Array sensor-based station building structure vibration source positioning method and device

Abstract

The application discloses a station building structure vibration source positioning method and device based on an array sensor, and relates to the technical field of structure health monitoring. The method comprises the following steps: collecting vibration signals of multiple station building structure points when multiple trains pass through a station building area through an array sensor, and obtaining multiple vibration signal sets; obtaining station building characteristics of the station building area, and indexing multiple positive sample vibration signal sets of a same-family station building; obtaining multiple train characteristics of the multiple trains, combining multiple positive sample train characteristics, station building characteristics and same-family station building characteristics, configuring vibration signal fitting conditions through intelligent sensor fitting configuration rules; and performing iterative fitting in the multiple vibration signal sets by using the multiple positive sample vibration signal sets according to the vibration signal fitting conditions, obtaining multiple vibration fitting confidence coefficients, and screening to obtain a vibration source positioning result. The application realizes rapid and accurate identification and early warning of abnormal vibration of a station building structure.

Description

Technical Field

[0001] This invention relates to the field of structural health monitoring technology, specifically to a method and apparatus for locating vibration sources in station structures based on array sensors. Background Technology

[0002] With the rapid development of my country's high-speed railway network, train operation density and speed are continuously increasing. The environmental vibrations caused by train operation have a long-term impact on the structural safety and service performance of large transportation hubs such as railway stations. Existing technologies mainly conduct static or offline safety assessments of station structures through regular manual inspections, structural health monitoring systems with limited measurement points, or simulation analysis based on a single vibration source model.

[0003] However, when the internal structure of a station building develops localized defects due to aging, damage, or loose connections, passing trains can induce abnormal secondary vibrations, posing hidden safety risks. Traditional identification methods, primarily based on structural inspection, struggle to accurately pinpoint the specific source of damage or abnormal vibration from complex vibration signals. This hinders rapid and precise location of vibration sources, failing to meet the urgent needs of modern large-scale station buildings for real-time, intelligent, and safe operation and maintenance. Summary of the Invention

[0004] This invention provides a method and device for locating vibration sources in station structures based on array sensors, aiming to solve the technical problem that existing technologies cannot achieve rapid and accurate location of vibration sources.

[0005] In view of the above problems, the present invention provides a method and apparatus for locating vibration sources in station structures based on array sensors.

[0006] In a first aspect, the present invention provides a method for locating vibration sources in station structures based on array sensors, comprising: By using array sensors, vibration signals from multiple station building structural points are collected when multiple trains pass through the station building area, resulting in multiple vibration signal sets. The station building features of the station building area are obtained, and multiple positive sample vibration signal sets of the same family of stations are indexed, wherein the multiple positive sample vibration signal sets are obtained by testing when a train with multiple positive sample train features passes by. Multiple train features of multiple trains are acquired, and combined with multiple positive sample train features, station building features, and features of stations in the same family, vibration signal fitting conditions are configured through intelligent sensing fitting configuration rules. The vibration signal fitting conditions include the number of fitting iterations and the fitting error threshold. According to the vibration signal fitting conditions, the multiple positive sample vibration signal sets are used to perform iterative fitting within the multiple vibration signal sets to obtain multiple vibration fitting confidence coefficients, and the vibration source location results are obtained by screening.

[0007] Secondly, the present invention provides a vibration source location device for station structures based on array sensors, comprising: The vibration signal acquisition module is used to acquire vibration signals from multiple station building structure points when multiple trains pass through the station building area through an array of sensors, and obtain multiple vibration signal sets. The positive sample vibration signal index module is used to obtain the station building characteristics of the station building area and index multiple positive sample vibration signal sets of the same family of station buildings. The multiple positive sample vibration signal sets are obtained by testing when a train with multiple positive sample train characteristics passes by. The fitting condition configuration module is used to acquire multiple train features of multiple trains, combine multiple positive sample train features, station building features and features of the same group of stations, and configure vibration signal fitting conditions through intelligent sensing fitting configuration rules. The vibration signal fitting conditions include the number of fitting iterations and the fitting error threshold. The fitting and localization filtering module is used to perform iterative fitting within the multiple positive sample vibration signal sets according to the vibration signal fitting conditions, obtain multiple vibration fitting confidence coefficients, and filter to obtain vibration source localization results.

[0008] One or more technical solutions provided in this invention have at least the following technical effects or advantages: This invention provides a method and device for locating vibration sources in station structures based on array sensors. It utilizes array sensors to comprehensively collect vibration signals from multiple station structure points, providing accurate data support for vibration source location. By indexing positive sample signals from similar station buildings, it solidifies the fitting foundation, and by intelligently configuring fitting conditions based on train and station characteristics, it ensures the scientific validity of the fitting. Finally, through iterative fitting and confidence coefficient screening, it efficiently locates the vibration source. This not only solves the pain points of traditional building structure detection methods, such as low efficiency and inability to provide real-time dynamic positioning, but also accurately identifies secondary vibration risk sources caused by train-induced structural damage. This provides timely and accurate decision-making basis for the safe operation and maintenance of station structures, improving the safety assurance capabilities and operational efficiency of station structures. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A flowchart illustrating the method for locating vibration sources in a station building structure based on an array sensor, provided in an embodiment of the present invention. Figure 2A schematic diagram of the structure of the station building vibration source locating device based on array sensors provided in an embodiment of the present invention; The components represented by each number in the attached diagram are explained below: Vibration signal acquisition module 11, positive sample vibration signal index module 12, fitting condition configuration module 13, fitting positioning screening module 14. Detailed Implementation

[0011] This invention provides a method and apparatus for locating vibration sources in station structures based on array sensors, which addresses the technical problem that existing technologies struggle to achieve rapid and accurate location of vibration sources.

[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0013] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0014] Example 1, as Figure 1 As shown, this invention provides a method for locating vibration sources in station structures based on array sensors. The method includes: S100: By using array sensors, it collects vibration signals from multiple station building structural points when multiple trains pass through the station building area, and obtains multiple vibration signal sets.

[0015] In this embodiment of the invention, an array of sensors is used to collect vibration signals from multiple structural points of the station building as multiple trains pass through the station area, obtaining multiple vibration signal sets. When a train passes through the station area, the wheel-rail force is transmitted to the station foundation through the track, causing vibration responses in key structural points such as beams and columns. Damaged areas may also experience secondary vibrations due to train vibrations. If only a single-point sensor is used to collect data, it can only reflect the local vibration state and cannot capture the vibration differences between multiple structural points, making it difficult to support the accurate location of subsequent vibration sources. Therefore, it is necessary to use an array of sensors to collect vibration signals from multiple structural points of the station building as multiple trains pass through and aggregate them into signal sets to provide comprehensive and continuous raw data for subsequent steps.

[0016] Step S100 in the method provided in this embodiment of the invention includes: Vibration signals from multiple station building structural points are collected using an array of sensors as multiple trains pass through the station building area. The array of sensors includes vibration sensors installed at multiple station building structural points. The collected vibration signals are combined to obtain multiple vibration signal sets.

[0017] First, an array sensor is used to collect vibration signals from multiple structural points of the station building as multiple trains pass through the station area. The array sensor consists of vibration sensors installed at multiple structural points. The array sensor is a sensor network formed by multiple independent vibration sensors arranged according to a preset rule at different structural points of the station building, enabling synchronous acquisition of vibration signals from multiple locations. Structural points of the station building refer to key load-bearing locations within the station building, such as nodes along the force transmission path like beams, columns, and slabs. Vibration signals are physical signals describing the vibration state of structural points within the station building; the x-axis represents time, and the y-axis represents vibration acceleration, providing a direct reflection of vibration intensity at different times. Multiple key structural points within the station building where vibration signals need to be collected are identified, prioritizing core load-bearing locations such as beams and columns. A vibration sensor is installed at each identified structural point, and all sensors together form an array sensor. When a train passes through the station area, all vibration sensors in the array sensor are activated synchronously, collecting vibration signals within a preset time period and recording the time-vibration acceleration data for each structural point.

[0018] For example, taking the waiting hall of a high-speed railway hub as a scenario, two key structural points of the station building are selected: the mid-span of main beam No. 1 and the top of column No. 2. Vibration sensor A is installed at the mid-span of main beam No. 1, and vibration sensor B is installed at the top of column No. 2. Sensors A and B form an array sensor. Two trains are selected as the data acquisition targets, namely train 1 and train 2. When train 1 passes through the station area at a speed of 300 km / h, sensors A and B are simultaneously activated and continuously collected for 10 seconds, obtaining the vibration signal at the mid-span of main beam No. 1: x-axis time range 0-10s, y-axis vibration acceleration range 0-4.8. Vibration signal at the top of column No. 2: x-axis time range 0-10s, y-axis vibration acceleration range 0-3.5. When train 2 passes through the station area at a speed of 280 km / h, vibration signals at the mid-span of main beam 1 are obtained by synchronously collecting data for 10 seconds using the same array of sensors: x-axis time range 0-10s, y-axis vibration acceleration range 0-4.5. Vibration signal at the top of column No. 2: x-axis time range 0-10s, y-axis vibration acceleration range 0-3.2. .

[0019] Secondly, the collected vibration signals are aggregated to obtain multiple vibration signal sets. A vibration signal set is a group of signals formed by classifying and aggregating vibration signals from multiple station structure points when the same train passes, or vibration signals from the same structure point when multiple trains pass, according to a preset dimension. The aggregation dimension of the vibration signals can be determined by selecting train type or station structure point as the aggregation basis; all collected vibration signals are sorted and classified according to the determined dimension; all vibration signals under the same dimension are integrated to generate multiple vibration signal sets, and each signal set is labeled with the corresponding dimension attribute.

[0020] For example, the four vibration signals collected above are aggregated: the signals from main beam 1 and column 2 when train 1 passes, and the signals from main beam 1 and column 2 when train 2 passes. Aggregated by train type: the two vibration signals from the mid-span of main beam 1 and the top of column 2 when train 1 passes are integrated to form the train 1-station building structure vibration signal set; the two vibration signals from the mid-span of main beam 1 and the top of column 2 when train 2 passes are integrated to form the train 2-station building structure vibration signal set.

[0021] In this embodiment of the invention, an array of sensors enables the synchronous and comprehensive acquisition of vibration signals from multiple key structural points such as beams and columns of the station building when multiple trains pass by. The acquired vibration signals contain complete time-vibration acceleration response information, which compensates for the one-sidedness of data acquired by single-point sensors. At the same time, by aggregating vibration signals by dimension, multiple vibration signal sets with clear structures are generated. This provides a rich, accurate, and directly callable original data foundation for subsequent steps such as indexing positive sample vibration signal sets of the same station building, configuring vibration signal fitting conditions, and iterative fitting and positioning, ensuring the continuity and accuracy of subsequent vibration source positioning work.

[0022] S200: Obtain the station building characteristics of the station building area, and index multiple positive sample vibration signal sets of the same family of station buildings, wherein the multiple positive sample vibration signal sets are obtained by testing when a train with multiple positive sample train characteristics passes by.

[0023] In this embodiment of the invention, station building characteristics of the station area are acquired, and multiple positive sample vibration signal sets of station buildings in the same family are indexed. These multiple positive sample vibration signal sets are obtained through testing when trains with multiple positive sample train characteristics pass by. Different station buildings have different structural dimensions and distances from railway lines, resulting in different vibration response characteristics when trains pass by. Directly using vibration signals from station buildings with significant structural or locational differences as references would significantly reduce the accuracy of subsequent vibration source localization. Therefore, it is necessary to first acquire station building characteristics, select the most similar station buildings in the same family from the sample station buildings that have passed structural inspection, and then retrieve the qualified vibration signals collected from those station buildings as positive samples, providing a highly matched reference benchmark for subsequent vibration signal fitting.

[0024] Step S200 in the method provided in this embodiment of the invention includes: Obtain the station building characteristics of the station building area; Based on the station building characteristics, the same family of stations is obtained by indexing within the set of sample stations that have passed structural inspection; Multiple positive sample vibration signal sets of multiple station building structural points are obtained when multiple trains pass by the same station building through array sensor testing. Among them, multiple trains have multiple positive sample train characteristics.

[0025] First, obtain the station building features of the station area. Station building features are core parameters describing the structural and locational attributes of the target station building, including dimensional features and track distance features. Dimensional features refer to the geometric parameters of the key structural elements of the station building, while track distance features refer to the horizontal distance from the key structural points of the station building to the railway line. Determine the key dimensional features of the station building, measure the dimensional parameters of the corresponding structures of the target station building, measure the horizontal distance from the key structural points to the railway line, and organize these parameters to form a complete station building feature set.

[0026] For example, regarding the waiting hall of the target high-speed rail hub station: Dimensional characteristics: The span of the main beam at the mid-span position is 30m, and the height of the column at the top position of column 2 is 12m; Track distance characteristics: The horizontal distance from the mid-span of the main beam to the railway line is 25m, and the horizontal distance from the top of column 2 to the railway line is 25m; The station building characteristics obtained by sorting are: Dimensional characteristics: Main beam span 30m, column height 12m; Track distance characteristics: 25m, 25m.

[0027] Secondly, based on the station building characteristics, the same family of stations is obtained by indexing within the set of sample stations that have passed structural inspection.

[0028] Among them, based on the station building characteristics, the same family of stations is obtained by indexing within the set of sample stations that have passed structural inspection, including: Based on the station building characteristics, a station building feature vector is constructed, wherein the station building characteristics include size characteristics and line distance characteristics; Obtain a set of sample station buildings with the same architectural structure as the station buildings; Obtain the sample station feature set of the sample station set, and construct the sample station feature vector set; Calculate the similarity between the station building feature vector and each sample station building feature vector, and select the sample station building corresponding to the sample station building feature vector with the highest similarity as the same family of stations.

[0029] First, based on the station building characteristics, a station building feature vector is constructed. These characteristics include dimensional features and track distance features. The station building feature vector is a numerical vector formed by arranging the quantified parameters of the station building characteristics in a fixed order, used to quantify the differences in station building characteristics. The dimensional and track distance features are converted into numerical forms, and the numerical parameters are arranged into a vector according to the order of main beam span, column height, and distance from the structural point to the track. For example, based on the characteristics of the target station building, the quantified parameters are arranged as follows: main beam span 30m, column height 12m, distance from structural point to track 25m, resulting in a station building feature vector of [30, 12, 25].

[0030] Secondly, a sample set of station buildings with the same architectural structure as the target station building is obtained. This sample set refers to a database of station buildings that have completed structural safety inspections and whose architectural structure type matches the target station building. For example, if the target station building is a reinforced concrete frame structure, then all stations in the set are qualified stations of the same structural type. The architectural structure type of the target station building is determined, and station buildings with the same structural type and that have passed inspections are selected from the station building structural safety inspection database to form the sample set. For example, if the target station building is a reinforced concrete frame structure, two qualified sample station buildings with the same structure are selected from the database: Sample Station 1: Waiting hall of an intercity railway station; Sample Station 2: Ticket hall of a high-speed railway station.

[0031] Next, the feature set of the sample station buildings is obtained, and a feature vector set of the sample station buildings is constructed. The feature vector set of the sample station buildings refers to the set formed by converting the features of each station building in the sample station building set into numerical vectors in the same order as the feature vectors of the target station buildings. The size features and track distance features of each station building in the sample station building set are retrieved; according to the arrangement order of the feature vectors of the target station buildings, the features of each sample station building are converted into numerical vectors, and all the feature vectors of the sample station buildings are integrated to form the feature set of the sample station buildings.

[0032] For example, the features of sample station 1 and sample station 2 are retrieved to construct the feature vectors of the sample station: Features of sample station 1: main beam span 28m, column height 12m, distance from structural point to line 24m, feature vector [28,12,24]; Features of sample station 2: main beam span 32m, column height 11m, distance from structural point to line 26m, feature vector [32,11,26]; forming the feature vector set of sample station: {[28,12,24],[32,11,26]}.

[0033] Subsequently, the similarity between the station building feature vector and the feature vector of each sample station building is calculated. The sample station building with the highest similarity feature vector is selected as a family member station. Vector similarity is an indicator used to measure the degree of difference between two feature vectors. For example, cosine similarity is used, with a value ranging from 0 to 1. The closer the value is to 1, the more similar the features are. Cosine similarity is used to calculate the similarity between the target station building feature vector and the feature vector of each sample station building. The sample station building with the highest similarity value is selected and defined as a family member station. For example, using cosine similarity, the similarity between the target station building feature vector [30,12,25] and the feature vector of sample station building 1 [28,12,24] is 0.98, and the similarity between the target station building feature vector and the feature vector of sample station building 2 [32,11,26] is 0.95. The sample station building with the highest similarity is selected as a family member station of the target station building.

[0034] Furthermore, multiple positive sample vibration signal sets are obtained for multiple structural points of the same type of station building as multiple trains pass by, tested by array sensors. These multiple trains exhibit multiple positive sample train characteristics. The positive sample vibration signal set refers to the set of vibration signals collected by array sensors when multiple trains pass by the same type of station building under structurally acceptable conditions; that is, standard vibration signals without damage or abnormalities. Array sensors are deployed within the same type of station building, corresponding to the structural points of the target station building. Vibration signals collected when multiple trains pass by this station building are retrieved and integrated according to the dimensions of structural points + trains to form the positive sample vibration signal set.

[0035] For example, taking station building 1 (a sample station building of the same type) as the scenario, two key structural points of the same type of station building are selected: the mid-span of main beam 1 and the top of column 2. Vibration sensor A is installed at the mid-span of main beam 1, and vibration sensor B is installed at the top of column 2. Sensors A and B form an array sensor. Train 1 and train 2 are selected as test objects. When train 1 passes through the station building area at a speed of 300 km / h, sensors A and B are simultaneously activated and continuously collected for 10 seconds, obtaining the positive sample vibration signal at the mid-span of main beam 1: x-axis time range 0-10s, y-axis vibration acceleration range 0-4.8. Vibration signal of the positive sample at the top of column No. 2: x-axis time range 0-10s, y-axis vibration acceleration range 0-3.5. The two sets together form positive sample vibration signal set 1. When train 2 passes through the station area at a speed of 280 km / h, positive sample vibration signals at the mid-span of main beam 1 are obtained by synchronously collecting data for 10 seconds using the same array of sensors: x-axis time range 0-10s, y-axis vibration acceleration range 0-4.5. Vibration signal of the positive sample at the top of column No. 2: x-axis time range 0-10s, y-axis vibration acceleration range 0-3.2. The two sets together form a positive sample vibration signal set 2, and finally two positive sample vibration signal sets are obtained.

[0036] In this embodiment of the invention, by acquiring the characteristics of the target station and filtering out stations of the same family, it is ensured that the positive sample vibration signal set is highly matched with the structural and positional characteristics of the target station. At the same time, the positive sample signal set is the vibration data of stations of the same family under qualified conditions, avoiding the error interference caused by station samples with different structures, providing a reliable reference benchmark for subsequent vibration signal fitting, and improving the accuracy of subsequent vibration source positioning.

[0037] S300: Acquire multiple train features from multiple trains, combine multiple positive sample train features, station building features, and features of stations from the same family, and configure vibration signal fitting conditions through intelligent sensing fitting configuration rules. The vibration signal fitting conditions include the number of fitting iterations and the fitting error threshold.

[0038] In this embodiment of the invention, multiple train features from multiple trains are acquired. These features are combined with features from multiple positive sample trains, station buildings, and similar station buildings. Vibration signal fitting conditions are configured using intelligent sensing fitting rules. These fitting conditions include the number of fitting iterations and a fitting error threshold. Train operating parameters directly affect the vibration intensity of the station building structure. When there are differences in features between the target station building and similar stations, or between the target train and positive sample trains, fixed fitting conditions can lead to decreased vibration signal fitting accuracy. For example, when the differences are large, a fixed number of iterations may result in insufficient fitting. Therefore, it is necessary to dynamically configure fitting conditions based on the similarity between train features, station building features, and corresponding positive samples: the smaller the similarity and the larger the difference, the more iterations are increased to improve fitting sufficiency, and the error threshold is relaxed to adapt to the differences, thereby ensuring the accuracy of subsequent vibration signal fitting.

[0039] Step S300 in the method provided in this embodiment of the invention includes: Obtain multiple train features from multiple trains; Calculate the similarity between multiple train features and multiple positive sample train features to obtain the similarity of multiple train features; Obtain the similarity between the station building features and the features of stations in the same family to obtain the station building feature similarity; Based on the similarity of train features and station features, vibration signal fitting conditions are configured using intelligent sensing fitting configuration rules. The vibration signal fitting conditions include the number of fitting iterations and a fitting error threshold.

