02-03 Measure Inspect 2017

ME 240
Manufacturing
Measurement & Inspection
Lehigh University ___________________________________________________ Mechanical

1. Lab Week 2
Report to CAD Lab.
Turn in Week 1 Tutorial drawings
Make measurements on Air Engine components (student shop) &
start solid modeling (CAD Lab).
Completely dimensioned engineering drawings plus assembly due
at beginning of lab week 3 meeting. Save all SW files to H:\ Drive
2. Project
Submit your Team Roster via Course Site Feedback by Wednesday
2/8 (first come, first served on 3 person teams).
Teams of 4. (32 Teams max.)
Kick-off meeting at Broughal Middle School probably WED 2/15.
Measurement and Inspection
Measurement
Comparison of an unknown quantity relative to

a known standard
Inspection
Examination to determine conformance with

design specifications
Gauging
Qualitative determination of whether a part is

acceptable or not (go/no go)
Metrology
The science of measurement
Concerned with six fundamental quantities:
Length
Mass
Time
Electric current
Temperature
Light radiation
Metrology (continued)
From these basic quantities, most other physical quantities

are derived, such as:
Area
Volume
Velocity and Acceleration
Force
Electric voltage
Heat energy
Manufacturing Metrology
The measurement of length as it is relevant to a part or product
Length and Width

Depth
Diameter
Straightness, Flatness, and Roundness, etc.
Surface roughness
Metrology: Tolerances
Typical Tolerance
Limits
(by Mfrg Process)
Worse
Better
Groover, M., Fundamentals of Modern Manufacturing,, 1996

Metrology: Tolerances
Typical Tolerance
Limits
(by Mfrg Process)
Cheaper
Expensive
Groover, M., Fundamentals of Modern Manufacturing,, 1996

Metrology Challenges
Example: Large Scale Metrology Airbus A-400M Transport
http://www.builditsoftware.com/files/buildit-success_airbus.pdf
Metrology Challenges
Example: Small Scale Metrology Atomic Force Microscopy
Nanotube tipped AFM probe

Accuracy
Degree of agreement with the true value
Precision
Degree of repeatability
Accurate Precise Accurate & Neither

Precise Accurate
nor Precise
Resolution (Sensitivity)
Smallest variation in a quantity that can be detected

by a particular measuring instrument
e.g. 0.001 for a caliper
Rule of
10 Resolution of a measuring device should
be at least 10 times smaller than the
resolution required for the measurement.
e.g. a caliper (resolution 0.001 )
appropriate for accuracy to 0.01

Calibration
Procedure by which a measuring instrument is compared

against a known standard
Stability
Capability of an instrument to retain its calibration
Drift
Tendency of a device to gradually lose its accuracy/calibration

Calibration
Procedure by which a measuring instrument is compared

against a known standard
Calibration curve
A = f (B)
A
time
Stability B
Capability of an instrument to retain its calibration
Drift
Tendency of a device to gradually lose its accuracy/calibration

Systematic Error
+/- variation (errors)

- are consistent from measurement to measurement
May be precise, but not necessarily accurate
Note: High Accuracy Low Systematic Error

Random Error
- Usually due to human error, temperature change, wear,

etc.
- Usually follows a normal distribution

(distributed around a mean or average value)
Note: High Precision Low Random Error

Normal Typical Bell
Distribution Curve
Two important descriptive statistics...
Mean
Probability
N
= x i
N
Average
Standard Deviation
N
(xi - )2
= Value (x)
N
Spread 66% of all values
-3 +3
99.73% of data is within

3
For a given part measurement: where should nominal and value & tolerance band be??
-3 +3
99.73% of data is within

3
For a given part measurement: where should nominal and value & tolerance band be??
Nominal- Nominal Nominal+
-3 +3
Want part tolerances broader than sample

+
/- n
Nominal
Nominal- Nominal+
-3 +3
Otherwise you can expect to make

garbage
Accurate Precise Accurate & Neither
Precise Accurate
nor Precise
Inspection by Variables
(quantitative)
Actual values of key dimensions are measured
Record of manufacturing process--good for SPC
(Statistical Process Control)
(quantitative)
Inspection by Attributes
(qualitative)
Pass/Fail
No record of actual dimension--not as good for SPC
Quicker & cheaper
(quantitative)
Inspection by Attributes
(qualitative)
Pass/Fail
No record of actual dimension--not as good for SPC
Quicker & cheaper
Contact or Non-Contact
Contact - most common, requires orientation
Non-contact - faster, less handling, avoids damage
Manual Inspection
Boring & monotonous
Time consuming
TWO OPTIONS:
- Sampling inspection OR
- 100% inspection
Sampling Inspection
Typically perform representative (statistical) sampling
in most cases <1% of parts actually inspected
- by time (e.g., hourly, each shift)

- or by throughput (e.g., every 100 parts)
risk that defective parts will get through
trade-off between 100% assurance and cost

100% Inspection
Has its benefits & drawbacks...
every single part is checked--only good parts pass
expense of inspection added to every part
human error is bound to occur

100% Inspection
Has its benefits & drawbacks...
every single part is checked--only good parts pass
expense of inspection added to every part
human error is bound to occur
100% inspection using manual methods

is no guarantee of 100% good quality product
Standard Measuring
Tools
GO/ NO GO
Gauges
Upper Bound
Lower Bound
Advanced Measurement and Inspection Technologies
Modern alternative for manual measuring and gauging

techniques
Include contact and noncontact sensing methods:

