Statistical Process Control (aka SPC) is simply using statistical methods to monitor and control a process. Statistics have been used in operations of all types for ages, just for the most part, in the past, they were only used to monitor the output, as a report and feedback tool. SPC takes statistics to the next level, namely it becomes a proactive tool, capable of identifying a potential problem before it occurs.
This new, proactive ability of statistics has come about due to the power and capability of today’s computer hardware and software. Since it is now very easy to collect and analyze large amounts of data in real time, statistical methods can be used to identify trends that will lead to problems in the future.
Knowing that a problem is going to occur, but as yet has not. We can cut down the defect rates by taking preventive action, thus saving our economies vast fortunes in wasted resources, while keeping costs down for the consumer.
Statistical Process Control is not a tool for getting your process under control, but can be extremely useful tool to keep your process under control.
On this page we are not going to go into it very deep, as Statistical Process Control takes some serious training and time to really learn. Instead, we will give you a simple explanation of what it is and what it can do for you. Then you will know why you need to use it and you can start to learn how to use it.
When is a process under control? When 99.7% of the time the only out of process limit products (see below an explanation of a process limit) are the result of some special identifiable cause. Additionally those special causes have to occur rarely.
Statistical Process Control Handles
Statistical Process Control deals with what is called Common Process Variation. For example, while operating under normal conditions, a machine drilling holes will drill a hole a certain depth, but each hole will vary within a certain, narrow range. This condition may be caused by the gradual wear of the drill, thus causing a smaller/shallower hole.
Once you have gotten your processes under relatively good control, and you know how to adjust them in order to keep them producing as required, how do you know when to make the required offsets, or when to change the tool?
If you wait until your process starts making bad product then you end up wasting resources by creating either scrap or rework for no valid reason. At the same time, in these tight economic conditions you cannot afford to waste dollars adjusting a process until it really needs it.
That is where Statistical Process Control comes to the rescue. Since, in modern operations, we can easily track and record what is happening in any process, and store that data in a database (and most business already store the data they need), why not use that data to help us control the process?
With the aid of basic statistical analysis software we can easily setup routines that monitor what is happening in our process. As long as that process remains safely in a range of acceptable results, we have no need to tinker with the process. Instead, we can use that time and those resources to tackle real problems.
Additionally, you can use this same software and these same programs to recognize a trend in the process that, if continued, would put production outside of the desire specifications. By spotting this trend we can make the necessary adjustment to the process in a plan able and controlled fashion without ever having created any defective product.
We start by identifying Key Process Variables, both input and output if possible, we collect data on these variables and determine if our process meets the in-control requirement of Statistical Process Control, using process control limits that are narrower than those required by our customer’s specifications.
A process control limit that is tighter then the customer’s specifications keeps us from ever producing defective parts due to our processes’ natural variation. This also allows use greater flexibility to plan for the adjustments so that they do not adversely affect production.
Determining which variables to track will depend on your process. For example we have two plastic injection molding companies:
- Company A can track the amount of plastic going into the moulds very easily so it may choose to use input data, so if their data shows a steady trend toward either using more or less plastic they know they could be facing a problem in the near future.
- Company B already weighs the parts as they come out of the moulds, thus for them it may be easier to track this output weight as the variable to find a trend that would indicate a problem was starting to develop.
- Although, in the long run, tracking the inputs gives you a faster response, especially if your processes use longer time frames, but not every process will allow for good input control.
Then, by using Key Process Variables we are tracking, we are able to see what is happening with our process. Although, the ideal situation is the input side as we get earlier process data, which means we can act sooner. When a trend develops that indicates we are heading out of the process control limits, we can take proactive action before creating defective product.
Let’s go back to our drill bit example for a second. As the drill wears, the holes will start to trend toward a shallower depth. By monitoring the depth of the hole and recording it we will see as the drill approaches the end of its life a steady trend toward shallower holes that does not occur when it is brand new. This trend shows us that we should plan to replace the drill bit at some convenient point in time before it starts making parts that no longer meet customer requirements.
You surely understand your need for Statistical Process Control and what it can do for you if you use it properly. If you want to start learning about it so you can start using it, we offer our Statistical Process Control presentation that will help you get started down the right path.