Ultrasonic sensor layout method for identifying spacecraft space micrometeoroid impact events

Abstract

The application provides a kind of ultrasonic sensor layout method for identifying spacecraft space micro debris impact event, first, continuous structure is discretized into point cloud set, in the process of sensor layout optimization, Delaunay triangulation and thomson polygon network diagram are introduced, so that the optimization layout process is quantitatively characterized and controlled, and the sensor layout is optimized based on genetic optimization algorithm, and finally the sensor layout that meets the high sensitivity constraint, distortion minimization constraint, fuzzy area minimization constraint condition is obtained, therefore, the application has the characteristics of high efficiency in identifying space micro debris impact event, and the recognition efficiency can be designed to efficiently support the perception and identification of space micro debris impact in orbit.

Description

Technical Field

[0001] This invention belongs to the field of spacecraft equipment safety, and in particular relates to an ultrasonic sensor layout method for identifying spacecraft impact events involving small debris in space. Background Technology

[0002] Space debris (millimeter-scale and below) has become a significant safety hazard to spacecraft operation due to its high density and difficulty in comprehensive monitoring, tracking, and protection. Especially for sealed cabins composed of several interconnected modules, impacts and cumulative effects of micro-debris can lead to surface degradation or even defects, accelerating the on-orbit consumption rate of sealed cabin structural resources and potentially causing leaks.

[0003] To mitigate the risks associated with its on-orbit operation and achieve a comprehensive understanding of the situation, several ultrasonic sensors are typically deployed across the sealed cabin. These sensors detect the elastic wave signals and propagation caused by impacts from small debris, gradually identifying impact events in specific areas and ultimately identifying all high-threat zones within the cabin for event monitoring and prediction. However, the identification capability of this approach is highly dependent on the location and distribution of the sensors. Previous sensor placement practices often relied on subjective assumptions based on engineering experience, meaning that if shortcomings in the identification strategy are discovered later, optimization becomes extremely costly. Summary of the Invention

[0004] To address the aforementioned issues, this invention provides an ultrasonic sensor layout method for identifying spacecraft impact events caused by small space debris. Based on Delaunay triangulation networks and Voronoi diagrams, the sensor layout is performed on the sealed space cabin structure, enabling highly efficient identification of spacecraft impact events caused by small space debris.

[0005] An ultrasonic sensor layout method for identifying spacecraft impact events involving small debris includes the following steps:

[0006] S1: Given the sensor capacity and initial position, obtain the Thiessen polygon network diagram and the Deloitte triangle network diagram of the sensor on the spacecraft. Each Thiessen polygon contains one sensor, and the three vertices of the Deloitte triangle are the sensor positions.

[0007] S2: Determine whether the Thiessen polygonal network graph satisfies the layout sensitivity maximization constraint and the Delaunay triangle network graph satisfies the distortion minimization constraint. If both are satisfied, proceed to step S3; otherwise, adjust the initial position of the sensor and re-execute steps S1 to S2.

[0008] S3: Determine whether the adjacent locations of each segment of the spacecraft satisfy the fuzzy region minimization constraint under the current sensor layout. If yes, the final sensor layout is obtained; otherwise, a genetic algorithm is used to adjust the sensor layout of the adjacent locations or increase the number of sensors at the adjacent locations until the fuzzy region minimization constraint is satisfied.

[0009] Furthermore, if the Thiessen polygon network graph satisfies the layout sensitivity maximization constraint, then every Thiessen polygon constituting the Thiessen polygon network graph also satisfies the layout sensitivity maximization constraint. The method for determining whether each Thiessen polygon satisfies the layout sensitivity maximization constraint is as follows:

[0010] For any sensor i, calculate the maximum distance d from it to the boundary of the Thiessen polygon containing itself. 0,i The maximum distance d between a point within a Thiessen polygon region containing itself and a sensor in an adjacent Thiessen polygon. s,i And determine max{d 0,i ,d s,i Does it not exceed the farthest distance d of a tiny debris impact event signal that can be sensed by a single sensor? max,cr If so, then the current Thiessen polygon satisfies the layout sensitivity maximization constraint.

