So far in our blog series about why it’s probably time your statistical process control (
SPC) legacy system needs an overhaul, we’ve discussed the impact of disparate “
islands” of data that can occur within legacy systems, as well as the constant battle against
inefficiency and how utilizing
real-time SPC can help provide the insights you need to overcome it and make a real difference in your manufacturing operations.
If your legacy SPC system is one of the many that has been kept in “maintenance” mode because you do not wish to update it and bring your manufacturing operations into the 21
st century, then this blog series is for you. Please continue on…
About Statistical Process Control
SPC is based on several fundamental principles and techniques. In manufacturing, we typically take a time-ordered data set consisting of measurements of a particular characteristic of a specific part, component, or product. We then perform several statistical calculations on that data to calculate things like mean and standard deviation, allowing you to calculate Cp/CpK, Pp/PpK, and more. We then
visualize these data values as plots, typically on a control chart so that we can differentiate and identify special sources of variation.
We can apply control limits to our control charts, and when a measurement value falls outside those limits, it indicates that something unexpected has happened. And we can use more advanced methods such as applying predefined “rules” (such as
Western Electric Rules) to detect excessive variations—such as X points rising or descending, X points above or below the center line, and so on. We can use other charts—such as Gaussian distribution or bell curves (to visualize the distribution of data values) or Pareto charts, for example.
While I do not want this article to turn in to an “SPC 101” course (we have more information on our
website, if you care to dig deeper), my point here is to reinforce that SPC is an
industry-standard tool that helps operators control quality during the manufacturing process by using a set of
standardized tools and methodologies. Therefore, if you are doing all the above, then you are “doing SPC.” And if your SPC system allows you to do all the above, then it qualifies as an SPC tool.
Let’s turn to one of my favorite things—an analogy—to explain…
Driving Manufacturing Processes
The automobile is based on some fundamental principles and tools (components) such as a source of power (the engine), a way to transmit that power to the wheels (drive shaft), a way to turn that power into forward motion (using wheels), mechanisms to steer (rack and pinion), speed up (accelerator), and slow down (brakes), and some other essential items that make it usable—like a seat to sit on, a frame to keep all this stuff connected (a chassis), and much more.
You get the point. That’s pretty much all we need to have what qualifies to be called a motorized vehicle. Are you picturing something like a Ford Model T (or “Tin Lizzie” as it is affectionately called)?
However, while all these components are essential, the modern automobile provides far more capabilities than that, such as: electronic engine management systems, cruise control, lane keeping assist, automatic transmission, advanced driver information management systems, air bags, satellite navigation, air conditioning, audio systems with Dolby surround sound, vanity mirror, and, of course, those handy little compartments to keep things in that you’re otherwise always losing in the trunk somewhere
Modern cars are based on the same fundamental principles as the Model T, but are now vastly more efficient, safer, faster, usable, and ultimately more pleasurable to drive. Admittedly, some of my gear-head colleagues may disagree with that last point! But that’s their hobby, and manufacturing is not a hobby; it’s serious business.
The point I am making with all this modern automobile fun is also rather serious: that while you may have an SPC system that provides you with those fundamental tools, more often than not it does not enable you to unlock the true potential power and value of SPC. Let’s look at some of those limitations and contrast them with next-generation SPC solutions—like
Enact® from InfinityQS.
Advanced SPC Capabilities
As mentioned earlier, SPC is based on several fundamental tools and methods. But that does not mean that SPC should be limited by only those tools. Not only do advanced SPC solutions provide a greater wealth of statistical calculations and KPIs, but they also provide many more advanced ways to
interrogate that statistical data and methods of putting that insight to work to add value to manufacturing operations.
Going beyond the traditional approach to statistical calculations and control charts, InfinityQS SPC solutions provide a much greater armory of tools that can be deployed to solve almost any conceivable use case, industry requirement, or problem-solving scenario—such as advanced data analytics and visualization capabilities, and completely new and innovative applications of statistical techniques. In the case of InfinityQS Enact, for example, our industry-first “
Stream Grading” capability helps organizations summarize manufacturing performance from very specific part-process-feature streams right up to plant-by-plant or region-by-region through a unique and simple 3x3 grading matrix.
More on this can be found in this
blog article by my colleague Eric Weisbrod, InfinityQS VP of Product Management.
Finally, capturing a wealth of information beyond just measurement, part, processes, or feature data such as operator, shift, and lot number (for example) enables those powerful statistical analysis capabilities to be enriched even further. Mapping the relationship between data and processes, such as through Enact’s
Process Models or via Lot Genealogy, for example, turns this data from discrete individual data points to a kind of “information fabric” woven throughout the manufacturing and quality operation providing a 360 degree view of the manufacturing operations.
Advanced SPC: Real-time Data Collection, Analytics, and Notifications
To put it simply, manufacturing is a real-time activity, and therefore SPC and advanced analytics should be, too.
Manufacturing processes happen in real-time and so does variability. If we collect data manually over a period of time (such as on pen and paper or spreadsheets) and then batch that data up, prepare it for importing to our SPC tool, import it, and then analyze the results (which is grossly inefficient as discussed in
Part 2 of this blog series), then a lot of time passes between the measurement being taken and any statistical insight we may glean, and that potentially means a lot of waste, scrap, or downtime in between.
If we can cut out the middleman, so to speak, and have those measurements enter the SPC solution
as they are taken, then each measurement can be evaluated in real-time against previous measurements.
If any specification limit or statistical violation (e.g., limits, zone rules, trends, etc.) is triggered by that measurement, then the operator and other responsible parties can receive immediate feedback of that event, enabling them to take remedial action immediately—before problems occur or escalate.
That real-time data can come through mechanisms (such as fully automated data collection) that require no operator interaction or be entered directly into Enact at the required time and in the required format.
This real-time data collection capability makes both operators and manufacturing processes much more efficient by the automatic monitoring and statistical analysis of process performance and variability.
Just as real-time data collection opens the possibility of real-time SPC, it also opens the door to real-time data analytics. This means interrogating data on-the-fly and using the right tool for the right question (and we’ll discuss those tools next). But, most importantly, it means knowing that the data being interrogated is
always accurate and up-to-date—right up to the point of the last measurement made on the shop floor.
No longer do supervisors, plant managers, or quality executives need to wait for manual reports to be produced and delivered (which are out-of-date the moment they are created). Instead, they can interrogate the data directly either using high level aggregate analysis dashboards comparing plant-by-plant or product-by-product performance, or dive right into the process-level raw data analysis with just a few clicks of the mouse.
Advanced Data Analysis and Visualization
Don’t get me wrong, SPC is a great and powerful tool (it has been at the heart of InfinityQS solutions for more than three decades). But SPC can only tell us half of the story; and by
half, I mean that figuratively not statistically! Indeed, it tells us when unexpected (or non-normal depending on your choice of term) variation occurs, indicates problems within the production process, and illuminates trends that may impact performance and quality.
However, a traditional approach to SPC often only focuses on a single part-process-feature combination. It tells us that there is an issue—but not necessarily what the issue might be, or where a particular issue is occurring more or less frequently than others.
Therefore, by capturing information beyond the measurement-part-process-feature (as discussed above), we can start to interrogate the data in even more ways and use a variety of
visualization tools (SPC control charts and distribution curves being only part of the answer) such as Box-and-Whisker plots, animated Bubble charts, and Stream Grading, for example.
Unified Data
When all data is standardized and stored within a single, unified repository, then we can begin to analyze and interrogate data differently.
- Want to know how a particular feature on a particular product is performing across different production lines or shifts? You might want to use a Box-and-Whisker plot for that.
- Want to compare different products with different specification limits or feature measurements? Good thing that those Box-and-Whisker charts are normalized for just that reason.
- Want to know how yield is performing across multiple plants? You might want to use a Grading matrix for that.
- Want to know how and when quality checks are being performed on time? You might want to use a data collection compliance tile for that.
- Want to know if your process and products are meeting standards? And how that is evolving over time? You might want to use an animated Bubble chart for that one.
- Want these questions answered for the last shift? The last day? Last week vs this week? Line 1 vs Line 2? All possible with just a few mouse clicks.
The list of scenarios is almost endless. Suffice it to say that having these advanced data and analysis capabilities goes far beyond what legacy SPC solutions can provide—and can generate operational improvements that should not (and cannot) be understated.
Beyond SPC
With the above advanced capabilities, SPC solutions become integral to manufacturing quality processes. In fact, they become more than just SPC solutions, and that is why InfinityQS uses the term SPC
Quality Intelligence. While SPC Quality Intelligence retains SPC at its heart, at its foundational core it provides a wealth of capabilities that far exceed traditional SPC:
It could be quite easy at this point to launch into a diatribe of all the unique and powerful capabilities of our solutions, but I’ll refrain from that here and leave you to the areas of our
website that contain a wealth of information on this subject—capabilities such as Lot Genealogy, Acceptance Sampling, Label Verification, industry standard Net Content Control capabilities (Total Negative Error (TNE), T1/T1, or MAV, for example), 21 CFR Part 11 compliance, and so much more highlight the breadth of capabilities that take you far beyond SPC.
Many of these capabilities are key requirements for manufacturers today. Yet if we stay wedded to our legacy SPC solutions, it means inevitably tacking on additional, entirely separate, systems to provide them—more islands of data, applications, and interfaces that operators need to master, switch between, and navigate. The alternative, as you might have guessed, is to turn to next-generation SPC solutions like
InfinityQS Enact that go beyond traditional SPC.
Universal Accessibility
These advanced capabilities—which can now be provided by next-generation SPC solutions, like Enact—can be very valuable in today’s modern manufacturing environments. And they are readily available to those that need them
when they need them.
These valuable capabilities can soon be cancelled out if they can only be used on a workstation where the software is installed. In today’s environment, this ultimately leads us to cloud-based SPC capabilities, which provide three fundamental benefits:
- They enable data to be collected and stored in a secure, unified, centralized system (see Part 1 on No Man (Data) is an Island for more on this topic).
- Cloud-based SPC, deployed as a Software-as-a-Service (SaaS), decouples the cost and resources required to manage and maintain the software and supporting infrastructure (more to be discussed on this topic in Part 5 – The Big Easy).
- These capabilities can be accessed securely and universally from any device, from any location, and at any time—using just a web browser on anything from a smartphone, tablet, or laptop to a shop-floor workstation.
Closing
Thanks for coming along with me for this journey through SPC and more. In our next article,
Part 4: The Complexity Problem, we’ll talk about why complex systems are
not the best way to manage complex environments. As more and more capabilities are added to the next generation of SPC solutions in response to today’s increasingly demanding and complex manufacturing environments, it’s important that the solutions we rely on do not become overly complex themselves. To achieve that, a radical new approach is required from user interface design, exception-based monitoring, and intuitive dashboard-based analytics. Please join me!
Take advantage of the technology at your fingertips today: contact one of our account managers (1.800.772.7978 or
via our website) for more information.
Read the other articles in this series: