Scribd
Upload
167 views33 pages

Statistical Analysis Outliers

This document discusses statistical analysis procedures for evaluating quality control test results from highway construction projects. It covers assumptions made in statistical analysis including normal distribution of test results. Key aspects summarized include: 1. Calculating the percent within limits (PWL) to estimate what percentage of a material's characteristics are within specification tolerances based on a limited number of test results. 2. Procedures for determining PWL values for density, air voids, and other material properties which impact payment for contractors. 3. Addressing outliers in test data and rules for resampling materials if initial quality control values do not meet standards.

Uploaded by

AHMAD YAGHI
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
167 views33 pages

Statistical Analysis Outliers

This document discusses statistical analysis procedures for evaluating quality control test results from highway construction projects. It covers assumptions made in statistical analysis including normal distribution of test results. Key aspects summarized include: 1. Calculating the percent within limits (PWL) to estimate what percentage of a material's characteristics are within specification tolerances based on a limited number of test results. 2. Procedures for determining PWL values for density, air voids, and other material properties which impact payment for contractors. 3. Addressing outliers in test data and rules for resampling materials if initial quality control values do not meet standards.

Uploaded by

AHMAD YAGHI
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 33

Statistical Analysis

Carl Steinhauer
Consultant

Overview
P 401 Test Results
Statistical Analysis; PWL Estimate
Verify Production Process: Payment

Theory
Assumptions
2. Normal Distribution
3. Tools: Average and Standard
Deviation
4. Percent Within Limits
(PWL) Concept
1.

Assumptions

4.

Limited # of test results


Statistical Analysis
characteristics of large amount of material
Test result variability
Components:
materials
sampling-ERLPM
testing-ERLPM
Same Process
Random sampling-Lot, Sublot

5.

Normal Distribution

1.
2.

3.

Quality

Specific Procedures
1.
2.
3.
4.

Sublots, Lots, Partial Lots


Calculations
Retesting
Outliers

Population
80

F req u en cy

70
60
50
40
30
20
10
0
4-7 4-9 5-1 5-3 5-5 5-7 5-9 5- 6-1 6-3 6-5 6-7 6-9 611
11
Organize Date

Analysis
% Taller than 5-5
% between 5-5 and 6-5
Average Height

Limit # of samples
Statistical Analysis
Estimate average
and % within limits

Asphalt Content %

15

6.55

6.45

6.35

6.25

6.15

6.05

5.95

5.85

5.75

5.65

5.55

5.45

Frequency

Figure 1

96 tests

25

20

n=100

10

30

20
Frequency

60

50

40

10

2000
2025
2050
2075
2100
2125
2150
2175
2200
2225
2250
2275
2300
2325
2350
2375
2400

Marshall Stability

x+/- 1Sn=68%
x+/-2Sn+95%
x+/-3Sn+99.7%

PWL Concept

PWL-% of
test result
exceeding L

L
L=spec lower tolerance limit
eg. Mat density 96.3 for P401

PWL Calculation Procedures


ERLPM-page 47

Section 110-AC 1505370-10A


Method for Computing PWL and Examples
Section 110-02
Spec. P401 Table 5-L and U Spec. Limits (page 24)

Given
x1=2
x2=4
x3=6
x4=8
X = 2+4+6+8 = 5
4

X=5
d1=2-5=-3
d2=4-5=-1

d1=9
d2=1

d3=6-5=1
d4=8-5=3

d3=1
d4=9

Sn =

d1+d2+d3=d4
n-1
Sn =
9+1+1+9 =
20
4-1
3
Sn =2.58
(calculator n-1)

Roundout Rules
ERLPM-page 47

Example-last digit to be kept-nearest 10 th


4.61
4.62
4.64
4.6500
4.66
4.67
4.68
4.69

becomes 4.7

Even Digit-same
Odd Digit-increase by 1
This case becomes 4.6
If it was 4.7500 it would become 4.8

