Digital monitoring system for cable-stayed bridge construction based on digital twin

Abstract

The present invention relates to the field of digital twin technology, and in particular to a digital monitoring system for cable-stayed bridge construction based on digital twins. The system comprises: a data acquisition module, which is used to acquire point cloud data of a cable-stayed bridge, monitoring time series data of each sensor, and various environmental time series data. A historical abnormality analysis module, which is used to correct the abnormality of the monitoring value in the monitoring time series data of each sensor according to the changes in the point cloud data, so as to obtain a more accurate probability of correction of abnormalities, and characterize the possibility of abnormalities occurring at each historical moment. An environmental factor analysis module, which is used to analyze the impact of changes in environmental factors on changes in sensor monitoring values, and obtain the impact weights of all environmental factors on each sensor at each moment. An abnormality prediction module, which is used to predict the abnormality of the monitoring value of each sensor at the next moment according to the multiple indicators corresponding to each moment in the monitoring time series data of each sensor.

Description

Technical Field

[0001] The present invention relates to the technical field of digital twins, and in particular to a digital monitoring system for cable-stayed bridge construction based on digital twins. Background Art

[0002] With the advancement of modern engineering technology, cable-stayed bridges, as a bridge structure with excellent spanning capacity and aesthetics, are widely used in the construction of large-span bridges; therefore, in the construction and operation and maintenance of cable-stayed bridges, it is very important to analyze the complex structure and variability of the external environment to ensure the long-term health and stability of the bridge. Digital twins can make full use of data such as physical models and sensors to describe and model the characteristics of physical objects in the real world, so they can be used in the construction monitoring process of cable-stayed bridges.

[0003] When monitoring the construction process of a cable-stayed bridge, the existing technology often relies on real-time data provided by sensors (cable tension, bridge tower displacement, main beam deformation, etc.) for abnormal monitoring, and can also predict abnormal conditions of the cable-stayed bridge at future times based on real-time data, thereby improving the timeliness of monitoring. However, in actual application scenarios, since the environment in which the cable-stayed bridge is located may change at different times (wind speed, temperature, vibration, etc.), it is impossible to accurately predict the abnormal conditions that may occur in the cable-stayed bridge at future times based solely on the real-time data provided by the sensors on the cable-stayed bridge, which will reduce the accuracy of predictive monitoring. Summary of the invention

[0004] In order to solve the technical problem that the environment of the cable-stayed bridge may change (wind speed, temperature, vibration, etc.) at different times, and thus the abnormal conditions that may occur in the cable-stayed bridge in the future cannot be accurately predicted based on the real-time data provided by the sensors on the cable-stayed bridge, which will lead to reduced accuracy of predictive monitoring, the purpose of the present invention is to provide a digital monitoring system for cable-stayed bridge construction based on digital twins, and the technical solutions adopted are as follows:

[0005] The present invention proposes a digital monitoring system for cable-stayed bridge construction based on digital twins, the system comprising:

[0006] The data acquisition module is used to obtain the point cloud data of each position in the cable-stayed bridge after each new segment of the cable-stayed bridge is completed; obtain the monitoring time series data of each sensor in the historical segment of the cable-stayed bridge and various environmental time series data in the same period;

[0007] The historical abnormal situation analysis module is used to determine the initial abnormal probability of the monitoring value at each moment for any sensor based on the numerical characteristics of the monitoring value in the monitoring time series data of the sensor; match the two adjacent point cloud data within the local range of the sensor and analyze the differences, so as to correct the initial abnormal probability of the monitoring value at each moment and obtain the corrected abnormal probability;

[0008] The environmental factor analysis module is used to analyze the correlation between the monitoring time series data of each sensor and each environmental time series data, and determine the influence coefficient of each environmental factor on the monitoring situation of each sensor; based on the influence coefficient corresponding to the environmental factor, the difference fluctuation between the data values ​​in the environmental time series data and the time series characteristics, determine the influence weight of all environmental factors on each sensor at each moment;

[0009] The abnormal situation prediction module is used to predict the abnormal situation of the monitoring value of any sensor at the next moment based on the corrected abnormal probability of the monitoring value at each moment, the influence weight of all environmental factors on the sensor, and the time series characteristics between moments.

[0010] Furthermore, the method for obtaining the initial abnormality probability includes:

[0011] For the monitoring value at any moment in the monitoring time series data of any sensor, when the monitoring value at that moment is within the preset monitoring value range, the absolute value of the difference between the monitoring value at that moment and each boundary value of the preset monitoring value range is calculated as the monitoring difference factor, and the ratio of the maximum value of the two monitoring difference factors to the interval length of the preset monitoring value range is taken as the initial abnormal probability of the monitoring value at that moment;

[0012] When the monitoring value at that moment is not within the preset monitoring value range, the initial abnormal probability of the monitoring value at that moment is recorded as the preset value.

[0013] Furthermore, the method for obtaining the modified abnormal probability includes:

[0014] Determine a preset neighborhood corresponding to the location of each sensor;

[0015] For any sensor, match the point cloud data of two adjacent points in the preset neighborhood corresponding to the sensor and analyze the differences to obtain the average error value corresponding to the monitoring value of the sensor;

[0016] The product of the average error value corresponding to the monitoring value of the sensor and the initial abnormal probability value of the monitoring value of the sensor at each moment is used as the corrected abnormal probability of the monitoring value at each moment.

[0017] Furthermore, the method for obtaining the average error value includes:

[0018] For any sensor, the point cloud data of the sensor in the corresponding preset neighborhood are registered for the last two times in time sequence based on the point cloud data registration algorithm to obtain all matching points;

[0019] The absolute value of the difference between the two point cloud data of each pair of matching points of the sensor in the preset neighborhood is calculated as the deviation factor; the average value of the deviation factors of all matching points of the sensor in the preset neighborhood is normalized as the average error value.

[0020] Furthermore, the method for obtaining the influence coefficient includes:

[0021] The Pearson correlation coefficient between the monitoring value in the monitoring time series data of each sensor and the data value of each environmental time series data is calculated, and the absolute value of the Pearson correlation coefficient is used as the influence coefficient of each environmental factor on the monitoring situation of each sensor.

[0022] Furthermore, the method for obtaining the influence weight includes:

[0023] In each environmental time series data, analyze the differences and fluctuations between data values ​​to determine the environmental difference characteristic values ​​corresponding to each moment;

[0024] The product of the influence coefficient of each environmental factor on the monitoring situation of each sensor, the negative correlation mapping and normalization of the time difference between each moment and the current moment in the time series, and the corresponding environmental difference characteristic value in each environmental time series data at each moment is used as the weight factor of each environmental factor on each sensor at each moment;

[0025] For any sensor, the accumulated value of the weight factors of the influence of all kinds of environmental factors on the sensor at each moment is normalized and used as the influence weight of all environmental factors on the sensor at each moment.

[0026] Furthermore, the method for obtaining the environmental difference characteristic value includes:

[0027] The absolute value of the difference between the data value at each moment and the current moment in each environmental time series data is used as the environmental difference factor at each moment, and the standard deviation of the data value at all moments in each environmental time series data is used as the environmental fluctuation factor;

[0028] In each type of environmental time series data, the environmental difference characteristic value corresponding to each moment is calculated according to the environmental difference factor and the environmental fluctuation factor, and the environmental difference factor is positively correlated with the environmental difference characteristic value, and the environmental fluctuation factor is negatively correlated with the environmental difference characteristic value.

[0029] Furthermore, for any sensor, according to the corrected abnormal probability of the monitoring value at each moment, the influence weight of all environmental factors on the sensor, and the time series characteristics between moments, predicting the abnormal situation of the monitoring value of the sensor at the next moment includes:

[0030] For any sensor, based on the time difference between the current moment and each moment in the time series, and the influence weight of all environmental factors on the sensor at each moment, calculate the prediction weight of all environmental factors on the sensor at each moment;

[0031] The product of the predicted weight of all environmental factors for the sensor at each moment and the corrected abnormal probability in the monitoring time series data of the sensor at each moment is used as the predicted abnormal factor of the monitoring value at each moment for the monitoring value of the sensor at the next moment;

[0032] The predicted abnormal factors of the sensor monitoring values ​​at all moments are integrated to obtain the abnormal characteristic value of the sensor monitoring value at the next moment;

[0033] When the abnormal characteristic value of the sensor at the next moment is greater than the preset abnormal threshold, an early warning is issued.

[0034] Furthermore, the method for obtaining the prediction weight includes:

[0035] In terms of time series, the time difference between the current time and each time is calculated as the time distance weight corresponding to each time;

[0036] The value after negative correlation mapping of the time distance weight corresponding to each moment is normalized by the product of the weight of the influence of all environmental factors on each sensor at each moment, and the predicted weight of all environmental factors on each sensor at each moment is obtained.

[0037] Furthermore, the point cloud data registration algorithm adopts an ICP algorithm.

[0038] The present invention has the following beneficial effects:

[0039] Whenever a section of a cable-stayed bridge is completed, it is usually necessary to update the BIM model, so it is necessary to obtain the point cloud data of each position in the cable-stayed bridge, and at the same time obtain the monitoring time series data of each sensor in the historical section of the cable-stayed bridge and the environmental time series data of various environmental factors in the same period. This all-round data collection method provides a solid foundation for subsequent analysis. Given that when an indicator of a cable-stayed bridge is abnormal, the data monitored by the corresponding sensor will change suddenly, so the initial abnormal probability at each moment can be preliminarily determined based on the numerical characteristics of the monitoring value in the monitoring time series data. Then, because the sensor cannot cover all points on the cable-stayed bridge, in order to improve the accuracy of the abnormal probability calculation, the initial abnormal probability can be corrected by combining the difference in the point cloud data within the local range of two adjacent sensors to obtain a more accurate corrected abnormal probability. Furthermore, since environmental factors will change during the construction of the cable-stayed bridge, the influence of environmental factors must also be considered when predicting abnormal conditions at future moments. Since different environmental factors have different effects on the monitoring conditions of different sensors, and the reference value of data at different times will also be different, we can first analyze the influence coefficient of each environmental factor on the sensor, and then integrate the influence of all environmental factors on the sensor, the difference fluctuation of the environmental time series data itself, and the time series characteristics, and determine the influence weight of all environmental factors at each moment on the prediction of abnormal conditions of the monitoring values ​​of each sensor at future moments. Finally, for any sensor, the corrected abnormal probability, time series characteristics, and the corresponding influence weights of all environmental factors at all times are combined to predict the abnormal conditions of the monitoring values ​​of the sensor at the next moment. In summary, based on the analysis of historical data, the present invention combines the influence of environmental factors, can more accurately predict the abnormal conditions that may occur in cable-stayed bridges at future moments, and effectively improves the accuracy of predictive monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 It is a system block diagram of a digital monitoring system for cable-stayed bridge construction based on digital twins provided by an embodiment of the present invention;

[0042] Figure 2 is a method flow chart of a method for obtaining a corrected abnormal probability provided by an embodiment of the present invention;

[0043] Figure 3is a method flow chart of a method for obtaining influence weights provided by an embodiment of the present invention;

[0044] Figure 4 It is a process flow chart of predicting abnormal situations provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the intelligent monitoring system and monitoring method of a building hanging basket proposed by the present invention, its specific implementation method, structure, characteristics and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0046] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0047] The specific scheme of a digital monitoring system for cable-stayed bridge construction based on digital twin provided by the present invention is described in detail below with reference to the accompanying drawings.

[0048] See also Figure 1 , which shows a system block diagram of a digital construction monitoring system for a cable-stayed bridge based on digital twins provided by an embodiment of the present invention. The system includes: a data acquisition module 101, a historical abnormal situation analysis module 102, an environmental factor analysis module 103, and an abnormal situation prediction module 104.

[0049] The data acquisition module 101 is used to acquire the point cloud data of each position in the cable-stayed bridge after each new segment of the cable-stayed bridge is completed; and to acquire the monitoring time series data of each sensor in the historical segment of the cable-stayed bridge and various environmental time series data in the same period.

[0050] In modern large-scale infrastructure construction, cable-stayed bridges are a complex and efficient form of bridge structure, and their construction process has extremely high requirements for accuracy and safety. Because there will be various dynamic changes during the construction process, especially environmental factors (such as wind, temperature, humidity, etc.) will cause structural deformation and changes in stress distribution, so it is necessary to predict abnormal conditions at future moments to help timely construction inspections and adjustments.

[0051] With the rapid development of information technology, digital twin technology, as an emerging technical means, provides a new solution for cable-stayed bridge construction monitoring. Digital twin technology creates a virtual mirror of the physical world, realizes real-time interaction and synchronous update of the physical structure and the virtual model, and provides a high-precision, all-round monitoring platform for the construction process of cable-stayed bridges. In addition, this technology can integrate multi-source data, including but not limited to construction operation records, bridge structure status data obtained by sensors, environmental monitoring data, etc. Through algorithm analysis, it can realize real-time early warning of the construction status of cable-stayed bridges.

[0052] The construction process of the cable-stayed bridge mainly includes pile foundation construction, bridge tower construction, main beam construction, cable tensioning, bridge deck paving and other processes. After each section of the cable-stayed bridge is completed, it is necessary to obtain the point cloud data of all the completed parts of the cable-stayed bridge, so that the BIM model can be updated to facilitate the monitoring of the overall construction of the cable-stayed bridge. Specifically, the cable-stayed bridge can be fully scanned using three-dimensional laser scanning technology or high-precision photogrammetry equipment carried by drones to obtain point cloud data. At the same time, in order to carry out digital monitoring of bridge construction, it is necessary to arrange a variety of sensors in the completed sections, such as using a static level to measure static settlement, using an inclinometer to monitor the inclination of the structure, and using a displacement sensor to measure the relative displacement between structural components. In addition, since environmental factors will affect the construction, it is also necessary to set up multiple environmental measurement facilities (such as anemometers, thermometers, hygrometers, etc.) at appropriate locations on the construction site to obtain multiple environmental factors that affect construction safety. Environmental factors can specifically include: temperature, wind speed, humidity, vibration, and rainfall. In this way, the monitoring time series data of each sensor in the historical segment of the cable-stayed bridge (excluding the cable-stayed bridge segment that has just been completed) and various environmental time series data in the same period (each environmental factor corresponds to a kind of environmental time series data) can be obtained.

[0053] It should be noted that in the embodiment of the present invention, the length of the monitoring time series data and the environmental time series data are set to 1 hour before the current moment, and the sampling time interval is set to 1s. The specific duration and sampling time interval can be adjusted according to the implementation scenario and are not limited here. In the embodiment of the present invention, since it is necessary to use the change of point cloud data under different measurement times in the subsequent process to predict the abnormal situation of the monitoring value of the sensor, and for each newly completed cable-stayed bridge segment, since it only has one point cloud data, in the embodiment of the present invention, each newly completed cable-stayed bridge segment during the construction process is not within the scope considered by the embodiment of the present invention; specifically, for example, the cable-stayed bridge segment is divided into 5 segments. When segment 1 is just completed, the point cloud data of segment 1 is obtained, and various required sensors are installed on segment 1. At this time, no abnormal prediction is performed; when segment 2 is completed, the point cloud data of segment 1 and segment 2 are obtained. At this time, segment 1 is the historical segment of the cable-stayed bridge, and the abnormal situation of the monitoring value of each sensor in segment 1 can be predicted; when segment 3 is completed, the point cloud data of segments 1, 2, and 3 are obtained. At this time, segment 1 and segment 2 are both historical segments of the cable-stayed bridge, and the abnormal situation of the monitoring value of each sensor in segments 1 and 2 can be predicted, and so on.

[0054] The historical abnormal situation analysis module 102 is used to determine the initial abnormal probability at each moment for any sensor based on the numerical characteristics of the monitoring value in the monitoring time series data of the sensor; match two adjacent point cloud data within the local range of the sensor and analyze the differences, so as to correct the initial abnormal probability at each moment and obtain the corrected abnormal probability.

[0055] Normally, during the construction of a cable-stayed bridge, due to factors such as wind load and construction load, the monitoring values ​​detected by various sensors are allowed to fluctuate within a normal range. However, it is not accurate enough to rely solely on the numerical characteristics of the sensor to judge abnormal conditions, because the sensor data may be affected by many factors, including environmental factors, equipment errors, etc., while point cloud data can provide more comprehensive structural change information from the perspective of three-dimensional space. Therefore, the initial abnormal probability at each moment can be preliminarily determined based on the numerical characteristics of the monitoring values ​​in the monitoring time series data of each sensor, and then the initial abnormal probability is corrected based on the difference in the change of two adjacent point cloud data in the local range where the sensor is located, and the corrected abnormal probability at each moment is obtained. The corrected abnormal probability can more accurately and comprehensively characterize the abnormal conditions of the cable-stayed bridge monitored by the sensor at each moment. This analysis result provides an important basis for subsequent abnormal warnings and construction adjustments at future moments.

