Trestle stress monitoring method and system based on mechanical sensor

Abstract

The invention relates to the technical field of sensors, and provides a trestle stress monitoring method and system based on a mechanical sensor. The method comprises the steps of obtaining trestle stress monitoring flow through a stress monitoring network, performing health risk prediction, and constructing a multi-node trestle risk curve; constructing a monitoring deployment diagram of each trestle area according to the stress monitoring network, performing reliability evaluation, and establishing a reliability evaluation space; on the basis of the space, optimization guide feature analysis is carried out on the stress monitoring network according to constraints, and optimization guide features are generated; performing deployment adjustment on the stress monitoring network according to the characteristics to obtain an adjustment space; and performing optimization analysis according to the adjustment space, generating a stress monitoring deployment optimization strategy, and performing optimization adjustment. The technical problems that the monitoring result is inaccurate and the health condition of the trestle cannot be reflected in real time due to insufficient deployment optimization of the mechanical sensor in trestle stress monitoring are solved, and the technical effect of improving the reliability and accuracy of monitoring deployment of each area of the trestle is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of sensors, in particular to a trestle stress monitoring method and system based on mechanical sensors. BACKGROUND

[0002] As an important infrastructure, trestles are widely used in coal mines, power, wind power and other fields, bearing huge weight and complex external working conditions. The stability of the trestle structure is directly related to the safe operation and economic benefits of related equipment. Traditional trestle health monitoring mainly relies on manual inspection or regular structural detection, which often has problems such as reaction lag, low efficiency and inability to real-time feedback of the trestle health status. With the development of sensor technology, especially the continuous progress of mechanical sensors, the structure health monitoring system based on mechanical sensors has gradually become a research hotspot. Mechanical sensors can monitor physical quantities such as stress and strain of the trestle in real time, and analyze through a data processing system, so as to effectively evaluate the health status of the trestle. However, the existing stress monitoring system has the following technical problems: on the one hand, the sensor deployment and data collection scheme of each area of the trestle lacks scientific optimization, resulting in poor monitoring effect; on the other hand, with the increase of monitoring data, how to process and analyze large-scale sensor data to improve the intelligence and reliability of the monitoring system is still a problem to be solved. SUMMARY

[0003] The present application provides a trestle stress monitoring method and system based on mechanical sensors, aiming to solve the technical problem of inaccurate monitoring results and inability to reflect the health status of the trestle in real time due to insufficient optimization of sensor deployment in trestle stress monitoring, and to achieve the technical effect of improving the reliability and accuracy of mechanical sensor deployment in each area of the trestle through multi-scale health risk prediction and optimization guided feature analysis, and realizing more effective trestle health status evaluation and risk warning.

[0004] In a first aspect, a stack stress monitoring method based on a mechanical sensor is provided, the method comprising: obtaining a stack stress monitoring stream through a distributed stress monitoring network, and performing multi-scale stack health risk prediction according to the stack stress monitoring stream to construct a multi-node stack risk curve; constructing a stack area monitoring deployment map according to the distributed stress monitoring network, and performing reliability evaluation on the stack area monitoring deployment map according to the multi-node stack risk curve to establish a deployment reliability evaluation space; based on the deployment reliability evaluation space, performing optimization guided feature analysis on the distributed stress monitoring network according to deployment reliability evaluation constraints to generate deployment optimization guided features; performing deployment adjustment on the distributed stress monitoring network according to the deployment optimization guided features to obtain a monitoring deployment adjustment space; performing multi-level optimization analysis according to the monitoring deployment adjustment space to generate a stress monitoring deployment optimization strategy, and performing optimization adjustment on the distributed stress monitoring network according to the stress monitoring deployment optimization strategy.

[0005] In another aspect, a stack stress monitoring system based on a mechanical sensor is provided, the system comprising: a health risk prediction module: obtaining a stack stress monitoring stream through a distributed stress monitoring network, and performing multi-scale stack health risk prediction according to the stack stress monitoring stream to construct a multi-node stack risk curve; a reliability evaluation module: constructing a stack area monitoring deployment map according to the distributed stress monitoring network, and performing reliability evaluation on the stack area monitoring deployment map according to the multi-node stack risk curve to establish a deployment reliability evaluation space; a feature analysis module: based on the deployment reliability evaluation space, performing optimization guided feature analysis on the distributed stress monitoring network according to deployment reliability evaluation constraints to generate deployment optimization guided features; a deployment adjustment module: performing deployment adjustment on the distributed stress monitoring network according to the deployment optimization guided features to obtain a monitoring deployment adjustment space; an optimization adjustment module: performing multi-level optimization analysis according to the monitoring deployment adjustment space to generate a stress monitoring deployment optimization strategy, and performing optimization adjustment on the distributed stress monitoring network according to the stress monitoring deployment optimization strategy.

[0006] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0007] The stack stress monitoring method based on the mechanical sensor first acquires a stress data stream of the stack in real time through a distributed stress monitoring network, and performs multi-scale health risk prediction based on the data stream to construct a risk curve of each node of the stack. Subsequently, a monitoring deployment graph of each region of the stack is constructed, and reliability evaluation is performed in combination with the risk curve to form a deployment reliability evaluation space. On this basis, the sensor deployment of the distributed monitoring network is optimized according to the constraint condition of the reliability evaluation, an optimized guiding feature is generated, and the network deployment is adjusted, and finally an optimized monitoring deployment space is obtained. Then, through multi-level optimization analysis, an optimal strategy of stress monitoring deployment is generated, and the distributed stress monitoring network is adjusted accordingly to improve the health monitoring effect and reliability of the stack.

[0008] The above description is only a summary of the technical solutions of the present application, in order to more clearly understand the technical means of the present application, the specific embodiments of the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creating laborious work.

[0010] Figure 1 A flowchart of a stack stress monitoring method based on a mechanical sensor in an embodiment.

[0011] Figure 2 A stack stress monitoring system architecture diagram based on a mechanical sensor in an embodiment.

[0012] Legend: health risk prediction module 11, reliability evaluation module 12, feature analysis module 13, deployment adjustment module 14, and optimization adjustment module 15. DETAILED DESCRIPTION

[0013] The embodiments of the present application provide a stack stress monitoring method and system based on a mechanical sensor, solve the technical problem of inaccurate monitoring results and inability to reflect the health status of the stack in real time due to insufficient optimization of sensor deployment in stack stress monitoring, and achieve the technical effect of improving the reliability and accuracy of mechanical sensor deployment in each region of the stack through multi-scale health risk prediction and optimization guided feature analysis, and realizing more effective stack health status evaluation and risk warning.

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0015] It should be noted that the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such process, method, product, or device.

[0016] Example 1, as Figure 1 As shown, this application provides a method for monitoring the stress of a trestle bridge based on a mechanical sensor, the method comprising:

[0017] The stress monitoring flow of the trestle is obtained through a distributed stress monitoring network, and multi-scale health risk prediction of the trestle is performed based on the stress monitoring flow to construct a multi-node trestle risk curve.

[0018] In this embodiment, multiple mechanical sensors (such as strain gauges and accelerometers) distributed across key parts of the trestle in a distributed stress monitoring network collect stress data streams of the trestle in real time. These data streams include stress and acceleration, reflecting stress changes at various points on the trestle under different operating conditions. Subsequently, based on the collected stress monitoring streams, a comprehensive assessment of the trestle's health status is performed using a preset multi-scale window, predicting potential stress changes over multiple future time windows. Then, the predicted stress changes are used to assess the trestle's health risk, generating a risk curve for each node. These curves are then aggregated to form a multi-node trestle risk curve. This multi-node risk curve accurately represents the risk level of each node, helping to assess the trestle's stability under various operating conditions. In this way, the stress state of the trestle is continuously tracked, and early warnings are provided before problems occur, thus providing real-time data support for maintenance and management.

[0019] Furthermore, this application provides a method for multi-scale trestle health risk prediction based on the trestle stress monitoring flow, and for constructing a multi-node trestle risk curve, including:

[0020] According to the pier stress monitoring flow, the n th node stress monitoring data corresponding to the n th mechanical sensor is extracted, n is a positive integer greater than 1; according to the multi-scale window, the trend of the n th node stress monitoring data is predicted, and the n th node stress prediction set is obtained, and the multi-scale window includes a plurality of future time windows; according to the n th node stress prediction set, the pier health risk prediction is carried out, and the n th node pier risk curve is generated.

[0021] Preferably, the stress data of the n th node of the pier is collected in real time through the distributed stress monitoring network, which is provided by the mechanical sensor (such as strain gauge, acceleration sensor, etc.) installed on the n th node. Each sensor continuously monitors the stress change of the position, generates real-time stress data stream, including stress value, acceleration and other parameters. After obtaining the real-time stress monitoring data of the n th node, the multi-scale window method is used for trend prediction. Specifically, the stress data is divided into a plurality of different time windows, each time window corresponds to a different prediction range, for example, short-term prediction (such as 1 hour to 6 hours), medium-term prediction (such as 1 day to 1 week) and long-term prediction (such as 1 month to 1 year), wherein the short-term prediction window focuses on instantaneous stress fluctuation, the medium-term prediction window focuses on longer-term fluctuation and trend change, and the long-term prediction window focuses on long-term trend of stress accumulation. Then, the divided plurality of monitoring data is respectively input into the corresponding long short-term memory network (LSTM) for stress prediction under the corresponding window. These long short-term memory networks are trained based on historical stress monitoring data. During training, the corresponding time window is used for dividing the training data, and the LSTM is gradually converged through forward propagation, loss calculation, back propagation and parameter optimization. Through the prediction of the LSTM corresponding to different scales, the stress prediction set of the n th node under each scale can be obtained to show the stress change in a plurality of future time windows. Then, the n th node stress prediction set is used to predict the health risk of the pier. In this process, the mean and standard deviation of the n th node stress prediction set in each future time window are calculated, and then the stress prediction value is subtracted from the mean and divided by the standard deviation to obtain the risk value at each time in each future time window. Then, the risk values of different future time windows at the same time are weighted and summed to obtain the health risk value at that time. By collecting these health risk values in chronological order, the pier health risk prediction set of the n th node is obtained. Finally, the pier health risk prediction set of the n th node is processed into a curve to generate a pier risk curve. The risk curve of each node shows the health risk level of the position at different time scales, helping to predict the possible structural problems (such as fatigue damage, stress concentration, etc.) that the pier may face in the future, and providing a basis for maintenance decision.

