Operations Management

Dr. Jinil Persis

Assistant Professor
IIM Kozhikode
 Process variables
Process variables affecting quality
characteristics Product
X1
characteristics
X2
P1 X3

P2 X4
Process
Product X5
Input P3
X6
Process variables Product
characteristics X7

Dependencies
Which variables affect a product characteristic?
What are operating conditions to get target characteristics? Explained part of variation
Robust Design using Design of Experiments
Controllable variables
• Robust design of a product is (𝑋1 , 𝑋2 ,……)
insensitive to variations in the ….
process environment or in customer
environment Input Process Output (𝑌1 , 𝑌2 ,……)
CTQ
• Design of experiments (DOE) or
….

experimental design or designed

experiments is the quality Un - Controllable variables
(𝑍1 , 𝑍2 ,……)
management tool used to identify
optimal levels for nominal Optimal setting (𝑋1 , 𝑋2 ,……)* for desirable
dimensions. values of (𝑌1 , 𝑌2 ,……) is identified during design
by conducting experiments/optimization
Monitoring process variability & Establishing
Statistical process control (SPC)
• If the process is capable by design (say, Cp>=1.66), start monitoring
the quality characteristic using control charts.
• Sampling inspection of outgoing units will be carried out and the quality data
will be plotted to see variation.
• If the process is not capable by design, then, 100% inspection of all
outgoing units has to be performed for avoiding delivery of defective
units.
• Process redesign should be initiated to control variability.
Monitoring process variability & Establishing
Statistical process control (SPC) (Contd.)
• monitoring a process to identify special causes of variation and signal the need to
take corrective action when appropriate during production
• Control chart - a run chart to which two horizontal lines, called control limits are
added: the upper control limit (UCL) and lower control limit (LCL)

• A point that plots outside of the control limits is

interpreted as evidence that the process is out of
control
• Sometimes some specific patterns in control chart
also gives out-of-control signal
• Investigation and corrective action are required to
find and eliminate the assignable cause or causes
responsible for this behaviour.
 Controlling a process with variable control charts
Take adequate number of samples

Compute 𝑋ത 𝑎𝑛𝑑 𝑅 for each sample

Determine 𝑋ധ 𝑎𝑛𝑑 𝑅ത

Compute control limits

Process
control?
Specs
met?
Identify and remove cause
Continue sampling Review specs Continue sampling
Detect trends any Change process design Update control limits
Determine m/c resetting frequency 100% inspection
Determining control limits
Start base period with atleast 25 samples

Compute control limits for 𝑋ത 𝑎𝑛𝑑 𝑅

Outliers?
Remove outlier data points
R

Outliers?
Set control limits
𝑋ത

Remove outlier data points

Remaining
>=15?
Determining control limits (Contd.)
ഥ chart:
Control limits of 𝒙
If 𝜎𝑥 − 𝑘𝑛𝑜𝑤𝑛 ℎ𝑖𝑠𝑡𝑜𝑟𝑖𝑐𝑎𝑙𝑙𝑦 ,
𝜎𝑥 𝜎𝑥
𝐸 𝑥ҧ ± 3𝜎𝑥ҧ ⇒ 𝜇 ± 3𝜎𝑥ҧ ⇒ 𝐸 𝑥ҧ ± 3 ⇒ 𝑥Ӗ ± 3 ന ± 𝑨𝝈𝒙
⇒𝒙
𝑛 𝑛
where n- sample size
𝑠ҧ 𝑅ത
If 𝜎𝑥 − 𝑢𝑛𝑘𝑛𝑜𝑤𝑛, 𝑡ℎ𝑒𝑛 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒, 𝜎 ෞ𝑥 = or
𝑐4 𝑑2
𝜎𝑥 𝑅ത
• 𝐸 𝑥ҧ ± 3𝜎𝑥ҧ ⇒ 𝜇 ± 3𝜎𝑥ҧ ⇒ 𝐸 𝑥ҧ ± 3 ⇒ 𝑥Ӗ ± 3 ഥ
ന ± 𝑨𝟐 𝑹
⇒𝒙
𝑛 𝑑2 𝑛
𝜎𝑥 𝑠ҧ
• 𝐸 𝑥ҧ ± 3𝜎𝑥ҧ ⇒ 𝜇 ± 3𝜎𝑥ҧ ⇒ 𝐸 𝑥ҧ ± 3 ⇒ 𝑥Ӗ ± 3 ന ± 𝑨𝟑 𝑺ത
⇒𝒙
𝑛 𝑐4 𝑛
Determining control limits (Contd.)
Control limits of R chart: 𝐸 𝑅 ± 3𝜎𝑅
𝑅ത
It is known that 𝜎
ෞ𝑥 = ; Therefore,𝐸 𝑅 = 𝜎𝑥 𝑑2
𝑑2
It is known that 𝜎𝑅 =𝜎𝑥 𝑑3
If 𝜎𝑥 − 𝑘𝑛𝑜𝑤𝑛 ℎ𝑖𝑠𝑡𝑜𝑟𝑖𝑐𝑎𝑙𝑙𝑦 ,
𝐸 𝑅 ± 3𝜎𝑅 ⇒ 𝜎𝑥 𝑑2 ± 3𝜎𝑥 𝑑3 ⇒ 𝜎𝑥 𝑑2 ± 3𝑑3
Let, 𝐷1 = 𝑑2 − 3𝑑3 and 𝐷2 = 𝑑2 + 3𝑑3 . Then, UCLR=𝐷2 𝜎𝑥 ; LCLR=𝐷1 𝜎𝑥
𝑅ത
If 𝜎𝑥 − 𝑢𝑛𝑘𝑛𝑜𝑤𝑛, then, estimate, 𝜎
ෞ𝑥 =
𝑑2
𝑑3 𝑅 ത
ത
𝐸 𝑅 ± 3 𝜎𝑥 𝑑3 ⇒ 𝑅 ± 3
𝑑2
𝑑3 𝑑3
Let, 𝐷3 = 1 − 3 and 𝐷4 = 1 + 3 , ത LCLR=𝐷3 𝑅ത
Then, UCLR=𝐷4 𝑅;
𝑑2 𝑑2
Determining control limits (Contd.)
Control limits of S chart: 𝐸 𝑠 ± 3𝜎𝑠
𝑠ҧ
It is known that 𝜎
ෞ𝑥 = ; Therefore,𝐸 𝑠 = 𝜎𝑥 𝑐4
𝑐4
It is known that 𝜎𝑠 =𝜎𝑥 1 − 𝑐42
If 𝜎𝑥 − 𝑘𝑛𝑜𝑤𝑛 ℎ𝑖𝑠𝑡𝑜𝑟𝑖𝑐𝑎𝑙𝑙𝑦 ,
𝐸 𝑠 ± 3𝜎𝑠 ⇒ 𝜎𝑥 𝑐4 ± 3𝜎𝑥 1 − 𝑐42 ⇒ 𝜎𝑥 𝑐4 ± 3 1 − 𝑐42
Let, 𝐵5 = 𝑐4 + 3 1 − 𝑐42 and 𝐵6 = 𝑐4 − 3 1 − 𝑐42 , 𝑡hen, UCLs=𝐵6 𝜎𝑥 ; LCLs=𝐵5 𝜎𝑥
𝑠ҧ
If 𝜎𝑥 − 𝑢𝑛𝑘𝑛𝑜𝑤𝑛, then, 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑒, 𝜎
ෞ𝑥 =
𝑐4
𝑠ҧ 1−𝑐42 1−𝑐42
𝐸 𝑠 ± 3 𝜎𝑥 𝑑3 ⇒ 𝑠ҧ ± 3 ⇒ 𝑠ҧ 1 ± 3
𝑐4 𝑐4
1−𝑐42 1−𝑐42
Let, 𝐵4 = 1 + 3 and 𝐵3 = 1 − 3 , then, UCLs=𝐵4 𝑠;ҧ LCLs=𝐵3 𝑠ҧ
𝑐4 𝑐4
 ഥ − 𝑪𝒉𝒂𝒓𝒕 and R chart
Why both 𝒙
• 𝑥ҧ − 𝐶ℎ𝑎𝑟𝑡 → Shift in process mean
• R & S - Charts → Variability in the process
Reading a control chart

Cycles Mixture Shift in level

Trend Stratification pattern

Example
A hard-bake process is used in conjunction with photolithography in USL 2
semiconductor manufacturing. We wish to establish statistical control of the LSL 1
flow width of the resist in this process using and R charts. Twenty-five samples Cp 1.192058 Capable
k 3.576173
each of size five wafers, have been taken when we think the process is in Fraction defective
0.999647 352.5044 ppm
control. The interval of time between samples or subgroups is one hour. The cpk 1.178682 Capable
flow width measurement data (in x microns) from these samples are shown. The
specification limits on flow width are 1.50 ± 0.50 microns.

Number 1 2 3 4 5 xbar R UCL USL Process mean

Target meanLSL LCL UCL Rbar LCL
1 1.3235 1.4128 1.6744 1.4573 1.6914 1.51188 0.3679 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
2 1.4314 1.3592 1.6075 1.4666 1.6109 1.49512 0.2517 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
3 1.4284 1.4871 1.4932 1.4324 1.5674 1.4817 0.139 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
4 1.5028 1.6352 1.3841 1.2831 1.5507 1.47118 0.3521 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
5 1.5604 1.2735 1.5265 1.4363 1.6441 1.48816 0.3706 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
6 1.5955 1.5451 1.3574 1.3281 1.4198 1.44918 0.2674 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
7 1.6274 1.5064 1.8366 1.4177 1.5144 1.5805 0.4189 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
8 1.419 1.4303 1.6637 1.6067 1.5519 1.53432 0.2447 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
9 1.3884 1.7277 1.5355 1.5176 1.3688 1.5076 0.3589 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
10 1.4039 1.6697 1.5089 1.4627R chart1.522 1.51344 0.2658 1.693191 2 1.50561 1.5 X 1
bar 1.31803
chart 0.68749 0.325208 0
11 1.4158 1.7667 1.4278 1.5928 1.4181 1.52424 0.3509 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
0.8 2.5
12 1.5821 1.3355 1.5777 1.3908 1.7559 1.5284 0.4204 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
0.7
13 1.2856 1.4106 1.4447 1.6398 1.1928 1.3947 0.447 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
14 0.6
1.4951 1.4036 1.5893 1.6458 1.4969 1.52614 0.2422 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
15 1.3589
0.5 1.2863 1.5996 1.2497 1.5471 1.40832 0.3499 1.693191 2
2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
16 1.5747
0.4 1.5301 1.5171 1.1839 1.8662 1.5344 0.6823 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
17 1.368
0.3 1.7269 1.3957 1.5014 1.4449 1.48738 0.3589 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
18 1.4163 1.3864 1.3057 1.621 1.5573 1.45734 0.3153 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
0.2
19 1.5796 1.4185 1.6541 1.5116 1.7247 1.5777 0.3062 1.693191 1.5 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
0.1
20 1.7106 1.4412 1.2361 1.382 1.7601 1.506 0.524 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
21 1.43710 1.5051 1.3485 1.567 1.488 1.46914 0.2185 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
22 1.4738 1.5936 1.6583 1.4973 1.472 1.539 0.1863 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
23 1.5917 1.4333 1.5551 R 1.5295
UCL 1.6866
Rbar 1.55924
LCL 0.2533 1.693191 1 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
24 1.6399 1.5243 1.5705 1.5563 1.553 1.5688 0.1156 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
25 1.5797 1.3663 1.624 1.3732 1.6887 1.52638 0.3224 1.693191 2 1.50561 1.5 1 1.31803 0.68749 0.325208 0
process
0.5
mean 1.50561 0.325208 Rbar

Process SD 0.139814
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

xbar UCL USL Process mean

Target mean LSL LCL