Although the origin of SPC can be traced to America, early American practitioners didn’t recognize or take advantage of the benefits Statistical Process Control made available to the system. But despite its slow reception in America, SPC found widespread use in Japan where it helped improved and perfect their production process, further boosting the Japanese economy.
Table of Content
- What is Statistical Process Control (SPC)
- How Statistical Process Control Started
- Processes Involved in Statistical Process Control
- Why Use Statistical Process Control
- Statistical Process Control Tools
- Other Statistical Process Control Tools
- How To Go About Statistical Process Control
What is Statistical Process Control (SPC)
Statistical Process Control, SPC, allows manufacturers to control the processes involved in production through the use of statistical data and techniques. Using these data generated from SPC has allowed manufacturers to measure, understand, and improve their service delivery.
Unlike other methods of improving productivity, SPC places the responsibility for quality right in the hands of the operator by providing ways with which the operator can determine if a process is in line with set guidelines. This way defects are reduced drastically.
SPC gives manufacturers total control of the manufacturing process by measuring and controlling quality. Now, quality data is generated by measuring the readings from the machines and instruments involved in the manufacture of a product.
This method of controlling product quality allows operators to function at their maximum potential.
How Statistical Process Control Started
Statistical process control was first introduced circa 1920, at Bell Laboratories by Walter A. Shewhart. The process was further improved by William Ernest Johnson, W. Edwards Deming, and several others who helped American industries and the US Army effectively use this new, effective technique in the efforts of the Second World War.
Deming would later travel to Japan shortly after the Second World War ended to introduce Statistical Process Control to Japanese engineers and industries at large. The introduction of SPC to the Japanese industries led to the production of high-quality products.
The high-quality products manufactured in Japan were enough to trigger American manufacturers to step up their game, and one way they did this was by incorporating statistical process control.
Now that you’ve gotten a brief glimpse of how it all started, let’s take a step further and see what processes are involved in these types of quality controls.
Processes Involved in Statistical Process Control
SPC makes it possible to measure the outputs of several processes by paying attention to several changes that are statistically important so that corrective measures can be taken to safeguard the products from defects.
In manufacturing, SPC can drastically cut down the waste, or scrap material generated. But the manufacturing industry isn’t the only industry where SPC has found widespread applications. For one, healthcare and service industries have begun taking advantage of SPC.
But before these industries began taking advantage of the benefits SPC had to offer, they had to follow these basic steps:
Planning begins by first identifying the root cause of the product’s defect and taking preventive steps that’ll help solve the problem. It’s crucial that manufacturers take corrective action immediately a problem is identified.
Research, Experimentation, and Analysis
Immediately a problem is identified, and the necessary data collected and analyzed to understand the full scope of the problem, these causes need to be further validated and verified through statistical tools like Brainstorming, Five Why’s, Fishbone, Time series plots and Scatter plots. Statistical process control software like; DataLyzer Spectrum, Intrastage, Analyse-it, etc. can be of great help.
The outcome of the analysis from these statistical tools is identified as the root cause, which can be directly altered to improve the product quality. In just the same way it’s important to identify the main problem, it’s also important to identify the root cause.
Using SPC software makes it a whole lot easier to get a complete view of the manufacturing process and spot areas with difficulties, making for higher product quality in manufacturing.
After results have been gotten from adequate research, experimentation, and analysis, the next step would be identifying and implementing the best potential solutions to the main problem. Stakeholders play an important role in implementing these new production standards and making sure those involved perform their duties.
Implementation should begin immediately the steps have been identified to avoid the root cause of a problem. Every preventive step should be carefully tested and if effective, it should be applied to the production process.
After its application to the production process, results should also be evaluated regularly to ascertain if any progress is achieved and if any changes are necessary.
To make sure that the solution stands the test of time, manufacturers should ensure that whatever plan they chose to implement should be regularly monitored, to ensure that the new performance standard is maintained.
This way, manufacturers can smoothly transition from a new system of production to another. Although it’s important to note that the most important part of this phase is to regularly provide training on new changes to everyone involved in the production process.
Why Use Statistical Process Control
The steady increase in the cost of raw materials hasn’t done much to temper the growing competition in the manufacturing industries, if anything; it has even made it tougher, as companies now have to navigate factors that they can barely control.
This need to improve their quality and manage the factors that they can control, like their production processes, has led them to use SPC because of the advantages it yields like:
- Waste reduction.
- It increases customer satisfaction.
- It reduces the amount of time taken to produce a product.
- Saves material cost by reducing the number of defective parts.
- It allows for a better quality of production.
- SPC makes it a whole lot easier to detect errors while inspecting.
- It also reduces the cost of the inspection.