[0022] According to the distributed stress monitoring network construction stack each area monitoring deployment diagram, and according to the multi-node stack risk curve, the reliability of the stack each area monitoring deployment diagram is evaluated, and a deployment reliability evaluation space is established.

[0023] In one embodiment, the specific deployment positions of each mechanical sensor are obtained from the distributed stress monitoring network, and the stack is labeled according to these deployment positions to form a stack each area monitoring deployment diagram. The stack each area monitoring deployment diagram shows the monitoring point positions and corresponding sensor configurations of different regions of the stack, ensuring effective monitoring of each critical part of the stack. Subsequently, the health risk level of each monitoring region is analyzed using the multi-node stack risk curve constructed earlier. These risk curves reflect the stress changes and potential risk factors of the stack each area under different working conditions. By combining these risk curves with the corresponding region monitoring deployment diagram, the reliability of each region is evaluated to determine whether the monitoring deployment can accurately reflect the health status of the region. Finally, the reliability coefficients obtained through the evaluation are stored to establish a deployment reliability evaluation space. This deployment reliability evaluation space helps to identify which regions of the stack have weak monitoring and which regions have relatively reliable deployment, thereby providing decision-making basis for subsequent optimization deployment and risk management.

[0024] Further, the application provides a method for evaluating the reliability of the stack each area monitoring deployment diagram according to the multi-node stack risk curve, and establishing a deployment reliability evaluation space, including:

[0025] According to the stack each area monitoring deployment diagram, a stack one area monitoring deployment diagram is extracted; according to the stack one area monitoring deployment diagram, the multi-node stack risk curve is arranged, and a stack one area risk diagram is obtained; according to the stack one area risk diagram, a risk position prediction is performed, and a stack one area risk position diagram is established; according to the stack one area risk position diagram, the reliability of the stack one area monitoring deployment diagram is evaluated, a first area deployment reliability coefficient is obtained, and the first area deployment reliability coefficient is added to the deployment reliability evaluation space.

[0026] Preferably, from the monitoring deployment map of each area of the entire pier, the monitoring deployment map of the first area of the pier is extracted, and different sensors are arranged in each area of the pier. The monitoring deployment map of the first area of the pier shows the positions of all monitoring points in the first area of the pier and the corresponding sensor information. The arrangement position of each sensor determines the collection range and monitoring accuracy of the stress of the area. Subsequently, the monitoring deployment map of the first area of the pier is used to extract the data related to the first area of the pier from the multi-node pier risk curve, that is, the pier risk curve of the node belonging to the first area of the pier is extracted. By sorting and classifying the risk information of each monitoring point in the first area of the pier, the corresponding risk information is matched for each monitoring point in the monitoring deployment map of the first area of the pier, thereby forming a first area of the pier risk map to reflect the stress distribution and possible potential problems of the first area of the pier under different working conditions. Then, the pier risk curve of each monitoring point is obtained from the first area of the pier risk map, and the pier risk curve is analyzed using a risk threshold to determine the predicted time length of each monitoring point exceeding the threshold. If the time length is greater than or equal to the risk duration threshold, it indicates that the monitoring point has a greater risk, and the monitored area is marked at this time. At the same time, it is also checked whether the maximum risk in the pier risk curve exceeds the risk tolerance threshold. If it exceeds, it is judged whether the duration exceeding the risk tolerance threshold is greater than or equal to the limit risk duration threshold. If it exceeds, the monitored area is also marked. By summarizing these marked areas, the first area of the pier risk position map is predicted, which details the risk situation of each position in the first area of the pier and the possible dangerous points. After obtaining the first area of the pier risk position map, the monitoring deployment map of the first area of the pier is evaluated for reliability. In this process, the first area of the pier risk position map and the monitoring deployment map of the first area of the pier are input into the deployment reliability evaluation global channel for analysis, and the first area deployment reliability coefficient of the pier is calculated. This reliability coefficient reflects the reliability of the monitoring of the first area of the pier and can reflect whether it is sufficient to detect all potential risks. Finally, the calculated first area deployment reliability coefficient of the pier is added to the deployment reliability evaluation space. This space summarizes the reliability coefficients of different areas to provide a comprehensive evaluation of the monitoring reliability of the entire pier, thereby helping to identify the weaknesses in the monitoring deployment of the pier in some areas and providing data support for subsequent optimization deployment. Figure 1

[0027] Further, the application provides a method for evaluating the reliability of the monitoring deployment map of the first area of the pier according to the first area of the pier risk position map to obtain a first area deployment reliability coefficient, comprising:

[0028] ​According to P local computing nodes, P deployment reliability evaluation history sets are locally trained to generate P deployment reliability evaluation models, P is a positive integer greater than 1; according to the central server, the P deployment reliability evaluation models are federated aggregated to establish a deployment reliability evaluation global channel; the stack bridge region one risk position map and the stack bridge region one monitoring deployment map are input into the deployment reliability evaluation global channel to generate the first region deployment reliability coefficient of the stack bridge region one.