MAT Density
Manual Appendix E, page 4

Sublot
1. = 96.0
2. = 97.0
3. = 99.0
4. = 100.0
x=98.0
Section 110-02f
Sn=1.8
QL=x-L
Spec Tables 5
Sn
QL=98.0-96.3 = .9444
1.8
Section 110-Table l, N=4
PL=82
page 50 ERLPM

MAT Density
PL = 82%

98.0

18%
96.3

Target Density 98.0 Achieved


Sn=1.8 versus 1.3
Acceptable QC Value

Effect of Quality Control


Sn = 1.3
Spec. pg 24

Sn = 1.8

82 PWL

90 PWL
96.3%

98%
x

Air Voids
App. D, page 3

Sublot
1= 2.1
2= 3.2
3=2.5
4=6.0
X= 3.4
Sn= 1.76
Spec. page 5 - 0.65

QL = x-L = 3.4-2.0 = .7955


Sn

1.76

PL(table 1) = 77%
n=4
QU= U-x = 5-3.4 = .909
Sn
1.76
PU(table 1) = 81%
n=4
PWL= PL + PU-100
PWL= 77+81-100 = 58

Air Voids
58%

23%
19%

3.4

PWL= PL + PU-100
PWL= 77 + 81-100 =58

Payment
Spec-par 401-8.1a-page 29
MAT Density PWL=82
Air Voids PWL= 58

Lower value

Lot Pay Factor


Air Voids- 1.4 x 58-12= 69.2%
Mat Density- 0.5 x 82+55= 96%
Use lower of 2 values- 69.2%

Joint Density
Appendix E, page 5

93.3
95.0
97.0
96.0
X= 95.3
Sn= 1.58

Table 5

QL= (95.3-93.3) = 1.2658


1.58
PL= 93
Spec. page 21 par. 401-5.2(b)(3) if < 71% there is a 5% penalty

Partial Lots
spec page 20

Section P-401-5.1c

Sample Problem
Flow-Appendix D, page 5
Partial lot situation-6 sublots
8.0, 8.2, 8.5, 8.2, 8.9, 9.1
X= 8.5
Sn= 0.44
QL= x-L = 8.5-8.0 = 1.1364; PL= 88
Sn
0.44
QU= U-X = 16-8.5 = 18.75; PL= 100
Sn
0.44
PWL= 88 + 100-100= 88<90
Corrective Action! 401-5.2(b)(2)

Outliers
Spec -pg 23 401-5.2d
-pg 25 401-5.3c
MAT Density and Air Voids

Test for Outliers


MAT Density
94.0
96.0
97.0
98.0
x= 96.2
Sn= 1.71

QL= 96.2-96.3 = -.0585


1.71
PL= <50%

ASTM E 178, par. 4


T1= (x-x1)/Sn
T1= 96.2-94 = 1.286
1.71

Table 1-ASTM E 178

N=4

Upper 5% significance level 1.463

Since 1.286<1.463 the 94.0 test value is


not considered an outlier and is retained!

Sample Problem-Outliers
Air Voids
2.0, 4.8, 4.9, 5.0
X=4.2
Sn=1.45
QL= 4.2-2.0 = 1.5172; PL= 100
1.45
QU= 5-4.2 = 0.5517; PU= 69
1.45
PWL= (100+69)-100=69
ASTM E 178 par. 4
Tn= (x-x1)/Sn = 4.2-2.0 = 1.517
1.45
Table 1, ASTM E 178, N=4, 5% significance
T= 1.463<1.517
therefore 2.0 is the outlier and it is discarded

Resampling
401-5.3 page 25

MAT Density ONLY


(Appendix E-pg 4)
Prior MAT Density- 96, 97, 99, 100
PWL 82
4 new cores 96, 96, 97, 98
AVG-all 8, 97.4
Sn= 1.51
QL= x-L = 97.4-96.3 = .7337
Sn
1.51
Table 1, N= 8
PWL= 77

Questions?

Thanks!

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy