Data monitoring method in symmetrical dismantling process of highway T-shaped rigid frame bridge

Abstract

The invention relates to the technical field of data processing, and provides a data monitoring method in a symmetrical dismantling process of a T-shaped rigid frame bridge of a road, comprising the following steps: acquiring a frequency amplitude sequence; obtaining the availability of the frequency according to the frequency amplitude sequence; obtaining the overall availability degree according to the availability of the frequency; obtaining demolition symmetry according to the amplitude of the frequency and the overall availability degree; obtaining the overall influence degree according to the frequency value and the amplitude value; obtaining adjacent similarity according to the dismantling symmetry and the overall influence degree of the sensor category; obtaining the similarity of the sensor categories according to the adjacent similarity; and obtaining an abnormal acceleration sensor according to the similarity of the sensor category, and further obtaining a symmetrical dismantling monitoring result of the T-shaped rigid frame bridge of the highway. According to the invention, the abnormal acceleration sensor is obtained through the similarity of reliable sensor categories, so that the monitoring result of the bridge symmetric demolition process is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly relates to a data monitoring method during the symmetrical demolition of a highway T-shaped rigid frame bridge. Background Art

[0002] A highway T-shaped rigid frame bridge is a bridge structure form, generally composed of a bridge deck, bridge piers, abutments and a rigid beam (T-shaped cross-section), and is commonly used for traffic on highways or urban roads. During its demolition process, in order to ensure uniform force during demolition, ensure the balance of the overall bridge structure, reduce the safety risks of the bridge tilting or collapsing, and avoid unnecessary deformation or damage caused by premature or unbalanced demolition on one side, a symmetrical demolition method is generally used for its demolition. And the vibration generated during the bridge demolition process may affect the stability of the bridge and even cause unexpected collapse of the structure. Therefore, during the symmetrical demolition of a highway T-shaped rigid frame bridge, it is necessary to monitor the vibration data.

[0003] When analyzing the vibration data during the symmetrical demolition of a highway T-shaped rigid frame bridge, the demolition symmetry of the sensor categories composed of symmetrical sensors will be analyzed, and then clustering will be performed through the similarity between sensor categories to obtain abnormal acceleration sensors. However, for the similarity between different sensor categories, the availability of the data corresponding to each frequency is different when calculating the demolition symmetry of the sensor categories, resulting in inaccurate calculation of the demolition symmetry. And since the propagation between two different sensor categories during the demolition process is affected not only by the distance but also by other intermediate sensor categories, using only the symmetrical demolition property as the similarity measure between sensor categories is inaccurate, leading to inaccurate monitoring results of the bridge demolition. Summary of the Invention

[0004] The present invention provides a data monitoring method during the symmetrical demolition of a highway T-shaped rigid frame bridge to solve the problem of inaccurate monitoring results during the existing symmetrical demolition process of the bridge. The specific technical solution adopted is as follows:

[0005] The present invention proposes a data monitoring method during the symmetrical demolition of a highway T-shaped rigid frame bridge, and the method includes the following steps:

[0006] Arrange a plurality of acceleration sensors on the bridge according to the positions of the bridge piers, and use the acceleration sensors to obtain the frequency amplitude sequence of each acceleration sensor; the frequency amplitude sequence includes the amplitude of the acceleration sensor at each frequency;

[0007] Based on the amplitude and frequency value of the frequency in the frequency amplitude sequence of the acceleration sensor, obtain the availability of each frequency in the frequency amplitude sequence of each acceleration sensor; divide all acceleration sensors into several sensor categories; based on the availability of each frequency in the frequency amplitude sequence of the acceleration sensors in the sensor category, obtain the overall availability of each sensor category at each frequency;

[0008] Based on the similarity relationship of the amplitudes of the frequencies in the frequency amplitude sequences of all acceleration sensors within the sensor category, and the overall availability of the sensor category at each frequency, obtain the removal symmetry of each sensor category;

[0009] Based on the frequency value and its corresponding amplitude in the frequency amplitude sequence of the acceleration sensor, obtain the overall influence degree of the sensor interval on similarity; obtain several groups of adjacent categories according to the positions of the acceleration sensors within the sensor category, and based on the removal symmetry of the sensor categories in each group of adjacent categories and the overall influence degree of the sensor interval on similarity, obtain the adjacent similarity of each group of adjacent categories; based on the adjacent similarity of each group of adjacent categories, obtain the similarity between every two sensor categories;

[0010] Based on the similarity of all sensor categories, obtain the abnormal acceleration sensors, and further obtain the monitoring results of the symmetrical removal of the highway T-shaped rigid frame bridge.

[0011] Furthermore, the specific method for obtaining the availability of each frequency in the frequency amplitude sequence of each acceleration sensor based on the amplitude and frequency value of the frequency in the frequency amplitude sequence of the acceleration sensor is as follows:

[0012]

[0013] where Y i,h is the availability of the h-th frequency in the frequency amplitude sequence of the i-th acceleration sensor; F i,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the i-th acceleration sensor; F i-1,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the (i - 1)-th acceleration sensor; F i+1,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the (i + 1)-th acceleration sensor; P i,h is the frequency value of the h-th frequency in the frequency amplitude sequence of the i-th acceleration sensor; norm() is the linear normalization function; || is the absolute value function.

[0014] Furthermore, the specific method for dividing all acceleration sensors into several sensor categories includes:

[0015] Record each pair of acceleration sensors symmetric about the pier position as a sensor category.

[0016] Further, obtaining the overall availability of each sensor category at each frequency according to the availability of each frequency in the frequency amplitude sequence of the acceleration sensor in the sensor category includes the following specific method:

[0017] For any sensor category and any frequency, the minimum value of the availability of this frequency in the frequency amplitude sequences of the two acceleration sensors in this sensor category is denoted as the availability index of this sensor category at this frequency;

[0018] The inverse proportional normalization result of the absolute value of the difference in the availability of this frequency in the frequency amplitude sequences of the two acceleration sensors within this sensor category is denoted as the reliability of this sensor category at this frequency;

[0019] The product of the reliability and the availability index of this sensor category at this frequency is denoted as the overall availability of this sensor category at this frequency.

[0020] Further, obtaining the removal symmetry of each sensor category according to the similarity relationship of the amplitudes of the frequencies in the frequency amplitude sequences of all acceleration sensors within the sensor category, and the overall availability of the sensor category at each frequency, the specific obtaining method is as follows:

[0021]

[0022] In the formula, D q is the removal symmetry of the q-th sensor category; W is the number of frequencies in the frequency amplitude sequence of each acceleration sensor; K' q,h is the overall availability of the q-th sensor category at the h-th frequency; F' q,1,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the first acceleration sensor within the q-th sensor category; F' q,2,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the second acceleration sensor within the q-th sensor category; ε is a hyperparameter; || is the absolute value function; softmax() is the weight normalization function.

[0023] Further, obtaining the overall influence degree of the sensor interval on the similarity according to the frequency values and their corresponding amplitudes in the frequency amplitude sequence of the acceleration sensor includes the following specific method:

[0024] The calculation method of the left influence degree of the sensor interval on the similarity is:

[0025]

[0026] Wherein, X is the left influence degree of the sensor interval on similarity; W is the number of frequencies in the frequency amplitude sequence of each acceleration sensor; P′ h is the frequency value of the h-th frequency in the frequency amplitude sequence of the acceleration sensor; N is the number of acceleration sensors on the left side of the bridge pier; F″ c,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the c-th acceleration sensor on the left side of the bridge pier; μ h is the mean value of the amplitudes of the h-th frequency in the frequency amplitude sequences of all acceleration sensors on the left side of the bridge pier.

[0027] Obtain the right influence degree of the sensor interval on similarity;

[0028] Denote the sum of the left influence degree and the right influence degree of the sensor interval on similarity as the overall influence degree of the sensor interval on similarity.

[0029] Furthermore, the method for obtaining several groups of adjacent categories according to the positions of the acceleration sensors within the sensor category, and obtaining the adjacent similarity of each group of adjacent categories according to the removal symmetry of the sensor categories in each group of adjacent categories and the overall influence degree of the sensor interval on similarity includes the following specific methods:

[0030] Number all sensor categories in ascending order of the Euclidean distance between the acceleration sensors in each sensor category and the bridge pier;

[0031] Regard any two sensor categories with adjacent serial numbers as a group of adjacent categories;

[0032] For any group of adjacent categories, denote the inverse proportional normalization result of the absolute value of the difference in removal symmetry between the two sensor categories in this group of adjacent categories as the basic similarity of this group of adjacent categories;

[0033] Denote the product of the inverse proportional normalization result of the overall influence degree of the sensor interval on similarity and the basic similarity of this group of adjacent categories as the adjacent similarity of this group of adjacent categories.

[0034] Furthermore, the method for obtaining the similarity between every two sensor categories according to the adjacent similarity of each group of adjacent categories includes the following specific methods:

[0035] For any two sensor categories, during the process from the serial number of any one sensor category of these two sensor categories to the serial number of the other sensor category, calculate the adjacent similarity between the sensor category corresponding to each serial number and the next sensor category in turn, and denote the product of the adjacent similarities between the sensor categories corresponding to all serial numbers in this process and the next sensor category as the similarity between these two sensor categories.

[0036] Further, the specific method for obtaining the abnormal acceleration sensors according to the similarities of all sensor categories is as follows:

[0037] Step 1) Among all sensor categories, merge the two sensor categories with the highest similarity as a new sensor category.

[0038] Step 2) Denote the average of the similarities between the two sensor categories corresponding to the new sensor category before merging and each of the other sensor categories except the corresponding two sensor categories as the similarity between the new sensor category and the other sensor category.

[0039] Step 3) Obtain a similarity threshold based on the similarity of the two sensor categories merged for the first time.

[0040] Step 4) Repeat Step 1 and Step 2 until the similarities between all sensor categories are less than the similarity threshold, and then stop repeating.

[0041] The specific method for obtaining the similarity threshold is as follows:

[0042] Z = (1 - β) × S

[0043] In the formula, Z is the similarity threshold; β is a preset similarity fluctuation index; S is the similarity of the two sensor categories merged for the first time.

[0044] Denote the sensor categories with two acceleration sensors among the finally obtained several sensor categories as abnormal sensor categories; denote all the acceleration sensors corresponding to all abnormal sensor categories as abnormal acceleration sensors.

[0045] Further, the specific method for obtaining the frequency amplitude sequence of each acceleration sensor by using the acceleration sensor is as follows:

[0046] During the demolition process of a highway T-shaped rigid frame bridge, use the acceleration sensors to record the acceleration values, and denote the time series sequence formed by the acceleration values collected by each acceleration sensor as the acceleration monitoring sequence of each acceleration sensor.

[0047] Perform Fourier transform on the acceleration monitoring sequence of each acceleration sensor to obtain the amplitude spectrum of each acceleration sensor; normalize the amplitude of each integer frequency in the amplitude spectrum of any one acceleration sensor according to the maximum value of the amplitudes of all integer frequencies in the amplitude spectra of all acceleration sensors, and arrange the resulting sequence in ascending order of integer frequencies, which is denoted as the frequency amplitude sequence of this acceleration sensor.

[0048] The beneficial effects of the present invention are as follows: When analyzing the frequency amplitude sequence during the symmetric demolition of a highway T-shaped rigid frame bridge, since the availability of each frequency in the frequency amplitude sequence of each acceleration sensor is different, the present invention obtains the availability of each frequency in the frequency amplitude sequence of each acceleration sensor according to the amplitude and frequency value of the frequency in the frequency amplitude sequence of the acceleration sensor, and then obtains the overall availability of each sensor category at each frequency, providing a calculation weight at each frequency for the demolition symmetry of the sensor category; since the similarity of sensor categories is mainly measured based on the demolition symmetry of sensor categories, the present invention uses the similarity relationship of the amplitudes of the frequencies in the frequency amplitude sequences of all acceleration sensors within the sensor category, and takes the overall availability of the sensor category at each frequency as the weight to obtain the demolition symmetry of each sensor category; since the contribution of the fluctuation degree of the amplitudes of different frequencies to the similarity is different, for vibration monitoring, since low-frequency signals can effectively cover a longer distance, the area reflected by low-frequency signals will be larger when calculating the influence degree on similarity. The present invention obtains the overall influence degree of the sensor interval on similarity, and based on the demolition symmetry of the sensor category and the overall influence degree of the sensor interval on similarity, obtains the similarity between each two sensor categories. Thus, the present invention obtains abnormal acceleration sensors through the reliable similarity between sensor categories, and obtains a more accurate data monitoring result during the symmetric demolition of the highway T-shaped rigid frame bridge. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following-described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0050] Figure 1 It is a schematic flow chart of a data monitoring method during the symmetric demolition of a highway T-shaped rigid frame bridge provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0051] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0052] Please refer to Figure 1, which shows a flowchart of a data monitoring method during the symmetric demolition of a highway T-shaped rigid frame bridge provided by an embodiment of the present invention. The method includes the following steps:

[0053] Step S001: Arrange a number of acceleration sensors on the bridge according to the pier positions, and use the acceleration sensors to obtain the frequency amplitude sequence of each acceleration sensor.

[0054] It should be noted that the purpose of this embodiment is to monitor the vibration data during the symmetric demolition of a highway T-shaped rigid frame bridge. Since the symmetric demolition method is generally used for highway T-shaped rigid frame bridges, it is necessary to analyze the vibration data to obtain the demolition symmetry of each symmetric position, and then combine the influence of the position interval to obtain abnormal points. Therefore, it is first necessary to use sensors to collect the vibration data of the bridge during the demolition process.

