TE-MSC
07/04/2016 On behalf to MSC-MDT section and Coil working group Jose Ferradas TE-MSC-MDT Alejandro Carlon TE-MSC-MDT Juan Carlos Perez TE-MSC-MDT
• Overview • CERN Lab 927 Equipment • Nominal models used • CERN MQXFS Measurements procedure • What’s next? • Conclusions
1. OVERVIEW • Measurements at one single stage: A fter impregnation • Currently CERN approach: • Short coils Mechanical measurements using CMM Portable Device (FARO Arm) • Long coils Tests with CMM machine Work in progress • Currently FNAL, BNL & LBNL approach: • Short coils Mechanical measurements using CMM machine • Long coils Tests with CMM Portable Device (FARO arm)
1. OVERVIEW What happens if we compare Faro Arm measurements to CMM measurements? • CMM agreement difference between CERN Faro Arm and BNL Faro Arm • See presentation from E.Holik – FNAL • The aim of this presentation is to explain the procedure followed at Lab.927 CERN • First draft delivered
1. OVERVIEW • Note that: • No topics different than the procedure itself will be addressed • Important topics were moved to next meeting • Procedure sets the actions and their order independently of reference used, software used…
2.EQUIPMENT • CMM Portable device: Faro Arm Edge 2.7 • 3 mm. Touch Probe • Faro Arm Edge 2.7 volumetric maximum deviation: 41 μ m • Faro Arm Edge 2.7 repeatability: 29 μ m • Periodically sent to calibration • Polyworks Inspection Software 2015
2.EQUIPMENT • Dedicated area for metrology inside the workshop • Marble for measurements (2.5 m x 1.5 m) • Arm fixed to the table using rigid support
2.EQUIPMENT • MQXFS coils clamped using a specially designed support • Inner cylinder placed at 255 mm height • Possible to probe inner diameter • Support parts are aligned and screwed before operations • Support fixed to the table using conventional clamps
3. MODELS • Analysis done by CAD comparison: Measured values compared with nominal ones (extracted from the CAD model). • Two different CAD models used: 1. Simplified CAD model for analysis • Solid block • Only external shape considered 2. Simplified CAD model for cross section alignment • Surface • Outer cylinder and keyway * Note that necessity of one model for alignment is due to the software used. If the software is different it might not be necessary to use it
5.PROCEDURE • Metrology inspection performed as following: • Coil clamping • Data acquisition • Analysis Coil Data Analysis clamping acquisition
5.PROCEDURE Objective: Avoid any possible movement Distance between supports : 90 cm Distance between each end and support : 30 cm Gravity working against banana shape deformation
5.PROCEDURE • Once the coil is correctly clamped: • Marble should be cleaned removing any possible disturbing object or tooling • Area should be delimited, restricting the access to other person different to operator • Data acquisition procedure: 1. Device calibration 2. Device validation 3. Pre-alignment 4. General geometry probing 5. Best-fit alignment 6. Cross section probing
5.PROCEDURE Device Device Pre- calibration validation alignment Cross General Alignment section geometry improving probing probing
5.PROCEDURE • Performed following FARO hardware calibration procedures: • Hole compensation : Probe calibration • Single point articulation test (SPAT): Arm calibration (Repeatability) In case of failing, error message will be displayed
5.PROCEDURE • Measurements over calibrated pieces • In order to assess good performance of the device
5.PROCEDURE • Usual results: • Hole compensation 2-sigma: Below 11μm • Calibrated pieces: Below 8μm
5.PROCEDURE • Starting point: Pre-alignment of real coil to CAD model • Using 6 “Surface point alignment” • 6 points are defined in the CAD model, operator must probe the same 6 points in real coil • For improving reproducibility, singular points are chosen (Corners, edges…)
5.PROCEDURE • Once pre-alignment is done, data is collected over: • Lead end plane • Return end plane • Outer cylinder • Inner cylinder • Left Mid-plane • Right Mid-plane • Left key plane • Right key plane • Bottom key plane • Necessary to be as perpendicular as possible while probing
5.PROCEDURE • After first part of the probing process, information is enough to perform a better alignment • Best-fit alignment of real point cloud to CAD model is performed using: • Lead end plane • Outer cylinder • Left key plane • Right key plane • All degrees of freedom are fixed • Coil is aligned to Lead end in order to define the cross section distances
5.PROCEDURE • With the new alignment, software will locate all the cross section. • Cross sections are defined in the CAD model using the following scheme: • Cross sections are probed: • Projection method – 5 mm. • Trying to be as perpendicular as possible with the probe
5.PROCEDURE • Analysis procedure is done as follows (Software requirement): Each CS is Aligned point Normal vectors aligned individually cloud is exported extraction to OD and Keyway as .TXT (Modified CAD) Results extracted .TXT point cloud is comparing imported as imported data to separate data full CAD
5.PROCEDURE • Example of cross section alignment
5.PROCEDURE • Example of cross section alignment Resulting CS Original CS CS aligned using special CAD Aligned points compared with Banana shape present full CAD
5.PROCEDURE • General geometry results • Banana shape • Coil length • Mid-planes angle • Keyway width … • Inner radius • Outer radius • Cross section results • Individual features defined in each cross section • Mid-plane excess • Inner / outer radius • Keyway information
6. NEXT • Very important topics have been removed from the scope of this presentation • These topics will be addressed in next meeting: • Reference for measurements: Discussion & standardization • Reference influence over results • Differences between CERN side and non-CERN side • Results extraction analysis: Differences over results extraction / Choosing right method • Coil fiducialization • CERN Faro arm VS CERN CMM machine study • Results for coils to be presented
Comparison study between: FARO Arm CMM • Results obtained in different relative positions Coil – Device • Those obtained at CERN main metrology service Initial results show good agreement on deviation vector map.
• Polyworks Inspection approach: • Reference system defined using nominal features (CAD) • Origin is set in Lead End Plane. • X & Z axis are perpendicular to each mid-plane. • Y axis going in longitudinal direction • This means, the coordinate system is attached to the CAD. It is not defined during measurements. • It is not a metrological reference. It is impossible to completely define the reference by measuring features. • To be able to use it, the CAD is best-fitted to our real measured geometry • Therefore, reference system relies on the good alignment with the real piece
• Polyworks Inspection approach: • For MQXFS coils: • First: All features of the coil are defined by probing • Then, the alignment to the CAD model is performed: • Longitudinal translation fixed by best-fitting measured lead plane to nominal lead end plane. Lead end measured plane is made coincident with the nominal plane from CAD (First degree of freedom) • Second translation fixed by best-fitting to outer cylinder (Second degree of freedom) • Rotation fixed by best-fitting to key planes (Third degree of freedom) • Final alignment is the global result minimising the deviation between all features used for it. It is an average position trying to get minimum deviation. • Once both entities are aligned. Each cross section is measured by probing.
• Polyworks Inspection approach: • INCONVENIENTS: • Reference system relies on the good fit of our real coil to the nominal model • If the features used for the alignment are out of tolerance, the alignment could be not good • Small alignment deviations in the first part of the coil become bigger with length • ADVANTAJES: • Time to carry out the measurements is decreased. It is possible to perform the inspection following a stablished inspection guide. • Everything is previously defined before starting the measurements. Operator should only probe the stablished points, decreasing probability of additional errors (coils are always measured in the same way). • For cross sections, points are projected by the software within a stablished distance of +/- X mm. This fact allows us to not discard this approach. • Results in deviation vector map from CMM show good agreement with our measurements (See previous or next slides)
Recommend
More recommend
