Embedded design meets low-code in Siemens integration plans

Thirty months ago, Siemens closed its acquisition of Mentor Graphics. A year ago, the German-owned software company snapped up Dutch “low code” development-tools specialist Mendix. Though they are not obvious bedfellows, these other acquisitions made by Siemens are leading to integrations that the parent expects will expand the use of technologies such as the IoT and the digital-twin concept.

Mendix’s software represented a running theme at a media and analysts conference organized by the recently renamed Siemens Digital Industries Software in New York last week (September 5). Formerly known as Siemens PLM, the company sees Mendix as a way of breaking down the boundaries between not just project- and service-management tools but the job functions themselves. The idea is that domain experts will develop applications for desktop and mobile devices themselves by wiring functions together graphically rather than have to get IT specialists to do it using conventional programming languages.

One example presented by Ralf-Michael Wagner, chief operating officer of Siemens’ MindSphere group, used the example of the Olli automated people-mover vehicle made by Local Motors. The engineering company uses a variety of Siemens design products such as MindSphere, NX for CAD, simulation tools and the TeamCenter PLM package. MindSphere is used to connect sensors in the running vehicles to servers in the cloud so that enginers can get real-time updates on operation. Rather than simply provide data for servicing, Olli is beginning to use the data for design changes to future versions of the vehicle. “They have just released Olli 2.0,” Wagner said.

Hot takes for hot brakes

Wagner showed a situation where vehicle sensors indicated a brake rotor running hot. An application built by one of the engineers using the Mendix tools shows this condition and, by pulling data out of the PLM and CAD, where the problem is appearing. In addition, it shows the results of simulations of the design before the vehicle was made. “The engineer can check: does the simulation behave the same way?” Wagner said. If not, the real-world data can be used to update the simulation and new materials chosen that dissipate the heat better. “At that point, you have identified, in simulation, a new material for the brake rotor.”

Wagner said the same Mendix-built app can be updated to show the change and where it needs to be made when the vehicle comes in for service. “It was all done in Mendix in a few weeks,” he claimed. “Mendix will certainly help us tremendously.”

Although the initial examples of Mendix integration are aligned with mechanical design, the company sees the integration of electrical and electronic functions with mechanical being increasingly important. And with electronics assembly planning and monitoring tools provided through Mentor’s Valor operation, feedback from those processes will become more common.

Embedded deployment

A second form of integration is with the IoT devices themselves. Mendix today works primarily with cloud-based data sources, Siemens expects to provide devices that run Mentor’s embedded Linux to also work with Mendix-built apps. Raymond Kok, senior director of architecture and innovation at Siemens DIS, identified a couple of clear use-cases.

“The first takes advantage of near-field communications [NFC]. When a field engineer gets close to a machine, they will get error codes and field information out of the machine. The second is a much broader use-case. This is using Mendix for edge-based app development. Mendix has been really focused on web apps and cloud. But we’ve started a journey to look at whether we can take a Mendix app and run it as an edge app,” Kok explained. “We can take hardware and turn it into free compute on the shopfloor, using Mendix for [building] analytics.

With Mendix supported by a runtime environment on an embedded Linux, the idea is that the use of Mendix would speed up the development of custom applications on edge-based devices, particularly those that need to combine local data with information obtained from other IoT systems and the cloud without having to go into a C-based programming environment.