United States Patent I
US007043970B2
(12)
(54)
United States Patent
(10) Patent N0.:
Ristea et a].
(45) Date of Patent:
METHOD FOR MONITORING JP
80016205 B
(75) Inventors: Catalin Ristea, Coquitlam (CA); James
*
M1980
OTHER PUBLICATIONS
ilTllllingzlsch?rigioiilagshizfisgR (US)’
Bramhall, G., “Meeting New Kiln Drying Standards,” 1975,
Vancouver (CA)
Canad1an Forest Industr1es, 95(9), pp. 33-35. Bramhall, G. and Warren, W.G., “Moisture Content Control in Drying Dimension Lumber,” 1977, Forest Products Jour
(73) Assigneez Wagner Electronic Products, Inc” Rogue River, OR (US)
Hal’ 27(7)’ PP' 26128' _
Subject to any disclaimer, the term Of this patent is extended or adjusted under 35
mal Distributions by Maximum Likelihood,” 1951, Journal of the American Statistical Associate, 46(254), pp. 206-212,
U.S.C. 154(1)) by 0 days.
1951 Cohen, A.C., Whitten, B.J., and Ding, Y., Modi?ed Moment Estimation for the Three Parameter Lognormal Distribution, 1985, Journal of Quality Technology, 17(2), pp. 92-99. Ferrell, E.B., “Control Charts for Log-Normal Universes,” 1958, Industrial Quality Control, 15(2), pp. 4-6.
Nov. 28, 2003
(65)
_
_
Cohen, A.C., “Est1mat1ng Parameters of Loganthmlc-Nor
(21) Appl, No.1 10/722,389 (22) Filed:
May 16, 2006
FOREIGN PATENT DOCUMENTS
WOOD-DRYING KILN STATE
( * ) Notice;
US 7,043,970 B2
Prior Publication Data
US 2005/0115307 A1
(51) Int‘ Cl‘ G01N 25/56
Jun. 2, 2005
(Commued) Primary ExamineriHeZron Williams Assistant Examinerilohn Fitzgerald (74) Attorney, Agent, or FirmiKlarquist Sparkman, LLP
(200601)
(52)
US. Cl. ........................................................ ..
(58)
Field of Classi?cation Search ................... .. 73/73
73/73
(57)
ABSTRACT
See application ?le for complete search history. (56)
References Cited
The method utilizes a known good set of moisture content data for Wood dried in a kiln under set conditions to estimate parameters for a Lognormal distribution Which approxi
U-S~ PATENT DOCUMENTS 4,107,599 A *
8/1978 Preikschat ................ .. 324/689
4,193,761 A * 3/1980 Mantegani 5,586,066 A * 12/1996 White et a1. 6,077,552 A T 6/2000 Chimenti 6t ill. 6,305,224 B1 *
10/2001 StanTSh et a1 ~~~~ ~~
6’308’57l Bl * 10/2001 Stamsh et a1 6,703,847 B1 *
3/2004 Venter et al
6,735,576
5/2004
Bl*
2003/0110007 A1 *
Kaji et a1.
mates the moisture Content data for those conditions- The
432/123 702/181 426/231
method creates control charts Which monitor some of these parameters. The method also takes data from subsequent Charges and transfonns to be represented on the
~~ 73/ 597
control charts. The method uses this representation to deter
" 73/597
mine if the kiln is operating under out-of-control conditions.
324/663 ......
. . . . . . ..