- Coordinate measuring machines
- Laser scanning
- Machine vision
- Other noncontact techniques
Levels of Automated Inspection
1. Automatic handling/human inspection
2. Manual Loading into automatic inspection equipment
3. Automatic handling/automatic inspection

On-line or Off-line
With or without feedback

Coordinate Measuring Machine
Measurements performed
under manual or computer
control
Consists of contact probe &

3-D positioning mechanism
Workpart is fixtured to work-

table
Location coordinates of probe

are recorded as it contacts
part surfaces
Electrical touch trigger

probe senses probe
deflection
Some use ultrasonic

detection (non-contact)
Advantages over manual

measurements:
- can perform more

complex
measurements
- higher productivity
- inherently more accurate
- reduced human error

Laser Measurement
Low powered gas lasers
Scanning system
- light sweeps past object D
- time that light is

interrupted
by object is converted to a
linear measurement
Applications
- wire
extrusion
- steel rolling
- grinding
Laser Measurement
Sensors
Laser Triangulation
- light shines on object

from know angle, A
- light spot is sensed by
detector located known
distance (H)above work
plane
- part surface must be

reflective, since spot is
determined by re-
radiated light Sensor distance from laser
Part thickness D = H - R = H - L cot(A)

Fixed sensor height
Machine Vision
- Acquisition, processing, and interpretation of image data by

computer
- 2-D system views the scene as a plane

Uses:
-dimensional measuring and gauging;
-verifying presence of components;
-checking for features on a flat surface.
- 3-D vision systems are required where contours or shapes are

involved
- The majority of current applications are 2-D

Machine Vision
Camera used to capture picture
of part
Picture digitized into pixels
- binary
- gray
scale
Image analyzed using edge detection or feature
recognition
http://www.pptvision.com/
Machine Vision
Original image Binary digitizing of Binary

image representation
Image Processing and Analysis
The data for each frame must be analyzed within one scan
(1/30 or 1/25 s)
Analysis of image data:
-Edge detection
- location of boundaries of an object
- identification of contrast in light intensity between
adjacent pixels at borders
-Feature extraction
- determination of feature values of an image,
eg. area, length, diameter, and perimeter
Interpretation of Processed Image
Recognition of an object in 2-D --- How?
Identification by comparison with a predefined model or

standard
- Accomplished using extracted features
-Template Matching
-Comparison of one or more features of an image

with corresponding features of a model (template)
stored in computer memory
Machine Vision
Example of commercial
software
http://www.ds-gmbh.com
Machine Vision
Some Video Examples:
- Animation of Bottle Inspection
- High Speed Cup Inspection
- Pharma Blister Pack Animation
http://www.ds-gmbh.com

02-03 Measure Inspect 2017

Uploaded by

Copyright:

Available Formats

02-03 Measure Inspect 2017

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

02-03 Measure Inspect 2017

Uploaded by

Copyright:

Available Formats

ME 240

Lehigh University ___________________________________________________ Mechanical

Comparison of an unknown quantity relative to

Examination to determine conformance with

Qualitative determination of whether a part is

The science of measurement

Concerned with six fundamental quantities:

From these basic quantities, most other physical quantities

The measurement of length as it is relevant to a part or product

Length and Width

Groover, M., Fundamentals of Modern Manufacturing,, 1996

Groover, M., Fundamentals of Modern Manufacturing,, 1996

Example: Large Scale Metrology Airbus A-400M Transport

Example: Small Scale Metrology Atomic Force Microscopy

Nanotube tipped AFM probe

Accurate Precise Accurate & Neither

Smallest variation in a quantity that can be detected

e.g. 0.001 for a caliper

e.g. a caliper (resolution 0.001 )

appropriate for accuracy to 0.01

Procedure by which a measuring instrument is compared

Capability of an instrument to retain its calibration

Tendency of a device to gradually lose its accuracy/calibration

Procedure by which a measuring instrument is compared

Tendency of a device to gradually lose its accuracy/calibration

+/- variation (errors)

May be precise, but not necessarily accurate

Note: High Accuracy Low Systematic Error

- Usually due to human error, temperature change, wear,

- Usually follows a normal distribution

Note: High Precision Low Random Error

Two important descriptive statistics...

99.73% of data is within

99.73% of data is within

Want part tolerances broader than sample

Otherwise you can expect to make

Boring & monotonous

Typically perform representative (statistical) sampling

in most cases <1% of parts actually inspected

- by time (e.g., hourly, each shift)

risk that defective parts will get through

trade-off between 100% assurance and cost

Has its benefits & drawbacks...

every single part is checked--only good parts pass

expense of inspection added to every part

human error is bound to occur

Has its benefits & drawbacks...

every single part is checked--only good parts pass

expense of inspection added to every part

human error is bound to occur

100% inspection using manual methods

Modern alternative for manual measuring and gauging

Include contact and noncontact sensing methods:

1. Automatic handling/human inspection

2. Manual Loading into automatic inspection equipment

3. Automatic handling/automatic inspection

1. Automatic handling/human inspection

2. Manual Loading into automatic inspection equipment

3. Automatic handling/automatic inspection