A method and system for adaptive string data acquisition based on dtu

Abstract

The application provides a DTU-based adaptive vibrating string data acquisition method and system, which comprises the following steps: predicting network state and evaluating data change and deeply integrating them into acquisition decision. The system monitors network indicators in real time, predicts future health degree by using an exponential smoothing model, calculates the change gradient of the sliding window of the vibrating string data, dynamically adjusts the acquisition frequency after the two are weighted and fused by a value function. In the off-network scenario, the system predicts the continuous time length based on historical records, enables an edge autoregressive model to generate a predicted value, and reports key events through a backup channel; low-value data is selectively compressed during cache emergency. After the network is restored, the transmission urgency is calculated by comprehensively considering the data gradient, time value and queuing position, and the transmission window is accurately predicted to realize priority transmission, and expired data is automatically eliminated.

Description

Technical Field

[0001] This invention relates to the field of sinusoidal data acquisition technology, and in particular to an adaptive vibrating string data acquisition method and system based on a DTU. Background Technology

[0002] Currently, traditional vibrating wire data acquisition systems generally adopt a fixed-time, equal-frequency acquisition strategy. This means that regardless of network conditions or data stability, data is acquired and uploaded according to a preset cycle. This fixed mode has significant drawbacks: when the network signal is good, the fixed frequency may not fully utilize the transmission bandwidth, resulting in insufficient monitoring information density; when the network is congested or the signal deteriorates, a large amount of acquired data accumulates in the local cache, and may even be lost due to cache overflow, seriously affecting the integrity and continuity of the monitoring data.

[0003] Other traditional technical solutions introduce simple network condition adaptive strategies, namely, reducing the acquisition frequency when the network is poor and retransmitting after the network recovers. However, existing adaptive solutions still have the following shortcomings: First, the adjustment of the acquisition frequency only passively depends on the current network condition, lacking proactive prediction of future network trends, resulting in decision-making lag; Second, the adjustment basis is singular, failing to consider the differentiated impact of the drastic changes in the vibrating string data itself on the acquisition value, using the same adjustment rules for calm periods and drastic periods, resulting in unreasonable resource allocation; Third, during network outages, data is simply cached, and a first-in-first-out retransmission method is used after recovery, failing to distinguish the timeliness and importance of data, resulting in a large amount of low-value historical data occupying valuable transmission resources and delaying the timely delivery of high-value data. Summary of the Invention

[0004] One of the objectives of this invention is to provide an adaptive vibrating wire data acquisition method and system based on DTU. The method and system construct an adaptive acquisition value function that includes a network state dominant term, a data change dominant term, and a synergistic gain term of the two. This enables the acquisition frequency to proactively predict future network trends to match transmission capacity, and to dynamically adjust the acquisition density according to the stable or drastic state of the vibrating wire data itself. High-frequency acquisition is used to capture key information when the network is smooth and the data is drastic, while the frequency is proactively reduced to save resources when the network is congested or the data is stable. This achieves synergistic optimization of acquisition strategy and transmission capacity, and improves the comprehensive value of monitoring data under limited resources.

[0005] Another objective of this invention is to provide an adaptive vibrating wire data acquisition method and system based on DTU. The method and system predict the duration of network outages and monitor the cache status. When the network outage exceeds the expected duration or the cache is critically low, alternative data is generated using a local autoregressive prediction model. The predicted values ​​or key alarm events are reported through a backup channel. This avoids the complete loss of data caused by cache overflow and ensures the basic early warning capability of the monitoring system under extreme network conditions, greatly improving the robustness of the system in weak network environments.

[0006] Another objective of this invention is to provide an adaptive vibrating wire data acquisition method and system based on DTU. The method and system provide a method for evaluating and prioritizing the transmission urgency by comprehensively considering data change gradient, timeliness value decay, and queue waiting position. After network recovery, it prioritizes the retransmission of critical data that changes drastically, has high timeliness, and has long waiting time. At the same time, it uses network health prediction to identify available transmission windows for precise delivery, avoiding repeated invalid retransmissions when the signal is unstable. It also employs a mechanism for compressing or discarding low-value expired data to ensure that the most urgent data is delivered first within a limited transmission window, thus comprehensively improving the timeliness and transmission efficiency of data retransmission after network recovery.

[0007] To achieve at least one of the above-mentioned objectives, the present invention further provides an adaptive vibrating wire data acquisition method based on a DTU, the method comprising: The DTU collects frequency data from the vibrating wire sensor in real time and continuously monitors the current network status indicators, which include at least signal strength, signal-to-noise ratio, packet loss rate, and round-trip time. The instantaneous network health is calculated based on the current network status indicators, and the network health at multiple future moments is predicted based on the historical network health sequence. The comprehensive network health prediction value used for data collection and decision-making is obtained by combining the prediction values ​​at each moment. The data change gradient is calculated based on the vibrating string frequency data at the current acquisition time and historical times. The data change gradient is used to characterize the drastic degree of change in the vibrating string data. The comprehensive network health prediction value is fused with the data change gradient to construct an adaptive acquisition value function. The adaptive acquisition value function includes a network state dominant term, a data change dominant term, and a synergistic gain term of the two. The acquisition frequency of the next acquisition cycle is dynamically adjusted or the current acquisition cycle is skipped based on the calculation result of the adaptive acquisition value function. When the network health is lower than the network disconnection threshold, the network disconnection mode is entered. During the network disconnection, the local edge prediction model is activated to generate the predicted value of the vibrating string frequency, and the cache status and network disconnection duration are monitored. When the preset conditions are met, the predicted value is reported through the backup channel or a critical event alarm is triggered. Once the network is restored to a usable state, the data packets cached during the network outage are sorted according to their transmission urgency and retransmitted in sequence. The transmission urgency is calculated by the data change gradient, the data timeliness value, and the queue waiting position. After the retransmission is completed, a data integrity verification report is reported.

[0008] According to a preferred embodiment of the present invention, the method for calculating the instantaneous network health includes: converting signal strength, signal-to-noise ratio, packet loss rate, and round-trip delay into component values ​​in the range of 0 to 1 through their respective normalization mapping functions, and then weighting and summing the component values ​​to obtain the instantaneous health; the prediction of network health at multiple future moments adopts a combination model of exponential smoothing and linear trend, and multi-step prediction is achieved through the recursive update of horizontal components and trend components; The method for calculating the data change gradient includes: calculating the instantaneous rate of change of the vibrating string frequency at the current moment relative to the previous moment; further taking the product of the maximum value and the average value of each instantaneous rate of change within the sliding time window as the original gradient value; and then normalizing the original gradient value relative to a preset threshold to obtain a normalized data change gradient with a value range of 0 to 1.

[0009] According to another preferred embodiment of the present invention, the adaptive acquisition value function is constructed by: multiplying the comprehensive network health prediction value and the data change gradient by taking powers of each, respectively, to form a coupled form of network state dominant term and data change dominant term, and adding a product term of comprehensive network health prediction value and data change gradient as a cooperative gain term on the basis of the product of the two terms; the acquisition frequency is adjusted by: linearly mapping between the preset minimum acquisition frequency and the maximum acquisition frequency according to the calculated value of the value function, the higher the value function value, the closer the acquisition frequency is to the maximum acquisition frequency.

[0010] According to another preferred embodiment of the present invention, in the network outage response and edge prediction step, the prediction method for the network outage duration is as follows: based on the duration records of historical network outage events, the average and standard deviation of the historical durations are calculated, and the sum of the average and the safety factor multiplied by the standard deviation is used as the predicted value of the current network outage duration; the local edge prediction model adopts an autoregressive model, and the autoregressive model uses the nearest effective vibrating string frequency data point before the network outage to update the model parameters online through the recursive least squares method; The preset conditions for triggering the reporting of predicted values ​​include: the duration of network outage exceeds the predicted network outage duration, the local cache space utilization exceeds the preset upper limit threshold, and the current data change gradient is lower than the preset stable threshold, indicating that the data is in a stable change range; when any of the above conditions are met, the system reports the predicted value generated by the edge prediction model or the predicted value exceeding the limit alarm event through the backup communication channel.

[0011] According to another preferred embodiment of the present invention, the method further includes a data compression step during network outage: when the total amount of data cached locally during network outage exceeds a preset cache threshold, lossy compression is performed on data packets that meet the conditions of low change gradient and low time value. The compression method is to take the arithmetic mean of multiple vibrating string frequency values ​​within a preset time window as the representative value after compression, and at the same time record the maximum and minimum values ​​within the window as compression error boundaries.

[0012] According to another preferred embodiment of the present invention, the method for calculating the transmission urgency of the data packet includes: assigning weights to three factors—data change gradient, data timeliness value, and queue waiting position—and then performing a weighted summation, wherein the data timeliness value adopts a function form that decays exponentially with time, and the queue waiting position is mapped to a factor value between 0 and 1 through an S-shaped function; the retransmission order is arranged from high to low transmission urgency, prioritizing the retransmission of data packets with drastic changes, high timeliness, and long waiting time in the queue; The process before retransmission includes an expired data eviction method: calculating the expiration value of each cached data packet, discarding expired data packets with expiration values ​​lower than a preset minimum retention threshold, retaining only their statistical summary information, and the discarded expired data packets do not participate in subsequent retransmission queuing.

[0013] According to another preferred embodiment of the present invention, the method further includes a transmission window prediction and precise delivery step: based on the predicted value of network health, it is determined whether the continuous stability condition is met at each future time to identify the available transmission window, and the duration of the available transmission window is predicted; during the retransmission process after network recovery, data delivery is only performed within the predicted available transmission window, and when the network health is detected to drop below the transmission threshold, delivery is paused and the system waits for the next available transmission window.

[0014] According to another preferred embodiment of the present invention, when the calculated value of the value function is lower than the preset acquisition trigger threshold, the system skips the data acquisition operation of the current acquisition cycle and extends the acquisition interval; when the calculated value of the value function is not lower than the preset acquisition trigger threshold, the acquisition operation is performed according to the adjusted acquisition frequency, and the acquired data is stored in the local cache queue in a hierarchical manner according to the magnitude of the data change gradient. Data with a change gradient higher than the first preset threshold is marked as high priority, and data with a change gradient lower than the second preset threshold is marked as low priority. After the retransmission is completed, the method further includes generating data integrity verification information based on the number of retransmitted data packets, the timestamp coverage range, and the compression error boundary, and reporting the verification information to the cloud along with the retransmitted data. The cloud verifies the integrity and accuracy of the vibrating string monitoring data during the network outage based on the verification information.

[0015] To achieve at least one of the above-mentioned objectives, the present invention further provides an adaptive vibrating wire data acquisition system based on a DTU, wherein the system executes the aforementioned adaptive vibrating wire data acquisition method based on a DTU.

[0016] The present invention further provides a computer-readable storage medium storing a computer program, which is executed by a processor to implement the above-described adaptive vibrating string data acquisition method based on a DTU. Attached Figure Description

[0017] Figure 1 The diagram shown is a flowchart of an adaptive vibrating wire data acquisition method based on DTU according to the present invention. Detailed Implementation

[0018] The following description is intended to disclose the present invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art. The basic principles of the invention defined in the following description can be applied to other embodiments, modifications, improvements, equivalents, and other technical solutions that do not depart from the spirit and scope of the invention.

[0019] It is understood that the term "a" should be understood as "at least one" or "one or more," that is, in one embodiment, the number of an element can be one, while in another embodiment, the number of the element can be multiple, and the term "a" should not be understood as a limitation on the number.

[0020] Please combine Figure 1 This invention provides an adaptive vibrating wire data acquisition method and system based on a DTU, wherein the method mainly includes the following steps: The DTU collects frequency data from the vibrating wire sensor in real time and continuously monitors the current network status indicators, which include at least signal strength, signal-to-noise ratio, packet loss rate, and round-trip time. The instantaneous network health is calculated based on the current network status indicators, and the network health at multiple future moments is predicted based on the historical network health sequence. The comprehensive network health prediction value used for data collection and decision-making is obtained by combining the prediction values ​​at each moment. The data change gradient is calculated based on the vibrating string frequency data at the current acquisition time and historical times. The data change gradient is used to characterize the drastic degree of change in the vibrating string data. The comprehensive network health prediction value is fused with the data change gradient to construct an adaptive acquisition value function. The adaptive acquisition value function includes a network state dominant term, a data change dominant term, and a synergistic gain term of the two. The acquisition frequency of the next acquisition cycle is dynamically adjusted or the current acquisition cycle is skipped based on the calculation result of the adaptive acquisition value function. When the network health is lower than the network disconnection threshold, the network disconnection mode is entered. During the network disconnection, the local edge prediction model is activated to generate the predicted value of the vibrating string frequency, and the cache status and network disconnection duration are monitored. When the preset conditions are met, the predicted value is reported through the backup channel or a critical event alarm is triggered. Once the network is restored to a usable state, the data packets cached during the network outage are sorted according to their transmission urgency and retransmitted in sequence. The transmission urgency is calculated by the data change gradient, the data timeliness value, and the queue waiting position. After the retransmission is completed, a data integrity verification report is reported.

[0021] Specifically, this invention requires the deployment of a vibrating wire sensor and a DTU (Data Transfer Unit) acquisition terminal for multi-channel vibrating wire signal acquisition. Furthermore, this invention requires the initialization of a monitoring module to pre-set the acquisition period t and the sliding window size K. Based on the vibrating wire sensor and DTU, this invention acquires the following multi-dimensional network status data of vibrating wire signal transmission within each acquisition period t: signal strength RSST(t), in dBm; signal-to-noise ratio SNR(t), in dB; packet loss rate PLR(t), in %; round-trip time RTT(t), in ms; and data transmission rate Rate(t), in bps. Based on the above multi-dimensional network status data, this invention performs the following normalization processing: The normalization method for signal strength RSST(t) is as follows: ; The normalization method for the signal-to-noise ratio (SNR) (t) is as follows: ; The normalization method for the packet loss rate PLR(t) is as follows: ; The normalization method for the round-trip time RTT(t) is as follows: ; The normalization method for the data transmission rate Rate(t) is as follows: ; Based on the above-mentioned normalized network state data, this invention performs the following instantaneous network health calculation: ; Where the above corresponds to , , , and These are the corresponding weight values, where the sum of all the above weight values ​​is 1.

[0022] Furthermore, this invention is based on the instantaneous network health sequence of k historical moments. A combined model of first-order exponential smoothing and linear trend is used to predict the future. The network health status at any given time is calculated as follows: The horizontal component is calculated based on the first-order exponential smoothing: ;in This represents the network health level component at time t. This represents the network health level component at the previous time t-1. The level component represents the baseline level of network health after smoothing and filtering, which is the long-term trend benchmark after removing random fluctuations. The smoothing coefficient for the horizontal component ranges from (0,1). The larger the value, the higher the current instantaneous network health observation value. The more significant the impact on the level estimate, the higher the sensitivity of the level component's response. This represents the trend component of the previous time step t-1.

[0023] The formula for calculating the corresponding trend component is: ;in The trend component at time t represents the monotonic change direction and rate of network health over time (positive values ​​indicate an upward trend, and negative values ​​indicate a downward trend), which can be understood as a slope change. This is the smoothing coefficient for the trend component, with a value ranging from (0,1). A larger value indicates a greater horizontal change. The greater the impact on trend estimation, the better. Further, proceed according to the following formula... Step prediction: ;in Let be the prediction step size, and represent the corresponding future data collection periods. The formula for calculating the comprehensive network health prediction value based on prediction windows across multiple data collection periods is as follows: Where P is the prediction window size, For the decay weight, where The aforementioned attenuation weights can effectively configure recent forecasts to have higher confidence levels, thereby effectively reducing the impact of forecast timeliness on the forecast results.

[0024] Furthermore, this invention performs data importance calculation on vibrating wire data. This invention requires calculating the gradient of vibrating wire data changes and calculating the attenuation of data time-related value. The method for calculating the gradient of vibrating wire data changes includes: obtaining the frequency value collected by the vibrating wire sensor at time t. Calculate the instantaneous rate of change of the frequency value: ; Obtain the frequency values ​​of m sampling points under the sliding window Then, the sliding window gradient is further calculated: ; Where i represents the offset index within the sliding window, and M represents the length of the sliding window. This means taking the maximum value among all adjacent rates of change within the window. The normalized gradient is then calculated using the following formula: ;in This represents the preset gradient threshold. .when The closer the value is to 1, the more drastic the data changes and the higher the value of the data collection.

[0025] Furthermore, this invention can also calculate the function of time-related value decay of collected data: ;in This is the time-related decay coefficient. The time the data remains in the local cache, wherein the time decay coefficient is... The settings will vary depending on the scenario. In one preferred embodiment of the present invention, the data collection timeliness value decay function can be included in the overall value function calculation, or it can be excluded from the separate calculation; this will not be elaborated upon further in this invention.

[0026] It is worth mentioning that this invention calculates the acquisition value function based on the comprehensive network health and data change gradient, and constructs an adaptive sampling decision based on the acquisition value function. The specific method includes: calculating the acquisition value function according to the following formula: In the above formula, , representing the network sensitivity coefficient. These are the cross-term weight coefficients. This represents the overall network health prediction value. The term indicates an exponential scaling of the overall network health prediction value. The term represents complementary exponential scaling of the data change gradient. The physical meaning of the above acquisition value function is: when α is large, the acquisition decision depends more on the network state (network-first strategy); when α is small, the acquisition decision depends more on data change (data-first strategy); where the interaction term... This is used to reflect the synergistic gain effect between network status and data. It should be noted that, in one preferred embodiment of the present invention, the acquisition value function... It can be used to calculate the degree of time-related value decay function. Linear multiplication yields a collection value function that is more timely.

[0027] Furthermore, based on the collection value function Dynamically adjust the sampling frequency for the next cycle. ;in and These are the system's preset highest and lowest sampling frequencies. Configure the sampling value function threshold. ;when The sampling is triggered when the current period is reached; otherwise, the current period is skipped. The specific adaptive sampling algorithm flow of this invention includes: S01. In each decision cycle t, collect network status indicators and calculate the instantaneous network health. And update the smoothing model parameters. and ; S02, Predicting the Future Network health at any given moment Calculate the comprehensive network health prediction value ; S03, Read the latest frequency value of the vibrating wire sensor Calculate the gradient of change ; S04. Calculate the acquisition value function ,like Then execute the data collection action, with a collection frequency of [value missing]. Otherwise, skip this data collection and enter the waiting state; S05. The collected data is stored in the local cache queue according to the following priority: High priority: Dramatic data changes; Medium priority: Moderate changes in data; Low priority: Stable data; Furthermore, this invention also performs data resumption and data synchronization for interrupted data. The specific method includes: firstly, predicting the duration of the disconnection; when the comprehensive network health prediction value is detected... hour, This invention represents a comprehensive network health threshold indicating an impending network outage. Based on historical outage patterns, the system predicts the duration of this outage. : ; in This represents the duration of the i-th historical network outage, where N is the number of outages. For safety reasons, This indicates the average duration of historical internet outages.

[0028] Furthermore, this invention requires the construction of data compression and queuing strategies, wherein this invention will consider the comprehensive transmission urgency based on data packet p. A data compression and arrangement strategy is constructed, wherein the overall transmission urgency of the data packet p is considered. The calculation formula is: ; in Let be the gradient of the data packet p at the acquisition time. This indicates the timeliness value of data packet p. This indicates the waiting position of data packet p in the queue. , , These represent the weight coefficients of the corresponding items. and These represent the shape parameters of the sigmoid function. Indicates the steepness coefficient. Indicates the horizontal center position.

[0029] The corresponding compression strategy is: when the amount of cached data... Exceeding the preset threshold At this time, lossy compression is required for low-priority data packets: ; This indicates the number of data points within a time window; that is, the average of the frequency values ​​within the time window is taken, and the maximum value within the time window is recorded. and minimum value As an error boundary.

[0030] The present invention further performs the following window availability determination: if and only if the combined network health prediction value of two consecutive sampling points is... If the data transmission availability exceeds a preset threshold, the system is considered to be entering a stable transmission window. Within this window, data transmission is prioritized based on the urgency of the data packets. From high to low priority, abnormal delivery is prioritized. For high-priority data that has not been delivered within the available window due to timeout, such as... When this happens, the edge prediction model is triggered to report predicted values ​​to replace the original data.

[0031] During the network outage, the system activates a local lightweight edge prediction model, using an autoregressive (AR) model to predict the vibrating string frequency in the short term. ;in The model order is... is the autoregressive coefficient, c is the constant term, and the edge prediction model parameters are updated online using the L most recent valid data points before the network outage. Since the autoregressive (AR) model itself is existing technology, this invention will not elaborate on it further.

[0032] This invention requires that any one of the following conditions be met to trigger the reporting of predicted values: 1. Duration of network outage ; 2. Cache space utilization > 90%; 3. Current data change gradient . Furthermore, the present invention implements the following retransmission strategy after network outage recovery: S01. Calculate the time value of each data packet in the cache. and the original gradient ; S02, Calculate the overall transmission urgency ,according to Sort in descending order; S03, to Expired data should be discarded directly. S04. For the remaining data packets, deliver them in descending order of overall transmission urgency, and dynamically monitor the overall network health prediction value during the delivery process. When the comprehensive network health prediction value If so, delivery will be paused and the system will wait for the next transmission window; S05. After the retransmission is completed, send a data integrity verification report to the cloud.

[0033] in This indicates the lower limit threshold for timeliness. This indicates the transmission availability threshold.

[0034] The processes described in the flowcharts above, as disclosed in the embodiments of this invention, can be implemented as computer software programs. The embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication component, and / or installed from a removable medium. When the computer program is executed by a central processing unit (CPU), the methods of this application are not limited to the aforementioned functions. It should be noted that the computer-readable medium described above in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more conductor segments, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0035] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation that may be implemented in systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0036] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functions and structural principles of the present invention have been shown and explained in the embodiments. Without departing from the stated principles, the implementation of the present invention may have any variations or modifications.

Claims

1. An adaptive vibrating wire data acquisition method based on DTU, characterized in that, Includes the following steps: The DTU collects frequency data from the vibrating wire sensor in real time and continuously monitors the current network status indicators, which include at least signal strength, signal-to-noise ratio, packet loss rate, and round-trip time. The instantaneous network health is calculated based on the current network status indicators, and the network health at multiple future moments is predicted based on the historical network health sequence. The comprehensive network health prediction value used for data collection and decision-making is obtained by combining the prediction values ​​at each moment. The data change gradient is calculated based on the vibrating string frequency data at the current acquisition time and historical times. The data change gradient is used to characterize the drastic degree of change in the vibrating string data. The comprehensive network health prediction value is fused with the data change gradient to construct an adaptive acquisition value function. The adaptive acquisition value function includes a network state dominant term, a data change dominant term, and a synergistic gain term of the two. The acquisition frequency of the next acquisition cycle is dynamically adjusted or the current acquisition cycle is skipped based on the calculation result of the adaptive acquisition value function. When the network health is lower than the network disconnection threshold, the network disconnection mode is entered. During the network disconnection, the local edge prediction model is activated to generate the predicted value of the vibrating string frequency, and the cache status and network disconnection duration are monitored. When the preset conditions are met, the predicted value is reported through the backup channel or a critical event alarm is triggered. Once the network is restored to a usable state, the data packets cached during the network outage are sorted according to their transmission urgency and retransmitted in sequence. The transmission urgency is calculated by the data change gradient, the data timeliness value, and the queue waiting position. After the retransmission is completed, a data integrity verification report is reported.

2. The adaptive vibrating wire data acquisition method based on DTU according to claim 1, characterized in that, The method for calculating the instantaneous network health includes: converting signal strength, signal-to-noise ratio, packet loss rate, and round-trip delay into component values ​​in the range of 0 to 1 through their respective normalization mapping functions, and then weighting and summing the component values ​​to obtain the instantaneous health; the prediction of network health at multiple future moments adopts a combination model of exponential smoothing and linear trend, and multi-step prediction is achieved through the recursive update of horizontal components and trend components. The method for calculating the data change gradient includes: calculating the instantaneous rate of change of the vibrating string frequency at the current moment relative to the previous moment; further taking the product of the maximum value and the average value of each instantaneous rate of change within the sliding time window as the original gradient value; and then normalizing the original gradient value relative to a preset threshold to obtain a normalized data change gradient with a value range of 0 to 1.

