2- Gage RnR Examples
2.1- RnR Gage analysis (no tolerance, no report tittle)
2.1.1- Introduction
This first example will be based on Ted Hessing article.
2.1.2- STEP 1: Get the measurement
[1]:
from pypetb import RnR
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
#set seaborn style to improve the figure sight
sns.set()
# Load measurement dataset
url = 'https://raw.githubusercontent.com/jgherruzo/myFreeDatasets/main/web.csv' # noqa
df=pd.read_csv(url,sep=';')
print(df.info()) # In order to check column names
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 90 entries, 0 to 89
Data columns (total 3 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Technician 90 non-null object
1 Parts 90 non-null int64
2 Valor 90 non-null float64
dtypes: float64(1), int64(1), object(1)
memory usage: 2.2+ KB
None
2.1.3- STEP 2: Make all calculation calling .RnRSolve()
[2]:
#Build up the model
dict_key={'1':'Technician','2':'Parts','3':'Valor'}
RnRModel=RnR.RnRNumeric(
mydf_Raw=df,
mydict_key=dict_key
)
#Solve it
RnRModel.RnRSolve()
#Check the calculation
print(RnRModel.getLog())
Model is created
== DATASET EVALUATION ==
Operator: 3
Trials: 3
Piezes: 10
== CALCULATION ==
Total data: 90
Max. measured value: 5.8900
Min. measured value: 1.7200
Avg. measured value: 3.0660
Avg. Control limits
UCL: 3.4575
LCL: 2.6745
Avg. Range measured: 0.3827
Range Control limits
UCL: 0.9850
LCL: 0.0000
Sum of deviation by operator: 0.003331
Total Operator Sum of deviation: 0.099927
Sum of deviation by part: 12.947714
Total Part Sum of deviation: 116.529427
Total squared deviation: 120.681960
Equipment squared deviation: 3.605800
Iteration sum of squared: 0.446807
2.1.4- STEP 3: Check Anova Result
Anova calculation is implemented into other python package. The only different here is that result is obtained as pandas dataframe and each parameter could be accesed using its column and index for further calculations
[3]:
df_Result=RnRModel.RnRAnova()
#Checking Anova table
print(df_Result)
#accesing one individual value
print(f"Degree of freedom for part: {df_Result['DF'].loc['Part']}")
DF SS MS F P
Source of variability
Technician 2 0.099927 0.049963 2.012817 1.625911e-01
Part 9 116.529427 12.947714 521.610063 1.110223e-16
TechxPart (iteration) 18 0.446807 0.024823 0.413044 9.799258e-01
Repeatability with 60 3.605800 0.060097 NaN NaN
Repeatability without 78 4.052607 0.051956 NaN 9.799258e-01
Total 89 120.681960 NaN NaN NaN
Degree of freedom for part: 9
2.1.5- STEP 4: Variance table
Table that summarize the impact of each parameter to the variance. It helps a lot to find where the effort must be focused
[4]:
df_Result=RnRModel.RnR_varTable()
#Checking var. table
print(df_Result)
#accesing one individual value
print('\nRnR RESULT:\n-------------------')
dbl_RnR=df_Result['% Contribution'].loc['Total Gage R&R']
print(f"Total Gage R&R: {dbl_RnR:.3f}")
if dbl_RnR<1:
print('<1% --> Acceptable measurement system')
elif dbl_RnR>=1 and dbl_RnR<=9:
print(
'1-9%--> It may be acceptable depending on application and cost'
)
else:
print(
'>9% --> Unacceptable measurement system, it must be improved'
)
Variance % Contribution
Source
Total Gage R&R 0.051956 3.499182
Eq.Var. (Repeatability) 0.051956 3.499182
Op.Var. (Reproducibility) 0.000000 0.000000
Technician 0.000000 0.000000
Technician x Part iter. 0.000000 0.000000
Part to Part 1.432862 96.500818
Total variation 1.484818 100.000000
RnR RESULT:
-------------------
Total Gage R&R: 3.499
1-9%--> It may be acceptable depending on application and cost
2.1.6- STEP 5: Standard Deviation table
