Course 1 Essay Example

Statistical Process Control

Unit 036: Statistical Process Control

  1. Two basic types of inspection used in sampling for process control

The two types of sampling used in inspection are:

  1. Sampling inspection by attributes

  2. Sampling inspection by variables

  1. Sampling inspection by attributes

This is a method used to evaluate the quality or the characteristics of an item and classifying them as nonconforming or conforming depending on whether it is conforming to the standard specification. They can be characterized in terms of quantity or quality. The numbers of items which have nonconforming attributes are counted and if the number has not been exceeded, the lot is accepted. The advantage of this method is it simple to use and is more robust (Das, 2008).

b) Sampling inspection by variables

This method is used to evaluate the quality of items by measuring the value of the variable characterizing the inspected commodity. This method begins by selecting a number of items and measuring the characteristics or the dimensions so as to know if the characteristics of the sample are within the particular limits but not the actual value of the characteristics. The acceptance of a lot is based on the calculations of the variability or the average of the measurements in accordance with the set standards. This method requires smaller sample size compared to attribute method. It also provides more information about the effects of the process or mean on the quality (Das, 2008; Ravindran, 2008).

  1. Significance of natural and assignable causes of variation

Causes of variation are important during a quality inspection of a product. Thus, an understanding variation is essential for the management and success of operation process. Variation can be classified as natural variation and assignable variation. Natural variation is common or chance causes of variability, which occurs and cannot be traced to a particular cause. It is created by number of influences of minor factors within a predictable range and little can be done other than revise the change the fundamental process. This variation is the sum of a number of effects of multifaceted interaction of random cause, which may be slight. A set of random or common causes that creates variation in the product quality may originate from the variation from the input to the process. An old machine, for example, has a higher degree of natural variability compared to a new machine. Sources of this variation include temperature changes, vibration of the equipment, changes in the emotion and physical conditions of the operator or electrical changes. The process is stable if the common variations are present (Harry, 2010; Ravindran, 2008).

Assignable variation represents large unsatisfactory interruptions to the normal performance process. They are those effects which can be detected and controlled. A process operating with the existence of assignable causes is said to be out of control. They may include defective raw materials, improperly adjusted machine or operator error. These effects can be traced in order to change the equipment, operation technique or the materials used. A control chart may be used for monitoring the process in order to detect the presence of assignable causes. Thus, if the plot point is outside the limits of the control charts, the assignable cause is likely to have occurred. The process of variability can be reduced by identifying these occurrences and working to remove the causes from their process (Harry, 2010).

  1. Frequency distribution and mean, range and standard deviation

The table for Frequency distribution

Frequency

(xmid — x)2

(xmid — x)2f

1667.089

1037.849

537.2642

2745.193

1171.057

6520.833

course 1

Mean is given by

course 1 1 = 840.61 ohmscourse 1 2=

Sample standard deviation,

course 1 3 ohmscourse 1 4=

The range is given by

861.9 — 826 = 35.9 ohms

  1. Characteristics of the normal curve to the distribution of the means of small samples

Small samples in distribution graph shows normal distribution. For example, column 1, 2, and 3 are almost symmetrical about the central point. The same observation is shown in columns 2, 3 and 4. The last three columns are skewed. For normal the distribution, the median position coincides with the mean and the mode. The calculated mean value, which is 840.61 ohms, and the median, 844.5, does not coincide in the data set.

  1. Appropriate control chart limits

The upper control limit = Average value + 3 x Standard deviation = 840.61 + (3×11.092) = 873.886 ohms

The lower control limit = Average value – 3 x Standard deviation = 840.61 — (3×11.092) = 807.334 ohms

  1. Control charts for variables, rejects per unit and percentage defectives per batch

course 1 5

course 1 6

  1. Control program for an application

Control Program

Defective motors

X = Subgroup Id (1 to 18)

Numdef = Number of defective items in sub-group

Size = Total number of items in sub-group

Serial read numdef

Serial read size

Let x = Sequence 1 to 18

Lines solid solid dot dot

Xlimits 0 18

Xtic offset 0 1

Ylimits 0 10

Ytic offset 18

P control chart numdef size x

  1. Ungrouped data (the mean, range and standard deviation)

Table showing the data set

Resistors (ohms)

Sample number

15126.87

course 1 7

course 1 8

, where dcourse 1 9Standard deviation, σ =n =2.059 (Hartley’s constant) (Das, 2008).

Sample size, n = 3

Therefore, σ = 0.867/3 = 0.289 ohms

= 0.5course 1 11 = 3 x 0.289/course 1 10Standard errors =

  1. Relationship between the normal curve and the mean values

The process mean is approximately at the median position, indicating that there generally normal distribution.

  1. Select and group the data based on variable and attribute inspection methodology.

Based on variables aspects of the data, it is assumed that the quality of the characteristics of the data set follow normal distribution with the standard deviation σ and the mean, x.

The attributes aspects suppose that random samples of fixed value of the number size n taken from a group with a given number of defective products. Therefore, the number of defects follows binomial distribution.

  1. Control charts, calculate the limits

0.5 ohms course 1 13 = 840.381course 1 12Control limits =

Upper control limit = 840.881 ohms

Lower control limit = 839.881 ohms

Range chart

= 2.57 x 0.867 = 2.228 ohms course 1 14Control line = 2.57

  1. Control chart for variable inspection methodology, showing rejects per unit and percentage defects per batch

Control chart

course 1 15

course 1 16

References

Walker, H. F. (2012). The certified quality inspector handbook. Milwaukee, Wis: ASQ Quality Press.

Das, (2008). Statistical Methods (Combined), Tata McGraw-Hill Education

Harry, M. J. (2010). Practitioner’s guide for statistics and lean six sigma for process improvements. Hoboken, N.J: John Wiley & Sons.

Ravindran A. R,, (2008). Operations Research Applications, Operations Research Series

CRC Press