A buoyancy-resistant and pressure-relief drainage system for basement floor slabs

Abstract

This application relates to the field of intelligent buildings, and specifically discloses an anti-buoyancy pressure relief drainage system for the basement floor of a building. It uses a deep learning-based neural network model to mine the time-dimensional variation characteristics of the water level value of the accumulated water, and adaptively adjusts the power value of the water pump based on the variation of the water level value, thereby ensuring safety while reducing energy consumption.

Description

Technical Field

[0001] This application relates to the field of intelligent buildings, and more specifically, to a buoyancy-resistant pressure relief drainage system for the basement floor slab of a building. Background Technology

[0002] A basement is a room whose floor level is more than half the room's net height below the outdoor ground level. Multi-story and high-rise buildings require deeper foundations, and to utilize this height, basements are built beneath the ground floor. This increases usable floor space and eliminates the need for backfilling, improving land use efficiency. However, basements form the foundation of a building, and during rainy seasons, they often experience poor drainage and water accumulation. Prolonged water accumulation can cause the basement structure to rise, posing a safety hazard.

[0003] To address the aforementioned issues, Chinese Patent No. CN210031867U discloses a buoyancy-resistant, pressure-reducing drainage system for building basements. This system utilizes the buoyancy of accumulated water to push a buoyancy plate, causing a second insulating rod and connecting piece to rise, forming a closed loop between the second and first electric spheres. The suction pump is then energized, allowing water to flow through pipes, nozzles, suction pipes, and drain pipes to drain the basement. However, during use, it was found that when the water level rises faster than the pump's pumping speed, and the water level exceeds the containment tank, water enters the tank, submerging both the first and second electric spheres. Since water is conductive, this poses a safety hazard.

[0004] Therefore, an optimized anti-buoyancy pressure relief drainage system for building basement floors is desired. Summary of the Invention

[0005] To address the aforementioned technical problems, this application is proposed. Embodiments of this application provide a buoyancy-resistant and pressure-relief drainage system for the basement floor slab of a building. This system employs a deep learning-based neural network model to mine the time-dimensional variation characteristics of the water level, adaptively adjusting the power of the water pump based on these variations, thereby ensuring safety while reducing energy consumption.

[0006] According to one aspect of this application, a buoyancy-resistant and pressure-relief drainage system for a building basement floor is provided, comprising:

[0007] The data acquisition module is used to obtain the water level values ​​of the accumulated water at multiple predetermined time points within a predetermined time period, collected by the water level gauge.

[0008] The absolute water level data arrangement module is used to arrange the water level values ​​of the accumulated water at the multiple predetermined time points into a water level input vector according to the time dimension.

[0009] The water level relative data distribution module is used to calculate the difference between the water level values ​​at every two adjacent positions in the water level input vector to obtain the water level change input vector;

[0010] The water level time series feature extraction module is used to pass the water level change input vector and the water level input vector through a time series encoder containing a fully connected layer and a one-dimensional convolutional layer to obtain the absolute water level time series feature vector and the relative water level time series feature vector, respectively.

[0011] The feature fusion module is used to fuse the time-series feature vectors of absolute water level and relative water level based on the Gaussian density map to obtain a water level decoding feature matrix; and

[0012] The water pump power recommendation module is used to obtain a decoded value by passing the water level decoding feature matrix through a decoder. The decoded value is used to represent the recommended power value of the water pump at the current time point.

[0013] In the aforementioned basement floor anti-buoyancy and pressure relief drainage system, the water level temporal feature extraction module includes: a first fully connected encoding unit, used to perform fully connected encoding on the water level change input vector using the fully connected layer of the temporal encoder according to the following formula to extract the high-dimensional latent features of the feature values ​​at each position in the water level change input vector, wherein the formula is: Where X is the input vector of water level change, Y is the output vector, W is the weight matrix, and B is the bias vector. Represents matrix multiplication; the first one-dimensional convolutional coding unit is used to perform one-dimensional convolutional coding on the water level change input vector using the one-dimensional convolutional layer of the temporal encoder according to the following formula to extract high-dimensional implicit correlation features between feature values ​​at each position in the water level change input vector, wherein the formula is:

[0014]

[0015] Where a is the width of the convolution kernel in the x direction, F is the convolution kernel parameter vector, G is the local vector matrix operated with the convolution kernel function, w is the size of the convolution kernel, X represents the water level change input vector, and Cov(X) represents performing one-dimensional convolution encoding on the water level change input vector.

[0016] In the aforementioned anti-buoyancy and pressure relief drainage system for the basement floor slab of a building, the water level temporal feature extraction module further includes: a second fully connected encoding unit, used to perform fully connected encoding on the water level input vector using the fully connected layer of the temporal encoder according to the following formula to extract the high-dimensional latent features of the feature values ​​at each position in the water level input vector, wherein the formula is: Where X is the water level input vector, Y is the output vector, W is the weight matrix, and B is the bias vector. Represents matrix multiplication; the second one-dimensional convolutional coding unit is used to perform one-dimensional convolutional coding on the water level input vector using the one-dimensional convolutional layer of the temporal encoder according to the following formula to extract high-dimensional implicit correlation features between feature values ​​at each position in the water level input vector, wherein the formula is:

[0017]

[0018] Where a is the width of the convolution kernel in the x direction, F is the convolution kernel parameter vector, G is the local vector matrix operated with the convolution kernel function, w is the size of the convolution kernel, X represents the water level input vector, and Cov(X) represents performing one-dimensional convolution encoding on the water level input vector.

[0019] In the aforementioned anti-buoyancy and pressure-relief drainage system for the basement floor slab of a building, the feature fusion module includes: a joint Gaussian density map construction unit, used to calculate the joint Gaussian density map of the absolute water level time-series feature vector and the relative water level time-series feature vector, wherein the mean vector of the joint Gaussian density map is the positional mean vector between the absolute water level time-series feature vector and the relative water level time-series feature vector, and the covariance matrix of the joint Gaussian density map is the covariance matrix formed by the positional variances of the absolute water level time-series feature vector and the relative water level time-series feature vector; and a weighting factor calculation unit, used to calculate the Gaussian probability density distribution distance index between the absolute water level time-series feature vector and the relative water level time-series feature vector and the joint Gaussian density map, respectively, to obtain... The system includes a first Gaussian probability density distribution distance index and a second Gaussian probability density distribution distance index; a weighted optimization unit, used as weights to weight the absolute water level time series feature vector and the relative water level time series feature vector to obtain corrected absolute water level time series feature vector and corrected relative water level time series feature vector; a Gaussian density map fusion unit, used to construct a Gaussian density map of the corrected absolute water level time series feature vector and the corrected relative water level time series feature vector to obtain a fused Gaussian density map; and a Gaussian discretization unit, used to Gaussian discretize the Gaussian distribution at each position in the fused Gaussian density map to obtain the classification feature matrix.