[0055] Specifically, before the demolition of the highway T-shaped rigid frame bridge, on the main girder within the range of the current bridge section to be demolished, with the current pier to be demolished as the center, arrange an acceleration sensor every 1 meter;

[0056] During the demolition of the highway T-shaped rigid frame bridge, use the acceleration sensors to record the acceleration values at a sampling rate of 1000 Hz, and record the time series sequence composed of the acceleration values collected by each acceleration sensor in the most recent 1 second as the acceleration monitoring sequence of each acceleration sensor;

[0057] Perform Fourier transform on the acceleration monitoring sequence of each acceleration sensor to obtain the amplitude spectrum of each acceleration sensor; among them, the Fourier transform is a well-known technology, and the specific method will not be introduced here; normalize the amplitude of each integer frequency in the amplitude spectrum of any one acceleration sensor according to the maximum value of the amplitudes of all integer frequencies in the amplitude spectra of all acceleration sensors, and arrange them in ascending order of integer frequency to form a sequence, which is recorded as the frequency amplitude sequence of the acceleration sensor.

[0058] Step S002: Obtain the availability of each frequency in the frequency amplitude sequence of each acceleration sensor according to the amplitude and frequency value of the frequency in the frequency amplitude sequence of the acceleration sensor; divide all acceleration sensors into several sensor categories; according to the availability of each frequency in the frequency amplitude sequence of the acceleration sensors in the sensor category, obtain the overall availability of each sensor category at each frequency.

[0059] It should be noted that when analyzing the frequency amplitude sequence of the acceleration sensor to obtain the demolition symmetry of each symmetric position, the availability of each frequency in the frequency amplitude sequence of each acceleration sensor is different. Specifically, for the vibration signal monitored by the acceleration sensor, the high-frequency signal has a short wavelength and is easily absorbed and scattered by the medium during propagation, resulting in more energy loss of the high-frequency signal during propagation. Therefore, its attenuation speed is faster. It is precisely due to the characteristic that high-frequency signals are prone to rapid attenuation that by monitoring the amplitude change of high-frequency signals of different acceleration sensors, the trend of signal attenuation can be captured more sensitively. That is, the higher the availability of the symmetry demolition of the high-frequency signal with a larger amplitude change, especially for adjacent acceleration sensors, the higher the availability with a larger frequency amplitude change.

[0060] Specifically, the calculation method of the availability of the h-th frequency in the frequency amplitude sequence of the i-th acceleration sensor is as follows:

[0061]

[0062] In the formula, Y i,h is the availability of the h-th frequency in the frequency amplitude sequence of the i-th acceleration sensor; F i,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the i-th acceleration sensor; F i-1,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the (i - 1)-th acceleration sensor; F i+1,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the (i + 1)-th acceleration sensor; P i,h is the frequency value of the h-th frequency in the frequency amplitude sequence of the i-th acceleration sensor; norm() is a linear normalization function, and the normalization object is the h-th frequency in the frequency amplitude sequences of all acceleration sensors || is the absolute value function; it should be especially noted that the availability of all frequencies of the first acceleration sensor and the last acceleration sensor is recorded as 0.

[0063] It should be noted that the larger P i,h is, the closer the h-th frequency in the frequency amplitude sequence of the i-th acceleration sensor is to the high-frequency signal; the larger it is, the greater the amplitude change of the h-th frequency in the frequency amplitude sequence of the i-th acceleration sensor.

[0064] It should be noted that since the highway T-shaped rigid frame bridge is demolished symmetrically, the distribution of acceleration sensors has symmetry. Therefore, each group of symmetric acceleration sensors is analyzed as a category.

[0065] Specifically, each pair of acceleration sensors symmetric about the pier position is denoted as a sensor category, and all sensor categories are numbered in ascending order of the Euclidean distance between the acceleration sensors in each sensor category and the pier.

[0066] It should be noted that for any sensor category at any frequency, when evaluating the availability of the sensor category at that frequency, it is required that the two acceleration sensors in the sensor category must be available at that frequency. Therefore, the minimum availability is selected as the availability index.

[0067] Specifically, for any sensor category and any frequency, the minimum value of the availability of that frequency in the frequency amplitude sequences of the two acceleration sensors in the sensor category is denoted as the availability index of the sensor category at that frequency.

[0068] It should be noted that when the difference in the availability of the frequency amplitude sequences of all acceleration sensors in a certain sensor category at a certain frequency is relatively large, it indicates that the signal quality and stability of the sensor category at that frequency vary greatly, and thus the availability index becomes unreliable.