3. The adaptive vibrating wire data acquisition method based on DTU according to claim 1, characterized in that, The adaptive acquisition value function is constructed by: multiplying the comprehensive network health prediction value and the data change gradient by taking powers of each, forming a coupled form of network state dominant term and data change dominant term, and adding a product term of comprehensive network health prediction value and data change gradient as a cooperative gain term to the product of the two terms; the acquisition frequency is adjusted by: linearly mapping between the preset minimum acquisition frequency and the maximum acquisition frequency according to the calculated value of the value function, the higher the value function value, the closer the acquisition frequency is to the maximum acquisition frequency.

4. The adaptive vibrating wire data acquisition method based on DTU according to claim 1, characterized in that, In the network outage response and edge prediction steps, the prediction method for network outage duration is as follows: based on the duration records of historical network outage events, calculate the average and standard deviation of the historical duration, and multiply the average and safety factor by the standard deviation to obtain the predicted value of the current network outage duration; the local edge prediction model adopts an autoregressive model, which uses the most recent effective vibrating string frequency data point before the network outage to update the model parameters online through the recursive least squares method; The preset conditions for triggering the reporting of predicted values ​​include: the duration of network outage exceeds the predicted network outage duration, the local cache space utilization exceeds the preset upper limit threshold, and the current data change gradient is lower than the preset stable threshold, indicating that the data is in a stable change range; when any of the above conditions are met, the system reports the predicted value generated by the edge prediction model or the predicted value exceeding the limit alarm event through the backup communication channel.

5. The adaptive vibrating wire data acquisition method based on DTU according to claim 1, characterized in that, The method also includes a data compression step during network outage: when the total amount of data cached locally during network outage exceeds a preset cache threshold, lossy compression is performed on data packets that meet the conditions of low change gradient and low time value. The compression method is to take the arithmetic mean of multiple vibrating string frequency values ​​within a preset time window as the representative value after compression, and at the same time record the maximum and minimum values ​​within the window as compression error boundaries.

6. The adaptive vibrating wire data acquisition method based on DTU according to claim 1, characterized in that, The method for calculating the transmission urgency of the data packets includes: assigning weights to three factors—data change gradient, data timeliness value, and queue waiting position—and then performing a weighted summation. The data timeliness value is expressed as a function that decays exponentially with time, and the queue waiting position is mapped to a factor value between 0 and 1 using an S-shaped function. The retransmission order is arranged from high to low transmission urgency, prioritizing the retransmission of data packets that change drastically, have high timeliness, and have been waiting in the queue for a long time. The process before retransmission includes an expired data eviction method: calculating the expiration value of each cached data packet, discarding expired data packets with expiration values ​​lower than a preset minimum retention threshold, retaining only their statistical summary information, and the discarded expired data packets do not participate in subsequent retransmission queuing.

7. The adaptive vibrating wire data acquisition method based on DTU according to claim 1, characterized in that, The method also includes transmission window prediction and precise delivery steps: based on the predicted value of network health, it is determined whether the continuous stability condition is met at each future time to identify available transmission windows, and the duration of available transmission windows is predicted. During the retransmission process after network recovery, data delivery is only performed within the predicted available transmission window. When the network health is detected to drop below the transmission threshold, delivery is paused and the system waits for the next available transmission window.

8. The adaptive vibrating wire data acquisition method based on DTU according to claim 1, characterized in that, When the calculated value of the value function is lower than the preset acquisition trigger threshold, the system skips the data acquisition operation of the current acquisition cycle and extends the acquisition interval; when the calculated value of the value function is not lower than the preset acquisition trigger threshold, the acquisition operation is performed according to the adjusted acquisition frequency, and the acquired data is stored in the local cache queue according to the magnitude of the data change gradient. Data with a change gradient higher than the first preset threshold is marked as high priority, and data with a change gradient lower than the second preset threshold is marked as low priority. After the retransmission is completed, the method further includes generating data integrity verification information based on the number of retransmitted data packets, the timestamp coverage range, and the compression error boundary, and reporting the verification information to the cloud along with the retransmitted data. The cloud verifies the integrity and accuracy of the vibrating string monitoring data during the network outage based on the verification information.

9. An adaptive vibrating wire data acquisition system based on a DTU, characterized in that, The system executes an adaptive vibrating wire data acquisition method based on a DTU as described in any one of claims 1-8.

10. A computer-readable storage medium storing a computer program that is executed by a processor to implement the adaptive vibrating wire data acquisition method based on a DTU as described in any one of claims 1-8.

Images