XPETS Module 2 Summary and User`s Manual
Contents
1. Thresholds i Phase3 Eliminate Overlaps m e The contrast projection looks like this Geo TopDownIntersection m Thresholds _Trianglelntersection m 0 degree 22 5 degree 45 degree 67 5 degree 90 degree 3 48 e The output contrast for the 5 images an Contrast 1 69 1 12 224 RUN_201SMMDD m m onetwo3D_scriptm Validator m Render_Pebbles m i m m Find_Contrast_apin m Bpen Phasel_Parallel_Find m S Find2 W Verifym y Phase2_Eliminate_Outsiders m Thresholds i Phase3 Eliminate Overlaps m Find Contrast Ipin m _ Contrast Data A Geo TopDownIntersection m i r Thresholds e Detailed view otter Multiple contrast samples account for uncertainty PHASE1 Summary e In summary phase 1 looks like this 0 2 0 15 0 1 m 9 05 m j i 05 0 1 0 25 UCB Nuclear Engineering XPETS Module 5 Thermal Hydraulics Lab Summary and User s Manual Phase2 Eliminate Outsiders m Trianglelntersection m e This phase eliminates any possible solutions that may have been incorrectly identified as correct solutions outside of the physical silo e It uses max and min endpoints from module 1 to set silo boundaries in the 5 different views e It also uses two custom set silo boundaries at custom2. as g fF UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 7 Defining O_z e This is the algebraically difficult part e Start from a top down view in order to calculate the radius of rotation of the thumb screw tip e Radius of rotation is calculated by the following equation Y _of xabs i O X E Y_of 0_y cos 0 abs i O _x sin e Where is the horizontal pixel location column index of the thumbscrew tip in one image This is usually taken when the screw Is far right or far left Theta in the above equation is theta of the turntable plus the theta offset Turntable Rotation Axis me Set as Origin Thumb Screw X Ray Point Source Sie i s Detector Plane fF UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual eo gt a See lat St eseneee 8 Defining O_z e We move to a side view in order to calculate O_z e Two images are picked usually the two where the thumbscrew tip is highest and lowest in the image R1 and R2 are calculated using R_rot sin theta such that R1 lt 0O R2 gt 0 O zis calculated as follows es A 8 ed Li Mk cs aed 1 R2 RI lt u ee eee X Ray Point Source etttee a as g Detector Plane UCB Nuclear Engineering XPETS Module 2 Sec Thermal Hydraulics Lab Summary and User s Manual 9 Geom
3. 1 angle nan j 1 J e Into the code XPREX_20141029_Full_Geo_Solver m e Example XPREX 20141029 Full Geo Solver m CoBIE 20150129 1 Ozman 2800 2673 CoBIE 20150125 2 22 5 nan 2763 2512 CoBIE 20150129 3 45 nan 2734 2217 CoBIE 20150125 4 67 5 nan 2719 1833 CoBIE 20150129 5 90 nan 2717 1409 CoBIE 20150129 6 112 5 nan 2728 999 CoBIE 20150125 7 135 nan 2751 652 CoBIE 20150129 8 157 5 man 2785 421 CoBIE 20150129 9 180 nan 2825 344 CoBIE_20150129 10 202 5 nan 2865 447 CoBIE 20150125 225 nan 2900 731 247 5 nan 2923 1166 270 nan 2934 1670 292 5 nan 2913 2139 2 15 315 nan 2881 2487 CoBIE 20150129 16 337 5 nan 2840 2671 fF UCB Nuclear Engineering XPETS Module2 Thermal Hydraulics Lab Summary and User s Manual Defining O_x e XPREX_20141029_Full_Geo_Solver m then utilizes the fact that the horizontal location of the screw tip will move in a sinusoidal path wi Cc a m oO oa Pd Cc oa lt 200 angles e XPREX_20141029_Full_Geo_Solver m will automatically calculate any angle offset and O_x Angle offset is caused by the screw not being placed perfectly at o degees when the turntable is at o degrees This is accounted for and It is okay to place the screw at a random angle Benet te eee
4. Summary e Inputs found in the RUN m file Elimination Thresholds Silo Boundary Theta Max Min e If you were to take an x ray image at the specified theta degrees the min and the max values in pixels specify left and right silo boundaries Any solutions found to be left of min and right of max will be eliminated Contrast Threshold e Average contrast of solution projected onto the image must be greater than the contrast threshold Number of Valid images e The solution must have contrast greater than the contrast threshold in the specified number of images 5 signifies that the solution must be validated in all 5 images Normalized overlap e Fraction of a pebble diameter that two pins can overlap 1 means two pebbles cannot overlap at all centroids must be 1 diameter apart 0 5 means that centroids can be 1 radius apart and not be eliminated UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 35 Output Summary e Module 2 saves output files with outputs from each phase 1 2 3 as well as the related diagnostics files Phase3 Is the final output e The output data structure is as follows Xl X2 PHASE3_OUT n Y1 Y2 Al Z2 e Where the 1 and 2 signify the two endpoints of a physical pin in 3 dimensions for n number of pins e And the corresponding diagnostics data structure Is Length PX PX SX PX PX SX C C PHASE3_DIAGNOSTICS n Colinearity PY