[0069] Specifically, for any sensor category and any frequency, the inverse proportional normalization result of the absolute value of the difference in the availability of that frequency in the frequency amplitude sequences of the two acceleration sensors in the sensor category is denoted as the reliability degree of the sensor category at that frequency;

[0070] The product of the reliability degree and the availability index of the sensor category at that frequency is denoted as the overall availability degree of the sensor category at that frequency.

[0071] Step S003: Obtain the demolition symmetry of each sensor category according to the similarity relationship of the amplitudes of the frequencies in the frequency amplitude sequences of all acceleration sensors in the sensor category and the overall availability degree of the sensor category at each frequency.

[0072] It should be noted that since the two acceleration sensors in each sensor category are symmetric about the pier, and since the highway T-shaped rigid frame bridge is symmetrically demolished, the more similar the amplitudes of the two acceleration sensors in a single sensor category are at each frequency, the higher the symmetry during demolition. Due to the different availabilities of each frequency in the frequency amplitude sequences of the two acceleration sensors in a single sensor category, the demolition symmetry of each sensor category is obtained by combining the availabilities of each frequency in the frequency amplitude sequences of the two acceleration sensors in each sensor category.

[0073] Specifically, the calculation method of the demolition symmetry of the q-th sensor category is as follows:

[0074]

[0075] where D q is the removal symmetry of the q-th sensor category; W is the number of frequencies in the frequency amplitude sequence of each acceleration sensor. It should be noted that the number of frequencies in the frequency amplitude sequences of all acceleration sensors is equal; K′ q,h is the overall availability of the q-th sensor category at the h-th frequency; F′ q,1,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the first acceleration sensor within the q-th sensor category; F′ q,2,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the second acceleration sensor within the q-th sensor category; ε is a hyperparameter to prevent the denominator from being zero. In this embodiment, ε = 0.00001 is taken as an example for description; || is the absolute value function; softmax() is the weight normalization function, and the normalization object is the overall availability of the q-th sensor category at all frequencies.

[0076] Step S004: Obtain the overall influence degree of the sensor interval on similarity according to the frequency values and their corresponding amplitudes in the frequency amplitude sequence of the acceleration sensor; obtain several groups of adjacent categories according to the positions of the acceleration sensors within the sensor category, and obtain the adjacent similarity of each group of adjacent categories according to the removal symmetry of the sensor categories in each group of adjacent categories and the overall influence degree of the sensor interval on similarity; obtain the similarity between every two sensor categories according to the adjacent similarity of each group of adjacent categories.

[0077] It should be noted that for any side of the bridge pier, a number of acceleration sensors are arranged, and each acceleration sensor has different amplitudes for each frequency in the frequency amplitude sequence. For the amplitude of any frequency in the frequency amplitude sequence of any acceleration sensor, if the fluctuation degree of this amplitude relative to the amplitude of the same frequency in the frequency amplitude sequences of other acceleration sensors is greater, it means that the influence degree of this amplitude on similarity is higher. However, the contribution of the fluctuation degree of amplitudes for different frequencies to the influence on similarity is different. For vibration monitoring, since low-frequency signals can effectively cover a longer distance, when calculating the influence degree on similarity, the area reflected by low-frequency signals will be larger, and more extensive spatial information can be provided. That is, the lower the frequency, the higher the weight of the fluctuation degree of the amplitude.

[0078] Specifically, the calculation method of the left influence degree of the sensor interval on similarity is:

[0079]

[0080] Wherein, X is the left influence degree of the sensor interval on similarity; W is the number of frequencies in the frequency amplitude sequence of each acceleration sensor; P′ h is the frequency value of the h-th frequency in the frequency amplitude sequence of the acceleration sensor; N is the number of acceleration sensors on the left side of the pier; F″ c,h is the amplitude of the h-th frequency in the frequency amplitude sequence of the c-th acceleration sensor on the left side of the pier; μ h is the mean value of the amplitudes of the h-th frequency in the frequency amplitude sequences of all acceleration sensors on the left side of the pier.

[0081] It should be noted that represents the fluctuation degree of the amplitude of the h-th frequency in the frequency amplitude sequence of the c-th acceleration sensor on the left side of the pier relative to the amplitude of the h-th frequency in the frequency amplitude sequences of other acceleration sensors.

[0082] Furthermore, according to the acquisition method of the left influence degree of the sensor interval on similarity, the right influence degree of the sensor interval on similarity is acquired;

[0083] The sum value of the left influence degree and the right influence degree of the sensor interval on similarity is denoted as the overall influence degree of the sensor interval on similarity.