[0040] First, acquire multiple train characteristics from multiple trains. Train characteristics are core parameters describing the train's operating status and load attributes, and are key factors affecting the station building's vibration response. These typically include train speed and total load. Identify train characteristics strongly correlated with station building vibration, such as speed and load. Obtain the characteristic parameters corresponding to the target train through the railway dispatching system or train monitoring equipment. For example, for two trains collecting vibration signals at the target station building: Train 1's characteristics: operating speed 300 km / h, total load 800t; Train 2's characteristics: operating speed 280 km / h, total load 750t.

[0041] Secondly, the similarity between multiple train features and multiple positive sample train features is calculated to obtain the similarity of multiple train features. Positive sample train features refer to the train features corresponding to the positive sample vibration signals collected from stations within the same family of stations; that is, the train parameters that passed through the same family of stations during detection. Train feature similarity is an index that measures the degree of difference between the target train features and the positive sample train features. Cosine similarity is used, with a value ranging from 0 to 1; the closer the value is to 1, the more consistent the features. The positive sample train features corresponding to the positive sample vibration signals from the same family of stations are retrieved, and the target train features and positive sample train features are converted into numerical vectors respectively. The similarity between the two is then calculated using the cosine similarity formula.

[0042] For example, the positive sample train features of sample station 1 are: operating speed 290km / h, total load 780t, corresponding to feature vector [290,780]. The similarity between the feature vector [300,800] of train 1 and the positive sample vector is 0.96; the similarity between the feature vector [280,750] of train 2 and the positive sample vector is 0.93; the train feature similarity is obtained as follows: train 1 feature similarity 0.96, train 2 feature similarity 0.93.

[0043] Further, the similarity between the station building features and the features of stations in the same family is obtained to obtain the station building feature similarity. The station building feature similarity refers to the similarity between the target station building feature vector calculated in S200 and the feature vector of stations in the same family. The station building feature similarity result calculated in S200 is directly retrieved. For example, the station building feature similarity between the target station building and sample station building 1 is 0.98.

[0044] Then, based on the similarity of train features and station features, vibration signal fitting conditions are configured through intelligent sensing fitting configuration rules, wherein the vibration signal fitting conditions include the number of fitting iterations and the fitting error threshold.

[0045] Specifically, based on the train feature similarity and station building feature similarity, vibration signal fitting conditions are configured using intelligent sensing fitting configuration rules, including: Get the preset number of iterations and the preset error threshold; Based on the similarity of the train features, calculate the first number adjustment coefficient and the first error adjustment coefficient; based on the similarity of the station building features, calculate the second number adjustment coefficient and the second error adjustment coefficient. Based on the first number adjustment coefficient, the first error adjustment coefficient, the second number adjustment coefficient, and the second error adjustment coefficient, the preset number of iterations and the preset error threshold are adjusted and calculated to obtain the fitting iteration number and the fitting error threshold, which are used as the fitting conditions for the vibration signal.

[0046] First, obtain the preset number of iterations and the preset error threshold. The preset number of iterations refers to the default base number of iterations, which is the initial parameter to ensure fitting convergence. The preset error threshold is the default vibration acceleration error threshold used to determine whether a signal point is an inlier. An inlier is a data point in the target vibration signal whose error between the data point and the fitting model of the positive sample vibration signal is less than the fitting error threshold. Retrieve the default parameter values ​​for the station vibration signal fitting scenario; for example, preset number of iterations: 50; preset error threshold: 0.02. This provides a benchmark for determining the error in vibration acceleration.

[0047] Secondly, based on the similarity of the train features, the first number adjustment coefficient and the first error adjustment coefficient are calculated, and based on the similarity of the station building features, the second number adjustment coefficient and the second error adjustment coefficient are calculated.

[0048] Specifically, based on the train feature similarity, a first number adjustment coefficient and a first error adjustment coefficient are calculated; based on the station building feature similarity, a second number adjustment coefficient and a second error adjustment coefficient are calculated, including: Calculate the reciprocal of the train feature similarity, and use it as the first number adjustment coefficient and the first error adjustment coefficient; The reciprocal of the similarity of the station building features is calculated and used as the second number adjustment coefficient and the second error adjustment coefficient.

[0049] First, the reciprocal of the train feature similarity is calculated and used as both the first number adjustment coefficient and the first error adjustment coefficient. The first number adjustment coefficient, calculated based on the train feature similarity, is used to correct for a preset iteration count. The first number adjustment coefficient is equal to the reciprocal of the train feature similarity; the smaller the similarity, the greater the train difference, and the larger the first number adjustment coefficient. The first error adjustment coefficient, also calculated based on the train feature similarity, is used to correct for a preset error threshold. The first error adjustment coefficient is identical to the first number adjustment coefficient, ensuring that the adjustment range of the iteration count matches the error threshold. The feature similarity between the target train and the positive sample trains is retrieved, and the reciprocal of this similarity is calculated. This reciprocal is used as both the first number adjustment coefficient and the first error adjustment coefficient.

[0050] For example, the feature similarity between train 1 and the positive sample train is 0.96. Calculating the reciprocal: 1 / 0.96 ≈ 1.04, therefore the first adjustment coefficient for train 1 is 1.04, and the first error adjustment coefficient is also 1.04. The feature similarity between train 2 and the positive sample train is 0.93. Calculating the reciprocal: 1 / 0.93 ≈ 1.10, therefore the first adjustment coefficient for train 2 is 1.10, and the first error adjustment coefficient is also 1.10. Train 2 has a greater difference from the positive sample train, and therefore a larger adjustment coefficient.

[0051] Secondly, the reciprocal of the station building feature similarity is calculated and used as the second iteration adjustment coefficient and the second error adjustment coefficient. The second iteration adjustment coefficient is a coefficient calculated based on the station building feature similarity and is used to assist in correcting the preset iteration number. The second iteration adjustment coefficient is equal to the reciprocal of the station building feature similarity; the smaller the similarity, the greater the difference between the station buildings, and the larger the second iteration adjustment coefficient. The second error adjustment coefficient is a coefficient calculated based on the station building feature similarity and is used to assist in correcting the preset error threshold. The second error adjustment coefficient is exactly the same as the second iteration adjustment coefficient, ensuring that the adjustment logic of the iteration number and the error threshold is consistent. The feature similarity between the target station building and other stations in the same family is retrieved, and the reciprocal of this similarity is calculated. The calculated reciprocal result is used as both the second iteration adjustment coefficient and the second error adjustment coefficient. For example, if the feature similarity between the target station building and sample station building 1 is 0.98, the reciprocal is calculated as: 1 / 0.98 ≈ 1.02. Therefore, the second iteration adjustment coefficient for the target station building is 1.02, and the second error adjustment coefficient is also 1.02. This coefficient is a fixed value and does not change with the train type; it is only used to reflect the difference between the station building itself and other stations in the same family.

[0052] Further, based on the first number adjustment coefficient, the first error adjustment coefficient, the second number adjustment coefficient, and the second error adjustment coefficient, the preset iteration number and preset error threshold are adjusted and calculated to obtain the fitting iteration number and fitting error threshold, which serve as the vibration signal fitting conditions. The fitting iteration number refers to the number of iterations after adapting to the characteristics of the target train and station. The fitting error threshold refers to the vibration acceleration error judgment threshold after adapting to the characteristic differences. The mean of the first number adjustment coefficient and the second number adjustment coefficient is calculated, multiplied by the preset iteration number, and rounded to obtain the fitting iteration number; the mean of the first error adjustment coefficient and the second error adjustment coefficient is calculated, multiplied by the preset error threshold, and the fitting error threshold is obtained.