5. level organization UCB Nuclear Engineering Thermal Hydraulics Lab XPETS Module 2 Summary and User s Manual 13 onetwo3D_script m e This file will execute the following Optional pre select a region of module1 data This is good for prototyping code Selects a small region of interest for faster runtime Place3D m Phase1_ Parallel_Find m Phase2_Eliminate_Outsiders m Phase3_Eliminate_Overlaps m e This file is largely organizational in nature e See details on each phase in later slides fF UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual Le Place3D m e This file is responsible for taking module 1 data and placing the endpoints on their respective detector planes in 3 dimensional space e Highly geometry dependent e Input modulez in pixels X n a RowIndex_1 Rowlndex_ 2 Collndex_1 Collndex_2 e Output 3D in meters pointl_x point2_x source_x X 0 pointl_y point2_y source_y pointl_z point2_z source_z e Where point 1 and point 2 should be the endpoints of a pin in an x ray image projected onto the detector plane in 3D space UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 15 Place3D m e Looks like this UCB Nuclear Engineering Thermal Hydraulics Lab DAR
6. set angles user input in the RUN file e It only eliminates horizontal outliers using vertical boundaries UCB Nuclear Engineering XPETS Module2 Thermal Hydraulics Lab Summary and User s Manual 7 Phase2 Eliminate Outsiders m e Looks like this ae Bede EN T Ae ee r r a TRN e Note this image was created with very liberal thresholds and the result is therefore considerably more messy than it should be Red lines are eliminated for being outside of the silo blue lines are solutions pink circles are module 1 results fF UCB Nuclear Engineering XPETS Module2 g Thermal Hydraulics Lab Summary and User s Manual Phase3_Eliminate_Overlaps m e This module uses the physical constraint that pebbles should contact but not overlap e An overlap threshold is set normalized to a pebble diameter in the RUN m file e If any two solutions overlap more than the threshold one is eliminated based on the uncertainty set forth by the find contrast function fF UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 29 RUN_201SMMDD m Phase3_Eliminate_Overlaps m Before Overlap Elimination UCB Nuclear Engineering g XPETS Module 2 o Thermal Hydraulics Lab Summary and User s Manual 3 Validator m e This simply projects the solutions onto the x ray image so the user can visually verify the results e Var
7. to focal source point O_x x axis distance from right hand side of detector to origin O_y y axis distance from detector plane to origin O_z z axis distance from top of detector to origin N Turntable Rotation Axis Set as Origin v X Ray Point Source Q _ _ _ fF UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 4 O_yand Y_of e O_yand Y_of have been defined using a machined reference geometry and Solidworks geometry solver O_y 0 2283 m Y_of 1 8453 m e These values should be considered constant unless The turntable is moved The columnator or detector are moved in the y direction but vertical shifts should not affect O_y and Y_of Machined Reference Piece SolidWorks Geo Solver Top Down View XY Plane Detector Plane Radius of Rotation B N C A Reference Piece Diameter gt a fF UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 5 Defining O_x e We define O_x using an x ray image set of a single point usually a thumb screw rotated through 360 degrees NOTE the screw should be placed as far off center as possible away from the center of the detector in z and x directions e The angle and the 1 j tip location in pixels i column J row for each image need to be entered in the format Ox_DATA
8. Department of NUCLEAR ENGINEERING University of California Berkeley Thermal Hydraulics Laboratory XPETS Module 2 Summary and User s Manual By Grant Buster May 27 2015 Nuclear Energy University Programs i iD j U A U S Department of Energy Temperature Reactor i U S Department of Energy XPETS Module 2 Calibration Geometry Solver X UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 2 Geometry Assumptions e Your system geometry is only defined by your x ray point source detector and turntable NOT the test section e The detector is assumed to be a perfectly vertical plane e The detector is assumed not to be rotated in the xz plane LEVEL IT e The origin Is set as 0 0 0 The xy origin Is set at the turntable rotational axis The z origin coordinate is set at the plane intersecting the x ray point source and perpendicular to the detector plane the z origin plane Is horizontal Turntable Rotation Axis Set as Origin r X Ray Point Source Q _ _ _ _ fF UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 3 Importance of Geometry e An accurate 3D reconstruction Is impossible without a good geometry solution What does this mean e 4 key geometric parameters need to be defined Y_of y axis distance from origin
9. PY SY PY PY SY C Spatial_Similarity PZ PZ SZ BZ PBZ S Z C U e Diagnostics definitions on the next slide fF UCB Nuclear Engineering XPETS Module2 4 Thermal Hydraulics Lab Summary and User s Manual 3 Output Diagnostics Format Details Length PX PX SX PBX BX SX C C PHASE3_DIAGNOSTICS n Colinearity PY PY SY PY PY SY C C Spatial_Similarity PZ PZ SZ PZ PZ Z C U e Length colinearity and spatial similarity are all threshold type variables set by the intersection algorithms e P isthe 3D module 1 point from Place3D m that the solution originated from e Sis the 3D source point corresponding to P e Cis the average contrast of the solution in each image 1 5 e Uisan uncertainty parameter set by the find contrast algorithm UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 37 DON T PANIC Always feel free to call 720 495 6245 or email grantbuster gmail com UCB Nuclear Engineering XPETS Module 2 g Thermal Hydraulics Lab Summary and User s Manual 3
10. _script m j Validator Render _Pebbles m TopDownIntersection m oo Phase _Parallel_Find m Find2_W_Verify m f Phase2_Eliminate_Outsiders m e The possible solutions look like this CE Top down ntersections 0 02 S Actual Solution Actual 0 Solution Top down Intersections possible solutions UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 22 Find_Contrast_1pin m e Every possible solution is projected onto the x ray contrast data for every detector angle usually 5 angles so that it can be verified that this is indeed a pin in each image e Vector math is used for the projections Similar to intersection algorithms e An average contrast threshold is set for each image angle set in the run file e 5 pixels of slop is considered and factored into an quantitative uncertainty in the output e This function ts called the most times and is therefore the most computationally expensive function to run It is on the other hand incredibly useful in making sure you actually found a solution fF UCB Nuclear Engineering XPETS Module2 Thermal Hydraulics Lab Summary and User s Manual 3 RUN_201SMMDD m m onetwo3D_scriptm Validator m Render_Pebbles m i E al Find_Contrast_1pin m os Phasel_Parallel_Find m S Find2 W Verify y Phase2_Eliminate_Outsiders m
11. anual 9 E T gt Senenee Trianglelntersection m e This is what a triangle solutions look like Endpoint Pair 1 Endpoint Pair 2 Possible solutions UCB Nuclear Engineering Thermal Hydraulics Lab RUN_201SMMDD m Phasel_Parallel_Find m Phase2_Eliminate_Outsiders m Phase3 Eliminate Overlaps m Find Contrast_lpin m O 8 g o a S Thresholds TopDownIntersection m TriangleIntersection m XPETS Module 2 Summary and User s Manual 20 TopDownIntersection m e As the pin approaches a horizontal orientation in the actual system the triangle intersection method approaches massive uncertainty think about the intersection of two near horizontal planes e The top down method addresses this issue by examining the intersection of the XY vector components and then examining where these projected intersections actually lie in the 3 space e The thresholds are Normalized length are the linear solutions the same length as the pin s true length pebble diameter Metric spatial similarity are there consistent solutions in 3 space e This method returns more possible results than the triangle Intersection method see next slide fF UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 21 RUN_2015MMDD m _ of onetwo3D
12. as oer S Se k Ro pps ce Vode Yay b he ot iN es X lt XPETS Module 2 Summary and User s Manual 16 Phase1 Parallel Find m e This phase sets up the parallel processing inputs in a way that MATLAB will accept e This phase iterates through possible comparisons of module 1 data running the Find2_W_Verify m algorithm For example module 1 will return endpoints for 5 different detector angles The find2_w_verify algorithm will take data sets from 2 detector angles and compare them for possible 3D results Phases_parallel_find m iterates through possible input pairs for find2_w_verify m e Phase1_Parallel_Find m is organizational and is mostly meant to setup the parallel processing structure e Parallel processing can be easily swapped with serial by replacing the PARFOR loop with a FOR loop Everything else should be kept the same setup is for parallel but gon can also work for serial not the other way around UCB Nuclear Engineering XPETS Module 2 See Thermal Hydraulics Lab Summary and User s Manual Find2_W_Verify m e This is the main reconstruction algorithm e The function takes in two sets of pairs of endpoints e The function iterates through both sets trying all possible comparisons between two pairs of endpoints e An initial midpoint vector calculation is done to see if they are at al
13. etry Solved e The geo mat file should now be good to go e Two other important parameters The detector has 6993 pixels per meter PPM Apin pebble diameter is 0 01257 m e The geo mat file should be placed in the active directory and loaded before modulez2 starts to run J F UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 10 XPETS Module 2 3D Bed Reconstruction X UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 11 Module 2 Code Architecture Output file Thresholds A Find Contrast 1pin m Contrast Data TopDownlIntersection m Trianglelntersection m UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 12 RUN_2015MMDD m e This file will execute the following Setup parallel toolbox Iterate through study timesteps Import modules data Take user selected threshold values Load geometry file Save output file lIterate through post processing Validator Render eo PHL co Place3D m Phase3_Eliminate_Overlaps m Phasel_Parallel_Findm 1 Find2 W Verifym Phase2 Eliminate Outsiders m Find Contrast_lpin m J TopDownIntersection m TriangleIntersection m e This file is intended to hold all of the user inputs and setup parameters High
14. ious options can be found in the code for labels image saving module1 plotting and other accessories fF UCB Nuclear Engineering XPETS Module 2 4 Thermal Hydraulics Lab Summary and User s Manual 3 Render_Pebbles m a e This function plots pebbles with the option to plot the pebble equator color and transparency e Very helpful for post processing modules Pebble Bed Render 333 Pebbles Found z d Z axis APebble ABlade y axis x axis XF UCB Nuclear Engineering XPETS Module2 45 Thermal Hydraulics Lab Summary and User s Manual 3 XPETS Module 2 Summary f UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 33 User Input Summary e Inputs found in the RUN m file Trianglelntersection Thresholds Normalized Length e oto 1 how close to the true pin length is the possible solution 1 is exactly the correct length Metric Spatial e In meters How close are the possible solutions in 3D metric coordinates Perfectly close together threshold o is a better solution TopDownlntersection Thresholds also uses normalized length Similar Z metric e In meters How close are possible solutions in the z direction Perfectly close together threshold o is a better solution UCB Nuclear Engineering XPETS Module 2 Thermal Hydraulics Lab Summary and User s Manual 34 User Input
15. l close vertically easy way to eliminate possibilities e If pairs of endpoints are close run intersection algorithms Trianglelntersection m TopDownlintersection m e If intersection algorithms find a solution within set thresholds use a contrast verification algorithm to check solution against contrast data for all five images e If contrast data is verified add solution to output fF UCB Nuclear Engineering XPETS Module2 4 9 Thermal Hydraulics Lab Summary and User s Manual Trianglelntersection m e Aendpoints belonging to the same pin in an image form a triangle with the third point being the x ray point source e Vector calculus can be used to compare two triangles and find their linear intersection e A good vector calculus reference can be found here http geomalgorithms com ao6 _intersect 2 html e The algorithm results in two line segment solutions Endpointsz2 as a triangle intersecting endpoints2 as a plane Endpoints2 as a triangle intersecting endpoints as a plane e Thresholds used to eliminate Normalized length are the linear solutions the same length as the pin s true length pebble diameter Metric spatial similarity are the midpoints of the two linear solutions close together The two linear solutions will always be perfectly parallel fF UCB Nuclear Engineering XPETS Module 24 Thermal Hydraulics Lab Summary and User s M
Download Pdf Manuals
Related Search