[0020] In the aforementioned anti-buoyancy and pressure relief drainage system for the basement floor slab of a building, the joint Gaussian density map construction unit is used to: calculate the time-series feature vector of the absolute water level and the time-series feature vector of the relative water level using the Gaussian density map according to the following formula to obtain the joint Gaussian density map; wherein, the formula is:

[0021]

[0022] Where μ(f1,f2) represents the mean vector of the joint Gaussian density map, ∑(f1,f2) represents the covariance matrix of the joint Gaussian density map, f1 represents the time series feature vector of the absolute water level, and f2 represents the time series feature vector of the relative water level.

[0023] In the aforementioned anti-buoyancy and pressure relief drainage system for the basement floor slab of a building, the weighting factor calculation unit includes: a water level absolute change weighting factor calculation subunit, used to calculate the Gaussian probability density distribution distance index between the water level absolute quantity time-series feature vector and the joint Gaussian density map using the following formula to obtain the first Gaussian probability density distribution distance index; wherein, the formula is:

[0024]

[0025] Where V1 is the time-series feature vector of the absolute water level, μ u and ∑ u These are the mean vector and covariance matrix of the joint Gaussian density plot. The time-series characteristic vector of the absolute water level is in column vector form. exp(·) represents the matrix exponentiation operation, which represents the natural exponential function value raised to the power of the eigenvalues ​​at each position in the matrix. This indicates subtraction by position. denoted as matrix multiplication, w1 represents the distance exponent of the first Gaussian probability density distribution.

[0026] In the aforementioned anti-buoyancy and pressure relief drainage system for the basement floor slab of a building, the weighting factor calculation unit further includes: a water level relative change weighting factor calculation subunit, used to calculate the Gaussian probability density distribution distance index between the water level relative quantity time series feature vector and the joint Gaussian density map using the following formula to obtain the second Gaussian probability density distribution distance index; wherein, the formula is:

[0027]

[0028] Where V2 is the time-series feature vector of the relative water level, μ u and ∑ u These are the mean vector and covariance matrix of the joint Gaussian density plot. The time-series feature vector of the relative water level is in column vector form. exp(·) represents the matrix exponentiation operation, which represents the natural exponential function value raised to the power of the eigenvalues ​​at each position in the matrix. This indicates subtraction by position. w1 represents matrix multiplication, and w2 represents the distance exponent of the second Gaussian probability density distribution.

[0029] In the aforementioned anti-buoyancy and pressure relief drainage system for the basement floor of a building, the Gaussian discretization unit includes: a discrete subunit, used to discretize the Gaussian distribution at each position in the fused Gaussian density map to obtain multiple one-dimensional vectors; and a two-dimensional arrangement subunit, used to arrange the multiple one-dimensional vectors in a two-dimensional arrangement to obtain the classification feature matrix.

[0030] In the aforementioned anti-buoyancy and pressure relief drainage system for the basement floor of a building, the pump power recommendation module is used to: use the decoder to perform decoding regression on the water level decoding feature matrix using the following formula to obtain a decoding value representing the difference in imaging quality; wherein, the formula is: Where X represents the water level decoding feature matrix, Y is the decoded value, and W is the weight matrix. This represents matrix multiplication.

[0031] According to another aspect of this application, a method for preventing buoyancy and relieving pressure in the floor slab of a building basement is provided, comprising:

[0032] Obtain the water level values ​​of accumulated water at multiple predetermined time points within a predetermined time period collected by a water level gauge;

[0033] Arrange the water level values ​​of the accumulated water at the multiple predetermined time points into a water level input vector according to the time dimension;

[0034] Calculate the difference between the water level values ​​at every two adjacent positions in the water level input vector to obtain the water level change input vector;

[0035] The water level change input vector and the water level input vector are respectively passed through a temporal encoder containing a fully connected layer and a one-dimensional convolutional layer to obtain the absolute water level temporal feature vector and the relative water level temporal feature vector.

[0036] The absolute water level temporal feature vector and the relative water level temporal feature vector are fused based on the Gaussian density map to obtain the water level decoding feature matrix; and

[0037] The water level decoding feature matrix is ​​processed by a decoder to obtain a decoded value, which is used to represent the recommended power value of the water pump at the current time point.

[0038] According to another aspect of this application, an electronic device is provided, comprising: a processor; and a memory storing computer program instructions, which, when executed by the processor, cause the processor to perform the anti-buoyancy pressure relief drainage method for the basement floor slab as described above.

[0039] According to another aspect of this application, a computer-readable medium is provided having computer program instructions stored thereon, which, when executed by a processor, cause the processor to perform the anti-buoyancy pressure relief drainage method for the basement floor slab described above.

[0040] Compared with the prior art, the anti-buoyancy and pressure relief drainage system for the basement floor of a building provided in this application uses a deep learning-based neural network model to mine the time-dimensional variation characteristics of the water level value of the accumulated water, and adaptively adjusts the power value of the water pump based on the variation of the water level value, thereby ensuring safety while reducing energy consumption. Attached Figure Description

[0041] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0042] Figure 1 This is an application scenario diagram of the anti-buoyancy pressure relief drainage system for the basement floor slab of a building according to an embodiment of this application;

[0043] Figure 2 This is a block diagram of a building basement floor anti-buoyancy pressure relief drainage system according to an embodiment of this application;

[0044] Figure 3 This is a system architecture diagram of a building basement floor anti-buoyancy pressure relief drainage system according to an embodiment of this application;

[0045] Figure 4 This is a block diagram of a feature fusion module in a building basement floor anti-buoyancy pressure relief drainage system according to an embodiment of this application;

[0046] Figure 5 This is a block diagram of a Gaussian discretized unit in a building basement floor anti-buoyancy pressure relief drainage system according to an embodiment of this application;

[0047] Figure 6 This is a flowchart of a method for preventing buoyancy and relieving pressure in the basement floor slab of a building according to an embodiment of this application;

[0048] Figure 7 This is a block diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0049] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0050] Application Overview

[0051] As mentioned above, a pressure-relief and anti-buoyancy drainage system for a building basement disclosed in Chinese Patent No. CN210031867U has been found to have a safety hazard during use. When the water level rises faster than the pump speed, and the water level exceeds the containment tank, water enters the tank, submerging the first and second electric shock spheres. Since water is conductive, this poses a safety risk. Therefore, an optimized pressure-relief and anti-buoyancy drainage system for the basement floor is desired.

[0052] Accordingly, considering the actual process of anti-buoyancy and pressure relief drainage of a building's basement floor slab, the pumping speed of the water pump should be adapted to the rising speed of the accumulated water level to reduce energy consumption while ensuring safety. In other words, the power of the water pump should be adaptively adjusted based on changes in the accumulated water level to ensure the electrical safety of the containment tank. However, since both the power of the water pump and the water level exhibit dynamic changes over time, and there is a temporal correlation between them, it is difficult to accurately control the pump power in real time. Therefore, the challenge lies in accurately identifying the time-varying characteristics of the water level to adaptively adjust the pump power based on these changes, thereby ensuring safety while reducing energy consumption.

[0053] In recent years, deep learning and neural networks have been widely applied in fields such as computer vision, natural language processing, and text signal processing. Furthermore, deep learning and neural networks have demonstrated near-human or even superior performance in areas such as image classification, object detection, semantic segmentation, and text translation.

