Siemens tries to close the circle

At first sight the recent purchases by Siemens’ industrial software operations do not seem to connect but the group thinks it can already show how to build a coherent whole out of the jigsaw pieces.

Having embarked on a buying spree that has over the past few years brought together companies such as electronics design specialist Mentor Graphics and Mendix, a Boston-based developer of tools for speeding up the creation of mobile and web apps, Siemens PLM Software last week said it was time to change its name to reflect its wider scope.

The group, now known as Siemens Digital Industries Software (DIS), used a conference in New York organised for analysts to make a case as to why all the pieces fit together.

The main components are the product-lifecycle management software TeamCenter that gave the group its former name; a collection of simulation and computer-aided design (CAD) tools that now encompass mechanical, electrical and electronic work; and the MindSphere cloud-based environment for co-ordinating IoT systems. Rather than build them into some kind of super-package, Siemens is unbundling some of the components, offering them as rented ‘microservices’ running on cloud servers. Instead, Hemmelgarn and his lieutenants expect Mendix to act as the glue, with customers deciding how they can best use the tools. Because Mendix is presented as numerous pre-cooked components that users wire together to build workflows and user interfaces, Siemens reckons more customers will seize upon a greater ability to customise.

“I call it the CIO pressure-relief valve,” Tony Hemmelgarn, Siemens DIS president and CEO said in his keynote, because they will not have to budget for additional IT staff. Instead, engineers and other domain experts at customers will build their own apps: “They can say: ‘I don’t like this app; I want to change it’.”

Many have already been doing it in an ad hoc manner by building spreadsheets and scripts that cull information pulled out of software like TeamCenter, such as bills of material. “They can do it without creating shadow databases,” Hemmelgarn said. “The IT guys know they’ve got guys developing shadow databases. This way, customers preserve revision and access control but they can put development in the hands of the users.”

How does that work in practice? Hemmelgarn claims it gives IoT deployments much greater power than many existing applications suggest. Siemens promotes the idea of a maturity model, similar to those used for software development, that users can follow. Simple monitoring of devices is right at the bottom of that maturity model.

“Everyone starts with preventive maintenance,” Hemmelgarn argues. “When we show customers the ability to make a round trip so they can fix problems [in the design] that makes a big difference.”

How does this work in practice? Siemens has some examples. Local Motors, which is developing the Olli autonomous people-mover, has an application built using the Mendix tools that does not just track the real-time behaviour of its systems but which supports changes to the design.

According to Ralf-Michael Wagner, chief operating officer of Siemens’ MindSphere group, the vehicle maker uses a variety of Siemens design products such as MindSphere, NX for CAD, simulation tools and the TeamCenter PLM package. Now they are wiring these tools together using the Mendix app builder. In his presentation of how Olli design engineers use the software, Wagner showed a situation where sensors indicated a brake rotor running hot. The app pulls data from the PLM and CAD tools to show 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 the simulation has not captured the problem, it almost needs certainly updating. If on the other hand it is a likely condition but not one that the product-engineering team originally expected to be met, they can opt to make a change to the vehicle design itself – either to the next generation or possibly when this machine comes in for a service. “At that point, you have identified, in simulation, a new material for the brake rotor,” Wagner said, adding: “It was all done in Mendix in a few weeks.”

Shipyard crane builder Konecranes was a customer that set out to use IoT sensors to do standard maintenance prediction around three years ago. Andreas Geiss, CTO for cloud app solutions at Siemens, said: “Over time Konecranes has developed. They said it’s great feedback we are getting [from the machines] but how can we feed the data back into crane designs?” 

Similar to Olli, Konecranes is now pulling data out of Mindsphere and combining it with functions in TeamCenter and SimCenter to try to streamline the way it creates new designs and potentially even change the way it sells the machinery. 

“Typically, in the past, when Konecranes got a request for a new crane setup they took customer requests and went back to cranes they had built already. Mechanical engineers don’t like to reinvent: you take what you have done already and optimise from that. However, this means you tend to over-engineer everything. You deliver cranes that are way more complex and may be delivered for different capacity than what is actually needed,” Geiss explained. 

Using the operational data from existing cranes, Konecranes was able to build better simulations that the company used to drive the new-product design process, with greater ability to fine-tune the characteristics of each order.  The next step is to offer customers the ability to rent lifting capacity rather than selling cranes to harbour logistics providers, who, according to Geiss want better uptime and lower capex. In principle, because the cranes can be more closely matched to the need and maintained that way through real-time data from the cranes themselves, the overall system will be more efficient. 

Referring to Siemens’ three-level concept of digital maturity, Geiss said: “Konecranes started at connect and monitor. They moved over to analyse and predict and now its digitalise and transform. But without the first step I don’t think they could have come up with this approach.”

Leave a Reply