[0029] Optionally, first, in P local computing nodes, each node will use the deployment reliability evaluation history set stored locally for training, which contains the past monitoring data (such as risk position map, monitoring deployment map) of each monitoring region of the stack bridge and the related reliability evaluation results (such as reliability coefficient). Each node learns from these data, iteratively trains the locally stored multi-layer perceptron (MLP) framework, gradually optimizes the MLP through forward propagation, loss calculation (mean square error), back propagation, parameter optimization (Adma) and other steps, until the maximum iteration number or loss convergence is reached, thereby constructing the deployment reliability evaluation model of each local computing node. After the deployment reliability evaluation models of the P local nodes are trained, the central server will federated aggregate all P local deployment reliability evaluation models. Federated aggregation means combining the knowledge and results of multiple local models to form a global model without uploading the original data of each local node to the central server. This way protects the privacy of the data while ensuring information sharing and integration between different nodes. Common aggregation methods include weighted average, model fusion, etc. Through the aggregated global model, a deployment reliability evaluation global channel is established, which can process and transmit reliability evaluation information from different regions and nodes. Then, the risk position map and the monitoring deployment map of the stack bridge region one are input into the deployment reliability evaluation global channel. The stack bridge region one risk position map shows the risk points in the region, and the monitoring deployment map shows the layout of the sensors in the region. By inputting these two maps into the global channel, the relationship between the risk region and the sensor deployment can be analyzed and compared, and the coverage ability and effectiveness of monitoring in the specific risk region can be evaluated. After the deployment reliability evaluation global channel processes the input data, the first region deployment reliability coefficient of the stack bridge region one is generated. This first region deployment reliability coefficient reflects the reliability level of the monitoring deployment in the stack bridge region one, including whether the arrangement of sensors in the region can effectively monitor the risk position, the stability of the deployment scheme and the adaptability to different working conditions. Finally, this coefficient will be used to evaluate the overall performance of the monitoring of the stack bridge region one and as part of the deployment reliability evaluation space for subsequent optimization and adjustment decisions.

[0030] Based on the deployment reliability evaluation space, the distributed stress monitoring network is analyzed according to deployment reliability evaluation constraints to generate deployment optimization guide features.

[0031] In one embodiment, in the deployment reliability evaluation space, a plurality of reliability coefficients have been formed by reliability evaluation of each monitoring area. By comparing the reliability coefficients with deployment reliability evaluation constraints, it is determined that there are abnormal areas in the deployment evaluation. Then, by weak backtracking and over-dense backtracking of these abnormal areas, the specific reasons for the abnormal areas are determined, such as insufficient number of sensors or unreasonable layout. Then, by fusing the results of the backtracking, deployment optimization guide features are configured for these abnormal areas, which provide corresponding adjustment directions and provide scientific basis for subsequent network optimization and deployment adjustment.

[0032] Further, the application provides that based on the deployment reliability evaluation space, the distributed stress monitoring network is analyzed according to deployment reliability evaluation constraints to generate deployment optimization guide features, including:

[0033] According to the deployment reliability evaluation constraints, the deployment reliability evaluation space is detected for abnormality to determine a deployment evaluation abnormal distribution. According to the deployment evaluation abnormal distribution, deployment weak correlation backtracking is performed on the distributed stress monitoring network to obtain a first backtracking result of deployment abnormality. According to the deployment evaluation abnormal distribution, deployment over-dense correlation backtracking is performed on the distributed stress monitoring network to obtain a second backtracking result of deployment abnormality. The first backtracking result of deployment abnormality and the second backtracking result of deployment abnormality are fused to generate the deployment optimization guide features.

[0034] Preferably, in the deployment reliability evaluation space, firstly, the deployment reliability coefficients of each monitoring area are checked based on the preset deployment reliability evaluation constraints. If the deployment reliability coefficient of a certain area does not meet the constraint condition, the deployment of this area is marked as abnormal. These areas that do not meet the standard are added to the deployment evaluation abnormal distribution to form an abnormal distribution set to identify which areas of the monitoring deployment need to be optimized. Subsequently, based on the deployment evaluation abnormal distribution, the deployment weak correlation of the distributed stress monitoring network is traced back. In this process, the abnormal area is located from the distributed stress monitoring network, and the sensor density around the abnormal area is counted. If the sensor density around the abnormal area is less than the preset density threshold (such as the number of nodes within 3m radius < 2), it indicates that the abnormality of this area is caused by weak deployment. At this time, the abnormal area is assigned a deployment weak label and recorded in the first tracing result of deployment abnormality. In addition, the deployment over-dense correlation of the distributed stress monitoring network is also traced back, that is, the sensor coverage overlap degree of the abnormal area is counted. For the sensor coverage overlap degree, the center distance can be calculated using the coordinates of adjacent sensor nodes, and then the center distance, major axis radius and minor axis radius are calculated through the approximate solution of ellipse intersection to obtain the overlapping area. Dividing the overlapping area by the effective monitoring area can obtain the sensor coverage overlap degree. If the sensor coverage overlap degree is greater than or equal to the preset overlap degree (such as the monitoring range overlap rate of nodes B2 and B3 > 60%), it indicates that the abnormality of this area is caused by over-dense deployment. At this time, the abnormal area is assigned an over-dense deployment label and recorded in the second tracing result of deployment abnormality. Then, the first tracing result of deployment abnormality and the second tracing result of deployment abnormality are fused, that is, the two tracing results are combined to form the final deployment optimization guide features. These features include the optimization direction of each abnormal area, which can ensure that the distributed stress monitoring network can efficiently and accurately perform the monitoring task in each area.

