i and the target y i are respectively expressed as:

[0039] X i = [F(t i ), F(t i+1 ), …, F(t i+n-1 )] ∈ R n×d ,

[0040] y i = [C(t i+n-1 ), S(t i+n-1 )] ∈ R 2 .

[0041] where i = 1, 2, …, τ, where τ = N - n + 1, N is the total number of data time instants, and R n×d is an n×d-dimensional channel vector feature map.

[0042] Preferably, in the S4 step, the multi-task time series model includes a shared encoder module and a dual-branch decoder module, where:

[0043] (1) The shared encoder, with the input being the time series sample X i ∈ R n×d , extracts the local and global dynamic features of the input time series data through convolutional layers and a recurrent neural network. The process is denoted as:

[0044] h i = f enc (X i )

[0045] where h i ∈ R p is the shared feature representation, and f enc represents an encoder function composed of a convolutional neural network and a recurrent neural network. R n×d is an n×d-dimensional channel vector feature map, and R p is a p-dimensional vector feature map;

[0046] (2) The dual-branch decoder is used to process the shared features respectively. The shared feature h i is then sent to two independent decoders respectively, and the predicted values of the crack and the steel bar stress are output independently. There are:

[0047]

[0048] Among them, f dec1 and f dec2 respectively represent the decoder modules for predicting crack and steel bar stress, and the decoder modules are composed of fully connected layers. and are the crack size and steel bar stress predicted by the model respectively.

[0049] Preferably, the training objective of the multi-task time series model in the S4 step is to minimize the mean square error of all samples, and its overall loss function is expressed as:

[0050]

[0051] where γ1 and γ2 are weight parameters to ensure a reasonable balance of the task contributions of the crack size C(t) and the steel bar stress S(t) in the loss function.

[0052] Preferably, the data processing operation flow of the multi-task time series model in the S5 step includes:

[0053] S51. For the i-th sample, the prediction target y i at the last moment of the output window and the predicted value after multiple rounds of prediction are:

[0054]

[0055] where C i and respectively represent the true and predicted crack sizes, and S i and respectively represent the true and predicted steel bar stresses;

[0056] S52. Use the root mean square error and the coefficient of determination R 2 as indicators to evaluate the model performance, and optimize the parameters of the multi-task time series model based on the evaluation results. The calculation formula for the indicator evaluation is:

[0057]

[0058]

[0059] where, is the mean vector of all true targets;

[0060] S53. After model training and parameter optimization, use the trained multi-task time series model for predicting real-time dam safety monitoring data. For the new input time series data X new , the prediction result is expressed as:

[0061]

[0062] Among them, f(·) represents the trained joint prediction model, and the predicted results of crack size and steel bar stress are respectively used to warn of abnormal dam cracks and steel bar stress.

[0063] On the other hand, the present invention also discloses a computer-readable storage medium storing a computer program, which when executed by a processor causes the processor to execute the steps of the above method.

[0064] On yet another aspect, the present invention also discloses a computer device including a memory and a processor, the memory storing a computer program, which when executed by the processor causes the processor to execute the steps of the above method.

[0065] As can be seen from the above technical solutions, the present invention provides a method for joint prediction of dam cracks and steel bar stress. Compared with the prior art, the present invention has the following advantages:

[0066] 1. By time alignment, missing value filling and weighted fusion of data, the present invention significantly improves the integrity and reliability of data, and ensures the accurate matching of multi-source monitoring data such as crack monitoring, steel bar stress and reservoir water level.

[0067] 2. By constructing multi-scale statistical features and adopting a multi-task time series model combining a shared encoder and a dual-branch decoder, the present invention can simultaneously predict the crack propagation trend and the change of steel bar stress. At the same time, local features are extracted by a convolutional neural network, and the long-term dependence of time series data is captured by combining a recurrent neural network, improving the model's ability to model complex dynamic changes.

[0068] 3. By using a shared encoder to extract local and global dynamic features of the input time series data and adopting a dual-branch decoder to separately predict cracks and steel bar stress, the present invention not only ensures the efficiency of feature sharing but also takes into account the characteristics of different physical quantities, improving the prediction accuracy and stability.

[0069] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Of course, any product implementing the present invention does not necessarily need to achieve all the above advantages simultaneously. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] The accompanying drawings forming a part of this application are used to provide a further understanding of the present invention. The schematic embodiments and descriptions thereof of the present invention are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0071] Figure 1 is a schematic diagram of the overall process of the present invention;

[0072] Figure 2 is a schematic diagram of the training optimization process of the multi-task timing model of the present invention;

[0073] Figure 3 is a schematic diagram of the operation method process of the multi-task timing model structure of the present invention;

[0074] Figure 4 is a schematic diagram of the model prediction experimental results of the monitoring point J01-01 [1# stack | crack] in the embodiment of the present invention;

[0075] Figure 5 is a schematic diagram of the model prediction experimental results of the monitoring point R02-02 [2# stack | steel bar stress] in the embodiment of the present invention. Detailed implementation manners

[0076] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0077] In the embodiment, refer in detail to Figures 1 to 5 .

[0078] As Figure 1 and Figure 2 shown. A method for jointly predicting dam cracks and steel bar stresses based on multi-scale feature fusion proposed in the embodiment of the present invention can improve the accuracy and reliability of prediction through a reasonable weighting mechanism and a multi-task learning model. The specific implementation steps include:

[0079] S1. Collect dam monitoring data. After preprocessing operations, the temperature data and reservoir water level data are weighted respectively according to preset weights to form weighted temperature data and water level data.

[0080] At this time, the specific operation process includes:

[0081] S11. Data reading and sorting: Collect crack data and steel bar stress data, and there are:

[0082] The crack data is where t i represents the acquisition time, M i represents the temperature, and C i represents the crack size;

[0083] The steel bar stress data is where S j represents the steel bar stress value;

[0084] The reservoir water level data is where H k represents the water level,

[0085] Due to the difference in the acquisition frequencies of different devices, there are the data volumes N of the crack data c , the data volumes N of the steel bar stress data s and the data volumes N of the water level data H ;

[0086] S12. Inner join and merge the crack data and the steel bar stress data based on the time t to form combined monitoring data:

[0087] JointData(t) = {(t, C(t), S(t)) | t ∈ T C ∩ T S}

[0088] where t is the data acquisition time, and T C and T S represent the sets of acquisition times of the crack data and the steel bar stress data respectively, C(t) is the data set of the crack size, and S(t) is the data set of the steel bar stress values;

[0089] S13. For each moment t in the combined monitoring data, use the approximate merge by time (merge_asof) method to align the water level data. The specific operation process includes:

[0090] Assume the water level data is H(t). For each combined monitoring moment t, find the moment t * in the water level data that is closest to t, and there is:

[0091]

[0092] where T H is the time set of the water level data;

[0093] Let the water level value H(t * ) at this moment be the water level of the combined monitoring data at moment t, and the reservoir water level data H(t) = H(t * );

[0094] Finally, it forms combined monitoring data including time t, crack size C(t), steel bar stress S(t), and the aligned water level H t of.

[0095] After completing the above steps, the following data table structure is obtained, with:

[0096] Time t Temperature M Crack size C Steel bar stress S Reservoir water level H <![CDATA[t1]]> <![CDATA[M1]]> <![CDATA[C1]]> <![CDATA[S1]]> <![CDATA[H1]]> <![CDATA[t2]]> <![CDATA[M2]]> <![CDATA[C2]]> <![CDATA[S2]]> <![CDATA[H2]]> <![CDATA[t3]]> <![CDATA[M3]]> <![CDATA[C3]]> <![CDATA[S3]]> <![CDATA[H3]]> … … … … …

[0097] At this time, through time alignment, missing value filling, and weighted fusion of the data, the integrity and reliability of the data are significantly improved, ensuring the accurate matching of multi-source monitoring data such as crack monitoring, steel bar stress, and reservoir water level.

[0098] S14. Set the preset temperature weight as α = 0.3 and the reservoir water level weight as β = 1 - α. Weight the temperature and reservoir water level data respectively to obtain the weighted temperature data M′(t) and water level data H′(t). The calculation formulas are:

[0099] M′(t) = α·M(t)

[0100] H′(t) = (1 - α)H(t)

[0101] where M(t) is the temperature data at the acquisition time t, and H(t) is the reservoir water level data at the acquisition time t.

[0102] S2. Use the rolling window technique to construct multi-scale features for the weighted temperature and water level data.