[0056] Preferably, in one embodiment of the present invention, the method for obtaining the initial abnormality probability includes:

[0057] For each sensor, a monitoring value range (open interval) is preset, and the range represents the normal fluctuation range of the monitoring value. For any sensor, the monitoring value at each moment in the monitoring time series data of the sensor is compared with the boundary value of the preset monitoring value range, and the proximity of the monitoring value to the boundary value of the normal range can be quantified: when the monitoring value at a certain moment is within the preset monitoring value range, the absolute value of the difference between the monitoring value at this moment and each boundary value of the preset monitoring value range is calculated as the monitoring difference factor, which represents the degree of proximity between the monitoring value at this moment and the boundary value of the preset monitoring range. The smaller the monitoring difference factor between the monitoring value at this moment and a certain boundary value, the closer it is to the boundary value on this side, and the higher the probability of an abnormality, then the larger the monitoring difference factor between the monitoring value at this moment and the boundary value on the other side; and the interval length of the preset monitoring value range represents the size of the allowed fluctuation, so the ratio of the maximum value of the two monitoring difference factors to the interval length of the preset monitoring value range is used as the initial abnormal probability at this moment. The abnormal probability reflects the relative position of the monitoring value in the preset monitoring value range, which can be used to describe its proximity to the abnormality. The larger the value of the initial abnormal probability, the closer the monitoring value at this moment is to a certain boundary value in the normal range, and the higher the probability of an abnormality. For example, the preset monitoring value range is (5, 8), and a certain monitoring value is 7, and the corresponding two monitoring difference factors are 1 and 2 respectively, then the initial abnormal probability value corresponding to the monitoring value is 2 / 3.

[0058] It should be noted that the preset monitoring value range of each sensor can be set according to the corresponding construction requirements and acceptance standards of the cable-stayed bridge, for example, set to ±5% of the standard value, etc., and can be adjusted according to the implementation scenario and is not limited here.

[0059] When the monitoring value at a certain moment is not within the preset monitoring value range, it means that it has exceeded the normal fluctuation range, and the initial abnormal probability at that moment is directly recorded as the preset value. It should be noted that the initial abnormal probability values ​​of the monitoring values ​​within the preset monitoring value range calculated based on the above method are all less than 1, so the preset value here needs to be greater than or equal to 1. In order to facilitate subsequent calculations, it is set to 1 here.

[0060] Considering that sensors cannot cover all areas during the construction of cable-stayed bridges, especially those with larger structures or complex shapes, when a local area cannot be completely covered by the sensors, the local structural problems of the cable-stayed bridge in this area will not be detected by the sensors. Since point cloud data can perform an all-round scan of the surface of the cable-stayed bridge, the initial abnormal probability of the sensor monitoring value at each moment can be corrected through the difference between the point cloud data of two adjacent measurements, thereby obtaining a more accurate corrected abnormal probability.

[0061] Preferably, in one embodiment of the present invention, the method for obtaining the modified abnormal probability includes:

[0062] See also Figure 2 , which shows a method flow chart of a method for obtaining a modified abnormal probability in one embodiment of the present invention, the method comprising the following steps:

[0063] Step S201: Determine a preset neighborhood corresponding to the location of each sensor.

[0064] For any sensor, a circle is drawn with the installation location of the sensor as the center and the preset constant as the radius, so as to obtain the preset neighborhood corresponding to the sensor. It should be noted that the preset constant can be 1m, and the specific value can be adjusted according to the implementation scenario, and the shape and size of the preset neighborhood can be adjusted according to the implementation scenario, which is not limited here.

[0065] Step S202: For any sensor, match the point cloud data of the preset neighborhood corresponding to the sensor at two adjacent times and analyze the differences to obtain the average error value corresponding to the monitoring value of the sensor.

[0066] Since point cloud data is obtained for each position in the cable-stayed bridge after each new section of the cable-stayed bridge is completed, the acquisition of point cloud data also has a time sequence. The later the point cloud data is in the time sequence, the closer the structural condition of the cable-stayed bridge it represents is to the current actual situation. Therefore, it is more accurate to use it for subsequent predictions of abnormal conditions in the future.

[0067] Therefore, for any sensor, the point cloud data of the sensor in the corresponding preset neighborhood for the last two times in time sequence are registered based on the ICP point cloud data registration algorithm to obtain all matching points.

[0068] Then, the absolute value of the difference between the point cloud data of each pair of matching points in the preset neighborhood of the sensor is calculated as the deviation factor. The larger the deviation factor, the greater the degree of change in the structure of the cable-stayed bridge at each matching point in the preset neighborhood corresponding to the sensor, and the higher the probability of an abnormality in the cable-stayed bridge at this point.

[0069] Finally, the average value of the deviation factor of all matching points of the sensor in the preset neighborhood is normalized as the average error value. The larger the average error value, the greater the structural change of the cable-stayed bridge in the preset neighborhood where the sensor is located, and the greater the possibility of abnormality. Normalization is a technical means well known to those skilled in the art, and the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0070] It should be noted that the ICP point cloud data registration algorithm is a well-known technology, and the specific process will not be described here.

[0071] Step S203: based on the average error value corresponding to each sensor monitoring value, the initial abnormal probability of the monitoring value of each sensor at each moment is corrected to obtain a corrected abnormal probability.

[0072] Based on the above analysis, it can be seen that the larger the value of the initial abnormal probability is, the higher the probability that the monitoring value at that moment is abnormal; the larger the average error value is, the greater the structure of the cable-stayed bridge has undergone major changes in the preset neighborhood of the sensor, and the greater the possibility of abnormality; therefore, the product of the average error value corresponding to the sensor and the initial abnormal probability value of the monitoring value of the sensor at each moment is used as the corrected abnormal probability of the monitoring value of the sensor at each moment. The larger the corrected abnormal probability value is, the higher the degree of abnormality of the monitoring value at that moment.

[0073] At this point, through the historical abnormal situation analysis module 102, the historical abnormal situation of the monitoring value under each sensor at each moment can be obtained (characterized by correcting the abnormal probability), which can be used in the subsequent prediction process of abnormal situations at future moments.

[0074] The environmental factor analysis module 103 is used to analyze the change correlation between the monitoring time series data of each sensor and each environmental time series data, and determine the influence coefficient of each environmental factor on the monitoring situation of each sensor; based on the influence coefficient corresponding to the environmental factor, the difference fluctuation between the data values ​​in the environmental time series data and the time series characteristics, determine the influence weight of all environmental factors on each sensor at each moment.

[0075] Due to the complex structure of the cable-stayed bridge, it is easily affected by environmental factors during the construction process, resulting in structural deformation and changes in stress distribution. Therefore, when predicting abnormal conditions at future moments, it is also necessary to consider the impact of environmental factors on the sensor monitoring values. Given that different environmental factors have different degrees of influence on different monitoring indicators of cable-stayed bridges, for example, temperature will more affect the inclination change of the cable-stayed bridge, because with the change of temperature, the cable-stayed bridge will expand and contract, resulting in changes in the overall shape and internal stress state; wind speed will affect the overall displacement of the cable-stayed bridge, etc. Therefore, it is necessary to first determine the influence coefficient of each environmental factor on the monitoring of each sensor; and the change of environmental conditions is also a factor in analyzing its influence, and the reference value of data at different moments is different. Therefore, when integrating the influence coefficients of all environmental factors, the difference fluctuations between the data values ​​of the environmental time series data and the time series characteristics between moments are also analyzed, thereby determining the influence weight of all environmental factors on each sensor at each moment, which can more accurately evaluate the influence of environmental factors on the monitoring values ​​in the sensor at each moment, and effectively improve the accuracy of subsequent abnormal situation predictions at future moments.

[0076] Preferably, in one embodiment of the present invention, the method for obtaining the influence coefficient includes:

[0077] Since the Pearson correlation coefficient can measure the correlation of changes between two sets of data, in an embodiment of the present invention, the Pearson correlation coefficient between the monitoring value in the monitoring time series data of each sensor and the data value of each environmental time series data is calculated. The Pearson correlation coefficient is a positive value, and the closer it is to 1, it indicates that there is a positive correlation between the environmental time series data and the monitoring time series data. Conversely, the Pearson correlation coefficient is a negative value, and the closer it is to -1, it indicates that there is a negative correlation between the environmental time series data and the monitoring time series data. However, since the embodiment of the present invention only cares about the impact of environmental factors on the changes in sensor monitoring data (both positive and negative relationships are considered to have an impact), the absolute value of the Pearson correlation coefficient is used as the impact coefficient of each environmental factor on the monitoring situation of each sensor. At this time, the larger the impact coefficient, the greater the degree of influence of the environmental factor on the change of the sensor monitoring value.

[0078] The fluctuation of data values ​​between moments in environmental time series data reflects the dynamic changes of environmental factors. By analyzing the difference fluctuations between data values ​​in environmental time series data, the changes of environmental factors at different time points can be captured, and the time series characteristics of the moments can indicate the reference value of the data values ​​at each moment. Therefore, the dynamic changes of environmental factors, the time series characteristics of the moments and the corresponding influence coefficients of environmental factors are combined to determine the influence weight of environmental factors on sensors at each moment. The influence weight can provide more accurate prediction results when predicting abnormal conditions of sensor monitoring values ​​at future moments, thereby effectively improving the accuracy of early warning.

[0079] Preferably, in one embodiment of the present invention, the method for obtaining the influence weight includes:

[0080] See also Figure 3 , which shows a method flow chart of a method for obtaining influence weights in one embodiment of the present invention, the method comprising the following steps:

[0081] Step S301: In each type of environmental time series data, analyze the differences and fluctuations between data values ​​to determine the environmental difference characteristic value corresponding to each moment.

[0082] Because the abnormal situation of the sensor monitoring value at a future moment needs to be predicted, and the environmental factors are usually similar in a short time, and the current moment is the moment closest to the future moment, when analyzing the changes in environmental factors, the present invention compares the data value at each moment in the environmental time series data with the data value at the current moment, and takes the absolute value of the difference between the data value at each moment and the current moment in each environmental time series data as the environmental difference factor at each moment. The larger the environmental difference factor, the more obvious the change in the environmental factors between the current moment.

[0083] The standard deviation of the data values ​​at all times in each environmental time series data is taken as the environmental fluctuation factor to measure the overall fluctuation degree of the environmental data. The larger the environmental fluctuation factor, the more obvious the changes in environmental factors are overall. Conversely, the smaller the environmental fluctuation factor, the more subtle the changes in environmental factors are overall.

[0084] When the overall change of environmental factors is more drastic, the significance of the environmental difference factor corresponding to a single moment will be reduced; on the contrary, when the overall change of environmental factors is relatively subtle, the significance of the environmental difference factor corresponding to a single moment will be increased, so the sum of the environmental fluctuation factor and the preset first parameter is used as the denominator, the environmental difference factor is used as the numerator, and the value after the ratio is normalized is used as the environmental difference characteristic value corresponding to each moment in each environmental time series data. At this time, the larger the environmental difference characteristic value in a certain environmental time series data at a certain moment, the more significant the change of this environmental factor at this moment. Wherein normalization is a technical means well known to those skilled in the art, and the choice of normalization function can be linear normalization or standard normalization, etc., and the specific normalization method is not limited here.

[0085] It should be noted that the purpose of presetting the first parameter is to prevent the denominator from being 0, and the value may be 0.001. The specific value may be adjusted according to the implementation scenario and is not limited here.

[0086] Step S302: Determine the weight factor of each environmental factor on each sensor at each moment by combining the environmental difference characteristic value, time series characteristics and the influence coefficient of each environmental factor on the monitoring situation of each sensor in each environmental time series data at each moment.

[0087] Because in the time series, the farther the moment is from the current moment, the lower the reference value of the change of its environmental factors on the prediction of the next moment monitoring value of the current moment. Therefore, in the environmental time series data of each environmental factor, the time difference between each moment and the current moment is calculated. The larger the time difference, the farther the time distance is, the lower the reference value is. On the contrary, the smaller the time difference, the closer the time distance is, the higher the reference value is. Therefore, the time difference is negatively correlated and normalized to achieve logical relationship correction and obtain the time weight. At this time, the larger the time weight of a certain moment, the higher the reference value of the change of the environmental factors at that moment. The negative correlation mapping here can use the exp(-x) function, where exp() represents an exponential function with the natural constant e as the base, and x represents the independent variable.

[0088] Then, since the larger the influence coefficient corresponding to the environmental factor is, the greater the influence of its change on the change of the sensor monitoring value is; similarly, the larger the environmental difference characteristic value corresponding to each moment in each environmental time series data is, the more significant the change of the environmental factor is. Therefore, the product of the influence coefficient of each environmental factor on the monitoring situation of each sensor, the time weight of each moment and the environmental difference characteristic value corresponding to each moment in each environmental time series data is used as the weight factor of each environmental factor on each sensor at each moment. At this time, the larger the weight factor corresponding to a certain environmental factor at a certain moment is, the greater the influence of the change of the environmental factor at that moment on the subsequent abnormal situation prediction is.

[0089] Step S303: At each moment, the weight factors of the impact of all environmental factors on each sensor are integrated to determine the impact weights of all environmental factors on the sensor at each moment.

[0090] Based on the above steps, the weight factor of each environmental factor affecting each sensor at each moment can be calculated. Since there are multiple environmental influencing factors, for any sensor, the weight factor of all types of environmental factors affecting the sensor at each moment can be integrated: the accumulated value of the weight factor of all types of environmental factors affecting the sensor at each moment is normalized, and used as the weight of all environmental factors affecting the sensor at each moment, recorded as XQ. The greater the weight of a certain moment, the higher the reference value of the environmental change at that moment for subsequent abnormal prediction. Normalization is a technical means well known to those skilled in the art, and the selection of the normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0091] At this point, the module can obtain the impact weight of the environmental factors at each moment by analyzing the impact of environmental factors on the sensor monitoring values, which can be used in subsequent analysis processes.

[0092] The abnormal situation prediction module 104 is used to predict the abnormal situation of the monitoring value of any sensor at the next moment according to the modified abnormal probability of the monitoring value at each moment, the influence weight of all environmental factors on the sensor and the time series characteristics between moments.

[0093] By predicting future abnormal situations, early warnings can be issued so that relevant personnel can take necessary measures to solve potential problems in a timely manner. For any sensor, the corrected abnormal probability of the sensor's monitoring value at each moment represents the abnormality of the monitoring value, and the weight of the influence of all environmental factors on the sensor at each moment reveals the degree of influence of changes in environmental factors on the sensor's monitoring value. At the same time, the time series characteristics between moments can represent the reference value of the abnormal situation at each moment for predicting abnormal situations at future moments. Therefore, combining all the above indicators, the abnormal situation of the sensor's monitoring value at the next moment can be predicted more accurately, thereby improving the safety of the cable-stayed bridge construction process and the timeliness of the adjustment of potential problems.

[0094] Preferably, in one embodiment of the present invention, for any sensor, based on the corrected abnormal probability of the monitoring value at each moment, the weight of the influence of all environmental factors on the sensor, and the time series characteristics between moments, predicting the abnormal situation of the monitoring value of the sensor at the next moment includes:

[0095] See also Figure 4 , which shows a flowchart of predicting abnormal situations in one embodiment of the present invention, the process includes the following steps:

[0096] Step S401: For any sensor, based on the time difference between the current time and each time in the time series, and the influence weight of all environmental factors on the sensor at each time, calculate the prediction weight of all environmental factors on the sensor at each time.

[0097] Since in time series data, recent data can often better reflect the changing state of current data than distant data, when making predictions for future moments, data closer to the current moment should be given a higher weight.

[0098] Therefore, in terms of time series, the time difference between the current moment and each moment is calculated as the time distance weight corresponding to each moment, denoted as ΔT. The time distance weight reflects the degree of temporal proximity between a certain moment and the current moment. The smaller the time distance weight, the closer the time is, and the greater the prediction weight of the moment should be.

[0099] Because when the weight of all environmental factors on the sensor at a certain moment is greater, it means that the environmental changes at that moment have a higher reference value for subsequent abnormal prediction, so the time distance weight corresponding to each moment is negatively correlated with the prediction weight, while the influence weight is positively correlated with the prediction weight. Therefore, the formula model of the prediction weight is constructed as follows:

[0100] YQ=norm(XQ×α ΔT )

[0101] Among them, YQ represents the prediction weight of all environmental factors on each sensor at each moment; XQ represents the influence weight of all environmental factors on each sensor at each moment; ΔT represents the time distance weight at each moment; α represents the preset second parameter; norm() represents the normalization function.

[0102] In the formula model of the prediction weight, the exponential function is used in the embodiment of the present invention to negatively correlate the time distance weight, so the value of the preset second parameter α should be between (0, 1) to achieve logical relationship correction, and obtain α ΔT , the smaller the time distance weight is, the closer the time distance between a certain moment and the current moment is, then α ΔT The larger the value, the higher the reference value of the moment. Finally, the α of each moment ΔT The product of the corresponding influence weight is normalized to obtain the prediction weight of all environmental factors for each sensor at each moment. The larger the prediction weight, the higher the reference value of the abnormal situation.

[0103] It should be noted that the preset second parameter α here is 0.9, and the specific value can be adjusted according to the implementation scenario and is not limited here.

[0104] Step S402: The prediction weights of all environmental factors for each sensor at each moment and the corrected abnormal probability of each sensor monitoring value at each moment are integrated to calculate the predicted abnormal factor of the monitoring value at each moment for the monitoring value of each sensor at the next moment.

[0105] In the monitoring time series data of each sensor, the corrected abnormal probability table of the monitoring value at each moment increases the degree of abnormality of the monitoring value at that moment, and the prediction weight corresponding to each moment integrates the time series characteristics and environmental influencing factors, and represents the reference value of all environmental factors at each moment for predicting the abnormal situation of the monitoring value at the next moment. Therefore, the product of the prediction weight corresponding to each moment and the corrected abnormal probability of the monitoring value at each moment is used as the prediction abnormality factor of the monitoring value at each moment for the monitoring value of the sensor at the next moment. At this time, the larger the prediction abnormality factor is, the greater the possibility of abnormal situation of the monitoring value at the next moment of the current moment when taking this moment as the benchmark.

[0106] Step S403: predicting the abnormal situation of the sensor monitoring value at the next moment based on the predicted abnormal factors of the sensor monitoring value at the next moment at all moments, and performing abnormal monitoring according to the prediction results.

[0107] For any sensor, the predicted abnormal factors corresponding to all moments are integrated to obtain the abnormal characteristic value of the sensor monitoring value at the next moment: the value after normalizing the mean value of the predicted abnormal factors of the sensor's monitoring value at the next moment at all moments is used as the abnormal characteristic value of the sensor monitoring value at the next moment of the current moment. The abnormal characteristic value is a comprehensive reflection of the predicted abnormal factors at all moments, which can more accurately predict the probability of abnormal monitoring value at the next moment, and the larger the value, the higher the possibility of abnormality. Therefore, when the abnormal characteristic value of the sensor at the next moment is greater than the preset abnormal threshold, an early warning is required to remind relevant personnel to perform maintenance or adjustments in time. It should be noted that the preset abnormal threshold is 0.7, and the specific value can be adjusted according to the implementation scenario, which is not limited here; wherein normalization is a technical means well known to technicians in this field, and the selection of the normalization function can be linear normalization or standard normalization, etc., and the specific normalization method is not limited here.

[0108] In summary, whenever a section of a cable-stayed bridge is completed, the BIM model usually needs to be updated, so it is necessary to obtain the point cloud data of each position in the cable-stayed bridge, and at the same time obtain the monitoring time series data of each sensor in the historical section of the cable-stayed bridge and the environmental time series data of various environmental factors in the same period. This all-round data collection method provides a solid foundation for subsequent analysis. Given that when an indicator of a cable-stayed bridge is abnormal, the data monitored by the corresponding sensor will change suddenly, so the initial abnormal probability at each moment can be preliminarily determined based on the numerical characteristics of the monitoring value in the monitoring time series data. Then, because the sensor cannot cover all points on the cable-stayed bridge, in order to improve the accuracy of the abnormal probability calculation, the initial abnormal probability can be corrected by combining the difference in the point cloud data within the local range of two adjacent sensors to obtain a more accurate corrected abnormal probability. Furthermore, since environmental factors will change during the construction of the cable-stayed bridge, the influence of environmental factors must also be considered when predicting abnormal conditions at future moments. Since different environmental factors have different impacts on the monitoring conditions of different sensors, and the reference value of data at different times will also be different, we can first analyze the influence coefficient of each environmental factor on the sensor, and then integrate the impact of all environmental factors on the sensor, the difference fluctuation of the environmental time series data itself, and the time series characteristics, to determine the influence weight of all environmental factors at each moment on the prediction of abnormal conditions of the monitoring values ​​of each sensor at future moments. Finally, for any sensor, the corrected abnormal probability, time series characteristics, and the corresponding influence weights of all environmental factors at all times are combined to predict the abnormal conditions of the monitoring values ​​of the sensor at the next moment. In summary, based on the analysis of historical data, the embodiment of the present invention combines the influence of environmental factors, can more accurately predict the abnormal conditions that may occur in cable-stayed bridges at future moments, and effectively improve the accuracy of predictive monitoring.

