Is The Digital Transformation Changing Industry’s Approach to Quality?

An update to the Automation World survey on the use of automation for quality inspections detects some surprising changes in the era of digitalization. Quality may be becoming associated more with process feedback than classical inspection technologies.

2006 F2 1

Most manufacturers work hard to produce quality products, but how much of their efforts are devoted to automating their inspection processes?

To find out, Automation World conducted a reader survey early this year. More than a third of respondents (37%) came from the process industries (e.g., food and beverage, chemical, petroleum, utilities); nearly half (46%) came from discrete manufacturing (e.g., automotive, aerospace, machinery and equipment); and the remaining 17% are associated with construction and mining, systems integration, consulting, and other businesses.

Their responses to the survey were, not surprisingly, mixed, with 52% of the respondents saying they inspect before production, and 67% saying they do so afterward. These numbers do not add up to 100%, as many respondents conduct quality checks at multiple points in the production process.

More surprising is how little of these inspections are automated in an age of increasing automation and digitization. Just under half of the respondents (45%) say that less than 10% of their inspections are automated. About a quarter (26%) report that 10-25% of inspections are automated, followed by 14% who report 25-50% automated inspections, and 15% who say 50-100% of inspections are automated. We were not expecting these results to so closely mirror those received the last time Automation World conducted this survey four years ago.

Automation vendors offer a number of reasons for this finding. One being that the survey was given to people in a broad range of industries—from process to discrete manufacturing.

“We see the adoption of automation growing in industries and applications where the term quality may not have been used to describe output standards,” says Siva Kanesvaran, manager of application design engineering for industry business at Schneider Electric. “Consequently, the different segments [polled in the survey] may use different terminology for referring to what they consider to be quality,” he says. “Not everyone may be looking at quality in the same way as it is understood on a manufacturing line.”

“The good news is that quality still appears to be an important topic, because 80% of the respondents [report that they] are performing in-process quality checks,” notes Alastair Orchard, vice president of digital enterprise for Siemens Digital Industries Software.

Alastair Orchard, vice president of digital enterprise for Siemens Digital Industries Software, speaking at Siemens Digital Industries Software Analyst and Media Conference 2019Alastair Orchard, vice president of digital enterprise for Siemens Digital Industries Software, speaking at Siemens Digital Industries Software Analyst and Media Conference 2019

IIoT and analytics effects
The overall trend toward greater connectivity and the growing use of data analytics is another explanation for the lack of automation advancement in the area of quality. Not only are operations increasingly able to act upon a richer trove of feedback from sensors and devices like drives, but they can also track and use that granular data for quality control (QC). “They are still measuring quality,” says Kanesvaran. “It’s just that it’s being spread out across the process.”

Greater connectivity and resolution into processes often reduces the need for post-production checks. “This makes it possible to simplify final, end-product inspection,” notes Kanesvaran. After all, if you don’t need to check as many points after production, then the engineering and expense of a vision system, for example, is less likely to be warranted.

“Perhaps the question [of the method of inspection being used] has become less relevant as IIoT [Industrial Internet of Things] adoption increases,” says Orchard at Siemens. “Instead of automating quality control in the PLC [programmable logic controller] and passing the data through industrial networks in dedicated QC servers, it’s quite probable that IIoT devices are leapfrogging legacy infrastructures and publishing data directly into an IIoT platform for visualization and analysis.”

These observations seem to jibe with some other curious findings in this year’s survey. For example, a significant decline was found in the reported use of technologies typically associated with automated inspection. Respondents claiming to use vision systems, for example, fell to 35% this year, down by nine points from the previous survey. Similarly, those using automated measurement fell by 11 points, down to 4%. The use of QC software declined even more, by 12 points to 31%, as did network communications between QC software and the controller, which fell by 14 points to 18%. Yet the use of manufacturing execution systems (MESs) and track-and-trace software remained about the same as it was four years ago.

Kanesvaran points to another finding in the survey that suggests a growing reliance on sensors and process feedback. As noted above, about 80% of facilities are conducting quality inspections at some point during production, whereas 67% are doing so afterward. Not only does this data reflect a bias toward in-process inspection, but the movement since the last survey suggests that this bias may be intensifying. In the earlier survey, roughly the same percentage, 79%, reported inspecting during production, while 72% indicated they were inspecting afterward in this year’s survey.

director of Connected Enterprise Operations at Rockwell Automationdirector of Connected Enterprise Operations at Rockwell Automation

A journey toward maturity
Despite showing a relatively low use of automated inspection, the survey did show a positive trend, according to Louis Columbus, a principal at IQMS, an enterprise resource planning (ERP) and MES provider that recently became part of Dassault Systemes. He saw this trend emerge after mapping the number of automated inspections to a maturity model that divides users into four categories: reacting, anticipating, collaborating, and orchestrating.

