ARTICLES

In Process Control

Dr Juergen Ude Dr Juergen Ude

The quality movement seems to agree that it is wastefully to inspect quality into the product, after it is manufactured, instead of getting-it-right-first. As we are advancing towards the twenty first century it is appropriate to ask whether we really have advanced from the old way of doing things. If we have, then we need to ask whether what we are doing is effective, is effective, and if not, then we must ask why not.

Whether industry, in general, does adopt the attitude of getting-it-right-first-is difficult to answer. There are many issues to consider. It is fair to say that all of industry spends an effort on getting-it-right-first, at least to some degree, and that this has always been the case. However, it is equally fair to say that the most notable area that little effort is spent on is in-process control. Other than exceptions, it is either not done at all, or tends to be done ineffectively. This observation is based on observations by myself in several hundred large corporations, in over fifteen countries during the last five years.

What do we mean by In-Process Control? Different people will have different interpretations. For the purpose of this article, in–process control is defined as the activity of statistically monitoring relevant critical stages of the process and taking whatever action is required, to maintain the process in a state of control, that economically minimizes variation and targets the process to its most economical level. By controlling the process through the various stages of manufacture we are often able to take corrective action before more damage is caused. This is simple common sense, yet why is it still not practiced extensively?

The reason are many. Effectively, controlling the process means measuring key characteristic of both the process and product. This takes time and effort and is often considered mundane. Over confidence is another reason. Process workers and management often do not understand the effect of excessive variation on the cost of the product, the performance of the product and ultimately the end customer. Since the effect is not understood the assumption is that there is no problem and that process control is unnecessary. The inability to take corrective action is a third reason. Process workers soon lose interest if their efforts to not lead to perceivable improvements.

In-process control often fails because inappropriate tools or procedures are supplied, that lead to no real improvement or costly over correction. Many of the popular control charts tend to be ineffective, and often lead to over adjustment. However, arguably the most significant reason for failure is that only product quality characteristics are statically monitored, not process variables, such as processing temperature, machine settings, and raw material batch numbers.

What are the alternatives to monitoring process variables? One alternative is to design a robust process, one that is reliable and does not require monitoring, besides occasional spot checking. In many situations such processes are impossible to economically achieve. A second alternative is to ignore process variables, as most organizations do.

I do not recommend this approach. It has little chance of leading to improvement. There are many reasons, which will be discussed in a future article. However, it suffices to say, that causes to problems cannot easily be identified. If causes cannot be easily identified then the process cannot be improved.

Some process workers believe that they usually know the cause of a problem, when they are made aware of a problem. Hence they feel there is no need to monitor process characteristics. I have not seen evidence of such abilities by anyone, and in fact note that absence of conclusive evidence leads to pet-theories.

There is only one way to perform effective in-process control and to achieve real quality improvements. Operators must enter process information directly into a computer, or preferably enter the information electronically. As data is entered into the computer it must be plotted. Key characteristics and process variables should be plotted on the same screen and the most modem tools used to identify when a problem had started and finished. In the presence of all reasonable information the system must point the operators into the right direction. The information entered may then also be analyzed by the specialist leading to further improvements. There are spin-off benefits for such computerized process control. Management will have instant access to information, instead of searching through log sheets that tend to get lost. Customers will be more confident in your process. Due to the higher frequency of sampling, management will have a better estimate of outgoing quality.

What are the costs of such a system? There are computer and networking costs, plus software, but these are small in relation to other costs. An initial outlay of say $30,000-50,000 can be expected, but this is nothing compared to the benefits that may be achieved. The biggest danger is to use the proverbial accountants approach and look at cost only. It is true, that the return cannot be easily quantified, if at all. However, the costs of having a system that cannot be effective is far greater. To justify such a system, management must believe in the benefits of basing decisions on fact and getting it right first. Management must understand that quality cannot be improved by looking at final product characteristics, after the event and serves no purpose, other than giving misleading confidence in the system.

To conclude, using in-process control, means that we increase our chance of doing something about quality problems. Not using it means that our SPC system will advise us of problems, but that is all. Most managers know they have problems, but the trick is to do something about them.

Download the Inferences APP, comprised of mainstream and machine-powered analytics for statistical analysis

Analytics as a Service (AaaS) for Quality

Drive quality improvement through actionable insights using analytics you can trust! Use up to 200 analytics tools downloadable through a suite of Apps!

FREE usage of the analytics Apps for quality improvement
  • Augmented with machine-powered smarts
  • Always updated with the latest tools and features
  • No licencing or fixed subscriptions - Pay ONLY for the analysis you run from 20 USD cents per analysis, billed monthly! Set a budget so you don't exceed!