706/1
6/2003 McGee et a1. ............ .. 702/179
502 __,_
33 Claims, 9 Drawing Sheets
GATHER MULTIPLE SAMPLES OF KNOWN
GOOD DATA FOR A SPECIFIC KILN AND SET OF CONDITIONS
500
504
DETERMINE IF MOISTURE CONTENT DATA FOLLOWS LOGNORMAL DISTRIBUTION
ESTIMATE A LOGNORMAL DIST IBUTION THAT APPROXIMATES THE DATA
508 ~'“
APPROXIMATE THE LOGNORMAL THRESHOLD PARAMETER
512 ‘’
CALCULATE THE LOGNORMAL SHAPE PARAMETER
CALCULATE THE
LOGNORMAL SCALE PARAMETER
I
DEVELOP CONTROL CHARTS CREATE LOGARITHMIC SCALE CHART CREATE 522-” LOGARITHMIC SHAPE CHART
CREATE
mmos SCALE CHART v
CREATE ANTILOG SHAPE CHART
w 514
US 7,043,970 B2 Page 2 OTHER PUBLICATIONS
of British Columbia, Department of Mechanical Engineer
JolTe, AD, and Sichel, H.S., “A Chart for Sequentially Testing Observed Arithmetic Means from Lognormal Popu lations Against a Given Standard,” 1968, Technometrics, 10(3), pp. 605-612. Maki, R.G., and Milota, M.R., “Statistical Quality Control
Pratt, W.E., “Some Applications of Statistical Quality Con trol to the Drying of Lumber,” 1953, Journal of FPRS, 3(5),
Applied to Lumber Drying,” 1993, Quality Progress, 26(12),
1ng.
pp. 28-31.
Pratt, W.E., “Estimating the Moisture Content of Lumber
During the Drying Process,” 1956, Forest Products Journal, 6(9), pp. 333:337. ZWick, R.L. and Cook, J.D., “The Modeling of Moisture
pp. 75-80.
Content Distributions Based on Censored Readings from a
McMahon, E.P., “Applying Cumulative Frequency Distri
Resistance Meter,” 1985, Technical paper presented at West ern Dry Kiln Association Meeting. Maki, “An Application of Statistical Process Control Mea sures for Maintaining Optimal Quality from Dry Kiln Operations,” Thesis submitted to Oregon State University,
bution in Moisture Control During Kiln Drying,” Forest Products Journal, 11(3), pp. 133-138. Morrison, J ., “The Lognormal Distribution in Quality Con
trol,” 1958, Applied Statistics, 7(3), pp. 160-172. Noghondarian, K., “Quality Control With Non-Normal, Cen sured and Truncated Data,” 1997, Ph.D. Thesis, University
Apr. 26, 1991, 120 pages. * cited by examiner
U.S. Patent
Fig. 1
May 16, 2006
Sheet 1 0f 9
US 7,043,970 B2
U.S. Patent
May 16, 2006
US 7,043,970 B2
Sheet 2 0f 9
208,
0 21
202
206 201
2 Fig.
204 I11
202
U.S. Patent
May 16, 2006
Sheet 3 0f 9
US 7,043,970 B2
U.S. Patent
May 16, 2006
Sheet 4 0f 9
US 7,043,970 B2
Fig. 4a
Normal distribution curve
Fig. 4b Lognormal distribution curve
U.S. Patent
May 16, 2006
Sheet 5 0f 9
US 7,043,970 B2
F I9. 5 502 w
GATHER MULTIPLE SAMPLES OF'KNOWN
GOOD DATA FOR A SPECIFIC KILN AND SET OF CONDITIONS
500
V
504 4‘
DETERMINE IF MOISTURE CONTENT DATA
FOLLOWS LOGNORMAL DISTRIBUTION
Y
ESTIMATE A LOGNORMAL DISTRIBUTION THAT APPROXIMATES 508 _/~
THE DATA 510 _,~ APPROXIMATE THE
LOGNORMAL THRESHOLD PARAMETER
512 O-
CALCULATE THE
CALCULATE THE
LOGNORMAL
wings?“
SCALE PARAMETER
PARAMETER
w 514
DEVELOP CONTROL CHARTS 516 I
w
518
CREATE
CREATE
LOGARITHMIC
ANTILOG
SCALE CHART
sCALE CHART
CREATE 522“ LOGARITHMIC
SHAPE CHART
T" 520
CREATE ANTILOG
SHAPE CHART
w 522
U.S. Patent
May 16, 2006
US 7,043,970 B2
Sheet 6 0f 9
Hg. 6
TRANSFORM EACH MEASUREMENT
604 a
‘f 602
SAMPLE CALCULATIONS 606 T\-
Y
'
CALCULATE
CALCULATE
SAMPLE MEAN
STANDARD
,
w 608
DEVIATION y
I, A
” S
PARAMETER CALCULATION v_______ 610 A,
CALCULATE MEAN
v
w 614
OF SAMPLE
STANDARD
OF SAMPLE MEANS “y bar bar” FOR ESTIMATE OF ‘
SC A "E PA RAME TE R
"8 bar, 4 ' INTRODUCE
CONSTANT (04) TO CREATE UNBIASED ESTIMATOR OF SHAPE PARAMETER
_,_ 616
U.S. Patent
May 16, 2006
Sheet 7 0f 9
US 7,043,970 B2
Hg. 7 516
DEVELOP CONTROL CHART
710
ESTABLISH SAMPLE SIZE
J‘ 704
DETERMINE CENTER LINES
_,_ 706
CALCULATE CONTROL LIMITS
~1~ 708
gfCLn?AJ \”
CALCULATE
FOR SCALE
PROBABILITY LIMITS L/\ 712
CHART
FOR SHAPE CI-IART
SCALE LIMITS AND
J‘ 714
CENTER LINE
-———-——> v
PLOT NEW DATA ON
CHART
T" 716
U.S. Patent
May 16, 2006
Sheet 8 0f 9
US 7,043,970 B2
Fi g . 8
716
802
GATHER NEW DATA SET
TRANSFORM DATA SET 808 ~/~
TRANSFORM EACH
MEASUREMENT
814 \f:
’\_/ 810
CALCULATE
CALCULATE
SAMPLE
SAMPLE ST.
MEAN FOR
DEv FOR
SCALE CHART
SHAPE CHART
CHANGE SCALE '
0
’\_, 816
/T\/ 818
Y
-> PLOT ON CONTROL CHART
(12)
(54)
United States Patent
(10) Patent N0.:
Ristea et a].
(45) Date of Patent:
METHOD FOR MONITORING JP
80016205 B
(75) Inventors: Catalin Ristea, Coquitlam (CA); James
*
M1980
OTHER PUBLICATIONS
ilTllllingzlsch?rigioiilagshizfisgR (US)’
Bramhall, G., “Meeting New Kiln Drying Standards,” 1975,
Vancouver (CA)
Canad1an Forest Industr1es, 95(9), pp. 33-35. Bramhall, G. and Warren, W.G., “Moisture Content Control in Drying Dimension Lumber,” 1977, Forest Products Jour
(73) Assigneez Wagner Electronic Products, Inc” Rogue River, OR (US)
Hal’ 27(7)’ PP' 26128' _
Subject to any disclaimer, the term Of this patent is extended or adjusted under 35
mal Distributions by Maximum Likelihood,” 1951, Journal of the American Statistical Associate, 46(254), pp. 206-212,
U.S.C. 154(1)) by 0 days.
1951 Cohen, A.C., Whitten, B.J., and Ding, Y., Modi?ed Moment Estimation for the Three Parameter Lognormal Distribution, 1985, Journal of Quality Technology, 17(2), pp. 92-99. Ferrell, E.B., “Control Charts for Log-Normal Universes,” 1958, Industrial Quality Control, 15(2), pp. 4-6.
Nov. 28, 2003
(65)
_
_
Cohen, A.C., “Est1mat1ng Parameters of Loganthmlc-Nor
(21) Appl, No.1 10/722,389 (22) Filed:
May 16, 2006
FOREIGN PATENT DOCUMENTS
WOOD-DRYING KILN STATE
( * ) Notice;
US 7,043,970 B2
Prior Publication Data
US 2005/0115307 A1
(51) Int‘ Cl‘ G01N 25/56
Jun. 2, 2005
(Commued) Primary ExamineriHeZron Williams Assistant Examinerilohn Fitzgerald (74) Attorney, Agent, or FirmiKlarquist Sparkman, LLP
(200601)
(52)
US. Cl. ........................................................ ..
(58)
Field of Classi?cation Search ................... .. 73/73
73/73
(57)
ABSTRACT
See application ?le for complete search history. (56)
References Cited
The method utilizes a known good set of moisture content data for Wood dried in a kiln under set conditions to estimate parameters for a Lognormal distribution Which approxi
U-S~ PATENT DOCUMENTS 4,107,599 A *
8/1978 Preikschat ................ .. 324/689
4,193,761 A * 3/1980 Mantegani 5,586,066 A * 12/1996 White et a1. 6,077,552 A T 6/2000 Chimenti 6t ill. 6,305,224 B1 *
10/2001 StanTSh et a1 ~~~~ ~~
6’308’57l Bl * 10/2001 Stamsh et a1 6,703,847 B1 *
3/2004 Venter et al
6,735,576
5/2004
Bl*
2003/0110007 A1 *
Kaji et a1.
mates the moisture Content data for those conditions- The
432/123 702/181 426/231
method creates control charts Which monitor some of these parameters. The method also takes data from subsequent Charges and transfonns to be represented on the
~~ 73/ 597
control charts. The method uses this representation to deter
" 73/597
mine if the kiln is operating under out-of-control conditions.
324/663 ......
. . . . . . ..
706/1
6/2003 McGee et a1. ............ .. 702/179
502 __,_
33 Claims, 9 Drawing Sheets
GATHER MULTIPLE SAMPLES OF KNOWN
GOOD DATA FOR A SPECIFIC KILN AND SET OF CONDITIONS
500
504
DETERMINE IF MOISTURE CONTENT DATA FOLLOWS LOGNORMAL DISTRIBUTION
ESTIMATE A LOGNORMAL DIST IBUTION THAT APPROXIMATES THE DATA
508 ~'“
APPROXIMATE THE LOGNORMAL THRESHOLD PARAMETER
512 ‘’
CALCULATE THE LOGNORMAL SHAPE PARAMETER
CALCULATE THE
LOGNORMAL SCALE PARAMETER
I
DEVELOP CONTROL CHARTS CREATE LOGARITHMIC SCALE CHART CREATE 522-” LOGARITHMIC SHAPE CHART
CREATE
mmos SCALE CHART v
CREATE ANTILOG SHAPE CHART
w 514
US 7,043,970 B2 Page 2 OTHER PUBLICATIONS
of British Columbia, Department of Mechanical Engineer
JolTe, AD, and Sichel, H.S., “A Chart for Sequentially Testing Observed Arithmetic Means from Lognormal Popu lations Against a Given Standard,” 1968, Technometrics, 10(3), pp. 605-612. Maki, R.G., and Milota, M.R., “Statistical Quality Control
Pratt, W.E., “Some Applications of Statistical Quality Con trol to the Drying of Lumber,” 1953, Journal of FPRS, 3(5),
Applied to Lumber Drying,” 1993, Quality Progress, 26(12),
1ng.
pp. 28-31.
Pratt, W.E., “Estimating the Moisture Content of Lumber
During the Drying Process,” 1956, Forest Products Journal, 6(9), pp. 333:337. ZWick, R.L. and Cook, J.D., “The Modeling of Moisture
pp. 75-80.