[0035] According to the deployment optimization guide features, the deployment of the distributed stress monitoring network is adjusted to obtain a monitoring deployment adjustment space.

[0036] In one embodiment, the deployment optimization guidance feature is generated to guide the adjustment of the distributed stress monitoring network. Specifically, the deployment optimization guidance feature provides specific directions on how to improve the monitoring deployment, such as which areas need to increase sensor layout, which areas need to reduce sensor layout, etc. Then, based on the specific directions in the deployment optimization guidance feature, a plurality of specific adjustment schemes are generated, for example, in areas where mechanical sensors need to be added, sensors are added in uncovered areas, in areas where sensors need to be removed, sensors with high overlap are removed and only one is kept, after removal, if the coverage cannot meet the actual needs, mechanical sensors are added in uncovered areas. By applying these adjustment schemes to the distributed stress monitoring network, a monitoring deployment adjustment space is formed, which contains all possible deployment adjustment schemes, aiming to improve the efficiency and reliability of monitoring by improving the layout and configuration, providing multiple possible paths for subsequent optimization process to ensure that the stress monitoring network of the trestle can work stably and accurately under various working conditions.

[0037] According to the monitoring deployment adjustment space, multi-level optimization analysis is performed to generate a stress monitoring deployment optimization strategy, and the distributed stress monitoring network is optimized and adjusted according to the stress monitoring deployment optimization strategy.

[0038] In one embodiment, according to the generated monitoring deployment adjustment space, multi-level optimization analysis is performed, first, according to the deployment reliability evaluation constraint, each scheme in the adjustment space is screened to obtain a deployment adjustment first optimization space, in the deployment adjustment first optimization space, the reliability performance of each deployment scheme is evaluated to ensure that it meets the minimum reliability requirement of the trestle monitoring. Then, the deployment adaptability of each scheme in the deployment adjustment first optimization space is calculated, this process generates a deployment adaptability evaluation space by analyzing the adaptability of different deployment schemes in actual operation to measure the performance of each scheme, thereby providing a basis for subsequent optimization. Subsequently, based on the deployment adaptability evaluation space, the schemes are further optimized according to the predetermined adaptability coefficient to obtain a deployment adjustment second optimization space, in this deployment adjustment second optimization space, the selection of each scheme is further refined to ensure that the final deployment scheme can still provide accurate monitoring effect under the condition of maximum adaptation to the complex conditions of the trestle. Finally, according to the deployment adjustment second optimization space, the most cost-effective monitoring deployment scheme is selected through cost minimization optimization, this scheme finally constitutes the stress monitoring deployment optimization strategy, which comprehensively considers reliability, adaptability and cost effectiveness, providing the optimal deployment scheme for the stress monitoring of the trestle, ensuring the accuracy and economy of the health monitoring of the trestle.

[0039] Further, the application provides multi-level optimization analysis on the monitoring deployment adjustment space according to the monitoring deployment, to generate a stress monitoring deployment optimization strategy, including:

[0040] According to the deployment reliability evaluation constraint, the deployment reliability optimization is performed on the monitoring deployment adjustment space, to obtain a deployment adjustment first optimization space; according to each deployment adjustment scheme in the deployment adjustment first optimization space, the deployment adaptability calculation is performed, to obtain a deployment adaptability evaluation space; based on the deployment adaptability evaluation space, according to a predetermined deployment adaptability coefficient, the deployment adaptability optimization is performed on the deployment adjustment first optimization space, to generate a deployment adjustment second optimization space; according to the deployment adjustment second optimization space, the cost minimization optimization is performed, to obtain the stress monitoring deployment optimization strategy.

[0041] Preferably, according to the deployment reliability evaluation constraint, the deployment reliability optimization is first performed on the monitoring deployment adjustment space, to generate a deployment adjustment first optimization space, and the purpose of this process is to screen out the best adjustment scheme that meets the deployment reliability constraint in the monitoring deployment adjustment space, to ensure that these schemes can meet the minimum reliability requirement. Subsequently, the deployment adaptability calculation is performed on each scheme in the deployment adjustment first optimization space, that is, the mean value calculation is performed on the regional deployment reliability coefficient of each adjustment scheme in the abnormal area, to obtain the deployment adaptability evaluation coefficient of each adjustment scheme, and this deployment adaptability evaluation coefficient reflects the adaptability performance of each adjustment scheme in the entire monitoring network. By summarizing these deployment adaptability evaluation coefficients, the deployment adaptability evaluation space is generated, and this deployment adaptability evaluation space can reflect the adaptability performance of different adjustment schemes in the stack bridge, to provide data support for the optimization decision in the next step. Then, based on the deployment adaptability evaluation space, according to a predetermined deployment adaptability coefficient, the deployment adaptability optimization is performed on the deployment adjustment first optimization space, and in this process, the adjustment scheme with the deployment adaptability evaluation coefficient greater than or equal to the predetermined deployment adaptability coefficient is screened out to form a deployment adjustment second optimization space, to ensure that the most adaptable adjustment scheme is selected on the basis of ensuring reliability. Finally, the cost minimization optimization is performed on the deployment adjustment second optimization space for further optimization, that is, the adjustment scheme with the least used mechanical sensors is screened out from the deployment adjustment second optimization space as the stress monitoring deployment optimization strategy, and this stress monitoring deployment optimization strategy can ensure that the final deployment can improve the monitoring accuracy and reliability while maximizing the reduction of implementation cost.

[0042] Further, the application provides that, according to the deployment reliability evaluation constraint, the deployment reliability optimization is performed on the monitoring deployment adjustment space, to obtain a deployment adjustment first optimization space, including:

[0043] According to the monitoring deployment adjustment space, a deployment adjustment first scheme is extracted; the deployment adjustment first scheme is evaluated for stack bridge multi-region deployment reliability according to the multi-node stack bridge risk curve, a first scheme reliability evaluation space is established; it is judged whether the first scheme reliability evaluation space meets the deployment reliability evaluation constraint; if the first scheme reliability evaluation space meets the deployment reliability evaluation constraint, the deployment adjustment first scheme is added to the deployment adjustment first optimization space.

[0044] Optionally, a specific adjustment scheme is selected from the monitoring deployment adjustment space as the deployment adjustment first scheme, which is generated based on the previous optimization guide features and represents a possible sensor layout, aiming to improve the monitoring effect of the stack bridge. Then, the deployment adjustment first scheme is simulated using the simulation platform, and the multi-node stack bridge risk curve of the deployment adjustment first scheme is obtained through the same way as described above, and the reliability evaluation is performed using the multi-node stack bridge risk curve, and a first scheme reliability evaluation space is established, which also includes a plurality of region deployment reliability coefficients corresponding to a plurality of regions. Then, the plurality of region deployment reliability coefficients in the first scheme reliability evaluation space are compared with the deployment reliability evaluation constraint. If each region in the first scheme reliability evaluation space meets the deployment reliability evaluation constraint, it means that the scheme meets the reliability requirement, and at this time, the scheme meeting the requirement is added to the deployment adjustment first optimization space to further participate in the subsequent optimization and selection process. Through the above steps, the selected optimization scheme can meet the monitoring reliability requirement and meet the deployment constraint, providing an efficient and reliable deployment scheme for stack bridge health monitoring.

[0045] In summary, the embodiments of the present application have at least the following technical effects:

[0046] This application embodiment first obtains the stress monitoring flow of the trestle through a distributed stress monitoring network, and performs multi-scale trestle health risk prediction based on the stress monitoring flow to construct a multi-node trestle risk curve. Then, it constructs a monitoring deployment map for each area of ​​the trestle based on the distributed stress monitoring network, and evaluates the reliability of the monitoring deployment map based on the multi-node trestle risk curve to establish a deployment reliability evaluation space. Next, based on the deployment reliability evaluation space, it performs optimization guidance feature analysis on the distributed stress monitoring network according to deployment reliability evaluation constraints to generate deployment optimization guidance features. Then, it adjusts the deployment of the distributed stress monitoring network according to the deployment optimization guidance features to obtain a monitoring deployment adjustment space. Finally, it performs multi-level optimization analysis based on the monitoring deployment adjustment space to generate a stress monitoring deployment optimization strategy, and optimizes and adjusts the distributed stress monitoring network according to the stress monitoring deployment optimization strategy. These technologies collectively address the technical problem of inadequate sensor deployment optimization in trestle stress monitoring, which leads to inaccurate monitoring results and an inability to reflect the trestle's health status in real time. By using multi-scale health risk prediction and optimized feature analysis, the reliability and accuracy of mechanical sensor deployment in various areas of the trestle are improved, resulting in more effective trestle health status assessment and risk warning.

[0047] Example 2 is based on the same inventive concept as the mechanical sensor-based method for monitoring bridge stress in the foregoing examples, such as... Figure 2 As shown, this application provides a trestle stress monitoring system based on mechanical sensors. The system includes: a health risk prediction module 11: obtaining trestle stress monitoring flow through a distributed stress monitoring network, and performing multi-scale trestle health risk prediction based on the trestle stress monitoring flow to construct a multi-node trestle risk curve; a reliability evaluation module 12: constructing a monitoring deployment map of each area of ​​the trestle based on the distributed stress monitoring network, and performing reliability evaluation on the monitoring deployment map of each area of ​​the trestle based on the multi-node trestle risk curve to establish a deployment reliability evaluation space; a feature parsing module 13: performing optimization guidance feature parsing on the distributed stress monitoring network based on the deployment reliability evaluation space and deployment reliability evaluation constraints to generate deployment optimization guidance features; a deployment adjustment module 14: adjusting the deployment of the distributed stress monitoring network based on the deployment optimization guidance features to obtain a monitoring deployment adjustment space; and an optimization adjustment module 15: performing multi-level optimization analysis based on the monitoring deployment adjustment space to generate a stress monitoring deployment optimization strategy, and optimizing the distributed stress monitoring network based on the stress monitoring deployment optimization strategy.

[0048] Furthermore, the health risk prediction module 11 is also used to perform the following methods:

[0049] According to the pier stress monitoring flow, the n th node stress monitoring data corresponding to the n th mechanical sensor is extracted, n is a positive integer greater than 1; according to the multi-scale window, the trend of the n th node stress monitoring data is predicted to obtain an n th node stress prediction set, and the multi-scale window includes a plurality of future time windows; according to the n th node stress prediction set, pier health risk prediction is performed to generate an n th node pier risk curve.

[0050] Further, the reliability evaluation module 12 is also used to execute the following method:

[0051] According to the pier area monitoring deployment map, a pier area one monitoring deployment map is extracted; according to the pier area one monitoring deployment map, the multi-node pier risk curve is arranged in the same area to obtain a pier area one risk map; according to the pier area one risk map, risk position prediction is performed to establish a pier area one risk position map; according to the pier area one risk position map, the pier area one monitoring deployment map is evaluated for reliability to obtain a first regional deployment reliability coefficient, and the first regional deployment reliability coefficient is added to the deployment reliability evaluation space.

[0052] Further, the reliability evaluation module 12 is also used to execute the following method:

[0053] According to P local computing nodes, P deployment reliability evaluation history sets are locally trained to generate P deployment reliability evaluation models, P is a positive integer greater than 1; according to the center server, the P deployment reliability evaluation models are federated aggregated to establish a deployment reliability evaluation global channel; the pier area one risk position map and the pier area one monitoring deployment map are input into the deployment reliability evaluation global channel to generate the first regional deployment reliability coefficient.

[0054] Further, the feature analysis module 13 is also used to execute the following method:

[0055] According to the deployment reliability evaluation constraint, the deployment reliability evaluation space is abnormally detected to determine a deployment evaluation abnormal distribution; according to the deployment evaluation abnormal distribution, the distributed stress monitoring network is deployed weakly associated with backtracking to obtain a deployment abnormal first backtracking result; according to the deployment evaluation abnormal distribution, the distributed stress monitoring network is deployed densely associated with backtracking to obtain a deployment abnormal second backtracking result; the deployment abnormal first backtracking result and the deployment abnormal second backtracking result are fused to generate the deployment optimization guidance feature.

[0056] Further, the optimization adjustment module 15 is also used to execute the following method:

[0057] According to the deployment reliability evaluation constraint, deployment reliability optimization is performed on the monitoring deployment adjustment space to obtain a deployment adjustment first optimization space; according to each deployment adjustment scheme in the deployment adjustment first optimization space, deployment adaptability calculation is performed to obtain a deployment adaptability evaluation space; based on the deployment adaptability evaluation space, according to a predetermined deployment adaptability coefficient, deployment adaptability optimization is performed on the deployment adjustment first optimization space to generate a deployment adjustment second optimization space; according to the deployment adjustment second optimization space, cost minimization optimization is performed to obtain the stress monitoring deployment optimization strategy.

[0058] Further, the optimization adjustment module 15 is further configured to perform the following method:

[0059] According to the monitoring deployment adjustment space, a deployment adjustment first scheme is extracted; according to the multi-node stack bridge risk curve, stack bridge multi-region deployment reliability evaluation is performed on the deployment adjustment first scheme to establish a first scheme reliability evaluation space; it is judged whether the first scheme reliability evaluation space satisfies the deployment reliability evaluation constraint; if the first scheme reliability evaluation space satisfies the deployment reliability evaluation constraint, the deployment adjustment first scheme is added to the deployment adjustment first optimization space.

[0060] It should be noted that the above sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above describes a specific embodiment of the present application. The processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0061] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0062] The present application and the drawings are only exemplary descriptions of the present application, and are considered to cover any and all modifications, changes, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the present application and its equivalents, the present application intends to include these modifications and changes.

Claims

1. A method for monitoring stress of a trestle based on a mechanical sensor, characterized by, The method comprises: obtaining a trestle stress monitoring stream through a distributed stress monitoring network, and performing multi-scale trestle health risk prediction according to the trestle stress monitoring stream, and constructing a multi-node trestle risk curve; constructing a trestle area monitoring deployment graph according to the distributed stress monitoring network, and performing reliability evaluation on the trestle area monitoring deployment graph according to the multi-node trestle risk curve, and establishing a deployment reliability evaluation space; based on the deployment reliability evaluation space, performing optimization guiding feature analysis on the distributed stress monitoring network according to the deployment reliability evaluation constraint, and generating a deployment optimization guiding feature; adjusting the deployment of the distributed stress monitoring network according to the deployment optimization guiding feature, and obtaining a monitoring deployment adjustment space; performing multi-level optimization analysis according to the monitoring deployment adjustment space, generating a stress monitoring deployment optimization strategy, and optimizing and adjusting the distributed stress monitoring network according to the stress monitoring deployment optimization strategy.

2. The method of claim 1, wherein the method further comprises: According to the trestle stress monitoring stream, a multi-scale trestle health risk prediction is performed to construct a multi-node trestle risk curve, which comprises: extracting the nth node stress monitoring data corresponding to the nth mechanical sensor from the trestle stress monitoring stream, wherein n is a positive integer greater than 1; performing trend prediction on the nth node stress monitoring data according to a multi-scale window to obtain an nth node stress prediction set, wherein the multi-scale window comprises a plurality of future time windows; performing trestle health risk prediction according to the nth node stress prediction set to generate an nth node trestle risk curve.