[0011] Furthermore, if the Delaunay triangle network graph satisfies the distortion minimization constraint, then every Delaunay triangle in the Delaunay triangle network graph also satisfies the distortion minimization constraint. The method for determining whether each Delaunay triangle satisfies the distortion minimization constraint is as follows:

[0012] For any Delaunay triangle, calculate its minimum interior angle and maximum aspect ratio. Determine if the minimum interior angle is not less than 25° and the maximum aspect ratio is less than 5. If both conditions are met, then the current Delaunay triangle satisfies the distortion minimization constraint.

[0013] Furthermore, the method for determining whether the adjacent areas of each spacecraft segment satisfy the fuzzy region minimization constraint under the current sensor layout is as follows:

[0014] The impact response is simulated at a designated location at the junction of each segment. For each designated location, the numbers and response delays of at least four nearby sensors with the fastest response are extracted as feature vectors. Then, cluster analysis is performed based on the feature vectors to obtain the predicted segment to which the current designated location belongs. It is then determined whether the actual segment to which the current designated location belongs is the same as the predicted segment. If they are the same, the current designated location is a non-fuzzy area; otherwise, the current designated location is a fuzzy area.

[0015] Calculate the area of ​​all designated locations that are identified as fuzzy regions, and determine whether the area is not greater than a set range. If it is, the adjacent locations of each segment of the spacecraft satisfy the fuzzy region minimization constraint under the current sensor layout.

[0016] Furthermore, when the adjacent areas of each module of the spacecraft do not meet the fuzzy region minimization constraint under the current sensor layout, a genetic algorithm is first used to adjust the sensor layout of the adjacent areas. If the genetic algorithm still fails to obtain a sensor layout that meets the fuzzy region minimization constraint after a set number of iterations, the number of sensors at the adjacent areas is increased under the condition that the number of sensors does not exceed a set upper limit, until a sensor layout that meets the fuzzy region minimization constraint is obtained.

[0017] Furthermore, the sensor capacity and initial location are determined based on the requirements for identifying micro-fragment impact events on the spacecraft's sealed cabin, the feasibility of sensor installation, and past engineering experience.

[0018] Beneficial effects:

[0019] 1. This invention provides an ultrasonic sensor layout method for identifying spacecraft impact events involving small debris. First, the continuous structure is discretized into a point cloud. During the sensor layout optimization process, Delaunay triangulation and Thiessen polygon network graph are introduced to quantitatively characterize and control the optimization process. Then, the sensor layout is optimized based on a genetic optimization algorithm, ultimately obtaining a sensor layout that simultaneously satisfies high sensitivity constraints, distortion minimization constraints, and fuzzy region minimization constraints. Therefore, this invention has the characteristics of high efficiency in identifying space debris impact events and can ensure that the identification performance is designable to efficiently support the sensing and identification of space debris impacts in orbit.

[0020] 2. This invention provides an ultrasonic sensor layout method for identifying spacecraft impact events involving small debris. When the feasibility of the sensor is limited by the cabin configuration, the genetic optimization results can be fine-tuned based on the actual conditions. The fine-tuned results should be verified by high sensitivity constraints, distortion minimization constraints, and fuzzy region minimization constraints to finally obtain a scientific sensor layout that simultaneously satisfies the above three constraints, so as to efficiently support the on-orbit perception and identification of space debris impacts.

[0021] 3. This invention provides an ultrasonic sensor layout method for identifying spacecraft impact events involving small debris. With slight adaptive adjustments to the sealed / unsealed shell combination structure of different types of spacecraft, this invention can provide a scientific sensor layout that simultaneously satisfies the above three constraints, and has good versatility. Attached Figure Description

[0022] Figure 1This is a flowchart of the ultrasonic sensor layout method of the present invention;

[0023] Figure 2 This is a schematic diagram illustrating the optimized layout of ultrasonic sensors used for impact monitoring in a sealed chamber. Detailed Implementation

[0024] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0025] This invention introduces graphical optimization and evaluation based on Delaunay triangle networks and Thiessen polygon networks in sensor layout optimization, and numerical optimization and evaluation based on genetic algorithms. Under specified engineering constraints, it aims to achieve high sensitivity of a single sensor layout, minimize distortion of local region identification strategies, and minimize ambiguity in inter-module identification strategies. It provides an optimal sensor layout strategy to efficiently support on-orbit sensing and identification of small space debris impacts. Specifically, such as... Figure 1 As shown, an ultrasonic sensor layout method for identifying spacecraft impact events involving small debris includes the following steps:

[0026] S1: Given the sensor capacity and initial position, obtain the Thiessen polygon network diagram and the Deloitte triangle network diagram of the sensor on the spacecraft. Each Thiessen polygon contains one sensor, and the three vertices of the Deloitte triangle represent the sensor positions.

[0027] It should be noted that the sensor capacity and initial location were determined based on the requirements for identifying micro-fragment impact events on the spacecraft's sealed cabin, the feasibility of sensor installation, and past engineering experience. When obtaining the Thiessen Voronoi polygon network diagram and the Delaunay triangle network diagram, the layout cabin was first converted into a discrete point cloud with a positioning accuracy of no more than one-tenth of the density. Then, based on the discrete point cloud, the Delaunay triangle network partitioning diagram of the sensor layout and the Voronoi polygon network diagram of all sensors were calculated step by step.

[0028] S2: Determine whether the Thiessen polygonal network graph satisfies the layout sensitivity maximization constraint and the Delaunay triangle network graph satisfies the distortion minimization constraint. If both are satisfied, proceed to step S3; otherwise, adjust the initial position of the sensor and re-execute steps S1 to S2.

[0029] It should be noted that each polygon region in the Thiessen polygon network graph contains a sensor, representing the area that is preferentially sensed by that sensor. For each sensor location (taking the i-th sensor as an example), the maximum distance d from it to the boundary of the Thiessen polygon containing itself is calculated. 0,iThe maximum distance d between a point within a Thiessen polygon region containing itself and a sensor in an adjacent Thiessen polygon. s,i And determine max{d 0,i ,d s,i Does it not exceed the farthest distance d of a tiny debris impact event signal that can be sensed by a single sensor? max,cr If so, then the current Thiessen polygon satisfies the layout sensitivity maximization constraint. If every Thiessen polygon constituting the Thiessen polygon network satisfies the layout sensitivity maximization constraint, then the entire Thiessen polygon network satisfies the layout sensitivity maximization constraint.

[0030] Meanwhile, the three vertices of each Deloitte triangle correspond to the three sensor positions. This area represents the local area identified by the three sensors. To ensure the robustness of the recognition algorithm, the triangular mesh area should be as close as possible to an equilateral triangle (ideally an equilateral triangle). Calculate the minimum interior angle and the maximum aspect ratio of each triangular mesh. The minimum interior angle should not exceed 60°. The closer it is to 60°, the closer the mesh is to the ideal situation. The maximum aspect ratio should be greater than 1. The closer it is to 1, the more ideal it is. Generally, the minimum interior angle is greater than 25° and the maximum aspect ratio is less than 5. The layout should ensure that these values ​​are within the above-mentioned range to control the distortion of the triangular mesh area (distortion minimization constraint). Based on this, if the DeLore triangle network graph satisfies the distortion minimization constraint, then every DeLore triangle in the DeLore triangle network graph satisfies the distortion minimization constraint. The method to determine whether each DeLore triangle satisfies the distortion minimization constraint is as follows: For any DeLore triangle, calculate its minimum interior angle and maximum aspect ratio, and determine whether the minimum interior angle is not less than 25° and the maximum aspect ratio is less than 5. If both are satisfied, then the current DeLore triangle satisfies the distortion minimization constraint.

[0031] Therefore, it can be seen that the present invention, for all possible discrete points monitored in the cabin, satisfies the condition that none of the sensors exceed the farthest distance d that a single sensor can detect for a micro-fragment impact event signal. max,cr All local identification areas meet the specified distortion requirements. If they do not meet the requirements, the initial position of the sensor layout can be adjusted or the number of sensors in the local non-compliant areas can be increased if permissible. The calculation process from step S1 to S2 can be repeated until the requirements are met.

[0032] S3: Determine whether the adjacent locations of each segment of the spacecraft satisfy the fuzzy region minimization constraint under the current sensor layout. If yes, the final sensor layout is obtained; otherwise, a genetic algorithm is used to adjust the sensor layout of the adjacent locations or increase the number of sensors at the adjacent locations until the fuzzy region minimization constraint is satisfied.

[0033] It should be noted that, because the configuration features and dimensions of each section of the hull are different, the algorithms used to identify impact events of each section in engineering (generally physical model method or data-driven model method) are also different. Therefore, before effective identification, clustering algorithms (such as artificial neural networks, nearest neighbor classification, etc.) should be used to partition the impact events by section. However, this may result in the area near the docking boundary between sections being assigned to the areas of two adjacent sections at the same time, that is, there is an ambiguous area between sections. Under a specific layout, when a collision event occurs at all possible monitoring points, the sensor numbers and relative delays that sense the event fastest are calculated in the simulation. These are then used as feature vectors for cluster analysis to ensure that when a collision occurs in a certain compartment, clustering based on the feature vectors can also classify it as a collision occurring in that compartment. If this condition is not met, under the constraints of high sensitivity and minimum distortion, the positions of the sensors used for identifying the area near the compartment are optimized and adjusted using a genetic algorithm, or the number of sensors is locally increased if permissible. The compartment to which all possible collision events belong within the docking area between compartments are iteratively calculated. The envelopes and feature sizes of points that are simultaneously identified as belonging to adjacent compartments are statistically analyzed and ensured to be within a specified range (e.g., less than the positioning accuracy or a percentage of the positioning accuracy) to control the ambiguity area between compartments (ambiguity area minimization constraint).

[0034] Specifically, the method for determining whether the adjacent areas of each spacecraft module satisfy the fuzzy region minimization constraint under the current sensor layout is as follows:

[0035] Impact responses are simulated at designated locations at the junctions of each module. For each designated location, the numbers and response delays of at least four nearby sensors with the fastest response are extracted as feature vectors. Cluster analysis is then performed based on the feature vectors to obtain the predicted module to which the current designated location belongs. It is then determined whether the actual module to which the current designated location belongs is the same as the predicted module. If they are the same, the current designated location is a non-fuzzy region; otherwise, the current designated location is a fuzzy region. The area of ​​all designated locations that are determined to be fuzzy regions is counted, and it is determined whether the area is not greater than a set range. If it is, the junctions of each module of the spacecraft satisfy the fuzzy region minimization constraint under the current sensor layout.

[0036] After optimization steps S1 to S3, the ultrasonic sensor position and capacity simultaneously satisfy the high sensitivity constraint, distortion minimization constraint, and fuzzy region minimization constraint are obtained, resulting in a high-efficiency sensor layout optimization. It should be noted that when sensor feasibility is limited by cabin configuration conditions, the final execution result can be fine-tuned based on actual conditions. The fine-tuned result should be verified against the high sensitivity constraint, distortion minimization constraint, and fuzzy region minimization constraint.

[0037] The ultrasonic sensor layout method of the present invention will be further explained below using the sensor layout of a certain sealed chamber as an example. Figure 2 As shown, a sealed chamber, composed of two connected sections, has undergone optimized layout of ultrasonic sensors for impact monitoring (discrete points on the chamber are cloudified, and the density of light gray points is set to no more than 1 / 10 of the monitoring and positioning accuracy). The optimization results are as follows: After optimization, a total of 16 sensors are arranged at the large black square points shown in the figure above. The maximum aspect ratio of the triangular projection area formed by the sensors is 2.86 (corresponding to a minimum interior angle of 36.9°), and the minimum interior angle of the triangular projection area is 26.2°. There is no misjudgment caused by the impact response to two adjacent sensors being too close due to distortion of the monitoring triangular area. The area enclosed by the corresponding small black square points around each sensor (large black square point) is the priority sensing domain of that sensor. The distance from the boundary of the domain to the corresponding sensor is less than 1.5m, and the distance from the priority sensing domain to the adjacent sensor is less than 1.5m, meeting the sensitivity requirements for effective sensing. For each light gray point, an impact response was simulated. The sensor numbers of the four nearest sensors with the fastest response and their response delays were extracted as feature vectors. Based on this, cluster analysis was performed to classify impact events by compartment. For example, when an impact point occurs... Figure 2 When the impact occurs in the left-side compartment, the extracted feature vectors can be classified and identified as impacts that occurred in the left-side compartment through clustering, meaning that the optimized layout meets the requirement of minimizing the fuzzy region. Figure 2 A schematic diagram illustrating the optimized layout of ultrasonic sensors for impact monitoring in a sealed chamber.

[0038] Therefore, this invention, in the sensor layout optimization process, cloudifies the discrete points of the cabin and introduces Delaunay triangulation and Voronoi diagram generation methods, enabling the optimization layout process to be quantitatively characterized and controlled. This leads to a sensor layout optimization method that simultaneously satisfies high sensitivity constraints, distortion minimization constraints, and fuzzy region minimization constraints. The sensor layout scheme proposed in this invention features high-efficiency recognition and ensures that the recognition performance is designable. Furthermore, for sealed / unsealed shell combinations of different spacecraft models, this invention, with slight adaptive adjustments, can provide a scientific sensor layout that simultaneously satisfies the above three constraints, demonstrating good versatility.

[0039] Of course, the present invention may have other various embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding changes and modifications should all fall within the protection scope of the appended claims.

Claims

1. A method for deploying ultrasonic sensors to identify spacecraft impact events involving small debris, characterized in that, Includes the following steps: S1: Given the sensor capacity and initial position, obtain the Thiessen polygon network diagram and the Deloitte triangle network diagram of the sensor on the spacecraft. Each Thiessen polygon contains one sensor, and the three vertices of the Deloitte triangle are the sensor positions. S2: Determine whether the Thiessen polygonal network graph satisfies the layout sensitivity maximization constraint and the Delaunay triangle network graph satisfies the distortion minimization constraint. If both are satisfied, proceed to step S3; otherwise, adjust the initial position of the sensor and re-execute steps S1~S2. S3: Determine whether the adjacent nodes of each spacecraft segment satisfy the fuzzy region minimization constraint under the current sensor layout. If yes, the final sensor layout is obtained; otherwise, a genetic algorithm is used to adjust the sensor layout of adjacent nodes or increase the number of sensors at adjacent nodes until the fuzzy region minimization constraint is satisfied. If the Thiessen polygon network graph satisfies the layout sensitivity maximization constraint, then each Thiessen polygon constituting the Thiessen polygon network graph satisfies the layout sensitivity maximization constraint. The method for determining whether each Thiessen polygon satisfies the layout sensitivity maximization constraint is as follows: For any sensor i Calculate the maximum distance from each of the three polygons to the boundary of the Thésen polygon containing itself. The maximum distance between a point within a Thiessen polygon region (including itself) and a sensor within an adjacent Thiessen polygon. and judge Does it not exceed the farthest distance of a tiny debris impact event signal that can be detected by a single sensor? If yes, then the current Thiessen polygon satisfies the layout sensitivity maximization constraint; if the Delaunay triangle network graph satisfies the distortion minimization constraint, then every Delaunay triangle in the Delaunay triangle network graph satisfies the distortion minimization constraint. The method for determining whether each Delaunay triangle satisfies the distortion minimization constraint is as follows: For any Delaunay triangle, calculate its minimum interior angle and maximum aspect ratio. Determine whether the minimum interior angle is not less than 25° and the maximum aspect ratio is less than 5. If both conditions are met, then the current Delaunay triangle satisfies the distortion minimization constraint. The method for determining whether the adjacent points of different spacecraft modules satisfy the fuzzy region minimization constraint under the current sensor layout is as follows: The impact response is simulated at a designated location at the junction of each segment. For each designated location, the numbers and response delays of at least four nearby sensors with the fastest response are extracted as feature vectors. Then, cluster analysis is performed based on the feature vectors to obtain the predicted segment to which the current designated location belongs. It is then determined whether the actual segment to which the current designated location belongs is the same as the predicted segment. If they are the same, the current designated location is a non-fuzzy area; otherwise, the current designated location is a fuzzy area. Calculate the area of ​​all designated locations that are identified as fuzzy regions, and determine whether the area of ​​each region is not greater than a set range. If it is, then the adjacent locations of each segment of the spacecraft satisfy the fuzzy region minimization constraint under the current sensor layout.

2. The ultrasonic sensor layout method for identifying spacecraft micro-debris impact events as described in claim 1, characterized in that, When the adjacent areas of different modules of the spacecraft do not meet the fuzzy region minimization constraint under the current sensor layout, a genetic algorithm is first used to adjust the sensor layout of the adjacent areas. If the genetic algorithm still fails to obtain a sensor layout that meets the fuzzy region minimization constraint after a set number of iterations, the number of sensors at the adjacent areas is increased, while ensuring that the number of sensors does not exceed a set upper limit, until a sensor layout that meets the fuzzy region minimization constraint is obtained.

3. The ultrasonic sensor layout method for identifying spacecraft micro-debris impact events as described in any one of claims 1 to 2, characterized in that, The sensor capacity and initial location are determined based on the need to identify micro-fragment impact events on the spacecraft's sealed cabin, the feasibility of sensor installation, and past engineering experience.

Images