Content Distributions Based on Censored Readings from a
McMahon, E.P., “Applying Cumulative Frequency Distri
Resistance Meter,” 1985, Technical paper presented at West ern Dry Kiln Association Meeting. Maki, “An Application of Statistical Process Control Mea sures for Maintaining Optimal Quality from Dry Kiln Operations,” Thesis submitted to Oregon State University,
bution in Moisture Control During Kiln Drying,” Forest Products Journal, 11(3), pp. 133-138. Morrison, J ., “The Lognormal Distribution in Quality Con
trol,” 1958, Applied Statistics, 7(3), pp. 160-172. Noghondarian, K., “Quality Control With Non-Normal, Cen sured and Truncated Data,” 1997, Ph.D. Thesis, University
Apr. 26, 1991, 120 pages. * cited by examiner
U.S. Patent
Fig. 1
May 16, 2006
Sheet 1 0f 9
US 7,043,970 B2
U.S. Patent
May 16, 2006
US 7,043,970 B2
Sheet 2 0f 9
208,
0 21
202
206 201
2 Fig.
204 I11
202
U.S. Patent
May 16, 2006
Sheet 3 0f 9
US 7,043,970 B2
U.S. Patent
May 16, 2006
Sheet 4 0f 9
US 7,043,970 B2
Fig. 4a
Normal distribution curve
Fig. 4b Lognormal distribution curve
U.S. Patent
May 16, 2006
Sheet 5 0f 9
US 7,043,970 B2
F I9. 5 502 w
GATHER MULTIPLE SAMPLES OF'KNOWN
GOOD DATA FOR A SPECIFIC KILN AND SET OF CONDITIONS
500
V
504 4‘
DETERMINE IF MOISTURE CONTENT DATA
FOLLOWS LOGNORMAL DISTRIBUTION
Y
ESTIMATE A LOGNORMAL DISTRIBUTION THAT APPROXIMATES 508 _/~
THE DATA 510 _,~ APPROXIMATE THE
LOGNORMAL THRESHOLD PARAMETER
512 O-
CALCULATE THE
CALCULATE THE
LOGNORMAL
wings?“
SCALE PARAMETER
PARAMETER
w 514
DEVELOP CONTROL CHARTS 516 I
w
518
CREATE
CREATE
LOGARITHMIC
ANTILOG
SCALE CHART
sCALE CHART
CREATE 522“ LOGARITHMIC
SHAPE CHART
T" 520
CREATE ANTILOG
SHAPE CHART
w 522
U.S. Patent
May 16, 2006
US 7,043,970 B2
Sheet 6 0f 9
Hg. 6
TRANSFORM EACH MEASUREMENT
604 a
‘f 602
SAMPLE CALCULATIONS 606 T\-
Y
'
CALCULATE
CALCULATE
SAMPLE MEAN
STANDARD
,
w 608
DEVIATION y
I, A
” S
PARAMETER CALCULATION v_______ 610 A,
CALCULATE MEAN
v
w 614
OF SAMPLE
STANDARD
OF SAMPLE MEANS “y bar bar” FOR ESTIMATE OF ‘
SC A "E PA RAME TE R
"8 bar, 4 ' INTRODUCE
CONSTANT (04) TO CREATE UNBIASED ESTIMATOR OF SHAPE PARAMETER
_,_ 616
U.S. Patent
May 16, 2006
Sheet 7 0f 9
US 7,043,970 B2
Hg. 7 516
DEVELOP CONTROL CHART
710
ESTABLISH SAMPLE SIZE
J‘ 704
DETERMINE CENTER LINES
_,_ 706
CALCULATE CONTROL LIMITS
~1~ 708
gfCLn?AJ \”
CALCULATE
FOR SCALE
PROBABILITY LIMITS L/\ 712
CHART
FOR SHAPE CI-IART
SCALE LIMITS AND
J‘ 714
CENTER LINE
-———-——> v
PLOT NEW DATA ON
CHART
T" 716
U.S. Patent
May 16, 2006
Sheet 8 0f 9
US 7,043,970 B2
Fi g . 8
716
802
GATHER NEW DATA SET
TRANSFORM DATA SET 808 ~/~
TRANSFORM EACH
MEASUREMENT
814 \f:
’\_/ 810
CALCULATE
CALCULATE
SAMPLE
SAMPLE ST.
MEAN FOR
DEv FOR
SCALE CHART
SHAPE CHART
CHANGE SCALE '
0
’\_, 816
/T\/ 818
Y
-> PLOT ON CONTROL CHART