[0084] It should be noted that for any two sensor categories, if the acceleration sensors on the same side in these two sensor categories are in adjacent positions, the removal symmetry of the sensor categories can be used to measure the similarity of these two sensor categories. However, due to the negative impact of the sensor interval on similarity, the smaller the overall influence degree of the sensor interval on similarity, the higher the credibility of the similarity.

[0085] Specifically, any two sensor categories with adjacent serial numbers are recorded as a group of adjacent categories;

[0086] For any group of adjacent categories, the inverse proportional normalization result of the absolute value of the difference in the removal symmetry of the two sensor categories in this group of adjacent categories is denoted as the basic similarity of this group of adjacent categories;

[0087] The product of the inverse proportional normalization result of the overall influence degree of the sensor interval on similarity and the basic similarity of this group of adjacent categories is denoted as the adjacent similarity of this group of adjacent categories.

[0088] It should be noted that since the bridge is an integral structure, for two sensor categories with non - adjacent acceleration sensor positions on the same side, if there is a set of adjacent categories between the acceleration sensors corresponding to the same side of these two sensor categories, and due to the appearance of cracks in the bridge during the symmetric demolition process, the mechanical properties, vibration modes, and load distributions of the bridge are affected, resulting in the non - similarity of this set of adjacent categories, then this non - similarity will be transmitted through the main girder. Therefore, the similarity between any two sensor categories is calculated.

[0089] Specifically, for any two sensor categories, during the process from the serial number of any one sensor category of these two sensor categories to the serial number of the other sensor category, the adjacent similarity between the sensor category corresponding to each serial number and the next sensor category is calculated in sequence. The product of the adjacent similarities between the sensor categories corresponding to all serial numbers and the next sensor category during this process is denoted as the similarity between these two sensor categories. It should be noted that if there is no next sensor category corresponding to a certain serial number, the adjacent similarity corresponding to it is recorded as 1.

[0090] Step S005: Based on the similarities of all sensor categories, obtain the abnormal acceleration sensors, and then obtain the monitoring results of the symmetric demolition of the highway T - shaped rigid - frame bridge.

[0091] It should be noted that after obtaining the similarity between each two sensor categories, when performing symmetric demolition on the bridge, the overall stability of the entire bridge needs to be maintained, that is, all sensor categories are relatively similar. Therefore, the sensors are clustered according to the similarities of the sensor categories to determine the abnormal acceleration sensors.

[0092] Specifically, the specific clustering process of the acceleration sensors is as follows:

[0093] Step 1) Among all sensor categories, merge the two sensor categories with the highest similarity as a new sensor category.

[0094] Step 2) Denote the mean of the similarities between the two sensor categories corresponding to the new sensor category before merging and each of the other sensor categories except the corresponding two sensor categories as the similarity between the new sensor category and the other sensor category.

[0095] Step 3) Obtain the similarity threshold according to the similarity between the two sensor categories merged for the first time.

[0096] Step 4) Repeat Step 1 and Step 2 until the similarities between all sensor categories are less than the similarity threshold, then stop repeating.

[0097] The specific method for obtaining the similarity threshold is as follows:

[0098] Z = (1 - β) × S

[0099] Wherein, Z is the similarity threshold; β is a preset similarity fluctuation index, and in this embodiment, β = 0.1 is described; S is the similarity between two sensor categories merged for the first time.

[0100] It should be noted that the demolition of the highway T-shaped steel structure bridge is carried out symmetrically. Therefore, the similarities of all acceleration sensors are relatively high. So during the clustering process, normal acceleration sensors will be merged, while abnormal acceleration sensors will not be merged because their similarities with other acceleration sensors are relatively low. Therefore, the abnormal acceleration sensors are screened based on this.

[0101] Specifically, the sensor categories with two acceleration sensors in the finally obtained several sensor categories are recorded as abnormal sensor categories; all acceleration sensors corresponding to all abnormal sensor categories are recorded as abnormal acceleration sensors; during the symmetrical demolition of the highway T-shaped steel structure bridge, when abnormal acceleration sensors appear, the abnormal acceleration sensors are alarmed and processed.

[0102] It should be noted that in this embodiment, the exp(-MX) model is used to present the inverse proportional relationship and normalization process. MX is the input of the model, and the implementer can set the inverse proportional function and normalization function according to the actual situation.