Table that summarize the impact of each parameter to the standard deviation. It helps a lot to find where the effort must be focused
[5]:
df_Result=RnRModel.RnR_SDTable()
#Checking sd table
print(df_Result)
#accesing one individual value
print('\nAutomotive Industry Action Group (AIAG) measurement system assessment:\n-------------------')
dbl_RnR=df_Result['% Study Var'].loc['Total Gage R&R']
print(f"Total Gage R&R: {dbl_RnR:.3f}")
if dbl_RnR<10:
print('<10% --> Acceptable measurement system')
elif dbl_RnR>=10 and dbl_RnR<=30:
print(
'10-30%--> It may be acceptable depending on application and cost'
)
else:
print(
'>30% --> Unacceptable measurement system, it must be improved'
)
StdDev (SD) StudyVar (6*SD) % Study Var
Source
Total Gage R&R 0.227940 1.367638 18.70610
Eq.Var. (Repeatability) 0.227940 1.367638 18.70610
Op.Var. (Reproducibility) 0.000000 0.000000 0.00000
Technician 0.000000 0.000000 0.00000
Technician x Part iter. 0.000000 0.000000 0.00000
Part to Part 1.197022 7.182133 98.23483
Total variation 1.218531 7.311188 100.00000
Automotive Industry Action Group (AIAG) measurement system assessment:
-------------------
Total Gage R&R: 18.706
10-30%--> It may be acceptable depending on application and cost
2.1.7- STEP 6: Check the run chart
This chart summaryze all the measurement by part and operator, so in case of big difference, it makes easier find what could be the source
[6]:
call=RnRModel.RnR_RunChart()
plt.show()
2.1.8- STEP 8: Check the final report
This figure is a mix of chart that will help to find where to be focused in case of being out the recommended limits. The report througth some advisement based on AIAG assesment standard and the number of categories showed by the system
[7]:
call=RnRModel.RnR_Report()
plt.show()
2.2- RnR Gage analysis (using tolerance, adding a report tittle)
2.2.1- Introduction
Second example is commonly used during Minitab tutorial: https://youtu.be/JbvXY7PeKxU?si=ZDlJerdh_SJM_vy6
2.2.2- STEP 1: Get the measurement
[8]:
# Load measurement dataset
url = 'https://raw.githubusercontent.com/jgherruzo/myFreeDatasets/main/RnR_Example.csv' # noqa
df=pd.read_csv(url,sep=';')
print(df.info()) # In order to check column names
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 90 entries, 0 to 89
Data columns (total 3 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Part 90 non-null int64
1 Operator 90 non-null object
2 Measurement 90 non-null float64
dtypes: float64(1), int64(1), object(1)
memory usage: 2.2+ KB
None
2.2.3- STEP 2: Make all calculation calling .RnRSolve()
[9]:
#Build up the model
dict_key={
'1':'Operator',
'2':'Part',
'3':'Measurement'
}
RnRModel=RnR.RnRNumeric(
mydf_Raw=df,
mydict_key=dict_key,
mydbl_tol=8
)
#Solve it
RnRModel.RnRSolve()
#Check the calculation
print(RnRModel.getLog())
Model is created
== DATASET EVALUATION ==
Operator: 3
Trials: 3
Piezes: 10
== CALCULATION ==
Total data: 90
Max. measured value: 2.2600
Min. measured value: -2.1600
Avg. measured value: 0.0014
Avg. Control limits
UCL: 0.3510
LCL: -0.3481
Avg. Range measured: 0.3417
Range Control limits
UCL: 0.8794
LCL: 0.0000
Sum of deviation by operator: 0.105575
Total Operator Sum of deviation: 3.167262
Sum of deviation by part: 9.817993
Total Part Sum of deviation: 88.361934
Total squared deviation: 94.647112
Equipment squared deviation: 2.758933
Iteration sum of squared: 0.358982
2.2.4- STEP 3: Check Anova Result
Anova calculation is implemented into other python package. The only different here is that result is obtained as pandas dataframe and each parameter could be accesed using its column and index for further calculations
[10]:
df_Result=RnRModel.RnRAnova()
#Checking Anova table