[0054] The development of deep learning and neural networks has provided new ideas and solutions for mining the characteristics of water level changes over time.

[0055] Specifically, in the technical solution of this application, firstly, the water level values ​​of accumulated water at multiple predetermined time points within a predetermined time period are obtained by a water level gauge. Next, considering the dynamic changing characteristics of the water level values ​​over time, in order to accurately extract this temporal change characteristic information, the water level values ​​at the multiple predetermined time points are arranged into a water level input vector according to the time dimension, thereby integrating the temporal distribution information of the water level values.

[0056] Furthermore, in order to extract the temporal dynamic change features of the accumulated water level, feature mining is required on the water level input vector. However, considering that the temporal change features of the accumulated water level are small-scale features relative to the actual water level value, they are difficult to capture and extract. Therefore, extracting the water level change features using absolute change information would not only result in excessive computation and overfitting, but also make it difficult to detect the subtle, small-scale features of the water level change, thus affecting the subsequent decoding accuracy. Therefore, in the technical solution of this application, it is desirable to use both the relative and absolute dynamic features of the accumulated water level to comprehensively mine the water level change features, thereby controlling the pump power at the current time point. Specifically, firstly, the difference between the water level values ​​at every two adjacent positions in the water level input vector is calculated to obtain the water level change input vector.

[0057] Next, considering that the absolute and relative changes in the water level have hidden dynamic change characteristics in the time dimension, in order to fully and accurately extract the dynamic change feature information of the water level value in the time series, in the technical solution of this application, the water level change input vector and the water level input vector are encoded by a temporal encoder containing fully connected layers and one-dimensional convolutional layers, respectively, to extract the temporal dynamic change features of the absolute water level and the temporal dynamic correlation features of the relative water level, thereby obtaining the temporal feature vector of the absolute water level and the temporal feature vector of the relative water level. Specifically, here, the temporal encoder consists of alternating fully connected layers and one-dimensional convolutional layers. It uses one-dimensional convolutional encoding through one-dimensional convolutional layers to extract the correlation features of the absolute and relative changes in the water level in the time series, and uses fully connected layers to perform fully connected encoding to extract the high-dimensional hidden features of the absolute and relative changes in the water level.

[0058] Then, considering that the absolute water level time-series feature vector and the relative water level time-series feature vector each correspond to a feature distribution manifold in the high-dimensional feature space, and that these feature distribution manifolds, due to their irregular shapes and scattered locations, would be equivalent to simply superimposing these feature distribution manifolds according to their original positions and shapes if the high-dimensional implicit dynamic feature information of the accumulated water level value in the time dimension is represented by concatenating the absolute water level time-series feature vector and the relative water level time-series feature vector. This would result in a very irregular and complex boundary for the newly obtained feature distribution manifold, making it very easy to get trapped in local extrema when searching for the optimal point using gradient descent, thus failing to obtain the global optimal point. Therefore, it is necessary to further fuse the absolute water level time-series feature vector and the relative water level time-series feature vector appropriately, so that each feature distribution can converge relative to each other on the shape surface.

[0059] It should be understood that, considering the widespread use of Gaussian density maps in deep learning for prior-based posterior estimation of targets, they can be used to correct data distribution, thereby achieving the aforementioned objective. Specifically, in the technical solution of this application, firstly, Gaussian density maps of the absolute water level time-series feature vector and the relative water level time-series feature vector are constructed to obtain a fused Gaussian density map; then, the fused Gaussian density map is subjected to Gaussian discretization to avoid information loss during data feature augmentation, thereby improving the accuracy of subsequent decoding and regression to obtain the water level decoding feature matrix.

[0060] Furthermore, the water level decoding feature matrix is ​​processed through a decoder to obtain a decoded value representing the recommended power value of the water pump at the current time point. In other words, the decoding process utilizes the correlation distribution information between the dynamic characteristics of the relative change and the dynamic characteristics of the absolute change in the water level over time to regress the real-time decoded power value of the water pump. This allows for adaptive adjustment of the water pump power value based on changes in the water level, thereby ensuring safety while reducing energy consumption.

[0061] Specifically, in the technical solution of this application, when fusing the absolute water level time-series feature vector and the relative water level time-series feature vector based on the Gaussian density map to obtain the water level decoding feature matrix, since the absolute water level time-series feature vector and the relative water level time-series feature vector respectively express the time-series distribution of the absolute water level value and the water level change value, they differ in both time sequence arrangement and numerical order. Thus, if the fusion is directly performed through the Gaussian density map, the resulting water level decoding feature matrix has poor consistency and correlation in the fused feature dimension of the absolute water level time-series feature vector and the relative water level time-series feature vector, which serves as the target classification dimension, thereby affecting the accuracy of the decoding result of the water level decoding feature matrix.

[0062] Therefore, it is desired to converge the difference between the time-series feature vectors of absolute water level and relative water level at the Gaussian probability density level. Specifically, when calculating the joint Gaussian density map of the time-series feature vectors of absolute water level and relative water level, the Gaussian probability density distribution distance index between the time-series feature vectors of absolute water level and relative water level and the joint Gaussian density map is further calculated, expressed as:

[0063]

[0064]

[0065] Where V1 and V2 are the time-series feature vectors of absolute water level and relative water level, respectively, μ u and ∑ u It is the mean vector and covariance matrix of the joint Gaussian density plot, i.e., μ u Let represent the mean vector of the absolute water level time series feature vector and the relative water level time series feature vector, and ∑ u The covariance matrix represents the time-series feature vectors of absolute water level and relative water level, formed by their positional variances, where all vectors are column vectors.

[0066] Here, by calculating the Gaussian probability density distribution distance index between the absolute water level time-series feature vector and the relative water level time-series feature vector and their corresponding joint Gaussian density map, the feature distribution distance of the target feature vector relative to the joint Gaussian probability density distribution can be represented. By weighting the absolute water level time-series feature vector and the relative water level time-series feature vector with this index and then obtaining the water level decoding feature matrix based on the Gaussian density map, the compatibility of the water level decoding feature matrix with the correlation transfer to the joint probability density distribution of the Gaussian probability density in the target domain can be improved. This enhances the consistency and correlation of its Gaussian probability density distribution on the fused feature dimension of the absolute water level time-series feature vector and the relative water level time-series feature vector, which serves as the target classification dimension, thereby improving the accuracy of the decoding result of the water level decoding feature matrix. In this way, the power value of the water pump can be adaptively adjusted in real time and accurately based on changes in water level values, thus ensuring safety while reducing energy consumption.

[0067] Based on this, this application proposes a buoyancy-resistant and pressure-relief drainage system for the basement floor of a building, comprising: a data acquisition module for acquiring water level values ​​of accumulated water at multiple predetermined time points within a predetermined time period collected by a water level gauge; an absolute water level data arrangement module for arranging the water level values ​​of the accumulated water at the multiple predetermined time points into a water level input vector according to the time dimension; a relative water level data distribution module for calculating the difference between the water level values ​​at each of the two adjacent positions in the water level input vector to obtain a water level change input vector; a water level temporal feature extraction module for passing the water level change input vector and the water level input vector through a temporal encoder containing a fully connected layer and a one-dimensional convolutional layer to obtain a water level absolute quantity temporal feature vector and a water level relative quantity temporal feature vector; a feature fusion module for fusing the water level absolute quantity temporal feature vector and the water level relative quantity temporal feature vector based on a Gaussian density map to obtain a water level decoding feature matrix; and a water pump power recommendation module for passing the water level decoding feature matrix through a decoder to obtain a decoded value, wherein the decoded value represents the recommended water pump power value at the current time point.

