It’s been a while since Scott Lowen was in school, but he still does “science projects.” That’s what Lowen jokingly called a proof-of-concept project to measure a part and have the resulting data determine corrections that are fed automatically to a turning center, mill or other machine tool.

While Lowen’s ongoing science project at Zeiss Industrial Quality Solutions, Wixom, Mich., where he is product manager for software and accessories, may not have fully supported the hypothesis that metrology data can easily produce fully automated machining corrections, Zeiss continues to work on automated solutions. Other companies, such as Capture 3D Inc. (a Zeiss company), Renishaw Inc., and Verisurf Software Inc., have solutions that use metrology data to do varying degrees of automated corrections with machine tools.

In the meantime, the ongoing science project is a continuing part of Lowen’s job.

“The time invested to create the types of connections in an automated feedback [system] far outweighs the payback to have this type of automated feedback to machine tools,” he said. “We still believe in having an engineer in place to look at the appropriate data and create the appropriate feedback or correction to the machine tools.”

This might be good news for manufacturing engineers’ job security, but it poses these questions: Can the measurement process fully plug into smart manufacturing? Or might metrology be largely relegated to the processes manufacturers used prior to the connected, data-driven, automated world of the new way of making things?

Renishaw, West Dundee, Ill., has a product in soft launch mode: Renishaw Central is a manufacturing data platform for information including machine status, alerts and measurement results from machine tool probes and toolsetters, Equator gaging systems and CMMs.

The platform has an application programming interface that enables users to interrogate the information and feed it into other software systems, such as ones for enterprise resource planning or a customized dashboard.

“I think the big enhancement we’re seeing [our customers make] is to then take that measurement data to generate pre-configured, automatic corrective actions, whether that’s a tool offset, or work offset, etc.,” said Brandon Golab, machine tool software manager at Renishaw. “The users of our system configure how to react to the measurements and which machine they want to communicate the corrective action to.”

Making that possible is a product enhancement called intelligent process control, which is a series of communication protocols that allow a customer to dictate how they want to treat measurement scenarios. This, in turn, can create offsets and intelligent adjustments to the manufacturing process or machine tools if necessary.

Renishaw Central also has a visualization function that  provides standard dashboards to display machine status and measurement results. It can also be used for historical data analysis.

During soft launch, the platform is set up for one machine tool for each measuring device, but Golab said he envisions no limit to how many machines Renishaw Central can interact with.

“In addition to Renishaw Central, we also have several products that would allow the machine [tool] itself to not just collect the data but do closed-loop manufacturing feedback within a single machine architecture,” he said.

Designed for new or less-experienced operators is Set and Inspect with Program Builder, which is used on-machine, and an off-line solution, Productivity+. Set and Inspect is used to do jobs like parts setup, part validation, fixture load verification and cut-measure-cut (where an operator does a semi-finish pass, measures the result and then does a finish pass based on the measurement data). “All that can be set up with those pieces of software,” Golab said. Set and Inspect also can be used with Renishaw’s Reporter software to feed the data back to Renishaw Central.

“Productivity+ uses CAD models to generate similar programs to what [Renishaw’s] Inspection+ does but because it’s CAD-driven, it’s easier for the programmer,” said Golab. “But it’s very similar to Inspection+ regarding what can be with data at the individual machine level.” Inspection+ is Renishaw’s macro-based metrology software for more experienced users who can write G-code that runs at the CNC controller level.

Renishaw measurement technology can even be used for predictive failure. To do this, the data from Productivity+ is paired with the company’s Sprint sensor that measures thousands of points per second to assess the surface condition of a part.

“As the tool wears, you typically would see a waviness [on the part surface],” Golab said. “Because this is a direct measurement of the tool’s performance, it can be a better indicator of tool wear than indirect methods, such as spindle load.”

Closed-loop manufacturing is also familiar to Rick White, general manager of Capture 3D, Santa Ana, Calif., but more specifically with a closed-loop technique called adaptive machining. “It creates a closed loop in a CNC machining center based on what the results are of scan data,” he said. “[Adaptive machining software] figures out what to do to get to the CAD model and automatically creates a new path for the machine to fix whatever needs to be fixed. That is done in real-time, without human intervention.”

White was CEO at Capture 3D, the leading U.S. partner for GOM 3D non-contact measuring solutions, until October 2021 when the company was acquired by Zeiss.

In addition to its use in adaptive machining, data from Capture 3D’s full-field ATOS 3D scanning technology can also feed into digital assembly analysis with GOM software. “Our aerospace and automotive customers have had good success using our technology to digitally assemble parts from all over the world to ensure fit, alignment and specifications before anything is physically shipped,” White said. “This helps companies correct issues or prepare for changes quicker while reducing costs to help with a faster product launch.”

For example, after an automotive OEM receives the hang-on parts’ metrology data from its plants, vendors, and tier suppliers—no matter where they are—they can determine in a collaborative, digital process if the components will fit the assembly by using a digital model of the body-in-white. In addition, the full-field data provides millions of accurate data points for a digital twin. “And the beautiful part is it actually works and it’s quite successful,” White said.

Another technique that takes advantage of the flow of data used by some manufacturers is model-based definition (MBD). “What [MBD] means is that measurement of callouts and processes and variances and GD&T [geometric dimensioning and tolerancing] are part of the CAD model from the beginning, and along with PMI [product manufacturing information] are imported into the GOM software,” White said. “As designers design parts, they’re also putting in the information on the tolerances of the part. And then a metrology system has that information. It knows what to inspect and it knows how to determine what components are in and out of tolerance.”

This shifts decisions on what to measure from an inspector who is downstream in the process to the designer. The belief is designers can know that information better than anyone else because their perspective is more global in nature. They are designing parts to work together, and tolerances within even the same part are not universal.

“MBD requires a universal standardization to be read properly inside metrology software,” White said. “Hopefully, [the] QIF [format] is that bridge,” he said, referring to the Quality Information Framework, a unified XML framework standard for sharing quality information digitally. He went on to note that “some organizations are moving faster than others; however, we have seen an overall upward trend and this is where we see the future.”

The Digital Metrology Standards Consortium (DMSC) created QIF to address the crucial need for a digital metrology standard to help advance dimensional metrology, White explained. The DMSC announced the QIF 3.0 metrology standard was harvested, approved, and published by ISO as the new ISO Standard ISO 23952:2020 in August 2020, according to the consortium’s website.

“We are on board with standardizing a format that makes sense for everyone,” White said. “While we can’t answer on behalf of the major CAD software makers, it’s likely that they will need to use it if it becomes the standard.”

“Boeing requires [MBD]; all the big aerospace companies require it,” said Ernie Husted, president and CEO of Verisurf, Anaheim, Calif. He called MBD a time and money saver. “A lot of the tier companies aren’t used to handling it—they’re used to drawings.

“It’s only in the last few years that all the major CAD companies implemented it as a standard feature. So now they can tolerance the model with the appropriate GD&T tolerances and send that file to their vendors and they don’t need drawings any longer.”

If the CAD model has MBD information, Verisurf software can use the embedded data to automate the inspection plan with any CMM, arm, tracker or scanner, Husted said. “There’s a process where you can adjust tool offsets if you’re doing machine tool probing,” Husted said. “With on machine tools, you have cutter compensation and tool wear compensation and when a tool wears you can adjust the offset to machine a feature bigger or smaller depending on whether it’s an internal or external feature. So you can direct automatic adjustments for tool wear, but once a tool breaks down you need to replace the tool.”

The user can also write special applications with Verisurf’s software development kit to help make decisions. For example, if a part is out of tolerance the app can help implement a tool change if there are extra tools in a machine tool carousel. Routines can also be written to control robots for in-process inspection or other automated processes.

Users of Verisurf software can also add MBD on scanned meshes if there’s no CAD model. “So when you scan the part, I can actually click on a surface of a mesh and pull out a thickness,” he said. In November 2021, Renishaw and Verisurf entered into an agreement that lets companies in North America access Verisurf CMM software from Renishaw.

Zeiss’ Lowen cautioned that while automation is certainly progressing quickly, manufacturers still need valuable human input in the process.

“[In our} experience, there are too many real-life engineering variables that make it very difficult to automate,” Lowen said. “We determined you still need a manufacturing engineer with the human brain to say, ‘OK, here’s why this one shape is out of tolerance, here’s actually what went wrong, and here’s what we’ve got to do to correct it.’ ”

Tool wear is one of the simpler reasons to understand but it’s seldom the reason why a dimension is out of tolerance, he said. If tool wear were the only reason why a dimension is out of tolerance, then Lowen agrees it could be automated. But the reality is there’s a dozen other reasons why a part doesn’t match the specs, he said.

“Maybe the holding fixture is distorting the part or not clamping it tight enough. Maybe a tool is deflecting inconsistently during cutting in one axis or stock material is out of tolerance,” he said. “Whatever it is, that’s what engineers go to school for: … how to make better parts.”

Further hampering any payoff to engineer an automated solution is the current trend for manufacturers to have shorter runs with more diversity of parts.

“You might be able to approach an automotive supplier and say, ‘You’re making this one part for the next seven years, let’s engineer this’,” he said. “‘Let’s invest the time because you’re going to make a billion parts of this one shape.’ But many of our customers are making small lots.”

Zeiss offers the kind of engineering solution Lowen used in his hypothetical auto supplier example. But for those with shorter runs of more diverse parts, and others, the company has PiWeb, an engineering tool that helps manufacturing engineers analyze and determine not only what parts are out of tolerance but gives them tools to determine why.

“If a part is out of tolerance, manufacturing engineers are usually the most educated one to start hypothesizing why the part is out of tolerance and what they need is evidence to support a hypothesis,” Lowen said. “PiWeb allows them to search on a parts database and specifically look at either what machine tools or dates and times or batch numbers to filter and look at statistical trends. And then PiWeb gives them great visual tools—not only of the data but parts themselves showing the dimensions and the coordinates to let them know what axis of the machine tool may need to be corrected.” PiWeb gets data directly from Calypso, Zeiss’ metrology software. However, if the data comes from another metrology software, then PiWeb can read common file formats, including QDAS, DMO or QIF, Lowen said.

Lowen understands the industry needs to move toward more automated solutions with measurement data and automated machine tools correction. These will be more widely accepted as they become easier and faster to implement, he said. For that, Lowen and his colleagues will certainly be conducting more “science projects.”