print(df_Result)
#accesing one individual value
print(f"Degree of freedom for part: {df_Result['DF'].loc['Part']}")
DF SS MS F P
Source of variability
Technician 2 3.167262 1.583631 79.406049 1.174478e-09
Part 9 88.361934 9.817993 492.291423 1.110223e-16
TechxPart (iteration) 18 0.358982 0.019943 0.433721 9.741064e-01
Repeatability with 60 2.758933 0.045982 NaN NaN
Repeatability without 78 3.117916 0.039973 NaN 9.741064e-01
Total 89 94.647112 NaN NaN NaN
Degree of freedom for part: 9
2.2.5- STEP 4: Variance table
Table that summarize the impact of each parameter to the variance. It helps a lot to find where the effort must be focused
[11]:
df_Result=RnRModel.RnR_varTable()
#Checking var. table
print(df_Result)
#accesing one individual value
print('\nRnR RESULT:\n-------------------')
dbl_RnR=df_Result['% Contribution'].loc['Total Gage R&R']
print(f"Total Gage R&R: {dbl_RnR:.3f}")
if dbl_RnR<1:
print('<1% --> Acceptable measurement system')
elif dbl_RnR>=1 and dbl_RnR<=9:
print(
'1-9%--> It may be acceptable depending on application and cost'
)
else:
print(
'>9% --> Unacceptable measurement system, it must be improved'
)
Variance % Contribution
Source
Total Gage R&R 0.091429 7.762159
Eq.Var. (Repeatability) 0.039973 3.393677
Op.Var. (Reproducibility) 0.051455 4.368482
Technician 0.051455 4.368482
Technician x Part iter. 0.000000 0.000000
Part to Part 1.086447 92.237841
Total variation 1.177875 100.000000
RnR RESULT:
-------------------
Total Gage R&R: 7.762
1-9%--> It may be acceptable depending on application and cost
2.2.6- STEP 5: Standard Deviation table
Table that summarize the impact of each parameter to the standard deviation. It helps a lot to find where the effort must be focused
[12]:
df_Result=RnRModel.RnR_SDTable()
#Checking sd table
print(df_Result)
#accesing one individual value
print('\nAutomotive Industry Action Group (AIAG) measurement system assessment:\n-------------------')
dbl_RnR=df_Result['% Study Var'].loc['Total Gage R&R']
print(f"Total Gage R&R: {dbl_RnR:.3f}")
if dbl_RnR<10:
print('<10% --> Acceptable measurement system')
elif dbl_RnR>=10 and dbl_RnR<=30:
print(
'10-30%--> It may be acceptable depending on application and cost'
)
else:
print(
'>30% --> Unacceptable measurement system, it must be improved'
)
StdDev (SD) StudyVar (6*SD) % Study Var \
Source
Total Gage R&R 0.302372 1.814229 27.860651
Eq.Var. (Repeatability) 0.199933 1.199599 18.421935
Op.Var. (Reproducibility) 0.226838 1.361025 20.900913
Technician 0.226838 1.361025 20.900913
Technician x Part iter. 0.000000 0.000000 0.000000
Part to Part 1.042327 6.253965 96.040534
Total variation 1.085300 6.511797 100.000000
% tol (VE/tol)
Source
Total Gage R&R 22.677864
Eq.Var. (Repeatability) 14.994988
Op.Var. (Reproducibility) 17.012814
Technician 17.012814
Technician x Part iter. 0.000000
Part to Part 78.174562
Total variation 81.397467
Automotive Industry Action Group (AIAG) measurement system assessment:
-------------------
Total Gage R&R: 27.861
10-30%--> It may be acceptable depending on application and cost
2.2.7- STEP 6: Check the run chart
This chart summaryze all the measurement by part and operator, so in case of big difference, it makes easier find what could be the source
[13]:
call=RnRModel.RnR_RunChart()
plt.show()
2.2.8- STEP 8: Check the final report
This figure is a mix of chart that will help to find where to be focused in case of being out the recommended limits. The report througth some advisement based on AIAG assesment standard and the number of categories showed by the system
[14]:
call=RnRModel.RnR_Report("my Report Tittle")
plt.show()