[0068] Figure 1 This is an application scenario diagram of the anti-buoyancy and pressure relief drainage system for the basement floor slab of a building according to an embodiment of this application. (See diagram below.) Figure 1 As shown, in this application scenario, firstly, through a water level gauge (e.g., such as...) Figure 1 As shown in G), the water level values ​​of accumulated water at multiple predetermined time points within a predetermined time period are obtained. Then, the obtained data is input into a server deployed with a building basement floor anti-buoyancy pressure relief drainage algorithm (e.g., such as...). Figure 1The cloud server S shown in the figure is capable of processing the input data using a building basement floor anti-buoyancy pressure relief drainage algorithm to generate a decoded value, which represents the recommended power value of the water pump at the current time point.

[0069] After introducing the basic principles of this application, various non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0070] Exemplary System

[0071] Figure 2 This is a block diagram of a building basement floor anti-buoyancy pressure relief drainage system according to an embodiment of this application. Figure 2 As shown, the anti-buoyancy and pressure relief drainage system 300 for the basement floor slab of a building according to an embodiment of this application includes: a data acquisition module 310; an absolute water level data arrangement module 320; a relative water level data distribution module 330; a water level time series feature extraction module 340; a feature fusion module 350; and a water pump power recommendation module 360.

[0072] The system includes the following modules: a data acquisition module 310, which acquires water level values ​​at multiple predetermined time points within a predetermined time period collected by a water level gauge; an absolute water level data arrangement module 320, which arranges the water level values ​​at the multiple predetermined time points into a water level input vector according to the time dimension; a relative water level data distribution module 330, which calculates the difference between water level values ​​at each adjacent two positions in the water level input vector to obtain a water level change input vector; a water level temporal feature extraction module 340, which passes the water level change input vector and the water level input vector through a temporal encoder containing a fully connected layer and a one-dimensional convolutional layer to obtain an absolute water level temporal feature vector and a relative water level temporal feature vector; a feature fusion module 350, which fuses the absolute water level temporal feature vector and the relative water level temporal feature vector based on a Gaussian density map to obtain a water level decoding feature matrix; and a pump power recommendation module 360, which passes the water level decoding feature matrix through a decoder to obtain a decoded value, the decoded value representing the recommended pump power value at the current time point.

[0073] Figure 3 This is a system architecture diagram of a building basement floor anti-buoyancy pressure relief drainage system according to an embodiment of this application. Figure 3As shown, in this network architecture, the data acquisition module 310 first acquires the water level values ​​of accumulated water at multiple predetermined time points within a predetermined time period collected by the water level gauge; then, the absolute water level data arrangement module 320 arranges the water level values ​​of accumulated water at multiple predetermined time points acquired by the data acquisition module 310 according to the time dimension into a water level input vector; the relative water level data distribution module 330 calculates the difference between the water level values ​​of every two adjacent positions in the water level input vector obtained by the absolute water level data arrangement module 320 to obtain the water level change input vector; then, the water level time series feature extraction module 340 extracts the water level values ​​calculated by the relative water level data distribution module 330. The changing input vector and the water level input vector obtained by the absolute water level data arrangement module 320 are respectively passed through a temporal encoder containing a fully connected layer and a one-dimensional convolutional layer to obtain a temporal feature vector of absolute water level and a temporal feature vector of relative water level. The feature fusion module 350 fuses the temporal feature vector of absolute water level and the temporal feature vector of relative water level obtained by the water level temporal feature extraction module 340 based on a Gaussian density map to obtain a water level decoding feature matrix. Then, the water pump power recommendation module 360 ​​passes the water level decoding feature matrix obtained by the feature fusion module 350 through a decoder to obtain a decoded value, which is used to represent the recommended power value of the water pump at the current time point.

[0074] Specifically, during the operation of the anti-buoyancy and pressure relief drainage system 300 for the basement floor slab, the data acquisition module 310 is used to acquire the water level values ​​of the accumulated water at multiple predetermined time points within a predetermined time period, collected by a water level gauge. It should be understood that when actually performing anti-buoyancy and pressure relief drainage for the basement floor slab, the pumping speed of the water pump should be adapted to the rising speed of the accumulated water level to reduce energy consumption while ensuring safety. Therefore, in the technical solution of this application, the power value of the water pump should be adaptively adjusted based on the changes in the accumulated water level. Thus, in a specific example of this application, firstly, the water level values ​​of the accumulated water at multiple predetermined time points within a predetermined time period can be acquired using a water level gauge.

[0075] Specifically, during the operation of the anti-buoyancy and pressure relief drainage system 300 in the basement floor of the building, the absolute water level data arrangement module 320 is used to arrange the water level values ​​of the accumulated water at multiple predetermined time points into a water level input vector according to the time dimension. Considering that the water level values ​​of the accumulated water have dynamic change characteristics in the time dimension, in order to accurately extract this temporal change characteristic information, the water level values ​​of the accumulated water at multiple predetermined time points are arranged into a water level input vector according to the time dimension, thereby integrating the temporal distribution information of the water level values.

[0076] Specifically, during the operation of the anti-buoyancy and pressure relief drainage system 300 in the basement floor of the building, the relative water level data distribution module 330 is used to calculate the difference between the water level values ​​at each of the two adjacent positions in the water level input vector to obtain the water level change input vector. To extract the temporal dynamic change characteristics of the accumulated water level, feature mining is required on the water level input vector. However, considering that the temporal change characteristics of the accumulated water level are small-scale features relative to the actual water level value, they are difficult to capture and extract. Therefore, using absolute change information to extract the water level change features would not only result in a large computational load and overfitting, but also make it difficult to detect the subtle small-scale features of the water level change, thus affecting the subsequent decoding accuracy. Therefore, in the technical solution of this application, it is desirable to use both the relative dynamic features and the absolute dynamic features of the accumulated water level to comprehensively mine the water level change features, thereby controlling the pump power at the current time point. Specifically, firstly, the difference between the water level values ​​at each of the two adjacent positions in the water level input vector is calculated to obtain the water level change input vector.