[0109] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0110] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0111] 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 principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A digital monitoring system for cable-stayed bridge construction based on digital twin, characterized in that: The system includes: The data acquisition module is used to obtain the point cloud data of each position in the cable-stayed bridge after each new segment of the cable-stayed bridge is completed; obtain the monitoring time series data of each sensor in the historical segment of the cable-stayed bridge and various environmental time series data in the same period; The historical abnormal situation analysis module is used to determine the initial abnormal probability of the monitoring value at each moment for any sensor based on the numerical characteristics of the monitoring value in the monitoring time series data of the sensor; match the two adjacent point cloud data within the local range of the sensor and analyze the differences, so as to correct the initial abnormal probability of the monitoring value at each moment and obtain the corrected abnormal probability; The environmental factor analysis module is used to analyze the correlation between the monitoring time series data of each sensor and each environmental time series data, and determine the influence coefficient of each environmental factor on the monitoring situation of each sensor; based on the influence coefficient corresponding to the environmental factor, the difference fluctuation between the data values ​​in the environmental time series data and the time series characteristics, determine the influence weight of all environmental factors on each sensor at each moment; The abnormal situation prediction module is used to predict the abnormal situation of the monitoring value of any sensor at the next moment according to the corrected abnormal probability of the monitoring value at each moment, the influence weight of all environmental factors on the sensor, and the time series characteristics between moments; The method for obtaining the influence weight includes: In each environmental time series data, analyze the differences and fluctuations between data values ​​to determine the environmental difference characteristic values ​​corresponding to each moment; The product of the influence coefficient of each environmental factor on the monitoring situation of each sensor, the negative correlation mapping and normalization of the time difference between each moment and the current moment in the time series, and the corresponding environmental difference characteristic value in each environmental time series data at each moment is used as the weight factor of each environmental factor on each sensor at each moment; For any sensor, the accumulated value of the weight factors of the influence of all kinds of environmental factors on the sensor at each moment is normalized and used as the influence weight of all environmental factors on the sensor at each moment.

2. According to claim 1, a digital twin-based cable-stayed bridge construction digital monitoring system is characterized in that: The method for obtaining the initial abnormal probability includes: For the monitoring value at any moment in the monitoring time series data of any sensor, when the monitoring value at that moment is within the preset monitoring value range, the absolute value of the difference between the monitoring value at that moment and each boundary value of the preset monitoring value range is calculated as the monitoring difference factor, and the ratio of the maximum value of the two monitoring difference factors to the interval length of the preset monitoring value range is taken as the initial abnormal probability of the monitoring value at that moment; When the monitoring value at that moment is not within the preset monitoring value range, the initial abnormal probability of the monitoring value at that moment is recorded as the preset value.

3. According to claim 1, a digital twin-based cable-stayed bridge construction digital monitoring system is characterized in that: The method for obtaining the modified abnormal probability includes: Determine a preset neighborhood corresponding to the location of each sensor; For any sensor, match the point cloud data of two adjacent points in the preset neighborhood corresponding to the sensor and analyze the differences to obtain the average error value corresponding to the monitoring value of the sensor; The product of the average error value corresponding to the monitoring value of the sensor and the initial abnormal probability value of the monitoring value of the sensor at each moment is used as the corrected abnormal probability of the monitoring value at each moment.

4. According to claim 3, a digital monitoring system for cable-stayed bridge construction based on digital twins is characterized in that: The method for obtaining the average error value includes: For any sensor, the point cloud data of the sensor in the corresponding preset neighborhood are registered for the last two times in time sequence based on the point cloud data registration algorithm to obtain all matching points; The absolute value of the difference between the two point cloud data of each pair of matching points of the sensor in the preset neighborhood is calculated as the deviation factor; the average value of the deviation factors of all matching points of the sensor in the preset neighborhood is normalized as the average error value.

5. According to claim 1, a digital monitoring system for cable-stayed bridge construction based on digital twins is characterized in that: The method for obtaining the influence coefficient includes: The Pearson correlation coefficient between the monitoring value in the monitoring time series data of each sensor and the data value of each environmental time series data is calculated, and the absolute value of the Pearson correlation coefficient is used as the influence coefficient of each environmental factor on the monitoring situation of each sensor.

6. According to claim 1, a digital monitoring system for cable-stayed bridge construction based on digital twins is characterized in that: The method for obtaining the environmental difference characteristic value includes: The absolute value of the difference between the data value at each moment and the current moment in each environmental time series data is used as the environmental difference factor at each moment, and the standard deviation of the data value at all moments in each environmental time series data is used as the environmental fluctuation factor; In each type of environmental time series data, the environmental difference characteristic value corresponding to each moment is calculated according to the environmental difference factor and the environmental fluctuation factor, and the environmental difference factor is positively correlated with the environmental difference characteristic value, and the environmental fluctuation factor is negatively correlated with the environmental difference characteristic value.

7. The digital twin-based cable-stayed bridge construction digital monitoring system according to claim 1 is characterized in that: For any sensor, according to the corrected abnormal probability of the monitoring value at each moment, the weight of the influence of all environmental factors on the sensor, and the time series characteristics between moments, predicting the abnormal situation of the monitoring value of the sensor at the next moment includes: For any sensor, based on the time difference between the current moment and each moment in the time series, and the influence weight of all environmental factors on the sensor at each moment, calculate the prediction weight of all environmental factors on the sensor at each moment; The product of the predicted weight of all environmental factors for the sensor at each moment and the corrected abnormal probability in the monitoring time series data of the sensor at each moment is used as the predicted abnormal factor of the monitoring value at each moment for the monitoring value of the sensor at the next moment; The predicted abnormal factors of the sensor monitoring values ​​at all moments are integrated to obtain the abnormal characteristic value of the sensor monitoring value at the next moment; When the abnormal characteristic value of the sensor at the next moment is greater than the preset abnormal threshold, an early warning is issued.

8. The digital twin-based cable-stayed bridge construction digital monitoring system according to claim 7 is characterized in that: The method for obtaining the prediction weight includes: In terms of time series, the time difference between the current time and each time is calculated as the time distance weight corresponding to each time; The value after negative correlation mapping of the time distance weight corresponding to each moment is normalized by the product of the weight of the influence of all environmental factors on each sensor at each moment, and the predicted weight of all environmental factors on each sensor at each moment is obtained.

9. The digital twin-based cable-stayed bridge construction digital monitoring system according to claim 4 is characterized in that: The point cloud data registration algorithm adopts the ICP algorithm.

Images