- It makes it easier to reduce process variability in the long run.
- It is a stable process and operates with less variability.
Statistical Process Control Tools
When it comes to the tools involved in statistical process control one person we’d turn to is Dr. Kaoru Ishikawa. The Japanese engineer mastered and improved SPC across several spheres by introducing several quality strategies like the quality circle, the Ishikawa diagram, also known as the cause-and-effect diagram, and his book The Guide to Quality Control is still widely used in industries today.
And it’s this book we’d turn to because, in them, the engineering professor lay down the seven key tools for statistical process control. So let’s take a look at them:
- Cause-and-effect diagram
- Check sheet
- Control chart
- Pareto chart
- Scatter diagram
The Cause and Effect Diagram (The Ishikawa Diagram)
The cause and effect diagram allows businesses or manufacturers to determine why something happened or would happen, identify bottlenecks in their business process, discover a problem’s root cause, or identify why a process isn’t working as intended.
The cause-and-effect diagram works better when the problem and possible causes have been identified so that the relationship between these contributing factors can be better understood.
What’s Involved in Statistical Process Control
There are several factors to consider in a cause-and-effect diagram and each of these factors depends on the industry in question.
For example, in the marketing industry, you’d need to pay attention to:
- The problem
While in the manufacturing industry you’d be looking at:
- The problem.
How to Create a Cause and Effect Diagram
Identify the Main Problem
Recognize what the main problem you’re currently facing is. Where it happened and everything involved. Write the problem in a box and draw out a long horizontal line from the box.
Identify the Factors Involved
Next, identify the factors that caused the problem, write down as many as possible, although this may differ depending on the industry in question. See our list of factors here.
From the horizontal line, draw out shorter lines resembling a fish’s bone. Write the factor you’ve identified here.
Identify Possible Causes
After spotting the factors that contributed to the main problem, then it’s time to brainstorm problems that are related to that particular factor.
From the factor lines, draw out shorter lines where every possible cause of the main problem relating to that particular factor is represented.
Analyze the Diagram
By now your diagram should resemble a typical fishbone, and it should show you all the possible causes of that particular problem. Now it’s time for you to test all these possible causes and pinpoint the exact one responsible for causing the problem, and you can achieve this by running surveys, investigations, etc.
One reason why the cause and effect diagram is often called a fishbone diagram is because when it’s completed it resembles the skeleton of a fish with the head/problem to one side, and the bones/factors drawn out behind it.
The Check Sheet
Unlike the complexity fishbone diagrams poses, check sheets are a simple and effective tool for gathering and analyzing data that allows business owners or manufacturers to find solutions to a wide variety of problems. It’s sometimes called the location plot or concentration diagram.
Check sheets are normal tables with well-defined rows and columns where the collected data is usually recorded in its designated slots. Check sheets aren’t popular anymore today because computer software has replaced them.
Control charts have found wide-spread applications in manufacturing, marketing, and even in stock trading algorithms. Control charts are used to set upper and lower limits of acceptable performance under normal circumstances so producers can clearly observe the processes involved in manufacturing, allowing them to optimize performance.
Control charts are graphs that study how processes change as time goes by. Control charts always have a central line representing the average, an upper line for the upper control limit, and a lower line for the lower control limit.
These three lines are determined from previous data collected by the business. Therefore by comparing the newly generated data to these previous ones, making it a whole lot easier to see how consistent your business’s process variation has been, whether it’s in control, out of control, or affected by special causes of variation.
How To Create Control Charts
Remember that control charts are designed to allow businesses to respond quickly to a shift in the process because the faster problems are identified, the faster the company would return to high-quality production.
This speed provided by control charts save time and production materials, and these are the steps you should follow while creating your control charts.
Understand that it’s not bad if your control chart doesn’t immediately spot any issues with your system in the first months, or even years, mostly if it’s measuring something entirely new to your organization. When a new control chart is started, the process may end up being out of control.
Update Limits Regularly
It’s important to know that your limits mustn’t remain in one position forever. At certain points, it’ll become necessary to update your control limits to keep up with the times.
Watch the Bumps
Variations are bound to happen. Minor variations aren’t a cause for alarm. Instead, always watch out for those sharp spikes and dives in the control charts because those are the exact spots where you’d need to focus on.
Investigate their cause. Find out why they happened, how it can be handled and the steps taken to correct them.
Watch Out For Results
Use your control chart to observe the progress your organization is making and act appropriately. And as your organization continues making progress in several aspects you’d need to adjust your control limits.
One more thing; understanding why your control chart spikes or dives, basically anything outside the control limits, irrespective of whether it was good or bad would help you understand how to take advantage of the good occurrences while controlling the bad ones. And this is really useful after making new changes in your production process.
Histograms are one of the most widely used tools when an organization needs statistical process control, and one reason why this is the case is because histograms are very easy to understand.
Histograms are very similar to bar graphs being that they’re based on the frequency distribution of how often certain items occur in a set of data. This visual representation allows anyone to easily understand and interpret the given information.
By eliminating the gaps between each bar, like we clearly see bar graphs have, histograms are able to depict their data as a continuous occurrence.
The bars within the histogram are in complete accordance with how the given data is measured, i.e. the data isn’t rearranged along the lines of increasing or decreasing frequency. This way the information provided by the histogram is a true representation of the data being measured.
When To Use Histograms
- The numerical data is involved.
- To analyze whether a process is capable of meeting customer requirements.
- Analyzing what the output from a supplier’s process looks like.
- To see whether a process has changed over a period of time.
- When determining if outputs from two or more processes are different.
- To communicate the distribution of data quickly and easily to others.
Other Statistical Process Control Tools
- Pareto Diagrams
- Scatter Diagrams
- Process Flowcharts
- Defect Maps
- Progress Centers
- Event Logs
How To Go About Statistical Process Control
Before an organization can fully utilize statistical process control in managing its affairs, it must be prepared to follow the necessary steps required to make the process as effective as possible, and it involves:
Select Measurement Values
Statistical process controls heavily depends on the comprehensive analysis of the assembled data, so it’s crucial for the organization to carefully select what type of data to use.
And when it comes to collecting data for control charts, they’re two main types of data: Attribute and Variable. Attribute focuses on data sets that are separate from each other such as the number of defective parts etc., while Variable data sets focuses on factors that require continuous measurements like distance, temperature, weight, etc.
The Measurement System Needs To Be Reviewed
When preparing for SPC, make sure whatever measurement systems you choose is properly reviewed to spot, and if possible, correct the errors in the system. The data set shouldn’t be used in SPC if has too many errors.
Use a Sampling Plan
Use a sampling plan to collect data in a random but focused frequency. Collect data sets in subgroups. That way you can be sure to capture any variation between these subgroups. Remember that the type of chart used should be based on what type of data was collected.
Develop a Reaction Plan
Getting a well-defined reaction plan would provide the necessary guidance in case a condition goes out of control. The best way to ensure this is by creating a flow chart that captures the steps to be taken by the team. Take a look at the reaction plan below.
Add Chart to Control Plan
Control plans help in the documentation of important elements that must be executed in order to achieve the correct standard necessary for the quality assurance of a particular product or service.
Any chart you choose to use in the statistical process control must be added to the control plan. Here are some things your control plan must have:
- The type of chart
- Chart preparation team – Names of teammates responsible for collecting data and preparing the chart.
- Chart location
- The measurement method used
- Measurement system analysis
- The reaction plan/flowchart
- Calibration system
- The sampling plan
- Process stability status
- Process capability indices
Calculate and Evaluate Control Limit
Ensure that every formula you intend to use in the chart is carefully calculated and evaluated to remove the possibility of errors and to ensure that the process is under control.
Watch out for these out-of-control situations like:
- One or more points falling outside the UCL, upper control limit, or LCL, lower control limit.
- Two out of three successive points fall in the area beyond two standard deviations, either above or below, the mean.
- Four out of five successive points fall in the area beyond one standard deviation, either above or below, the mean.
- Six or more points are either successively higher or successively lower.
- 8 or more points fall on either side of the mean. It’s important to note that some organizations use either 7 or 9 points.
- 15 points in a row fall within the area on either side of the mean that is one standard deviation from the mean.
Identify Root Causes
After spotting a condition that’s out of control, make sure the condition is analyzed and the main cause of the problem is identified. Always review old control charts. Who knows, there may be notes from earlier incidents that would illuminate the current condition.
Design and Implement Actions to Improve Process Capability
When the root cause of a problem has been identified, you’ll need to plan and execute corrective actions that eliminate these problems and improve the general stability of your process.
Calculate Cp and Cpk and Compare to Benchmark
A process’s ability to satisfy requirements (customer expectations) is characterized as process capability, and it is determined by indexes comparing the spread (variability) and process centering to the upper and lower requirements. The disparity between the top and bottom parameters is called tolerance.
Always Monitor The Chart
Now that you’ve done all these continue to monitor the process, tackle every problem that comes along. This way your chart would remain up to date.
Statistical process control has been of great help when organizations need to maximize their production system. Knowing that we have the opportunity to respond when we come into problems gives us a sense of confidence.
This comfortable feeling in a manufacturing phase is supported by cost savings, customer loyalty, and the opportunity to confidently meet official inspections.