[0053] For example, the fitting conditions for train 1 are: mean of the number of iterations adjustment coefficient = (1.04 + 1.02) / 2 = 1.03, number of fitting iterations = 50 × 1.03 ≈ 52; mean of the error adjustment coefficient = (1.04 + 1.02) / 2 = 1.03, fitting error threshold = 0.02 × 1.03 ≈ 0.0206 The fitting conditions for Train 2 are as follows: Mean value of the iteration adjustment coefficient = (1.10 + 1.02) / 2 = 1.06, Number of fitting iterations: 50 × 1.06 ≈ 53; Mean value of the error adjustment coefficient = (1.10 + 1.02) / 2 = 1.06, Fitting error threshold: 0.02 × 1.06 ≈ 0.0212 .

[0054] In this embodiment of the invention, by combining the feature similarity between the train, station building, and corresponding positive samples, the number of fitting iterations and the error threshold are dynamically adjusted. The greater the feature difference and the smaller the similarity, the more iterations are needed and the more lenient the error threshold is. This ensures both the sufficiency of the fitting and adapts to the signal fluctuations caused by the differences. The final configured fitting conditions can accurately match the vibration signal characteristics of the target scene, providing a highly adaptable parameter benchmark for subsequent vibration signal iterative fitting and effectively improving the accuracy of the fitting results.

[0055] S400: According to the vibration signal fitting conditions, the multiple positive sample vibration signal sets are used to perform iterative fitting within the multiple vibration signal sets to obtain multiple vibration fitting confidence coefficients, and the vibration source location results are obtained by screening.

[0056] In this embodiment of the invention, according to the vibration signal fitting conditions, multiple positive sample vibration signal sets are used, and iterative fitting is performed within these sets to obtain multiple vibration fitting confidence coefficients. The vibration source location results are then selected. If the target building has structural defects such as corroded columns or cracked main beams, the vibration signal at the defective location will show a significant difference from the positive sample vibration signal without structural defects. However, directly comparing the signal waveforms with the naked eye cannot quantify the degree of difference, nor can it accurately locate the vibration source corresponding to the defect. By iteratively fitting the vibration signals of each structural point of the target building with the positive sample vibration signal without defects according to preset fitting conditions, the signal difference can be quantified into vibration fitting confidence coefficients. The smaller the confidence coefficient, the greater the difference between the signal at that structural point and the positive sample, which may be the location of the defect causing abnormal vibration. This solves the problems of low detection efficiency and poor positioning accuracy in traditional methods, achieving accurate positioning of the vibration source.

[0057] Step S400 in the method provided in this embodiment of the invention includes: Within the plurality of positive sample vibration signal sets, the first positive sample vibration signal set of the first family of station building structure points is selected, and the first positive sample vibration signal is randomly selected. Within multiple vibration signal sets, a first vibration signal set for the first station building structure point is selected, and a first vibration signal is randomly selected, wherein the first family of station building structure points corresponds to the first station building structure point; According to the vibration signal fitting conditions, the first positive sample vibration signal is used to iteratively fit within the first vibration signal to obtain the first fitting confidence coefficient. Continue fitting multiple positive sample vibration signal sets and multiple vibration signal sets to obtain multiple sets of fitting confidence coefficients. Calculate the mean of each set to obtain multiple vibration fitting confidence coefficients. The station building structure point corresponding to the smallest vibration fitting confidence coefficient is selected as the vibration source location result.

[0058] First, within the multiple sets of positive sample vibration signals, the first set of positive sample vibration signals for the first family of station building structural points is selected, and a first positive sample vibration signal is randomly selected. The first family of station building structural points are structural points without architectural defects in the family of station buildings that correspond one-to-one with the location of the first structural point of the target station building. For example, if the first structural point of the target station building is selected as the mid-span of main beam No. 1, the corresponding mid-span of main beams in the family of station buildings is also selected. The first set of positive sample vibration signals refers to the collection of vibration signals from multiple trains passing by the first structural point of the family of station buildings under defect-free conditions, collected by array sensors. The first positive sample vibration signal refers to a single defect-free vibration signal randomly selected from the first set of positive sample vibration signals. To determine the first structural point of the target station building to be analyzed, priority is given to core stress points such as beams and columns. From the set of positive sample vibration signals of the family of station buildings, the set of positive sample signals corresponding to the location of this structural point is retrieved. One positive sample vibration signal is randomly selected from this set as the reference signal for fitting.

[0059] For example, the mid-span of the No. 1 main beam of the target station building is selected as the first station building structural point, and the mid-span of the No. 1 main beam of the corresponding family station building, namely sample station building 1, is also selected as the first family station building structural point; the positive sample vibration signal set of this structural point is retrieved, which includes two defect-free signals when train 1 and train 2 pass by; the positive sample signal corresponding to train 1 is randomly selected as the first positive sample vibration signal.

[0060] Secondly, within multiple vibration signal sets, the first vibration signal set of the first station building structural point is selected, and the first vibration signal is randomly selected. The first family of station building structural points corresponds to the first station building structural point. The first station building structural point is the core structural point to be analyzed in the target station building, and its position completely corresponds to the first family of station building structural points. The first vibration signal set refers to the set of vibration signals collected from the first structural point of the target station building when multiple trains pass by, which may include abnormal signals caused by defects. The first vibration signal refers to a single target vibration signal randomly selected from the first vibration signal set, which is the analysis object to be fitted. The vibration signal set of the first station building structural point of the target station building collected in S100 is retrieved; one vibration signal from this set is randomly extracted as the target signal to be fitted. For example, the vibration signal set of the mid-span of the No. 1 main beam of the target station building is retrieved, containing two signals when trains 1 and 2 pass by; the signal corresponding to train 1 is randomly selected as the first vibration signal.

[0061] Furthermore, according to the vibration signal fitting conditions, the first positive sample vibration signal is used to iteratively fit within the first vibration signal to obtain the first fitting confidence coefficient.

[0062] Specifically, according to the vibration signal fitting conditions, the first positive sample vibration signal is used to iteratively fit within the first vibration signal to obtain the first fitting confidence coefficient, including: Randomly match the time axis of the first positive sample vibration signal with the time axis of the first vibration signal, and count the proportion of signal data points of the first vibration signal whose error with the signal data points of the first positive sample vibration signal is less than the fitting error threshold, and use it as the first iterative fitting confidence coefficient. Continue random matching and fitting until the number of fitting iterations is reached, and output the first fitting confidence coefficient with the largest value during the iterative fitting process to obtain the first fitting confidence coefficient.

