a true picture of the process center and the random (i.e., common cause) variability inherent in the current process. If special, or assignable, causes of performance variability are present, as in Figure 1, this picture is skewed and the baseline model becomes invalid. When special cause variation is
absent, the process is performing to its full capability. T at is, when only random variation exists, the output gives a true picture of the performance of the current process. Moreover, it cannot get any bet- ter without a conscious, permanent change to the settings, materials, procedures, etc. Figure 2 shows X-bar and R
charts of a signifi cant casting dimension for a process exhibiting an out-of-control condition. T is simply means its output diff ers from what is expected from purely random variation. T is can manifest itself by points outside the +/- 3 sigma control limits or by any of a number of non-random patterns (good SPC software automatically tests for vari- ous types of non-random behavior). T is case shows several suspect behaviors, including points outside the control limits and trends. T e key point is that, by defi nition, non- random behavior is unexpected and typically has an identifi able causal factor that deviates from the normal process. If those deviations are not eliminated, they get lumped into the baseline assessment and give a false picture of the process capability and performance over time. So, if you are considering a PI
project, the fi rst step is to identify the key outputs and metrics of the process and demonstrate they are statistically stable. T is diff ers from saying the output values are desirable or within specifi cation. It means if you sample the process in the future, you have a high confi dence it will be performing much the same as it is today, good or bad. Once the process is demonstrated to be in statistical control, a capabil- ity analysis becomes meaningful. An analysis, like the one shown in Figure 3 for a significant casting
When special cause variation is absent, the process is performing to its full capability.
Fig. 2. This X-bar and R control chart shows a process dimension not in statistical control.
Fig. 3. Capability analysis is conducted on a stable process. April 2013 MODERN CASTING | 35
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60