[0103] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A data monitoring method for the symmetrical demolition process of a highway T-shaped rigid frame bridge, characterized in that: The method comprises the following steps: Arrange a number of acceleration sensors on the bridge according to the positions of the bridge piers, and use the acceleration sensors to obtain a frequency amplitude sequence of each acceleration sensor; the frequency amplitude sequence includes the amplitude of the acceleration sensor at each frequency; According to the amplitude and frequency value of the frequency in the frequency amplitude sequence of the acceleration sensor, the availability of each frequency in the frequency amplitude sequence of each acceleration sensor is obtained; all acceleration sensors are divided into several sensor categories; according to the availability of each frequency in the frequency amplitude sequence of the acceleration sensors in the sensor category, the overall availability of each sensor category at each frequency is obtained; According to the similarity relationship of the amplitudes of the frequencies in the frequency amplitude sequences of all acceleration sensors in the sensor category and the overall availability of the sensor category at each frequency, the demolition symmetry of each sensor category is obtained; According to the frequency values ​​and their corresponding amplitudes in the frequency amplitude sequence of the acceleration sensor, the overall influence of the sensor interval on the similarity is obtained; according to the position of the acceleration sensor in the sensor category, several groups of adjacent categories are obtained, and according to the removal symmetry of the sensor categories in each group of adjacent categories and the overall influence of the sensor interval on the similarity, the adjacent similarity of each group of adjacent categories is obtained; according to the adjacent similarity of each group of adjacent categories, the similarity of every two sensor categories is obtained; According to the similarity of all sensor categories, the abnormal acceleration sensors are obtained, and then the symmetrical demolition monitoring results of the highway T-shaped rigid frame bridge are obtained.

2. The data monitoring method for the symmetrical demolition process of a highway T-shaped rigid frame bridge according to claim 1 is characterized in that: The availability of each frequency in the frequency amplitude sequence of each acceleration sensor is obtained according to the amplitude and frequency value of the frequency in the frequency amplitude sequence of the acceleration sensor. The specific acquisition method is: Where Y i,h is the availability of the hth frequency in the frequency amplitude sequence of the i-th acceleration sensor; F i,h is the amplitude of the hth frequency in the frequency amplitude sequence of the i-th acceleration sensor; F i-1,h is the amplitude of the hth frequency in the frequency amplitude sequence of the i-1th acceleration sensor; F i+1,h is the amplitude of the hth frequency in the frequency amplitude sequence of the i+1th acceleration sensor; P i,h is the frequency value of the hth frequency in the frequency amplitude sequence of the i-th acceleration sensor; norm() is the linear normalization function; || is the absolute value function.

3. The data monitoring method for the symmetrical demolition process of a highway T-shaped rigid frame bridge according to claim 1 is characterized in that: The specific method of classifying all acceleration sensors into several sensor categories is as follows: Each pair of acceleration sensors that are symmetrical about the bridge pier position is recorded as a sensor category.

4. The data monitoring method for the symmetrical demolition process of a highway T-shaped rigid frame bridge according to claim 1 is characterized in that: The method of obtaining the overall availability of each sensor category at each frequency according to the availability of each frequency in the frequency amplitude sequence of the acceleration sensor in the sensor category includes the following specific methods: For any sensor category and any frequency, the minimum value of the availability of the frequency in the frequency amplitude sequence of the two acceleration sensors in the sensor category is recorded as the availability index of the sensor category at the frequency; The inverse proportional normalization result of the absolute value of the difference in the availability of the frequency in the frequency amplitude sequences of two acceleration sensors within the sensor category is recorded as the reliability of the sensor category at the frequency; The product of the reliability of the sensor type at the frequency and the availability index is recorded as the overall availability of the sensor type at the frequency.

5. The data monitoring method for the symmetrical demolition process of a highway T-shaped rigid frame bridge according to claim 1 is characterized in that: According to the similarity relationship of the frequency amplitudes in the frequency amplitude sequences of all acceleration sensors in the sensor category and the overall availability of the sensor category at each frequency, the removal symmetry of each sensor category is obtained. The specific acquisition method is: Where D q is the removal symmetry of the qth sensor category; W is the number of frequencies in the frequency amplitude sequence of each acceleration sensor; K q ′ ,h is the overall availability of the qth sensor category at the hth frequency; F q ′ ,1,h is the amplitude of the hth frequency in the frequency amplitude sequence of the first acceleration sensor in the qth sensor category; F q ′ ,2,h is the amplitude of the hth frequency in the frequency amplitude sequence of the second acceleration sensor in the qth sensor category; ε is a hyperparameter; || is the absolute value function; softmax() is the weight normalization function.

6. The data monitoring method for the symmetrical demolition process of a highway T-shaped rigid frame bridge according to claim 1 is characterized in that: The method of obtaining the overall influence of the sensor interval on the similarity according to the frequency values ​​and their corresponding amplitudes in the frequency amplitude sequence of the acceleration sensor includes the following specific methods: The left-side influence of sensor interval on similarity is calculated as: Where X is the left-side influence of sensor spacing on similarity; W is the number of frequencies in the frequency amplitude sequence of each acceleration sensor; P h ′ is the frequency value of the hth frequency in the frequency amplitude sequence of the acceleration sensor; N is the number of acceleration sensors on the left side of the pier; F c ′,′ h is the amplitude of the hth frequency in the frequency amplitude sequence of the cth acceleration sensor on the left side of the pier; μ h is the mean value of the amplitude of the hth frequency in the frequency amplitude sequence of all acceleration sensors on the left side of the pier; Get the right-side influence of sensor interval on similarity; The sum of the left-side influence and the right-side influence of the sensor interval on the similarity is recorded as the overall influence of the sensor interval on the similarity.

7. The data monitoring method for the symmetrical demolition process of a highway T-shaped rigid frame bridge according to claim 1 is characterized in that: The method of obtaining several groups of adjacent categories according to the positions of acceleration sensors within the sensor category, and obtaining the adjacent similarity of each group of adjacent categories according to the removal symmetry of the sensor category in each group of adjacent categories and the overall influence of the sensor interval on the similarity, includes the following specific methods: All sensor categories are numbered in order from small to large according to the Euclidean distance between the acceleration sensor and the bridge pier in each sensor category; Any two sensor categories with adjacent serial numbers are recorded as a group of adjacent categories; For any group of adjacent categories, the inverse proportional normalization result of the absolute value of the difference between the demolition symmetry of two sensor categories in the group of adjacent categories is recorded as the basic similarity of the group of adjacent categories; The product of the inversely proportional normalized result of the overall impact of the sensor interval on the similarity and the basic similarity of the group of adjacent categories is recorded as the adjacent similarity of the group of adjacent categories.

8. The data monitoring method for the symmetrical dismantling process of a highway T-shaped rigid frame bridge according to claim 7 is characterized in that: The method of obtaining the similarity between each two sensor categories according to the adjacent similarity of each group of adjacent categories includes: For any two sensor categories, in the process from the serial number of any sensor category of the two sensor categories to the serial number of the other sensor category, the adjacent similarity between the sensor category corresponding to each serial number and the next sensor category is calculated in turn, and the product of the adjacent similarities between the sensor category corresponding to all the serial numbers in the process and the next sensor category is recorded as the similarity between the two sensor categories.

9. The data monitoring method for the symmetrical demolition process of a highway T-shaped rigid frame bridge according to claim 1 is characterized in that: The specific method of obtaining abnormal acceleration sensors according to the similarity of all sensor categories is as follows: Step 1) Among all sensor categories, the two sensor categories with the highest similarity are merged as a new sensor category; Step 2) The average of the similarities between the two sensor categories corresponding to the new sensor category before merging and another sensor category other than the two corresponding sensor categories is recorded as the similarity between the new sensor category and the another sensor category; Step 3) obtaining a similarity threshold according to the similarity of the two sensor categories merged for the first time; Step 4) Repeat steps 1 and 2 until the similarities between all sensor categories are less than the similarity threshold, then stop repeating; The specific method for obtaining the similarity threshold is: Z=(1-β)×S Where Z is the similarity threshold; β is the preset similarity fluctuation index; S is the similarity of the two sensor categories merged for the first time; Among the finally obtained sensor categories, a sensor category including two acceleration sensors is recorded as an abnormal sensor category; and all acceleration sensors corresponding to all abnormal sensor categories are recorded as abnormal acceleration sensors.

10. The data monitoring method for the symmetrical demolition process of a highway T-shaped rigid frame bridge according to claim 1 is characterized in that: The method of using the acceleration sensor to obtain the frequency amplitude sequence of each acceleration sensor includes the following specific methods: During the demolition process of the highway T-shaped rigid frame bridge, acceleration sensors are used to record acceleration values, and the time series consisting of the acceleration values ​​collected by each acceleration sensor is recorded as the acceleration monitoring sequence of each acceleration sensor; Performing Fourier transform on the acceleration monitoring sequence of each acceleration sensor to obtain the amplitude spectrum of each acceleration sensor; The amplitude of each integer frequency in the amplitude spectrum of any acceleration sensor is normalized according to the maximum amplitude of all integer frequencies in the amplitude spectrum of all acceleration sensors, and the sequence formed by arranging the integer frequencies from small to large is recorded as the frequency amplitude sequence of the acceleration sensor.

Images