Intelligent deformation measurement method and device

Abstract

The application discloses an intelligent deformation measurement method and device, the method comprises the following steps: establishing a deformation measurement sensor network, determining a network node working strategy through an artificial intelligence algorithm based on a greedy strategy of an ant colony algorithm, acquiring signals collected by active sensors and performing multimodal fusion, calculating deformation results of a plurality of monitoring points corresponding to a measurement object, and adaptively optimizing the network working strategy according to the deformation rate of the plurality of monitoring points. The application realizes intelligent working strategy determination of the deformation measurement sensor network, reduces the overall energy consumption of the sensor network, and improves the reliability of the system; the current sensor network working strategy is adaptively adjusted according to the deformation rate of the monitoring points, the dynamic adjustment of the monitoring network is realized, and the adaptability of the system is improved.

Description

Technical Field

[0001] This invention relates to the field of deformation measurement, and in particular to an intelligent deformation measurement method and apparatus. Background Technology

[0002] Deformation measurement technology is widely used in the safety monitoring of various engineering structures, providing crucial data support for infrastructure health assessment and disaster early warning. In the monitoring of bridges, tunnels, and other engineering projects, deformation is measured using sensors such as inclinometers, fiber optic sensors, and laser displacement sensors. However, current technologies for deformation measurement in engineering structures still suffer from problems such as high power consumption and the inability of multiple sensors to work collaboratively. Summary of the Invention

[0003] (a) Technical problems to be solved

[0004] To address the aforementioned technical problems, this invention provides an intelligent deformation measurement method and apparatus.

[0005] (II) Technical Solution

[0006] To solve the aforementioned technical problems and achieve the invention's objective, the present invention is implemented through the following technical solution:

[0007] A smart deformation measurement method includes the following steps:

[0008] S1: Establish a deformation measurement sensor network, including several different types of sensors;

[0009] S2: Based on different types of sensor information and the location of the point to be measured, an artificial intelligence algorithm based on a greedy strategy using ant colony algorithm is used to determine the working strategy of network nodes.

[0010] S3: Acquire signals from active sensors and perform multimodal fusion;

[0011] S4: Calculate the deformation results of several monitoring points corresponding to the measured object;

[0012] S5: Adaptive optimization of the network working strategy based on the deformation change rate of several monitoring points.

[0013] Furthermore, the different types of sensors include MEMS tiltmeters, fiber Bragg grating sensors, and laser displacement sensors.

[0014] Furthermore, in step S2, the artificial intelligence algorithm based on the greedy strategy of the ant colony algorithm is input into the set of information S = {s1, s2, ..., s} of different types of sensors. N} and the locations of N monitoring points P = {p1, p2, ..., p N}; where s iIt is a multi-dimensional vector, including type, expected lifetime, power consumption, acquisition range, location coordinates, and type fitness coefficient; it determines the maximum number of algorithm iterations and algorithm initialization parameters.

[0015] Furthermore, the objective function of the artificial intelligence algorithm based on the greedy strategy of the ant colony algorithm is as follows:

[0016]

[0017] Among them, E k Let ∑E be the energy consumption of the k-th sensor. k r represents the total energy consumption of the active sensor. k T is the type fitness coefficient for the k-th sensor. k For the lifetime of the k-th sensor, d k Let w1, w2, w3, and w4 be the sum of the distances between the k-th sensor and the measurement points within its range, and w1, w2, w3, and w4 be the corresponding weights.

[0018] The constraints are as follows:

[0019] a. The measurement range covers all measurement points;

[0020] b. The sensor's energy capacity can meet the energy consumption requirements;

[0021] c. The sensor's acquisition frequency is less than or equal to its maximum set frequency.

[0022] Furthermore, step S3 also includes generating a joint representation based on a variational autoencoder, and mining deep features in the data through variational inference and parameter reconstruction.

[0023] Furthermore, step S5 includes:

[0024] S51: Calculate the rate of change of deformation at the monitoring point;

[0025] S52: If the rate of change of deformation at the monitoring point is greater than the set threshold, the working strategy of the sensor network associated with that monitoring point will be adjusted.

[0026] Furthermore, the adjustment of the sensor network operating strategy related to this monitoring point is as follows:

[0027] a. Obtain the set of monitoring points EX{ex1,ex2,…ex} whose deformation rate of change is greater than the threshold. v}, where v is the total number of monitoring points where the rate of change of deformation is greater than the threshold;

[0028] b. Cluster the above set EX according to its position;

[0029] c. Determine the location of the centroid within each group;

[0030] d. Determine the sensor that needs adjustment based on the centroid position; select the sensor within the set threshold range around the centroid as the sensor that needs adjustment;

[0031] e. Determine the acquisition frequency of the sensor that needs to be adjusted based on the rate of change of deformation within each group; if the adjusted acquisition frequency exceeds the maximum acquisition frequency of the sensor, then activate the sensor with the smallest effective distance from the sensor; the sensor with the smallest effective distance is the sensor with the smallest distance from the sensor within a set threshold range around the centroid.

[0032] f. Output the updated active sensor number and acquisition frequency.

[0033] Furthermore, in step e, the acquisition frequency of the sensor that needs to be adjusted is determined based on the average rate of change of deformation within each group.

[0034] The present invention also provides an intelligent deformation measurement device, comprising:

[0035] The deformation measurement sensor network establishment module is used to select sensors based on the characteristics of the monitored object and obtain the expected lifespan, power consumption, and acquisition range of the corresponding sensors.

[0036] The sensor network node working strategy determination module is used to determine the active sensors and their acquisition frequency in the sensor network using an artificial intelligence algorithm based on a greedy strategy of ant colony algorithm.

[0037] The sensor signal multimodal fusion module is used to acquire multiple types of signals collected by active sensors, perform spatiotemporal data alignment, and generate joint representations based on variational autoencoders.

[0038] The deformation calculation module for monitoring points is used to calculate the deformation of monitoring points based on the results of multi-sensor fusion.

[0039] The adaptive strategy optimization module is used to adaptively optimize the network working strategy based on the rate of change of deformation at the root monitoring point.

[0040] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium storing program instructions for an intelligent deformation measurement method, wherein the program instructions for intelligent deformation measurement can be executed by one or more processors to implement the steps of the intelligent deformation measurement method as described above.

[0041] (III) Beneficial Effects

[0042] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0043] (1) The present invention realizes the determination of intelligent working strategy of deformation measurement sensor network, reduces the energy consumption of the overall sensor network, and improves the reliability of the system.

[0044] (2) The present invention adaptively adjusts the current sensor network working strategy according to the deformation change rate of the monitoring point, thereby realizing the dynamic adjustment of the monitoring network and improving the system adaptability.

[0045] (3) This invention generates joint representations based on variational autoencoders, and mines deep features in the data through variational inference and parameter reconstruction, thereby improving the ability to capture key features of the data. Attached Figure Description

[0046] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0047] Figure 1 This is a schematic flowchart of an intelligent deformation measurement method according to an embodiment of this application. Detailed Implementation

[0048] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0049] The following specific examples illustrate the implementation of this disclosure. Those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0050] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this disclosure. The drawings only show the components related to this disclosure and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0051] See Figure 1 A smart deformation measurement method includes the following steps:

[0052] S1: Establish a deformation measurement sensor network;

[0053] Select MEMS inclinometers, fiber optic grating sensors, laser displacement sensors, etc., based on the characteristics of the monitored object; alternatively, select appropriate sensors for different measurement locations.

[0054] For example, in bridge monitoring, FBG sensors are placed at key points on the main beam, and MEMS inclinometers are deployed on the piers.

[0055] Furthermore, the expected lifespan, power consumption, and acquisition range of the corresponding sensor are obtained;

[0056] S2: Determine the working strategy of network nodes based on information from different types of sensors and the location of the points to be measured;

[0057] In deformation monitoring scenarios for infrastructure such as highway bridges, pre-installed sensor networks typically employ redundancy to ensure system reliability. However, in actual operation, keeping all nodes active for extended periods leads to increased energy consumption and reduced equipment lifespan. Based on multi-dimensional optimization objectives, this invention proposes dynamically adjusting the operating modes of network nodes. An intelligent decision-making algorithm selects active sensor combinations and adaptively adjusts the sampling frequency, allowing redundant nodes to enter a low-power sleep state. This achieves synergistic optimization of monitoring accuracy, system energy consumption, and equipment durability. This strategy effectively solves the problems of low resource utilization and poor sustainability in traditional monitoring systems.

[0058] The sensor information includes expected lifespan, power consumption, acquisition range, location coordinates, and type adaptation coefficient.

[0059] The type adaptation coefficient is determined according to the different types of sensors. Since different types of sensors have different measurement principles, their accuracy and applicability vary, which is reflected in the adaptation coefficient. Optionally, the coefficient is determined based on experience. The higher the accuracy, the stronger the applicability, and the larger the type adaptation coefficient.

[0060] Specifically, the determination of the artificial intelligence algorithm based on the greedy strategy of ant colony optimization includes the following steps:

[0061] S21: Input a set of information from different types of sensors, S = {s1, s2, ... s...} N} and the locations of N monitoring points P = {p1, p2, ..., p N}; where s i This is a multi-dimensional vector, optional, including type, expected lifetime, power consumption, acquisition range, location coordinates, and type fitness coefficient; it determines the maximum number of algorithm iterations and algorithm initialization parameters;

[0062] S22: Initialize the sensor network state, randomly assign the initial network state, and determine the active sensor combination and its acquisition frequency;

[0063] S23: Calculate the measurement coverage area and objective function value under the current sensor network state; where the objective function is as follows:

[0064]

[0065] Among them, E k Let ∑E be the energy consumption of the k-th sensor. k r represents the total energy consumption of the active sensor. k T is the type fitness coefficient for the k-th sensor. k For the lifetime of the k-th sensor, d k Let w1, w2, w3, and w4 be the sum of the distances between the k-th sensor and the measurement points within its range, and w1, w2, w3, and w4 be the corresponding weights.

[0066] The constraints are as follows:

[0067] a. The measurement range covers all measurement points;

[0068] b. The sensor's energy capacity can meet the energy consumption requirements;

[0069] c. The sensor's acquisition frequency is less than or equal to its maximum set frequency.

[0070] S24: Update the ant colony position and calculate the objective function value at the current position, specifically as follows:

[0071] If the objective function value at the current position is better than that at the previous position, then the objective function value generated in this iteration is set as the global optimal objective function value, and the current position is set as the global optimal position; otherwise, it remains unchanged.

[0072] S25: If the maximum number of iterations is reached or the objective function value meets the requirements, the global optimal position is output, i.e., the active sensor number and its acquisition frequency.

[0073] S3: Acquire signals from active sensors and perform multimodal fusion;

[0074] S31: Acquire multiple types of signals collected by active sensors and perform spatiotemporal data alignment;

[0075] This includes unifying time and coordinates;

[0076] S32: Data multimodal fusion, specifically including:

[0077] To achieve multimodal data fusion and address the imbalance of original data samples, this invention normalizes the data and generates joint representations based on variational autoencoders, then mines deep features in the data through variational inference and parameter reconstruction.

[0078] The variational autoencoder takes as input the time series of multiple types of signals acquired and outputs the reconstructed joint data feature matrix after processing.

[0079] Optionally, in order to achieve the ability to capture key features of data, the present invention improves the variational autoencoder by introducing an attention mechanism. Preferably, a self-attention layer is added between the encoder and decoder of the variational autoencoder.

[0080] S4: Calculate the deformation results of N monitoring points corresponding to the measured object; including:

[0081] S41: For the i-th monitoring point out of N monitoring points, obtain its corresponding set of valid sensors S. i ;

[0082] The effective sensor set S i This refers to the set of all active sensors located within the sensor's acquisition range at the monitoring point;

[0083] S42: Obtain the effective sensor set S from the joint data feature matrix. i The corresponding data;

[0084] S43: Calculate the deformation u at the i-th monitoring point. i ;

[0085] Specifically, firstly, the deformation at monitoring point i is calculated based on sensor data from the effective sensor set. in, The deformation amount at the i-th monitoring point is calculated for the k-th sensor, thus obtaining the set of deformation results for monitoring point i. K represents the total number of sensor data points in the current effective sensor set; furthermore, the deformation g is calculated based on a weighted fusion method. i :

[0086] Where w k The weight of the k-th sensor data.