[0063] First, the time axis of the first positive sample vibration signal and the time axis of the first vibration signal are randomly matched. The proportion of signal data points of the first vibration signal whose error with the signal data points of the first positive sample vibration signal is less than the fitting error threshold is counted and used as the confidence coefficient for the first iteration fitting. Random time axis matching means randomly aligning the time axes of the first positive sample vibration signal and the first vibration signal. It is not required that the start time and time interval of the two signals be completely consistent, only that the time axis lengths of the two signals are the same after matching, which is used to simulate the phase difference of random vibration response when a train passes through a station. The confidence coefficient for the first iteration fitting refers to the proportion of inlier data points in the first vibration signal after a single random time axis matching. The value ranges from 0 to 1 and is the core indicator for measuring the similarity of a single fitting. The higher the proportion of inlier data points, the higher the similarity between the target signal and the positive sample signal in this matching. Signal data points refer to the corresponding coordinate points of the time axis and vibration acceleration axis in the vibration signal, in the format of (time t, vibration acceleration a), containing complete time-acceleration information. The fitting error threshold configured in S300 is retrieved, and the time axis of the first positive sample vibration signal and the first vibration signal are randomly matched and overlapped to ensure that the time axis lengths of the two signals are consistent after matching. The absolute value of the difference in vibration acceleration between the first vibration signal data point and the first positive sample vibration signal data point at the same time point after matching is calculated point by point. The difference is compared with the fitting error threshold. The difference is less than the threshold and is the inside point. The difference is greater than or equal to the threshold and is the outside point. The number of inside point data points is counted and the proportion of the number of inside points to the total number of data points of the first vibration signal is calculated. This proportion is the confidence coefficient of the first iteration fitting.

[0064] For example, the fitting object is the first vibration signal in the middle of the span of the No. 1 main beam of the target station building, and the first positive sample vibration signal is the defect-free signal in the middle of the span of the No. 1 main beam of the same family of station buildings. The fitting error threshold is 0.0206. The two signal segments each have a total of 1000 data points. The two signals are randomly matched along their time axes, with a random overlap start time of 1.2 seconds. After matching, the time axis length is 8 seconds, containing 800 data points. The acceleration difference is calculated point-by-point: the difference at 780 data points is less than 0.0206. The difference between the 20 data points is greater than or equal to 0.0206. The number of in-circuit points is 780, and the number of out-circuit points is 20. 780 / 800 = 0.975, therefore the confidence coefficient for the first iteration is 0.975. If the object being fitted is the first vibration signal at the top of column 2 of the target station building, and column 2 has a rust defect at its top, under the same positive sample signal and the same fitting error threshold, the number of in-circuit points after a single random matching is 520, and the total number of data points is 800. 520 / 800 = 0.65, then the confidence coefficient for the first iteration is 0.65.

[0065] Next, random matching and fitting continues until the predetermined number of fitting iterations is reached. The highest first-iteration fitting confidence coefficient during the iterative fitting process is then output, yielding the first fitting confidence coefficient. The number of fitting iterations refers to the preset number of iterations configured in S300, representing the total number of random matching and fitting operations required for this fitting. The first fitting confidence coefficient is the highest first-iteration fitting confidence coefficient among multiple iterative fittings, used to comprehensively measure the overall similarity between the first positive sample vibration signal and the first vibration signal. Its value ranges from 0 to 1; the larger the value, the higher the similarity between the two signals. The number of fitting iterations configured in S300 is retrieved, and the random matching and fitting process is repeated. After each iteration, the corresponding first-iteration fitting confidence coefficient is recorded. When the preset number of fitting iterations is reached, the iteration stops. From all recorded first-iteration fitting confidence coefficients, the largest value is selected; this value is the final first fitting confidence coefficient.

[0066] For example, the fitting iteration number corresponding to train 1 is 52: For the first vibration signal at the mid-span of main beam 1: 52 random matching fittings are performed, resulting in 52 first-iteration fitting confidence coefficients, the largest of which is 0.98. Therefore, the first fitting confidence coefficient for this structural point is 0.98. For the first vibration signal at the top of column 2: 52 random matching fittings are performed, resulting in 52 first-iteration fitting confidence coefficients, the largest of which is 0.65. Therefore, the first fitting confidence coefficient for this structural point is 0.65.

[0067] Based on this, we continue to fit multiple sets of positive sample vibration signals and multiple sets of vibration signals to obtain multiple sets of fitted confidence coefficients. The mean of each set is calculated to obtain multiple vibration fitted confidence coefficients. A fitted confidence coefficient set refers to the collection of multiple first-fit confidence coefficients obtained from multiple fittings of the same structural point. The vibration fitted confidence coefficient is the mean of the fitted confidence coefficient set, used to comprehensively evaluate the overall difference between the structural point and the positive sample signals, avoiding the random errors of a single fitting. For each core structural point of the target station, we repeat the above process, fitting all positive sample signals of its corresponding family of station structural points and all its own target signals; generating a unique fitted confidence coefficient set for each structural point; calculating the arithmetic mean of each set to obtain the final vibration fitted confidence coefficient for that structural point.

[0068] For example, batch fitting is performed on two core structural points of the target station building: Mid-span of main beam No. 1: two sets of signals from train 1 and train 2 are fitted, resulting in a set of fitting confidence coefficients {0.98, 0.97}, with a mean of (0.98 + 0.97) / 2 = 0.975, which is the vibration fitting confidence coefficient of this structural point; Top of column No. 2: two sets of signals from train 1 and train 2 are fitted, resulting in a set of fitting confidence coefficients {0.65, 0.63}, with a mean of (0.65 + 0.63) / 2 = 0.64, which is the vibration fitting confidence coefficient of this structural point.

[0069] Finally, the structural point in the station building corresponding to the smallest vibration fitting confidence coefficient is selected as the vibration source location result. The vibration source location result refers to the structural point in the target station building with the smallest vibration fitting confidence coefficient, that is, the location with the largest difference from the defect-free positive sample signal, which is likely to have a building defect, and is also the source of abnormal vibration when the train passes by. The vibration fitting confidence coefficients of all core structural points in the target station building are compiled, and the structural point corresponding to the smallest confidence coefficient is selected as the vibration source.

[0070] For example, the vibration fitting confidence coefficients of each structural point of the target station building are sorted out: mid-span of main beam No. 1: 0.975; top of column No. 2: 0.64; the top of column No. 2 corresponding to the smallest confidence coefficient of 0.64 is selected and determined to be the vibration source. The column has rust defects and is the source of abnormal vibration of the station building when the train passes.

[0071] In this embodiment of the invention, the difference between the defect-free positive sample signal and the target signal is quantified into a vibration fitting confidence coefficient through iterative fitting, thus achieving quantifiable analysis of signal differences. Batch fitting and averaging reduce the random error of single fitting, ultimately selecting the structural point with the smallest confidence coefficient as the vibration source, accurately locating the defect in the station structure that causes abnormal vibrations. Compared to traditional building structure inspection methods, this not only improves the accuracy of vibration source location but also realizes the transformation from manual experience-based judgment to quantitative data-based judgment, improving the efficiency and reliability of vibration source location in station structures.

[0072] Through the specific implementation methods described above, the embodiments of the present invention achieve the following technical effects: This invention provides a method and apparatus for locating vibration sources in station structures based on array sensors. The array sensors comprehensively collect and aggregate vibration signals from multiple key structural points of the target station as multiple trains pass by, providing rich and accurate raw data for subsequent analysis. By acquiring station characteristics and selecting highly matching family of stations, the positive sample vibration signal set of these stations in their defect-free state is retrieved, providing a reliable reference benchmark for fitting analysis. The fitting parameters are dynamically configured by combining the similarity between train and station characteristics and corresponding positive samples, effectively adapting to signal differences in different scenarios and improving the scientific rigor and adaptability of the fitting. Finally, by quantifying the difference between the target signal and the positive sample signal into a vibration fitting confidence coefficient, the structural point with the smallest confidence coefficient is accurately selected as the vibration source, achieving efficient and accurate location of vibration sources from station structure defects. This invention constructs a complete technical system from signal acquisition, benchmark acquisition, parameter configuration to precise location, effectively improving the automation and reliability of station structure vibration source location, and providing strong technical support for the safe operation and maintenance of station structures.

[0073] Example 2, as Figure 2 As shown, the present invention provides a vibration source location device for station buildings based on array sensors, the device comprising: The vibration signal acquisition module 11 is used to acquire vibration signals from multiple station building structure points when multiple trains pass through the station building area through array sensors, and obtain multiple vibration signal sets. Positive sample vibration signal index module 12 is used to obtain the station building characteristics of the station building area and index multiple positive sample vibration signal sets of the same family of station buildings, wherein the multiple positive sample vibration signal sets are obtained by testing when a train with multiple positive sample train characteristics passes by. The fitting condition configuration module 13 is used to acquire multiple train features of multiple trains, combine multiple positive sample train features, station building features and features of the same group of stations, and configure vibration signal fitting conditions through intelligent sensing fitting configuration rules. The vibration signal fitting conditions include the number of fitting iterations and the fitting error threshold. The fitting and localization screening module 14 is used to perform iterative fitting within the multiple positive sample vibration signal sets according to the vibration signal fitting conditions, obtain multiple vibration fitting confidence coefficients, and screen to obtain vibration source localization results.

[0074] In one embodiment, the vibration signal acquisition module 11 is further configured to: Vibration signals from multiple station building structural points are collected using an array of sensors as multiple trains pass through the station building area. The array of sensors includes vibration sensors installed at multiple station building structural points. The collected vibration signals are combined to obtain multiple vibration signal sets.

[0075] In one embodiment, the positive sample vibration signal indexing module 12 is further configured to: Obtain the station building characteristics of the station building area; Based on the station building characteristics, the same family of stations is obtained by indexing within the set of sample stations that have passed structural inspection; Multiple positive sample vibration signal sets of multiple station building structural points are obtained when multiple trains pass by the same station building through array sensor testing. Among them, multiple trains have multiple positive sample train characteristics.

[0076] Among them, based on the station building characteristics, the same family of stations is obtained by indexing within the set of sample stations that have passed structural inspection, including: Based on the station building characteristics, a station building feature vector is constructed, wherein the station building characteristics include size characteristics and line distance characteristics; Obtain a set of sample station buildings with the same architectural structure as the station buildings; Obtain the sample station feature set of the sample station set, and construct the sample station feature vector set; Calculate the similarity between the station building feature vector and each sample station building feature vector, and select the sample station building corresponding to the sample station building feature vector with the highest similarity as the same family of stations.

[0077] In one embodiment, the fitting condition configuration module 13 is further configured to: Obtain multiple train features from multiple trains; Calculate the similarity between multiple train features and multiple positive sample train features to obtain the similarity of multiple train features; Obtain the similarity between the station building features and the features of stations in the same family to obtain the station building feature similarity; Based on the similarity of train features and station features, vibration signal fitting conditions are configured using intelligent sensing fitting configuration rules. The vibration signal fitting conditions include the number of fitting iterations and a fitting error threshold.

[0078] Specifically, based on the train feature similarity and station building feature similarity, vibration signal fitting conditions are configured using intelligent sensing fitting configuration rules, including: Get the preset number of iterations and the preset error threshold; Based on the similarity of the train features, calculate the first number adjustment coefficient and the first error adjustment coefficient; based on the similarity of the station building features, calculate the second number adjustment coefficient and the second error adjustment coefficient. Based on the first number adjustment coefficient, the first error adjustment coefficient, the second number adjustment coefficient, and the second error adjustment coefficient, the preset number of iterations and the preset error threshold are adjusted and calculated to obtain the fitting iteration number and the fitting error threshold, which are used as the fitting conditions for the vibration signal.

[0079] Specifically, based on the train feature similarity, a first number adjustment coefficient and a first error adjustment coefficient are calculated; based on the station building feature similarity, a second number adjustment coefficient and a second error adjustment coefficient are calculated, including: Calculate the reciprocal of the train feature similarity, and use it as the first number adjustment coefficient and the first error adjustment coefficient; The reciprocal of the similarity of the station building features is calculated and used as the second number adjustment coefficient and the second error adjustment coefficient.

[0080] In one embodiment, the fitting and localization screening module 14 is further configured to: Within the plurality of positive sample vibration signal sets, the first positive sample vibration signal set of the first family of station building structure points is selected, and the first positive sample vibration signal is randomly selected. Within multiple vibration signal sets, a first vibration signal set for the first station building structure point is selected, and a first vibration signal is randomly selected, wherein the first family of station building structure points corresponds to the first station building structure point; According to the vibration signal fitting conditions, the first positive sample vibration signal is used to iteratively fit within the first vibration signal to obtain the first fitting confidence coefficient. Continue fitting multiple positive sample vibration signal sets and multiple vibration signal sets to obtain multiple sets of fitting confidence coefficients. Calculate the mean of each set to obtain multiple vibration fitting confidence coefficients. The station building structure point corresponding to the smallest vibration fitting confidence coefficient is selected as the vibration source location result.

[0081] Specifically, according to the vibration signal fitting conditions, the first positive sample vibration signal is used to iteratively fit within the first vibration signal to obtain the first fitting confidence coefficient, including: Randomly match the time axis of the first positive sample vibration signal with the time axis of the first vibration signal, and count the proportion of signal data points of the first vibration signal whose error with the signal data points of the first positive sample vibration signal is less than the fitting error threshold, and use it as the first iterative fitting confidence coefficient. Continue random matching and fitting until the number of fitting iterations is reached, and output the first fitting confidence coefficient with the largest value during the iterative fitting process to obtain the first fitting confidence coefficient.

[0082] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0083] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0084] This specification and accompanying drawings are merely illustrative examples of the invention and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its scope. Therefore, if such modifications and modifications fall within the scope of the invention and its equivalents, the invention is intended to include these modifications and modifications.