[0103] The multi-scale features at this time include, but are not limited to, the cumulative values, means, and standard deviations of short-term features (the last 3 moments), medium-term features (the last 6 moments), and long-term features (the last 21 moments) as the input feature set;

[0104] To fully capture the changing trends of the data on the time scale, for the weighted temperature M′(t) and reservoir water level H′(t), the rolling window technique is used to calculate the statistical features at different scales. Let the rolling window size be n (n = 3 for short-term, n = 6 for medium-term, n = 21 for long-term). For any weighted data X′(t) (where X represents M or H), the statistical features within the corresponding window at time t are defined as:

[0105]

[0106]

[0107]

[0108] where Sum n (t) is the cumulative value (accumulated value) of the multi-scale feature, n is the rolling window size, μn (t) is the mean of the multi-scale features, and σ n (t) is the standard deviation of the multi-scale features;

[0109] Apply the above formulas to the weighted temperature data M′(t) and water level data H′(t) at the acquisition time t respectively to construct a multi-scale temperature and water level statistical feature set for subsequent joint prediction model training.

[0110] S3. Input the multi-scale features at several consecutive moments, use the sliding window method to construct time series samples, and use the crack data and steel bar stress data at the last moment of the window as the target variables for joint prediction. For example, use the environmental features in a certain past period (such as 30 time steps, about 10 days or longer) as the model input, and the target is the crack and steel bar stress data corresponding to the end of the window.

[0111] The specific operation process of constructing time series samples using the sliding window method includes:

[0112] S31. Set the multi-scale input feature vector obtained at each time point t in the joint monitoring data as F(t)∈R d , where F(t) includes temperature, reservoir water level and various scale statistical features (accumulated value, mean, standard deviation) calculated by its rolling window, and R d is a d-dimensional vector feature map;

[0113] S32. At time t, set the target variables as the crack size C(t) and the steel bar stress S(t);

[0114] S33. Use the sliding window method to construct time series samples. Let the window size be n (n = 21 represents the observed data of the past 21 moments), then the input X i and the target y i are respectively expressed as:

[0115] X i =[F(t i ),F(t i+1 ),…,F(t i+n-1 )]∈R n×d ,

[0116] y i =[C(t i+n-1 ),S(t i+n-1 )]∈R 2 .

[0117] Among them, i = 1, 2, …, τ, where τ = N - n + 1, N is the total number of data moments, and R n×d is an n×d-dimensional channel vector feature map.

[0118] S4. Construct a multi-task time series model and train the time series samples. As Figure 3 shown, after normalization, the multi-task time series model is used to jointly predict the monitoring data. The multi-task time series model is used to independently output the predicted values of cracks and steel bar stresses through the time series samples.

[0119] Specifically, the multi-task time series model includes a shared encoder module and a dual-branch decoder module, where:[[]]

[0120] (1) The shared encoder, with the input being the time series sample X i ∈R n×d , extracts the local and global dynamic features of the input time series data through convolutional layers and a recurrent neural network (LSTM or GRU). The process is denoted as:

[0121] h i =f enc (X i )

[0122] where h i ∈R p is the shared feature representation, and f enc represents an encoder function composed of a convolutional neural network (CNN) and a recurrent neural network (LSTM or GRU). R n×d is an n×d-dimensional channel vector feature map, and R p is a p-dimensional vector feature map;

[0123] (2) The dual-branch decoder is used to process the shared features respectively. The shared feature h i is then sent into two independent decoders respectively, independently outputting the predicted values of cracks and steel bar stresses. There are:

[0124]

[0125] where f dec1 and f dec2 represent the decoder modules for predicting cracks and steel bar stresses respectively. The decoder modules are composed of fully connected layers, and are the crack size and steel bar stress predicted by the model respectively.

[0126] Therefore, local and global dynamic features of the input time-series data are extracted through a shared encoder, and a dual-branch decoder is used to predict crack and steel bar stress respectively, which not only ensures the efficiency of feature sharing but also takes into account the characteristics of different physical quantities, improving the prediction accuracy and stability. Furthermore, by constructing multi-scale statistical features and adopting a multi-task time-series model combining a shared encoder and a dual-branch decoder, the crack propagation trend and the change of steel bar stress can be predicted simultaneously. At this time, local features are extracted through a convolutional neural network, and the long-term dependence of time-series data is captured by combining with a recurrent neural network, improving the model's ability to model complex dynamic changes.

[0127] At this time, the training objective of the multi-task time-series model is to minimize the mean square error of all samples, and its overall loss function is expressed as:

[0128]

[0129] where γ1 = γ2 = 1 are weight parameters to ensure a reasonable balance of the task contributions of crack size C(t) and steel bar stress S(t) in the loss function.

[0130] In summary, the sliding window method is used to construct time-series samples, and the joint monitoring data is trained through a multi-task time-series model (shared encoder and dual-branch decoder structure) to realize the joint prediction of dam cracks and steel bar stress, providing an accurate early warning basis for dam safety monitoring.

[0131] S5. Use the multi-task time-series model to predict the entire dataset and display the prediction results in the form of a time axis. Then, use specified performance evaluation indicators such as mean square error to evaluate and optimize the parameters of the multi-task time-series model, and apply the trained model to the prediction of real-time dam safety monitoring data to provide early warning information on dam cracks and steel bar stress through the jointly predicted target variables.

[0132] The data processing operation flow of the multi-task time-series model includes:

[0133] S51. Model output and true value representation: For the i-th sample, output its target value (the predicted target y at the last moment of the window i ) and the predicted value after multiple rounds of prediction There are:

[0134]

[0135] where C i and respectively represent the true and predicted crack sizes, and S i and respectively represent the true and predicted steel bar stresses;

[0136] S52. Performance evaluation: In the model evaluation stage, the root mean square error (RMSE) and the coefficient of determination R 2 are used as indicators to evaluate the model performance, and the parameters of the multi-task time series model are optimized based on the evaluation results. The calculation formula for indicator evaluation is as follows:

[0137]

[0138]

[0139] where, is the mean vector of all true targets;

[0140] S53. Real-time prediction application: After model training and parameter optimization, the trained multi-task time series model is used to predict real-time dam safety monitoring data. For the new input time series data X new , the prediction result is expressed as:

[0141]

[0142] where f(·) represents the trained joint prediction model. The prediction result of crack size and the prediction result of steel bar stress are respectively used to warn of abnormal dam cracks and steel bar stresses, and assist in dam safety management and risk warning.

[0143] In a specific application of this embodiment, the present invention applies the trained model to the prediction task of real-time data. The experimental results in Reservoir B of City A are as shown in Figure 3 and Figure 4 . The R 2 values of the monitoring points J01-01 [1# stack | crack] and R02-02 [2# stack | steel bar stress] reach 0.9747 and 0.9968 respectively. This result shows that this method can achieve the joint prediction of the two types of monitoring data and obtain better effects, providing a scientific and reliable basis for dam safety management and risk warning.

[0144] In summary, the joint prediction method for dam cracks and steel bar stresses based on multi-scale feature fusion disclosed by the present invention uses indicators such as the coefficient of determination (R 2 ) to evaluate the model performance and optimize the parameters, ensuring that the model has good robustness and prediction accuracy. Through the accurate joint prediction of dam monitoring data, it effectively fuses multi-source monitoring data, realizes the joint prediction of key dam safety indicators, provides a scientific decision-making basis for dam safety management and maintenance, and realizes the real-time warning of dam cracks and steel bar stresses, thus greatly improving the safety and risk prevention and control capabilities of dam operation.

[0145] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to execute the steps of the above method.

[0146] In yet another aspect, the present invention also discloses a computer device including a memory and a processor, where the memory stores a computer program, which, when executed by the processor, causes the processor to execute the steps of the above method.

[0147] In yet another embodiment provided by the present application, there is also provided a computer program product containing instructions, which, when running on a computer, causes the computer to execute the method for jointly predicting dam cracks and steel bar stresses in any of the above embodiments.

[0148] It can be understood that the system provided by the embodiments of the present invention corresponds to the method provided by the embodiments of the present invention. For the explanations, examples, and beneficial effects of related content, reference can be made to the corresponding parts in the above method.

[0149] The embodiments of the present application also provide an electronic device, including a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus.

[0150] The memory is used to store a computer program.

[0151] The processor is used to implement the method for jointly predicting dam cracks and steel bar stresses when executing the program stored on the memory.

[0152] The communication bus mentioned in the above electronic device may be a peripheral component interconnect standard bus or an extended industry standard architecture bus, etc. This communication bus can be divided into an address bus, a data bus, a control bus, etc.

