3

I will give a brief explanation to my scenario. The company mass produces components like valves/nuts/bolts etc which need to measured for dimensions (like length,radius,thickness etc) for quality purposes. As it is not feasible to inspect all the pieces, they are chosen in a batch style. Foe eg: from a batch of every 100 pieces, 5 will be randomly selected & mean of their dimensions measured & noted for drawing SPC control charts (plots mean dimension on y axis & batch number on x axis).

Even though there are a number of factors (like operator efficiency, machine/tool condition etc) which affect the quality of the product, they don't seem to be measurable. My objective is to develop a machine learning model to predict the product dimensions of the coming batch samples(mean). This will help the operator to forecast if there is going to be any significant dimensional variation so that he can pause working & figure out potential reasons & thus prevent the wastage of the product/material.

I have some idea about R programming & machine learning techniques like decision trees/regression etc but couldn't land on a proper model for this. Mainly because I couldn't think of the independent variables for this situation. I don't have much idea about time series modelling though.

Will someone throw some insights/ideas/suggestions about how to tackle this. I am sorry that I had to write a long story but just wanted to make things as clear as possible.

Thanks in advance. Sreenath

Sreenath1986
  • 167
  • 4
  • 16
  • 2
    This approach--*predicting* future means so an operator can stop working and look for ways to improve--is kind of backward for statistical process control. Instead, I'd expect to measure variation, plot it, determine from the plot whether the process is in control, and if it's not, look for special causes of variation. What am I missing? – Mike Sherrill 'Cat Recall' Aug 22 '16 at 15:55
  • 1
    Hi Mike, thanks for the comment which is really valid. With the SPC charts, we can see whether the process is in control. But by the time we notice a variation in process via control chart, some defective parts will be already made (which is captured in SPC chart). But if we can find some pattern for the SPC from the historic data (for eg: after every 5 hours of machine operation there tends to be a spike in no of defectives, probably due to machine getting heated up), we can better control the process. Sorry, if I am not clear, I am just getting into SPC from conventional data analysis. – Sreenath1986 Aug 23 '16 at 05:11
  • 1
    *"But if we can find some pattern for the SPC from the historic data..."* then that pattern would be from things you've actually measured and recorded *in the past*. And you'd be expected to determine whether that variation comes from common causes or from special causes. And if it comes from special causes, you'd be expected to eliminate those special causes. (Like overheating.) No prediction is necessary, and is probably unwise. I think you should do more research into statistical quality control and into control charts before you go down this path. – Mike Sherrill 'Cat Recall' Aug 23 '16 at 15:39
  • Thanks Mike. I will take your advice & get a better understanding of the process & concepts. – Sreenath1986 Aug 23 '16 at 16:11
  • @Sreenath1986 it would be nice to hear your findings about your predictive model. What did you learn? Did it help? – ozkary Apr 08 '18 at 17:00
  • @ozkary unfortunately I couldn't proceed with task that as I left the organization. – Sreenath1986 Jun 26 '18 at 05:21

1 Answers1

3

Your requirement may apply with three level by steps:

1.Fundamental

Automatic apply SPC rule with machine learning, ex. identify SPC chart pattern with Nelson rule, and extend to new pattern of variation in specific process.

2.Supplemental

Predicate Cp and SPC trend with multivariant collection and machine learning. For example, particle of smoke will impact wafer yield rate, it may earlier to found if data analysis model link SPC and worker shift arrangement

3.Intelligent agent

Automatic process event through integration between SPC and reaction plan. The agent model by link SPC and FMEA and build with CEP engine in BAM architecture.

Jesse
  • 3,243
  • 1
  • 22
  • 29