[0077] Specifically, during the operation of the anti-buoyancy and pressure relief drainage system 300 in the basement floor of the building, the water level time-series feature extraction module 340 is used to pass the water level change input vector and the water level input vector through a time-series encoder containing a fully connected layer and a one-dimensional convolutional layer to obtain a water level absolute quantity time-series feature vector and a water level relative quantity time-series feature vector, respectively. Considering that the absolute quantity and relative change of the accumulated water level have hidden dynamic change characteristics in the time dimension, in order to fully and accurately extract the dynamic change feature information of the water level value in the time series, in the technical solution of this application, the water level change input vector and the water level input vector are encoded through a time-series encoder containing a fully connected layer and a one-dimensional convolutional layer, respectively, to extract the time-series dynamic change features of the absolute quantity of the accumulated water level and the time-series dynamic correlation features of the relative quantity of the relative quantity of the accumulated water level, thereby obtaining the water level absolute quantity time-series feature vector and the water level relative quantity time-series feature vector. Specifically, the temporal encoder here consists of alternating fully connected layers and one-dimensional convolutional layers. It uses one-dimensional convolutional encoding through the one-dimensional convolutional layers to extract the temporal correlation features of the absolute and relative changes in water level, and uses fully connected layers to perform fully connected encoding to extract the high-dimensional latent features of the absolute and relative changes in water level. In a specific example of this application, firstly, the fully connected layer of the temporal encoder performs fully connected encoding on the water level change input vector using the following formula to extract the high-dimensional latent features of the feature values ​​at each position in the water level change input vector, wherein the formula is: Where X is the input vector of water level change, Y is the output vector, W is the weight matrix, and B is the bias vector. This represents matrix multiplication; then, the one-dimensional convolutional layer of the temporal encoder is used to perform one-dimensional convolutional encoding on the water level change input vector using the following formula to extract high-dimensional implicit correlation features between feature values ​​at each position in the water level change input vector, wherein the formula is:

[0078]

[0079] Where a is the width of the convolution kernel in the x-direction, F is the convolution kernel parameter vector, G is the local vector matrix operated with the convolution kernel function, w is the size of the convolution kernel, X represents the water level change input vector, and Cov(X) represents one-dimensional convolutional encoding of the water level change input vector. Furthermore, the fully connected layer of the temporal encoder is used to fully encode the water level input vector using the following formula to extract high-dimensional latent features of the feature values ​​at each position in the water level input vector, wherein the formula is: Where X is the water level input vector, Y is the output vector, W is the weight matrix, and B is the bias vector. This represents matrix multiplication; the one-dimensional convolutional layer of the temporal encoder is used to perform one-dimensional convolutional encoding on the water level input vector using the following formula to extract high-dimensional implicit correlation features between feature values ​​at each position in the water level input vector, wherein the formula is:

[0080]

[0081] Where a is the width of the convolution kernel in the x direction, F is the convolution kernel parameter vector, G is the local vector matrix operated with the convolution kernel function, w is the size of the convolution kernel, X represents the water level input vector, and Cov(X) represents performing one-dimensional convolution encoding on the water level input vector.

[0082] Specifically, during the operation of the anti-buoyancy and pressure relief drainage system 300 in the basement floor of the building, the feature fusion module 350 is used to fuse the time-series feature vectors of absolute water level and relative water level based on the Gaussian density map to obtain a water level decoding feature matrix. Considering that the time-series feature vectors of absolute water level and relative water level each correspond to a feature distribution manifold in the high-dimensional feature space, and that these feature distribution manifolds, due to their irregular shapes and scattered locations, represent the high-dimensional implicit dynamic feature information of the water level value in the time dimension simply by concatenating the time-series feature vectors of absolute water level and relative water level, it would be equivalent to simply superimposing these feature distribution manifolds according to their original positions and shapes. This would make the boundaries of the newly obtained feature distribution manifolds very irregular and complex, and when searching for the optimal point using gradient descent, it would be very easy to get stuck in local extrema and fail to obtain the global optimal point. Therefore, it is necessary to further fuse the absolute water level time-series feature vector and the relative water level time-series feature vector appropriately, so that the distributions of each feature can converge relative to each other on the shape surface. It should be understood that, considering that Gaussian density maps are widely used in deep learning for prior-based posterior estimation of targets, they can be used to correct the data distribution, thereby achieving the above objective. Specifically, in the technical solution of this application, firstly, Gaussian density maps of the absolute water level time-series feature vector and the relative water level time-series feature vector are constructed to obtain a fused Gaussian density map; then, the fused Gaussian density map is subjected to Gaussian discretization to avoid information loss during data feature augmentation, thereby improving the accuracy of subsequent decoding and regression to obtain the water level decoding feature matrix. Specifically, in the technical solution of this application, when fusing the absolute water level time-series feature vector and the relative water level time-series feature vector based on the Gaussian density map to obtain the water level decoding feature matrix, since the absolute water level time-series feature vector and the relative water level time-series feature vector respectively express the time-series distribution of the absolute water level value and the water level change value, they differ in both time sequence arrangement and numerical order. Thus, if the fusion is directly performed through the Gaussian density map, the resulting water level decoding feature matrix has poor consistency and correlation in the fused feature dimension of the absolute water level time-series feature vector and the relative water level time-series feature vector, which serves as the target classification dimension, thereby affecting the accuracy of the decoding result of the water level decoding feature matrix. Therefore, it is desired to converge the difference between the time-series feature vectors of absolute water level and relative water level at the Gaussian probability density level. Specifically, when calculating the joint Gaussian density map of the time-series feature vectors of absolute water level and relative water level, the Gaussian probability density distribution distance index between the time-series feature vectors of absolute water level and relative water level and the joint Gaussian density map is further calculated, expressed as:

[0083]

[0084]

[0085] Where V1 and V2 are the time-series feature vectors of absolute water level and relative water level, respectively, μ u and ∑ u It is the mean vector and covariance matrix of the joint Gaussian density plot, i.e., μ u Let represent the mean vector of the absolute water level time series feature vector and the relative water level time series feature vector, and ∑ u The covariance matrix represents the positional variance of the absolute water level time-series feature vector and the relative water level time-series feature vector, where all vectors are column vectors. Here, by calculating the Gaussian probability density distribution distance index between the absolute water level time-series feature vector and the relative water level time-series feature vector and their corresponding joint Gaussian density map, the feature distribution distance of the target feature vector relative to the joint Gaussian probability density distribution can be represented. By weighting the absolute water level time-series feature vector and the relative water level time-series feature vector respectively with this index and then obtaining the water level decoding feature matrix based on the Gaussian density map, the compatibility of the water level decoding feature matrix with the correlation transfer to the joint probability density distribution of the Gaussian probability density in the target domain can be improved. This enhances the consistency and correlation of its Gaussian probability density distribution on the fused feature dimension of the absolute water level time-series feature vector and the relative water level time-series feature vector, which serves as the target classification dimension, thereby improving the accuracy of the decoding result of the water level decoding feature matrix. In this way, the power value of the water pump can be adaptively adjusted in real time and accurately based on changes in water level values, thereby ensuring safety while reducing energy consumption.

[0086] Figure 4 This is a block diagram of a feature fusion module in a building basement floor anti-buoyancy and pressure relief drainage system according to an embodiment of this application. Figure 4As shown, the feature fusion module 350 includes: a joint Gaussian density map construction unit 351, used to calculate the joint Gaussian density map of the absolute water level time series feature vector and the relative water level time series feature vector, wherein the mean vector of the joint Gaussian density map is the positional mean vector between the absolute water level time series feature vector and the relative water level time series feature vector, and the covariance matrix of the joint Gaussian density map is the covariance matrix formed by the positional variances of the absolute water level time series feature vector and the relative water level time series feature vector; and a weighting factor calculation unit 352, used to calculate the Gaussian probability density distribution distance index between the absolute water level time series feature vector and the relative water level time series feature vector and the joint Gaussian density map respectively to obtain a first Gaussian probability density. The system comprises: a distribution distance index and a second Gaussian probability density distribution distance index; a weighted optimization unit 353, used to weight the absolute water level time series feature vector and the relative water level time series feature vector using the first Gaussian probability density distribution distance index and the second Gaussian probability density distribution distance index as weights respectively, to obtain corrected absolute water level time series feature vector and corrected relative water level time series feature vector; a Gaussian density map fusion unit 354, used to construct a Gaussian density map of the corrected absolute water level time series feature vector and the corrected relative water level time series feature vector to obtain a fused Gaussian density map; and a Gaussian discretization unit 355, used to perform Gaussian discretization on the Gaussian distribution at each position in the fused Gaussian density map to obtain the classification feature matrix. The joint Gaussian density map construction unit 351 includes: using the Gaussian density map to calculate the absolute water level time series feature vector and the relative water level time series feature vector using the following formula to obtain the joint Gaussian density map; wherein the formula is:

[0087]

[0088] Where μ(f1,f2) represents the mean vector of the joint Gaussian density map, ∑(f1,f2) represents the covariance matrix of the joint Gaussian density map, f1 represents the time series feature vector of the absolute water level, and f2 represents the time series feature vector of the relative water level.

[0089] Figure 5 This is a block diagram of a Gaussian discretized unit in a building basement floor anti-buoyancy and pressure relief drainage system according to an embodiment of this application. Figure 5 As shown, the Gaussian discretization unit 355 includes: a discretization subunit 3551, used to discretize the Gaussian distribution at each position in the fused Gaussian density map to obtain multiple one-dimensional vectors; and a two-dimensional arrangement subunit 3552, used to arrange the multiple one-dimensional vectors in two dimensions to obtain the classification feature matrix.

[0090] Specifically, during the operation of the anti-buoyancy and pressure relief drainage system 300 in the basement floor of the building, the pump power recommendation module 360 ​​is used to pass the water level decoding feature matrix through a decoder to obtain a decoded value, which represents the recommended pump power value at the current time point. That is, the water level decoding feature matrix is ​​used as the decoding feature matrix and decoded and regressed by the decoder to obtain a decoded value representing the recommended pump power value at the current time point. In a specific example of this application, the decoder is used to perform decoding and regression on the water level decoding feature matrix using the following formula to obtain a decoded value representing the image quality difference; wherein, the formula is: Where X represents the water level decoding feature matrix, Y is the decoded value, and W is the weight matrix. This represents matrix multiplication. Specifically, it involves decoding the dynamic characteristics of the relative and absolute changes in the water level over time to regress the real-time power value of the water pump. This allows for adaptive adjustment of the pump's power based on changes in the water level, ensuring safety while reducing energy consumption.

[0091] In summary, the anti-buoyancy and pressure relief drainage system 300 for the basement floor slab of a building according to the embodiments of this application is explained. It uses a deep learning-based neural network model to mine the change characteristics of the water level value in the time dimension, and adaptively adjusts the power value of the water pump based on the change of the water level value, thereby ensuring safety while reducing energy consumption.

[0092] As described above, the anti-buoyancy and pressure relief drainage system for basement floors according to embodiments of this application can be implemented in various terminal devices. In one example, the anti-buoyancy and pressure relief drainage system 300 for basement floors according to embodiments of this application can be integrated into a terminal device as a software module and / or a hardware module. For example, the anti-buoyancy and pressure relief drainage system 300 can be a software module in the operating system of the terminal device, or it can be an application developed for the terminal device; of course, the anti-buoyancy and pressure relief drainage system 300 can also be one of many hardware modules of the terminal device.

[0093] Alternatively, in another example, the building basement floor anti-buoyancy pressure relief drainage system 300 and the terminal device can also be separate devices, and the building basement floor anti-buoyancy pressure relief drainage system 300 can be connected to the terminal device via wired and / or wireless network, and transmit interactive information in accordance with the agreed data format.

[0094] Exemplary methods

[0095] Figure 6This is a flowchart of a method for preventing buoyancy and relieving pressure in the basement floor slab of a building according to an embodiment of this application. Figure 6 As shown, the anti-buoyancy and pressure relief drainage method for the basement floor slab of a building according to an embodiment of this application includes the following steps: S110, acquiring water level values ​​of accumulated water at multiple predetermined time points within a predetermined time period collected by a water level gauge; S120, arranging the water level values ​​of the accumulated water at the multiple predetermined time points according to the time dimension into a water level input vector; S130, calculating the difference between the water level values ​​at each of the two adjacent positions in the water level input vector to obtain a water level change input vector; S140, passing the water level change input vector and the water level input vector through a temporal encoder containing a fully connected layer and a one-dimensional convolutional layer respectively to obtain a water level absolute quantity temporal feature vector and a water level relative quantity temporal feature vector; S150, fusing the water level absolute quantity temporal feature vector and the water level relative quantity temporal feature vector based on a Gaussian density map to obtain a water level decoding feature matrix; and S160, passing the water level decoding feature matrix through a decoder to obtain a decoded value, wherein the decoded value is used to represent the recommended power value of the water pump at the current time point.

[0096] In one example, in the above-mentioned anti-buoyancy and pressure relief drainage method for the basement floor of a building, step S140 includes: using the fully connected layer of the time encoder to perform fully connected encoding on the water level change input vector according to the following formula to extract high-dimensional latent features of the feature values ​​at each position in the water level change input vector, wherein the formula is: Where X is the input vector of water level change, Y is the output vector, W is the weight matrix, and B is the bias vector. This represents matrix multiplication; the one-dimensional convolutional layer of the temporal encoder is used to perform one-dimensional convolutional encoding on the water level change input vector using the following formula to extract high-dimensional implicit correlation features between feature values ​​at each position in the water level change input vector, wherein the formula is:

[0097]

[0098] Where a is the width of the convolution kernel in the x-direction, F is the convolution kernel parameter vector, G is the local vector matrix operated with the convolution kernel function, w is the size of the convolution kernel, X represents the water level change input vector, and Cov(X) represents one-dimensional convolutional encoding of the water level change input vector. Furthermore, the fully connected layer of the temporal encoder is used to fully encode the water level input vector using the following formula to extract high-dimensional latent features of the feature values ​​at each position in the water level input vector, wherein the formula is: Where X is the water level input vector, Y is the output vector, W is the weight matrix, and B is the bias vector. This represents matrix multiplication; the one-dimensional convolutional layer of the temporal encoder is used to perform one-dimensional convolutional encoding on the water level input vector using the following formula to extract high-dimensional implicit correlation features between feature values ​​at each position in the water level input vector, wherein the formula is:

[0099]

[0100] Where a is the width of the convolution kernel in the x direction, F is the convolution kernel parameter vector, G is the local vector matrix operated with the convolution kernel function, w is the size of the convolution kernel, X represents the water level input vector, and Cov(X) represents performing one-dimensional convolution encoding on the water level input vector.

[0101] In one example, in the above-mentioned anti-buoyancy and pressure relief drainage method for the basement floor of a building, step S150 includes: calculating the joint Gaussian density map of the absolute water level time-series feature vector and the relative water level time-series feature vector, wherein the mean vector of the joint Gaussian density map is the positional mean vector between the absolute water level time-series feature vector and the relative water level time-series feature vector, and the covariance matrix of the joint Gaussian density map is the covariance matrix formed by the positional variances of the absolute water level time-series feature vector and the relative water level time-series feature vector; and calculating the Gaussian probability density values ​​of the absolute water level time-series feature vector and the relative water level time-series feature vector and the joint Gaussian density map, respectively. The distance index is used to obtain the first Gaussian probability density distribution distance index and the second Gaussian probability density distribution distance index; the first Gaussian probability density distribution distance index and the second Gaussian probability density distribution distance index are used as weights to weight the time series feature vector of the absolute water level and the time series feature vector of the relative water level to obtain the corrected time series feature vector of the absolute water level and the corrected time series feature vector of the relative water level; a Gaussian density map of the corrected time series feature vector of the absolute water level and the corrected time series feature vector of the relative water level is constructed to obtain a fused Gaussian density map; the Gaussian distribution at each position in the fused Gaussian density map is discretized using Gaussian to obtain the classification feature matrix. The calculation of the joint Gaussian density map of the time series feature vector of the absolute water level and the time series feature vector of the relative water level includes: using the Gaussian density map to calculate the time series feature vector of the absolute water level and the time series feature vector of the relative water level using the following formula to obtain the joint Gaussian density map; wherein the formula is:

[0102]

[0103] Where μ(f1,f2) represents the mean vector of the joint Gaussian density map, ∑(f1,f2) represents the covariance matrix of the joint Gaussian density map, f1 represents the time-series feature vector of the absolute water level, and f2 represents the time-series feature vector of the relative water level. Furthermore, calculating the Gaussian probability density distribution distance index between the time-series feature vector of the absolute water level and the time-series feature vector of the relative water level and the joint Gaussian density map to obtain the first Gaussian probability density distribution distance index and the second Gaussian probability density distribution distance index includes: calculating the time-series feature vector of the absolute water level and the time-series feature vector of the relative water level to obtain the joint Gaussian density map using the following formula; wherein the formula is:

[0104]

[0105] Where μ(f1,f2) represents the mean vector of the joint Gaussian density map, ∑(f1,f2) represents the covariance matrix of the joint Gaussian density map, f1 represents the time-series feature vector of the absolute water level, and f2 represents the time-series feature vector of the relative water level; and the Gaussian probability density distribution distance index between the time-series feature vector of the absolute water level and the joint Gaussian density map is calculated using the following formula to obtain the first Gaussian probability density distribution distance index; wherein the formula is:

[0106]

[0107] Where V1 is the time-series feature vector of the absolute water level, μ u and ∑ u These are the mean vector and covariance matrix of the joint Gaussian density plot. The time-series characteristic vector of the absolute water level is in column vector form. exp(·) represents the matrix exponentiation operation, which represents the natural exponential function value raised to the power of the eigenvalues ​​at each position in the matrix. This indicates subtraction by position. The matrix multiplication is represented by w1, which represents the distance index of the first Gaussian probability density distribution. More specifically, the Gaussian distribution at each position in the fused Gaussian density map is discretized to obtain the classification feature matrix, including: discretizing the Gaussian distribution at each position in the fused Gaussian density map to obtain multiple one-dimensional vectors; and arranging the multiple one-dimensional vectors in a two-dimensional manner to obtain the classification feature matrix.

[0108] In one example, in the above-mentioned anti-buoyancy pressure relief drainage method for the basement floor of a building, step S160 includes: using the decoder to perform decoding regression on the water level decoding feature matrix using the following formula to obtain a decoding value representing the difference in imaging quality; wherein, the formula is: Where X represents the water level decoding feature matrix, Y is the decoded value, and W is the weight matrix. This represents matrix multiplication.

[0109] In summary, the anti-buoyancy and pressure relief drainage method for the basement floor slab of a building according to the embodiments of this application is explained. It uses a deep learning-based neural network model to mine the change characteristics of the water level value in the time dimension, and adaptively adjusts the power value of the water pump based on the change of the water level value, thereby ensuring safety while reducing energy consumption.

[0110] Exemplary electronic devices

[0111] Below, for reference Figure 7 This describes an electronic device according to embodiments of the present application.

[0112] Figure 7 A block diagram of an electronic device according to an embodiment of this application is illustrated.

[0113] like Figure 7 As shown, the electronic device 10 includes one or more processors 11 and memory 12.

[0114] The processor 11 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 10 to perform desired functions.

[0115] The memory 12 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 11 may execute the program instructions to implement the functions and / or other desired functions of the building basement floor anti-buoyancy pressure relief drainage system described in the various embodiments of this application above. Various contents, such as water level decoding feature matrices, may also be stored in the computer-readable storage medium.

[0116] In one example, the electronic device 10 may also include an input device 13 and an output device 14, which are interconnected via a bus system and / or other forms of connection mechanism (not shown).

[0117] The input device 13 may include, for example, a keyboard, a mouse, etc.

[0118] The output device 14 can output various information to the outside, including decoded values. The output device 14 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0119] Of course, for the sake of simplicity, Figure 7 Only some of the components of the electronic device 10 relevant to this application are shown in this illustration; components such as buses, input / output interfaces, etc., are omitted. In addition, the electronic device 10 may include any other suitable components depending on the specific application.

[0120] Exemplary computer program products and computer-readable storage media

[0121] In addition to the methods and devices described above, embodiments of this application may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps of the functions in the building basement floor anti-buoyancy pressure relief drainage method according to various embodiments of this application as described in the "Exemplary Systems" section above.

[0122] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0123] Furthermore, embodiments of this application may also be computer-readable storage media storing computer program instructions that, when executed by a processor, cause the processor to perform the steps of the functionalities in the building basement floor anti-buoyancy pressure relief drainage method according to various embodiments of this application as described in the "Exemplary System" section above.

[0124] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may, for example, include, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0125] The basic principles of this application have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the application to the necessity of employing the aforementioned specific details for implementation.

[0126] The block diagrams of devices, apparatuses, devices, and systems involved in this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0127] It should also be noted that in the apparatus, equipment, and methods of this application, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered as equivalent solutions of this application.

[0128] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of this application. Therefore, this application is not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0129] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A buoyancy-resistant and pressure-relief drainage system for the basement floor slab of a building, characterized in that, include: The data acquisition module is used to obtain the water level values ​​of the accumulated water at multiple predetermined time points within a predetermined time period, collected by the water level gauge. The absolute water level data arrangement module is used to arrange the water level values ​​of the accumulated water at the multiple predetermined time points into a water level input vector according to the time dimension. The water level relative data distribution module is used to calculate the difference between the water level values ​​at every two adjacent positions in the water level input vector to obtain the water level change input vector; The water level time series feature extraction module is used to pass the water level change input vector and the water level input vector through a time series encoder containing a fully connected layer and a one-dimensional convolutional layer to obtain the absolute water level time series feature vector and the relative water level time series feature vector, respectively. The feature fusion module is used to fuse the time-series feature vectors of absolute water level and relative water level based on a Gaussian density map to obtain a water level decoding feature matrix. It includes: a joint Gaussian density map construction unit, used to calculate a joint Gaussian density map of the time-series feature vectors of absolute water level and relative water level, wherein the mean vector of the joint Gaussian density map is the positional mean vector between the time-series feature vectors of absolute water level and relative water level, and the covariance matrix of the joint Gaussian density map is the covariance matrix formed by the positional variances of the time-series feature vectors of absolute water level and relative water level; and a weighting factor calculation unit, used to calculate the weighting factor between the time-series feature vectors of absolute water level and relative water level and the joint Gaussian density map. A Gaussian probability density distribution distance index is used to obtain a first Gaussian probability density distribution distance index and a second Gaussian probability density distribution distance index; a weighted optimization unit is used to use the first Gaussian probability density distribution distance index and the second Gaussian probability density distribution distance index as weights to weight the time series feature vector of the absolute water level and the time series feature vector of the relative water level to obtain the corrected time series feature vector of the absolute water level and the corrected time series feature vector of the relative water level; a Gaussian density map fusion unit is used to construct a Gaussian density map of the corrected time series feature vector of the absolute water level and the corrected time series feature vector of the relative water level to obtain a fused Gaussian density map; a Gaussian discretization unit is used to discretize the Gaussian distribution at each position in the fused Gaussian density map to obtain the water level decoding feature matrix. The water pump power recommendation module is used to obtain a decoded value by passing the water level decoding feature matrix through a decoder. The decoded value is used to represent the recommended power value of the water pump at the current time point.

2. The anti-buoyancy and pressure relief drainage system for the basement floor slab of a building according to claim 1, characterized in that, The water level time-series feature extraction module includes: The first fully connected encoding unit is used to perform fully connected encoding on the water level change input vector using the fully connected layer of the temporal encoder according to the following formula to extract high-dimensional latent features of the feature values ​​at each position in the water level change input vector, wherein the formula is: ,in It is the water level change input vector. It is the output vector. It is a weight matrix. It is a bias vector. Represents matrix multiplication; The first one-dimensional convolutional coding unit is used to perform one-dimensional convolutional coding on the water level change input vector using the one-dimensional convolutional layer of the temporal encoder according to the following formula to extract high-dimensional implicit correlation features between feature values ​​at each position in the water level change input vector, wherein the formula is: Where a is the width of the convolution kernel in the x-direction, F is the convolution kernel parameter vector, G is the local vector matrix operated on by the convolution kernel function, and w is the size of the convolution kernel. This represents the water level change input vector. This represents performing one-dimensional convolutional encoding on the input vector of the water level change.

3. The anti-buoyancy and pressure relief drainage system for the basement floor slab of a building according to claim 2, characterized in that, The water level time series feature extraction module also includes: The second fully connected encoding unit is used to perform fully connected encoding on the water level input vector using the fully connected layer of the temporal encoder to extract high-dimensional latent features of the feature values ​​at each position in the water level input vector, wherein the formula is: ,in It is the water level input vector. It is the output vector. It is a weight matrix. It is a bias vector. Represents matrix multiplication; The second one-dimensional convolutional coding unit is used to perform one-dimensional convolutional coding on the water level input vector using the one-dimensional convolutional layer of the temporal encoder according to the following formula to extract high-dimensional implicit correlation features between feature values ​​at each position in the water level input vector, wherein the formula is: Where a is the width of the convolution kernel in the x-direction, F is the convolution kernel parameter vector, G is the local vector matrix operated on by the convolution kernel function, and w is the size of the convolution kernel. This represents the water level input vector. This represents performing one-dimensional convolutional encoding on the water level input vector.

4. The anti-buoyancy and pressure relief drainage system for the basement floor slab of a building according to claim 3, characterized in that, The joint Gaussian density map construction unit is used to: use the Gaussian density map to calculate the time series feature vector of the absolute water level and the time series feature vector of the relative water level using the following formula to obtain the joint Gaussian density map; The formula is as follows: ,in, This represents the mean vector of the joint Gaussian density map. The covariance matrix of the joint Gaussian density map is represented. This represents the time-series feature vector of the absolute water level. This represents the time-series feature vector of the relative water level.

5. The anti-buoyancy and pressure relief drainage system for the basement floor slab according to claim 4, characterized in that, The weighting factor calculation unit includes: The water level absolute change weighting factor calculation subunit is used to calculate the Gaussian probability density distribution distance index between the water level absolute quantity time series feature vector and the joint Gaussian density map using the following formula to obtain the first Gaussian probability density distribution distance index. The formula is as follows: ,in, It is the time-series feature vector of the absolute water level. and These are the mean vector and covariance matrix of the joint Gaussian density plot, and the time-series feature vector of the absolute water level is in column vector form. The term represents the exponentiation operation on a matrix, where the exponentiation operation is expressed as the value of a natural exponential function raised to the power of the eigenvalues ​​at each position in the matrix. This indicates subtraction by position. Represents matrix multiplication. This represents the distance index of the first Gaussian probability density distribution.

6. The anti-buoyancy and pressure relief drainage system for the basement floor slab of a building according to claim 5, characterized in that, The weighting factor calculation unit further includes: The water level relative change weighting factor calculation subunit is used to calculate the Gaussian probability density distribution distance index between the water level relative quantity time series feature vector and the joint Gaussian density map using the following formula to obtain the second Gaussian probability density distribution distance index. The formula is as follows: ,in, It is the time-series feature vector of the relative water level. and These are the mean vector and covariance matrix of the joint Gaussian density map, and the time-series feature vector of the relative water level is in column vector form. The term represents the exponentiation operation on a matrix, where the exponentiation operation is expressed as the value of a natural exponential function raised to the power of the eigenvalues ​​at each position in the matrix. This indicates subtraction by position. Represents matrix multiplication. This represents the distance exponent of the second Gaussian probability density distribution.

7. The anti-buoyancy and pressure relief drainage system for the basement floor slab according to claim 6, characterized in that, The Gaussian discretization unit includes: Discrete subunits are used to discretize the Gaussian distribution at each location in the fused Gaussian density map to obtain multiple one-dimensional vectors; and A two-dimensional arrangement subunit is used to arrange the plurality of one-dimensional vectors in a two-dimensional manner to obtain the water level decoding feature matrix.

8. The anti-buoyancy and pressure relief drainage system for the basement floor slab according to claim 7, characterized in that, The water pump power recommendation module is used to: use the decoder to perform decoding regression on the water level decoding feature matrix using the following formula to obtain a decoding value representing the difference in imaging quality; The formula is as follows: ,in This represents the water level decoding feature matrix. It is the decoded value, It is a weight matrix. This represents matrix multiplication.

Images