[0153] The communication interface is used for communication between the above electronic device and other devices.

[0154] The memory may include a random access memory and may also include a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located far from the aforementioned processor.

[0155] The above-mentioned processor may be a general-purpose processor, including a central processing unit, a network processor, etc.; it may also be a digital signal processor, an application-specific integrated circuit, a field-programmable gate array, or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0156] It should also be noted that the electronic device further includes a terminal device, which can also be referred to as a terminal, user equipment, mobile station, mobile terminal, etc. The terminal device can be a mobile phone, smart TV, wearable device, tablet computer, computer with wireless transceiver function, virtual reality terminal device, augmented reality terminal device, wireless terminal in industrial control, wireless terminal in unmanned driving, wireless terminal in remote surgery, wireless terminal in smart grid, wireless terminal in transportation safety, wireless terminal in smart city, wireless terminal in smart home, and so on. The embodiments of the present application do not limit the specific technologies and specific device forms adopted by the terminal device.

[0157] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line) or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that includes one or more available media integrated. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium, or a semiconductor medium (such as a solid-state drive), etc.

[0158] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

[0159] In addition, it should be noted that if there are directional indications (such as up, down, left, right, front, back...) involved in the embodiments of the present invention, the directional indications are only used to explain the relative position relationship and movement conditions between components in a specific posture. If the specific posture changes, the directional indications will also change accordingly.

[0160] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only for descriptive purposes and should not be construed as indicating or implying their relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In addition, the meaning of "and / or" appearing throughout the text includes three parallel scenarios. Taking "A and / or B" as an example, it includes Scenario A, or Scenario B, or the scenario where both A and B are satisfied simultaneously. In addition, in the embodiments of the present invention, "a plurality of" means more than two. In addition, the technical solutions between various embodiments can be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions results in contradictions or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

Claims

1. A method for jointly predicting dam cracks and steel bar stresses, characterized in that Including: S1. Collect dam monitoring data. After preprocessing operations, perform weighted processing on the temperature data and reservoir water level data respectively according to preset weights. S2. Use the rolling window technique to construct multi-scale features for the weighted temperature and water level data. S3. Input the multi-scale features at consecutive moments. Adopt the sliding window method to construct time series samples, and use the crack data and steel bar stress data at the last moment of the window as the target variables for joint prediction. S4. Construct a multi-task time series model and train the time series samples. The multi-task time series model is used to independently output the predicted values of cracks and steel bar stresses through the time series samples. S5. Use specified performance evaluation indicators to evaluate and optimize the parameters of the multi-task time series model, and apply the trained model to the prediction of real-time dam safety monitoring data to provide early warning information on dam cracks and steel bar stresses through the target variables of joint prediction.

2. The method for jointly predicting dam cracks and steel bar stresses as described in claim 1, characterized in that, The specific operation process of step S1 includes: S11. Collect crack data and steel bar stress data. The amount of crack data N within the specified time c , Steel bar stress data volume N s And water level data volume N H ; S12. Inner join and merge the crack data and steel bar stress data based on time t to form combined monitoring data: JointData(t) = {(t, C(t), S(t)) | t ∈ T C ∩ T S} where t is the data acquisition time, T C and T S respectively represent the sets of acquisition times of crack data and steel bar stress data, C(t) is the set of crack size data, and S(t) is the set of steel bar stress values; S13. For each moment t in the combined monitoring data, use the approximate merging method by time to align the water level data. S14. Preset the temperature weight as α and the reservoir water level weight as β = 1 - α. Weight the temperature and reservoir water level data respectively to obtain the weighted temperature data M′(t) and water level data H′(t). The calculation formulas are: M′(t) = α·M(t) H′(t) = (1 - α)H(t) where M(t) is the temperature data at the collection time t, and H(t) is the reservoir water level data at the collection time t.

3. The method for jointly predicting dam cracks and steel bar stresses according to claim 2, characterized in that, The specific operation process of aligning the water level data in step S13 includes: For each joint monitoring time t, find the time t in the water level data that is closest to t * , there is: Among them, T H is the time set of water level data; Let the water level value H(t * ) at this moment be the water level of the combined monitoring data at time t, and the reservoir water level data H(t) = H(t * ).