[0087] Optionally, the weights are determined based on the Euclidean distance between the sensor location and the measurement point.

[0088] S44: Repeat steps S41-S43, sequentially calculating the deformation results G={g1,g2…g…g…} at N monitoring points. N}

[0089] S5: Adaptively optimize the network working strategy based on the deformation change rate of N monitoring points.

[0090] S51: Calculate the rate of change of deformation at the monitoring point; specifically, compare the calculated deformation at time t with the deformation at time tp; where p is a set time threshold; the calculation method is as follows:

[0091] y i =g i t -g i t-p

[0092] Among them, y i Let u be the rate of change of deformation at the i-th monitoring point. i t Let u be the deformation at the i-th monitoring point at time t. i t-p Let be the deformation of the i-th monitoring point at time tp.

[0093] S52: If y i If the value exceeds the set threshold, the working strategy of the sensor network associated with the monitoring point will be adjusted; wherein, the sensor network associated with the monitoring point is a sensor network consisting of a set of all sensors whose acquisition range includes the monitoring point.

[0094] Optionally, the sensor network operating strategy can be adjusted in the following ways:

[0095] a. Obtain the set of monitoring points EX{ex1,ex2,…ex} whose deformation rate of change is greater than the threshold. v}, where v is the total number of monitoring points where the rate of change of deformation is greater than the threshold;

[0096] b. Cluster the above set EX according to its position;

[0097] c. Determine the location of the centroid within each group;

[0098] d. Determine the sensor that needs to be adjusted based on the position of the centroid; optionally, select the sensor within a set threshold range around the centroid as the sensor that needs to be adjusted;

[0099] e. Determine the acquisition frequency of the sensor that needs to be adjusted based on the rate of change of deformation within each group; if the adjusted acquisition frequency exceeds the maximum acquisition frequency of the sensor, then activate the sensor with the smallest effective distance from the sensor; the sensor with the smallest effective distance is the sensor with the smallest distance from the sensor within a set threshold range around the centroid.

[0100] Optionally, the sampling frequency of the sensor to be adjusted can be determined based on the average rate of change of deformation within each group.

[0101] f. Output the updated active sensor number and acquisition frequency.

[0102] In this implementation, based on multi-dimensional optimization objectives, a dynamic adjustment mode for network node operation is proposed. An intelligent decision-making algorithm selects active sensor combinations and adaptively adjusts the sampling frequency, allowing redundant nodes to enter a low-power sleep state. This achieves synergistic optimization of monitoring accuracy, system energy consumption, and equipment durability. This strategy effectively solves the problems of low resource utilization and poor sustainability in traditional monitoring systems.

[0103] This invention also proposes a deformation measuring device, comprising:

[0104] The deformation measurement sensor network establishment module is used to select sensors based on the characteristics of the monitored object and obtain the expected lifespan, power consumption, and acquisition range of the corresponding sensors.

[0105] The sensor network node working strategy determination module is used to determine the active sensors and their acquisition frequency in the sensor network using an artificial intelligence algorithm based on a greedy strategy of ant colony algorithm.

[0106] The sensor signal multimodal fusion module is used to acquire multiple types of signals collected by active sensors, perform spatiotemporal data alignment, and generate joint representations based on variational autoencoders.

[0107] The deformation calculation module for monitoring points is used to calculate the deformation of monitoring points based on the results of multi-sensor fusion.

[0108] The adaptive strategy optimization module is used to adaptively optimize the network working strategy based on the rate of change of deformation at the root monitoring point.

[0109] Furthermore, embodiments of the present invention also propose a computer-readable storage medium storing program instructions for a deformation measurement method. These program instructions can be executed by one or more processors to implement the steps of the deformation measurement method as described above.

[0110] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A smart deformation measurement method, characterized in that, Includes the following steps: S1: Establish a deformation measurement sensor network, including several different types of sensors; S2: Based on information from different types of sensors and the location of the point to be measured, an artificial intelligence algorithm based on a greedy strategy using an ant colony algorithm is used to determine the working strategy of the network nodes; the different types of sensors include MEMS inclinometers, fiber optic grating sensors, and laser displacement sensors. The artificial intelligence algorithm based on the ant colony algorithm and its greedy strategy is input into different types of sensor information sets. and the locations of N monitoring points ;in, It is a multi-dimensional vector, including type, expected lifetime, power consumption, acquisition range, location coordinates, and type fitness coefficient, i=1…M, where M is the number of sensors; determine the maximum number of algorithm iterations and algorithm initialization parameters; The objective function of the artificial intelligence algorithm based on the greedy strategy of ant colony algorithm is as follows: ； in, Let k be the energy consumption of the k-th sensor. The total energy consumption of the active sensor, Let k be the type fitness coefficient of the k-th sensor. For the lifetime of the k-th sensor, Let be the sum of the distances between the k-th sensor and the measurement points within its range. , , , These are the corresponding weights; The constraints are as follows: a. The measurement range covers all measurement points; b. The sensor's energy capacity can meet the energy consumption requirements; c. The sensor's acquisition frequency is less than or equal to its maximum set frequency; S3: Acquire signals from active sensors and perform multimodal fusion; S4: Calculate the deformation results of several monitoring points corresponding to the measured object; S5: Adaptive optimization of the network working strategy based on the deformation change rate of several monitoring points; Step S5 includes: S51: Calculate the rate of change of deformation at the monitoring point; S52: If the rate of change of deformation at the monitoring point is greater than the set threshold, the working strategy of the sensor network associated with that monitoring point will be adjusted. The adjustment strategy for the sensor network associated with this monitoring point is as follows: a. Obtain the set of monitoring points where the rate of change of deformation is greater than a threshold. , where v is the total number of monitoring points where the rate of change of deformation is greater than the threshold; b. Cluster the above set EX according to its position; c. Determine the location of the centroid within each group; d. Determine the sensor that needs adjustment based on the centroid position; select the sensor within the set threshold range around the centroid as the sensor that needs adjustment; e. Determine the acquisition frequency of the sensor that needs to be adjusted based on the rate of change of deformation within each group; if the adjusted acquisition frequency exceeds the maximum acquisition frequency of the sensor, then activate the sensor with the smallest effective distance from the sensor; the sensor with the smallest effective distance is the sensor with the smallest distance from the sensor within a set threshold range around the centroid. f. Output the updated active sensor number and acquisition frequency.

2. The intelligent deformation measurement method according to claim 1, characterized in that, Step S3 further includes generating joint representations based on variational autoencoders, and mining deep features in the data through variational inference and parameter reconstruction.

3. The intelligent deformation measurement method according to claim 2, characterized in that, In step e, the acquisition frequency of the sensor that needs to be adjusted is determined based on the average rate of change of deformation within each group.

4. An intelligent deformation measurement device, based on the intelligent deformation measurement method as described in any one of claims 1 to 3, comprising: The deformation measurement sensor network establishment module is used to select sensors based on the characteristics of the monitored object and obtain the expected lifespan, power consumption, and acquisition range of the corresponding sensors. The sensor network node working strategy determination module is used to determine the active sensors and their acquisition frequency in the sensor network using an artificial intelligence algorithm based on a greedy strategy of ant colony algorithm. The sensor signal multimodal fusion module is used to acquire multiple types of signals collected by active sensors, perform spatiotemporal data alignment, and generate joint representations based on variational autoencoders. The deformation calculation module for monitoring points is used to calculate the deformation of monitoring points based on the results of multi-sensor fusion. The adaptive strategy optimization module is used to adaptively optimize the network working strategy based on the rate of change of deformation at the root monitoring point.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores program instructions for an intelligent deformation measurement method, which can be executed by one or more processors to implement the steps of the intelligent deformation measurement method as described in any one of claims 1-3.

Images