Claims

1. A method for locating vibration sources in station structures based on array sensors, characterized in that, The method includes: By using array sensors, vibration signals from multiple station building structural points are collected when multiple trains pass through the station building area, resulting in multiple vibration signal sets. The station building features of the station building area are obtained, and multiple positive sample vibration signal sets of the same family of stations are indexed, wherein the multiple positive sample vibration signal sets are obtained by testing when a train with multiple positive sample train features passes by. Multiple train features of multiple trains are acquired, and combined with multiple positive sample train features, station building features, and features of stations in the same family, vibration signal fitting conditions are configured through intelligent sensing fitting configuration rules. The vibration signal fitting conditions include the number of fitting iterations and the fitting error threshold. According to the vibration signal fitting conditions, the multiple positive sample vibration signal sets are used to perform iterative fitting within the multiple vibration signal sets to obtain multiple vibration fitting confidence coefficients, and the vibration source location results are obtained by screening.

2. The method for locating vibration sources in station structures based on array sensors according to claim 1, characterized in that, Vibration signals from multiple points on the station building structure were collected using an array of sensors as multiple trains passed through the station area, resulting in multiple vibration signal sets, including: Vibration signals from multiple station building structural points are collected using an array of sensors as multiple trains pass through the station building area. The array of sensors includes vibration sensors installed at multiple station building structural points. The collected vibration signals are combined to obtain multiple vibration signal sets.

3. The method for locating vibration sources in station structures based on array sensors according to claim 1, characterized in that, Obtain the station building characteristics of the station building area, and index multiple positive sample vibration signal sets of stations in the same family, including: Obtain the station building characteristics of the station building area; Based on the station building characteristics, the same family of stations is obtained by indexing within the set of sample stations that have passed structural inspection; Multiple positive sample vibration signal sets of multiple station building structural points are obtained when multiple trains pass by the same station building through array sensor testing. Among them, multiple trains have multiple positive sample train characteristics.

4. The method for locating vibration sources in station structures based on array sensors according to claim 3, characterized in that, Based on the station building characteristics, similar station buildings are indexed within the set of sample station buildings that have passed structural inspection, including: Based on the station building characteristics, a station building feature vector is constructed, wherein the station building characteristics include size characteristics and line distance characteristics; Obtain a set of sample station buildings with the same architectural structure as the station buildings; Obtain the sample station feature set of the sample station set, and construct the sample station feature vector set; Calculate the similarity between the station building feature vector and each sample station building feature vector, and select the sample station building corresponding to the sample station building feature vector with the highest similarity as the same family of stations.

5. The method for locating vibration sources in station structures based on array sensors according to claim 1, characterized in that, Multiple train features from multiple trains are acquired, and combined with multiple positive sample train features, station building features, and features of stations from the same family, vibration signal fitting conditions are configured through intelligent sensing fitting rules, including: Obtain multiple train features from multiple trains; Calculate the similarity between multiple train features and multiple positive sample train features to obtain the similarity of multiple train features; Obtain the similarity between the station building features and the features of stations in the same family to obtain the station building feature similarity; Based on the similarity of train features and station features, vibration signal fitting conditions are configured using intelligent sensing fitting configuration rules. The vibration signal fitting conditions include the number of fitting iterations and a fitting error threshold.

6. The method for locating vibration sources in station structures based on array sensors according to claim 5, characterized in that, Based on the train feature similarity and station building feature similarity, vibration signal fitting conditions are configured using intelligent sensing fitting configuration rules, including: Get the preset number of iterations and the preset error threshold; Based on the similarity of the train features, calculate the first number adjustment coefficient and the first error adjustment coefficient; based on the similarity of the station building features, calculate the second number adjustment coefficient and the second error adjustment coefficient. Based on the first number adjustment coefficient, the first error adjustment coefficient, the second number adjustment coefficient, and the second error adjustment coefficient, the preset number of iterations and the preset error threshold are adjusted and calculated to obtain the fitting iteration number and the fitting error threshold, which are used as the fitting conditions for the vibration signal.

7. The method for locating vibration sources in station structures based on array sensors according to claim 6, characterized in that, Based on the train feature similarity, calculate the first number adjustment coefficient and the first error adjustment coefficient; based on the station building feature similarity, calculate the second number adjustment coefficient and the second error adjustment coefficient, including: Calculate the reciprocal of the train feature similarity, and use it as the first number adjustment coefficient and the first error adjustment coefficient; The reciprocal of the similarity of the station building features is calculated and used as the second number adjustment coefficient and the second error adjustment coefficient.

8. The method for locating vibration sources in station structures based on array sensors according to claim 1, characterized in that, According to the vibration signal fitting conditions, using the multiple positive sample vibration signal sets, iterative fitting is performed within the multiple vibration signal sets to obtain multiple vibration fitting confidence coefficients. The vibration source localization results are then selected, including: Within the plurality of positive sample vibration signal sets, the first positive sample vibration signal set of the first family of station building structure points is selected, and the first positive sample vibration signal is randomly selected. Within multiple vibration signal sets, a first vibration signal set for the first station building structure point is selected, and a first vibration signal is randomly selected, wherein the first family of station building structure points corresponds to the first station building structure point; According to the vibration signal fitting conditions, the first positive sample vibration signal is used to iteratively fit within the first vibration signal to obtain the first fitting confidence coefficient. Continue fitting multiple positive sample vibration signal sets and multiple vibration signal sets to obtain multiple sets of fitting confidence coefficients. Calculate the mean of each set to obtain multiple vibration fitting confidence coefficients. The station building structure point corresponding to the smallest vibration fitting confidence coefficient is selected as the vibration source location result.

9. The method for locating vibration sources in station structures based on array sensors according to claim 8, characterized in that, According to the vibration signal fitting conditions, the first positive sample vibration signal is used to iteratively fit within the first vibration signal to obtain the first fitting confidence coefficient, including: Randomly match the time axis of the first positive sample vibration signal with the time axis of the first vibration signal, and count the proportion of signal data points of the first vibration signal whose error with the signal data points of the first positive sample vibration signal is less than the fitting error threshold, and use it as the first iterative fitting confidence coefficient. Continue random matching and fitting until the number of fitting iterations is reached, and output the first fitting confidence coefficient with the largest value during the iterative fitting process to obtain the first fitting confidence coefficient.

10. A vibration source location device for station buildings based on array sensors, characterized in that, For implementing the method for locating vibration sources of station structures based on array sensors according to any one of claims 1-9, the apparatus comprises: The vibration signal acquisition module is used to acquire vibration signals from multiple station building structure points when multiple trains pass through the station building area through an array of sensors, and obtain multiple vibration signal sets. The positive sample vibration signal index module is used to obtain the station building characteristics of the station building area and index multiple positive sample vibration signal sets of the same family of station buildings. The multiple positive sample vibration signal sets are obtained by testing when a train with multiple positive sample train characteristics passes by. The fitting condition configuration module is used to acquire multiple train features of multiple trains, combine multiple positive sample train features, station building features and features of the same group of stations, and configure vibration signal fitting conditions through intelligent sensing fitting configuration rules. The vibration signal fitting conditions include the number of fitting iterations and the fitting error threshold. The fitting and localization filtering module is used to perform iterative fitting within the multiple positive sample vibration signal sets according to the vibration signal fitting conditions, obtain multiple vibration fitting confidence coefficients, and filter to obtain vibration source localization results.

Images