4. The method for jointly predicting dam cracks and steel bar stresses as described in claim 1, characterized in that The multi-scale features in step S2 include, but are not limited to, the cumulative values, means, and standard deviations of short-term features, medium-term features, and long-term features as the input feature set. For any weighted data X′(t), the statistical features within the corresponding window at time t are defined as: Among them, Sum n (t) is the cumulative value of the multi-scale features, n is the rolling window size, μ n (t) is the mean value of the multi-scale features, σ n (t) is the standard deviation of the multi-scale features; Apply the above formulas to the weighted temperature data M′(t) and water level data H′(t) at the collection time t respectively to construct a multi-scale temperature and water level statistical feature set for subsequent joint prediction model training.

5. The method for jointly predicting dam cracks and steel bar stresses as described in claim 1, characterized in that, The specific operation process of constructing time series samples using the sliding window method in step S3 includes: S31. Set the multi-scale input feature vector obtained from the joint monitoring data at each time point t as F(t) ∈ R d , where F(t) includes temperature, reservoir water level, and various scale statistical features calculated from their rolling windows, and R d is a d-dimensional vector feature map; S32. At time t, set the target variables as the crack size C(t) and the steel bar stress S(t). S33. Use the sliding window method to construct time series samples. Let the window size be n, then the input X of the i-th time series sample i and the target y i are respectively expressed as: X i = [F(t i ), F(t i+1 ), …, F(t i+n-1 )] ∈ R n×d , y i = [C(t i+n-1 ), S(t i+n-1 )] ∈ R 2 . where \(i = 1, 2, \ldots, \tau\) and \(\tau=N - n + 1\), \(N\) being the total number of data time instances, and \(R\) n×d is an \(n\times d\) - dimensional channel - vector feature map.

6. The method for jointly predicting dam cracks and steel bar stresses according to claim 1, characterized in that, The multi-task time series model in step S4 includes a shared encoder module and a dual-branch decoder module, where: (1) Shared encoder, with the input being the time-series sample X i ∈R n×d , and the local and global dynamic features of the input time-series data are extracted through convolutional layers and recurrent neural networks, and the process is denoted as: h i = f enc (X i ) where h i ∈R p is the shared feature representation, f enc represents the encoder function composed of a convolutional neural network and a recurrent neural network, R n×d is an n×d-dimensional channel vector feature map, and R p is a p-dimensional vector feature map; (2) Dual-branch decoder, which is used to process the shared features respectively, and the shared feature h i Subsequently, they are respectively fed into two independent decoders to independently output the predicted values of crack and steel bar stress, and we have: Among them, f dec1 and f dec2 respectively represent the decoder modules for predicting crack and steel bar stress, and the decoder module is composed of fully connected layers, and are respectively the crack size and steel bar stress predicted by the model.

7. The method for jointly predicting dam cracks and steel bar stresses according to claim 6, characterized in that, The training objective of the multi-task time series model in step S4 is to minimize the mean square error of all samples. Its overall loss function is expressed as: where γ1 and γ2 are weight parameters to ensure a reasonable balance of the task contributions of the crack size C(t) and the steel bar stress S(t) in the loss function.

8. The method for jointly predicting dam cracks and steel bar stresses according to claim 1, wherein, The data processing operation process of the multi-task time series model in step S5 includes: S51. For the i-th sample, the predicted target y at the last moment of the output window i and the predicted value after multiple rounds of prediction are as follows: where C i and represent the true and predicted crack sizes respectively, and S i and represent the true and predicted steel bar stresses respectively; S52. Use the root mean square error and the coefficient of determination R 2 as indicators to evaluate the model performance, and optimize the parameters of the multi-task time series model based on the evaluation results. The calculation formula for the indicator evaluation is as follows: wherein, is the mean vector of all true targets; After model training and parameter optimization, the trained multi-task time series model is used to predict real-time dam safety monitoring data. For the new input time series data X new , the prediction result is expressed as: where f(·) represents the trained joint prediction model, and the predicted results of crack size and steel bar stress are respectively used to warn of abnormal dam cracks and steel bar stress.

9. A computer-readable storage medium, characterized in that, There is a computer program stored. When the computer program is executed by a processor, the processor executes the steps of the method according to any one of claims 1 to 8.

10. A computer device, characterized in that, Comprising a memory and a processor, the memory stores a computer program, and when the computer program is executed by the processor, the processor is caused to execute the steps of the method according to any one of claims 1 to 8.

Images