Method for improving ray travel tracking success rate
By judging whether the ray direction vector is surrounded by the incident point of the control body and the surface element node, the problems of ray tracking failure and uneven distribution in the reverse Monte Carlo method are solved, and the infrared radiation simulation accuracy of aircraft engines is improved.
Patent Information
- Application Number
- CN202411129545.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-11-29
AI Technical Summary
In the reverse Monte Carlo method, due to the truncation error during the ray tracing process, the failure rays are too many and unevenly distributed, which reduces the accuracy of infrared radiation simulation.
By determining whether the ray direction vector is surrounded by a vector formed by the incident point of the control body and the surface element node, the condition for the intersecting rays and surface element is defined as the normal unit vector value of all nodes ≥0, avoiding the influence of truncation errors and improving judgment accuracy.
The success rate of ray tracing is improved, the number of failed rays is reduced, and the accuracy of infrared radiation simulation is significantly improved.
Smart Images

Figure CN119046597A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method for improving the success rate of ray travel tracing, and belongs to the technical field of aeroengines. Background Art
[0002] When using the principle of the reverse Monte Carlo method to calculate the infrared radiation intensity of targets such as aircraft / engines, it is necessary to randomly emit rays within the solid angle of the field of view, and the direction of the random rays is determined by the probability density function. In order to simulate the imaging of the target infrared radiation, it is necessary to grid the projection plane of the target in the direction of the field of view, and then generate some uniformly distributed random points inside each pixel grid area. The direction vector of the random ray is the vector formed by the detection point to the random point, and the starting point of the random ray is the detection point. The target is usually composed of a CFD discrete grid structure. Before calculating the target radiation intensity, the target must first perform CFD flow field calculations.
[0003] The ray tracing method (RTM) is a method for searching the control volume that the characteristic ray passes through on the transmission path based on the topological structure of the unstructured grid system of the flow field (CFD). Based on RTM, the seamless connection between the flow field calculation data and the infrared calculation is realized, which greatly improves the calculation efficiency of the reverse Monte Carlo method (RMCM). However, since all numerical calculations have truncation errors, the truncation error may cause ray tracing failure. Too many failed rays or uneven distribution will increase the statistical error of RMCM, thereby reducing the simulation accuracy. Summary of the invention
[0004] In order to solve the above technical problems, the present invention provides a method for improving the success rate of ray travel tracing.
[0005] The present invention is achieved through the following technical solutions.
[0006] The present invention provides a method for improving the success rate of ray travel tracing, wherein:
[0007] Determine whether the direction vector of the ray is surrounded by the n vectors formed by the incident point P0 of the control volume and each node of the surface element.
[0008] It also includes: defining the unit vector formed by the control volume incident point P0 to the surface element node i as The ray direction unit vector is definition:
[0009]
[0010] in is the surface element normal unit vector; κ i ≥0(i=1,L,n) rays intersect with the surface element.
[0011] The beneficial effect of the present invention is that: since the direction vector of the ray is surrounded by n vectors formed by the incident point P0 of the control body and each node of the surface element, the direction of the ray is not affected by the truncation error during the tracking process, κ i The calculations are all performed at the order of magnitude of the characteristic size of the control body. The calculation accuracy of the P position has no relationship with the final judgment result, which improves the judgment accuracy and solves the problem of too many failed rays or uneven distribution increasing the RMCM statistical error and reducing the simulation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] Figure 1 The improved ray tracing principle of the present invention is shown in FIG. Figure 1 ;
[0013] Figure 2 The improved ray tracing principle of the present invention is shown in FIG. Figure 2 ;
[0014] Figure 3 It is a schematic diagram of the intersection relationship between rays and microelement surfaces;
[0015] Figure 4 It is a schematic diagram of the incident relationship between rays and volume unit boundaries;
[0016] Figure 5 It is a schematic diagram of the failure of the intersection relationship between the ray and the microelement surface;
[0017] Figure 6 It is a schematic diagram of the failure of the incidence relationship between rays and volume unit boundaries. DETAILED DESCRIPTION
[0018] The technical solution of the present invention is further described below, but the scope of protection claimed is not limited to the description.
[0019] like Figures 1 to 6 shown.
[0020] 1. Basic principles of RTM method
[0021] The ray starts from the detection point and begins the ray tracing process after entering the first face element of the calculation area grid: the ray enters the body unit pointed to by the emitting micro-element surface, and searches for the boundary face element intersecting with the ray among all the face elements that make up the body unit according to the type of body unit. At this time, the ray is emitted from the face element and enters the body unit whose normal is facing away from the face element. The above process is repeated until the ray emits the transparent boundary of the calculation area grid or is absorbed, completing the ray travel tracking.
[0022] The intersection relationship between the ray and the boundary element is determined as follows: Figure 2 As shown. Define the vector is the vector from the intersection point P of the plane where the ray and the facet are located to each vertex of the facet, and n is the number of nodes of the facet. The intersection relationship between the ray and the facet is determined as follows:
[0023]
[0024] In the formula, % is the remainder sign. is the unit vector of the surface element normal. i When ≥0 (i=1...n) is true for all nodes, it means that the ray intersects the face element. When it is not true for any node, it means that the ray has no intersection with the face element.
[0025] For any volume element, the incident direction of the ray should be less than 90° from the normal direction of the incident surface element. The angle between the normal and the exit surface element is greater than 90°. like Figure 3 As shown. To this end, add Judgment condition: if the ray intersects the face element and the angle between the normal direction and the ray direction is greater than 90°, the ray will pass through the face element, otherwise it will continue to search for other face elements until the exit face element is found. Adding this judgment condition can prevent the ray from being trapped inside certain body elements when passing through the edges or vertices of the volume mesh and being unable to complete the ray tracing.
[0026] 2. Analysis of the causes of failed rays and their impact on RMCM simulation accuracy
[0027] When the ray searches for the boundary elements in the control volume one by one, it is necessary to calculate the intersection point P of the ray and the plane where the boundary element is located. However, due to the existence of calculation truncation error, the coordinate calculation of P cannot be accurate. In the flow field simulation of aircraft engines, in order to accurately simulate the heat transfer of the flow field and some subtle flow field structures (such as the shock wave-expansion wave system at the nozzle outlet), a very fine CFD grid is required, especially inside the engine. The grid needs to fit some small and complex components, whose characteristic dimensions are often on the order of centimeters or even millimeters, and the starting point of the ray, that is, the detection point, is often on the order of kilometers away from the engine. In this case, to accurately calculate the intersection point P on the microelement surface, the difficulty is like using a pair of extremely long chopsticks to pick up a bean on the ground in a high-rise building of dozens of floors.
[0028] Due to truncation error, the calculated intersection point P C It will appear at a certain position along the ray path, centered at the true intersection point P and bounded by points P1 and P2, that is, P C ∈[P1,P2]. There are two cases that may lead to ray tracing errors: the first case is Figure 4 As shown, the actual intersection point P is near the surface element boundary 23. CWhen it appears on or near P2, its projection on the plane where the face element is located will appear outside the face element range.
[0029]
[0030] The program cannot determine the exit face of the ray in the current control volume, and the ray fails.
[0031] The second case is Figure 5 As shown in Figure 2, when the angle between the surface element 1 and the adjacent surface element 2 is small, and the calculated intersection point PC appears on P1 or near P2, both surface element 1 and surface element 2 satisfy β i ≥0(i=1...n), if the program's search order is to first determine face element 2 and then face element 1, the dot product of the ray's direction vector and the face element 2's normal vector is less than 0, satisfying the condition that face element 2 is the exit face element, so the ray will take face element 2 as the exit face element. Then, in the next control volume, the ray will not be able to find the exact exit face element, and the ray will fail.
[0032] From the above analysis, it can be seen that there is a probability of failure in each control volume, and this failure probability is cumulative. The more control volumes the ray passes through, the greater the probability of failure, which can be expressed as
[0033]
[0034] Where P f represents the probability that the ray fails before reaching the absorption point, i = 1, L, n represents the sequence number of all control bodies that the ray passes through, P i represents the probability that the ray fails in the i-th control volume. The worse the orthogonality of the control volume traversed by the ray, the smaller the volume, and P i The bigger it is.
[0035] The principle of RMCM is to randomly emit N rays in the solid angle space of the field of view, use the probability density function to determine the absorption point of the rays, and use these N statistically significant black body radiation sources to replace the actual radiation source. The average brightness of the target at the detector entrance pupil is
[0036]
[0037] Where L b (j) is the blackbody radiation brightness at the absorption point of the jth ray, j=1, L, N represents the serial number of the ray.
[0038] For aircraft engines, the temperature distribution gradient effect in various parts of the aircraft engine is very obvious. Generally speaking, the closer to the combustion chamber, the higher the temperature is. In addition, since the CFD grid inside the exhaust system cavity is generally finer than the grid in the tail flame flow area and the orthogonality is relatively poor, the probability of failure of the ray shot into the cavity inside the exhaust system is relatively large.
[0039] Since the high-temperature components are mainly concentrated inside the engine exhaust system cavity, the failed rays will inevitably lead to a smaller number of high-temperature rays, making the calculated infrared radiation intensity value smaller. s (j) The higher the value, the closer the statistical average brightness is to the true value. Therefore, the ray tracing method is improved to increase the success rate of each ray P. s (j) It is of great value to improve the accuracy of infrared radiation simulation and evaluation of aircraft engines.
[0040] 3. Improved methods
[0041] A method for improving the success rate of ray travel tracing in the present application does not need to determine whether the intersection point P is enclosed within the surface element when determining whether the ray intersects with the surface element, but determines whether the direction vector of the ray is surrounded by n vectors formed by the control volume incident point P0 and each surface element node, where n is the number of surface element nodes, such as Figure 1 As shown. Define the unit vector formed by the control volume incident point P0 to the surface element node i as The ray direction unit vector is definition:
[0042]
[0043] in is the unit vector normal to the surface element. i When ≥0(i=1,L,n) is true for all nodes, it means that the ray intersects the face element. When it is not true for any node, it means that the ray has no intersection with the face element.
[0044] Since the direction of the ray is not affected by the truncation error during the tracing process, κ i The calculations are all performed at the order of magnitude of the characteristic size of the control body. The calculation accuracy of the P position has no relationship with the final judgment result, so its judgment accuracy can be greatly improved. The comparison of the results of the aircraft engine simulation calculation shows that in general, the number of failed rays is reduced by 1 to 2 orders of magnitude.
Claims
1. A method for improving the success rate of ray travel tracing, characterized in that: include: Determine whether the direction vector of the ray is surrounded by the n vectors formed by the incident point P0 of the control volume and each node of the surface element.
2. The method for improving the success rate of ray travel tracing according to claim 1, characterized in that: in, The unit vector formed from the incident point P0 of the control volume to the surface node i is defined as The ray direction unit vector is definition: in is the surface element normal unit vector; κ i ≥0(i=1,L,n) rays intersect with the surface element.