Columbus put those automating less than 10% of inspection into the reacting category, which comprises those “reacting” to customer demands, often relying on expeditors to compensate for a lack of automation. Into the next category, “anticipating,” Columbus put those automating 10-25% of inspection. “Anticipators” have used automation to put enough order into their production to need expeditors only for exceptions. The third category, “collaborating,” contains manufacturers who have integrated all departments on a common ERP system so that each can contribute to fulfilling orders. Columbus associated this category with those automating 25-50% of inspection. The fourth and final category, “orchestrating,” contains those automating 50-100%. “Here, a truly integrated system has reached a critical mass of adoption and manufacturers begin to see its benefits in efficiency,” he says.

Based on this model, more than a quarter (29%) have reached the mature levels of collaborating and orchestrating. In this light, more than half (55%) of the respondents have either embarked upon the journey to maturity or reached it.

For this reason, Columbus interprets the results as a consequence of maturation. “Instead of thinking that they have to automate the collection of data, these respondents are recognizing that it’s really the up-feed and interpretation of data that is important,” he offers. “They are moving up the chain to analytic tools and fine-tuning production scheduling.”

Taking a slightly different perspective, Orchard at Siemens suspects that most in the 50-100% bracket are either high-volume manufacturers like automakers, bulk chemical producers (who require little to no flexibility), or part manufacturers that have been using CAD/CAM for decades. “The other respondents are probably on a digital transformation journey that must be completed before quality inspections can be automated,” he says.

He further suspects that most of those on this journey are either producing several types of products or personalizing their products for customers. Because these kinds of production lines require a flexibility that has not been available from conventional automation technologies used in facilities devoted to one product, the lines have had to rely heavily on manual interventions, including manual inspection. Respondents dealing with these kinds of operations may be having difficulty bringing them into the digital enterprise.

Siva Kanesvaran, manager of application design engineering for Industry Business at Schneider ElectricSiva Kanesvaran, manager of application design engineering for Industry Business at Schneider Electric

Why experience is a must
For those manufacturers looking to automate quality control, Orchard recommends doing so at the same rate as in production. “An imbalance in either direction will lead to lower overall KPIs [key performance indicators],” he says.

In fact, he believes that an imbalance could be a reason for a particularly puzzling finding that came from comparing this year’s survey to the one conducted four years ago. The percentage of respondents who reported benefits from automating inspection—better quality, less waste and rework, lower expenses, greater regulatory compliance, faster throughput, fewer recalls—fell by roughly 10 points each.

Doubting that the objective benefits have actually worsened, Orchard thinks it more likely that digitalization in quality has not moved as fast as it has in production. “It could be that improvements seen through automation four years ago have not led to continuous improvement at the same rate,” he says.

In many cases where users fail to get the desired benefits from automation, the culprit is a “point solution” that solved a specific problem but did not contribute to the overall process. An example might be a torque-controlling tool. “You get the right torque for screws you are installing,” explains Mick Mancuso, director of Connected Enterprise Operations at Rockwell Automation. “However, that system does not tie into a work instruction system and does not give feedback to the manufacturing process.”

Other reasons Mancuso gives for the slow progress in automating quality inspections and not getting the desired results lie in the lack of definition of the right return on investment and not orchestrating collaboration among engineering, operations, and information technology. “If it’s a project that can be justified, then the chances that it will get funding go up,” says Mancuso. “Then, if focus goes up, so does the success rate.” To get results, he advises seeking the right partners who can help you consider the entire end-to-end process and to use platforms like MES.

This advice to seek help could be of particular benefit in light of the change in response we saw to the number of respondents indicating that automation has played a key role in quality inspections for less than a year. In 2016, 5% indicated this to be the case. This year, that number jumped to 33%.

“Although some of this change could be due to new subscribers, the jump is so significant that it’s got to be companies who have just automated,” says Columbus at IQMS.

Those companies might not have the necessary experts yet, notes Mancuso at Rockwell Automation. Then, add to that the generational shift in the workforce. “As more experienced employees retire, they are being replaced by a new batch of employees who might just be starting their careers,” he says.

Not only does this generation lack the experience of their mentors, but they are also more apt to deviate from traditional PLC architectures and move down the path toward the digitalization. “Clarity around quality may be suffering because, under digital transformations, there is certainly a blending of these domains under combined digital threads,” observes Orchard at Siemens. “Perhaps quality is being recategorized as an attribute of production and logistics.”

More in IIoT