Scribd
Upload
72 views

Monitoreo de Perfiles: Rubén Darío Guevara González

The document summarizes two research papers on monitoring linear profiles. It discusses measuring process capability for linear profiles, which involves preserving an algebraic equation rather than single quantities. Classic single-variable equations for process capability ratios do not apply to multiple variable forms of profiles. It also covers capability indices for control charts based on regression models. Formulas are presented for control limits of regression charts, specification limits, and capability ratios for processes with symmetric specification limits. An example is provided to illustrate calculating these values.

Uploaded by

Rocegu
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
72 views

Monitoreo de Perfiles: Rubén Darío Guevara González

The document summarizes two research papers on monitoring linear profiles. It discusses measuring process capability for linear profiles, which involves preserving an algebraic equation rather than single quantities. Classic single-variable equations for process capability ratios do not apply to multiple variable forms of profiles. It also covers capability indices for control charts based on regression models. Formulas are presented for control limits of regression charts, specification limits, and capability ratios for processes with symmetric specification limits. An example is provided to illustrate calculating these values.

Uploaded by

Rocegu
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 17

Temas de la presentación RAZAVI, H. y ASADZADEH, A.

SOUZA, PEDRINI y TEN CATEN

MONITOREO DE PERFILES

RUBÉN DARÍO GUEVARA GONZÁLEZ

Universidad Nacional de Colombia Sede Bogotá


Seminario I

27 de marzo de 2011
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

Temas de la presentación

RAZAVI, H. y ASADZADEH, A. (2010). Measuring Process


Capability for Linear profiles. INTERNATIONAL JOURNAL
OF INDUSTRIAL ENGINEERING AND PRODUCTION
MANAGEMENT (IJIE), 21(3):33-43.
SOUZA, F.S.; PEDRINI, D. y TEN CATEN, C. S. (2011).
Capability Indices for Control Charts Based on Regression
Models.
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

Temas de la presentación

RAZAVI, H. y ASADZADEH, A. (2010). Measuring Process


Capability for Linear profiles. INTERNATIONAL JOURNAL
OF INDUSTRIAL ENGINEERING AND PRODUCTION
MANAGEMENT (IJIE), 21(3):33-43.
SOUZA, F.S.; PEDRINI, D. y TEN CATEN, C. S. (2011).
Capability Indices for Control Charts Based on Regression
Models.
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

Process capability indices such as Cp , Cpk y Cpm


represent PCRs while dividing allowable process spread
by actual process spread. The numerical result is generally
compared to a minimum acceptable limit in order to decide
on the process performance.
The ratio can also be interpreted as the capability of the
process to produce its scalar quality characteristic within
its specification limits. This argument is slightly different for
linear profiles, since the process must preserve an
algebraic equation rather than single quantities. In
addition, classic single-variable equations for PCRs do not
apply on the multiple variable forms of the profiles.
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

Limites de Control de la Carta de Regresion - Mandel (1969)


√
LCS = Y bi + L CMR
LC = Ybi
√
LCI = Ybi − L CMR
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

The slope measures in Table 2, are tested for statistical control


using EWMA chart with r = 0,8. The large r makes the method
sensitive to small shifts. If the initial trial shows out of control
points, those points are omitted assuming assignable causes
and the calculations are repeated.
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

Límites de Especificación

USL = UCL + s (1)


LSL = LCL − s (2)
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

Límites de Especificación

Ti = b0T + b1T x1i + b2T x2i + ... + bkT xki (3)


USLi = b0U + b1U x1i + b2U x2i + ... + bkU xki (4)
LSLi = b0L + b1L x1i + b2L x2i + ... + bkL xki (5)
(6)

donde i = 1, 2, ..., n. b0T , b0U y b0L son los interceptos del valor
objetivo, USL y del LSL. En la mayoría de los casos se admite
que T , LSL y USL son paralelos, luego bjT = bjU = bjL = bj .
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

Indices de capacidad GR para procesos con límites de


especificación simétricos

Pn
i=1 USLi − LSLi
USL − LSL = (7)
n
Asumiendo que los límites de especificación sean paralelos,
Pn
USLi − LSLi n(b0U − bOL )
USL − LSL = i=1 = = b0U − bOL
n n
(8)

entonces

b pR = b0U − bOL
C (9)
6b
σR
Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

Ejemplo
En FASE I, Sea CMR = 70,06

ybi = 89,84 + 9,49x1 − 5,13x2 − 6,48x3 + 12,59x4 − 7,94x12


(10)

En FASE II se toman 100 nuevas muestras,


Temas de la presentación RAZAVI, H. y ASADZADEH, A. SOUZA, PEDRINI y TEN CATEN

Ejemplo
Consíderese

LSU = 160, LSI = 50, T = 105 (11)

Luego

LSUi = 160 + 9,49x1 − 5,13x2 − 6,48x3 + 12,59x4 − 7,94x12


(12)
LSIi = 50 + 9,49x1 − 5,13x2 − 6,48x3 + 12,59x4 − 7,94x12
(13)
Ti = 105 + 9,49x1 − 5,13x2 − 6,48x3 + 12,59x4 − 7,94x12
(14)
(15)

You might also like