3. The method of claim 1, wherein the method further comprises: According to the multi-node trestle risk curve, the reliability evaluation of the trestle area monitoring deployment graph is performed to establish a deployment reliability evaluation space, which comprises: extracting a trestle area one monitoring deployment graph from the trestle area monitoring deployment graph; performing same-area data arrangement on the multi-node trestle risk curve according to the trestle area one monitoring deployment graph to obtain a trestle area one risk graph; performing risk position prediction according to the trestle area one risk graph to establish a trestle area one risk position graph; performing reliability evaluation on the trestle area one monitoring deployment graph according to the trestle area one risk position graph to obtain a first area deployment reliability coefficient, and adding the first area deployment reliability coefficient to the deployment reliability evaluation space.

4. The method of claim 3, wherein the method further comprises: According to the trestle area one risk position graph, the reliability evaluation of the trestle area one monitoring deployment graph is performed to obtain a first area deployment reliability coefficient, which comprises: performing local training on P deployment reliability evaluation history sets according to P local computing nodes to generate P deployment reliability evaluation models, wherein P is a positive integer greater than 1; performing federated aggregation on the P deployment reliability evaluation models according to a center server to establish a deployment reliability evaluation global channel; inputting the trestle area one risk position graph and the trestle area one monitoring deployment graph into the deployment reliability evaluation global channel to generate the first area deployment reliability coefficient.

5. The method of claim 1, wherein the method further comprises: determining a stress of the stack bridge based on the force measured by the force sensor. Based on the deployment reliability evaluation space, the optimization guiding feature analysis is performed on the distributed stress monitoring network according to the deployment reliability evaluation constraint to generate a deployment optimization guiding feature, which comprises: According to the deployment reliability evaluation constraint, the deployment reliability evaluation space is abnormity detected to determine a deployment evaluation abnormity distribution; According to the deployment evaluation abnormity distribution, a deployment abnormity first tracing result is obtained by tracing a weak correlation of the deployment of the distributed stress monitoring network; According to the deployment evaluation abnormity distribution, a deployment abnormity second tracing result is obtained by tracing an over-dense correlation of the deployment of the distributed stress monitoring network; The deployment abnormity first tracing result and the deployment abnormity second tracing result are fused to generate the deployment optimization guide feature.

6. The method of wharf stress monitoring based on mechanical sensors according to claim 1, characterized in that, According to the monitoring deployment adjustment space, a multi-level optimization analysis is performed to generate a stress monitoring deployment optimization strategy, including: According to the deployment reliability evaluation constraint, a deployment reliability optimization is performed on the monitoring deployment adjustment space to obtain a deployment adjustment first optimization space; According to the deployment adjustment first optimization space, a deployment adaptability calculation is performed on each deployment adjustment scheme to obtain a deployment adaptability evaluation space; Based on the deployment adaptability evaluation space, a deployment adaptability optimization is performed on the deployment adjustment first optimization space according to a predetermined deployment adaptability coefficient to generate a deployment adjustment second optimization space; According to the deployment adjustment second optimization space, a cost minimization optimization is performed to obtain the stress monitoring deployment optimization strategy.

7. The method of wharf stress monitoring based on mechanical sensors according to claim 6, characterized in that, According to the deployment reliability evaluation constraint, a deployment reliability optimization is performed on the monitoring deployment adjustment space to obtain a deployment adjustment first optimization space, including: A deployment adjustment first scheme is extracted from the monitoring deployment adjustment space; According to the multi-node stack risk curve, a stack multi-region deployment reliability evaluation is performed on the deployment adjustment first scheme to establish a first scheme reliability evaluation space; It is judged whether the first scheme reliability evaluation space satisfies the deployment reliability evaluation constraint; If the first scheme reliability evaluation space satisfies the deployment reliability evaluation constraint, the deployment adjustment first scheme is added to the deployment adjustment first optimization space.

8. A mechanical sensor based trestle stress monitoring system, characterized in that, The system is used to execute the stack stress monitoring method based on the mechanical sensor according to any one of claims 1-7, and the system includes: A health risk prediction module: a stack stress monitoring flow is obtained through the distributed stress monitoring network, and a multi-scale stack health risk prediction is performed according to the stack stress monitoring flow to construct a multi-node stack risk curve; A reliability evaluation module: a monitoring deployment diagram of each area of the stack is constructed according to the distributed stress monitoring network, and a reliability evaluation is performed on the monitoring deployment diagram of each area of the stack according to the multi-node stack risk curve to establish a deployment reliability evaluation space; A feature analysis module: based on the deployment reliability evaluation space, an optimization guide feature analysis is performed on the distributed stress monitoring network according to a deployment reliability evaluation constraint to generate a deployment optimization guide feature; A deployment adjustment module: according to the deployment optimization guide feature, a deployment adjustment is performed on the distributed stress monitoring network to obtain a monitoring deployment adjustment space; The optimization adjustment module performs multi-level optimization analysis according to the monitoring deployment adjustment space, generates a stress monitoring deployment optimization strategy, and optimizes and adjusts the distributed stress monitoring network according to the stress monitoring deployment optimization strategy.

Images