ANSYS & SAP create insights across engineering & operations
In all of the noise a few weeks ago about PTC’s partnership with ANSYS to embed ANSYS Discovery Live into Creo, you might have missed the news that ANSYS has also partnered with SAP to bring physics-based analytics to the foreground in manufacturing operations.
ANSYS is a fixture in out PLMish universe, a purveyor of many solvers, workbenches, pre- and post-processors — many of the technologies one needs to simulate complex modern products.
We think we know SAP — but, odds are, we don’t. I’ve always thought of SAP as the HR, payroll and manufacturing execution software supplier who happened to market a PLM that moved some of the logic from business operations into product manufacturing. That’s not wrong but it is insufficient, as I learned at last year’s SAPPHIRE user conference. In fact, SAP has been building out manufacturing-focused capabilities, especially as related to IoT readiness. SAP’s manufacturing offerings enter an enterprise from the C-suite, via ERP, CFOs, CIOs and the like; our PLMish vendors typically enter through engineering or manufacturing and have long struggled to grow their visibility and reach.
So when a business software supplier like SAP and a CAE vendor like ANSYS get together, it’s worth a longer look. Why are they choosing to do this? What’s in it for the customer?
This relationship is all about combining ANSYS’ expertise in engineering with SAP’s in operations, to create what the companies say will “optimize the entire enterprise value chain”. To do this, SAP and ANSYS will create solutions that gather insights across a product’s lifecycle. SAP already does this, for a single physics, via Predictive Engineering Insights — aka Fedem, modeling static and flexible structures and moving machinery — but the ANSYS partnership adds its full suite of solvers to the solution. SAP feels that creating multiphysics on its own would be prohibitive, and sees a partnership with ANSYS as getting more simulation capability into the hands of its customers, more quickly. ANSYS, for its part, sees this partnership as giving it access to the C-level, something it has struggled to do on its own.
The companies see customers adopting SAP Predictive Engineering in four phases:
First, at level 1, customers will gain real-time insights for predictive maintenance. For many customers, this is already a huge step. Simply understanding the state of in-service equipment enables line managers and production engineers to predict when maintenance should be performed. This lets them factor in their risk tolerance, and can show significant savings over time-based maintenance protocols, because the activity is performed only when needed.
At level 2, that continues into prescriptive maintenance via in-service analytics. Prescriptive maintenance uses analytics to make predictions about maintenance, as in level 1, but also to recommend action steps. This is advanced stuff, sitting at the crossroads of masses of sensor readings, physics, analytics, machine learning and business logic from asset management and maintenance systems. The best example I’ve heard to explain predictive versus prescriptive is this: physics tells us that a piece of equipment will vibrate to failure sometime soon. Predictive maintenance (level 1) recommends an overhaul before that date. A prescriptive (level 2) solution might consider that timeline, available people and spares (and labor and parts costs) and might even ensure a closes loop for that workflow. Level 2 will be hard; the sheer number of IT systems and departments to integrate make this a multi-year target for most manufacturers.
Level 3 will add these insights into the management of business operations, to optimize costs across an entire product ideation-production-maintenance chain, as well as enabling what ANSYS and SAP call “design to reality”, or taking real-world usage into account during the design phase. To me, this isn’t so much a separate level as a baseline; many manufacturers have tried this using warranty data and other sources of after-the-fact data to affect design iterations. But it, too, is hard to do — especially for companies that are new to having insight into product usage.
Finally, level 4 is about optimizing the entire ecosystem with autonomous simulation, using all of the accumulated information about working products for the next generation offer. Big, big, picture.
OK. That’s the goal. What does this mean in real life?
At this year’s SAPPHIRE, the companies announced “SAP Predictive Engineering Insights enabled by ANSYS” (I don’t even know how to turn that into an acronym —SAPPEIEBA??— so we’ll just refer to it as the first joint offering). This runs on SAP’s cloud platform to appeal to SAP’s installed base, and is targeted at operations and maintenance.
This first joint offering connects real-world operating data to pre-defined simulations through ANSYS’ Twin Builder platform. This is real-time simulation fed by operating data, meaning that the resulting simulations are only as good as the CAE models and the fidelity of the data stream. The companies have created a video to show it in action, here and pasted below:
This video shows maintenance workers clustered around terminals, looking at the results of the simulation. That’s a bit of artistic license — SAP’s Vatsan Govindarajan, who heads global product development related to digital twins at SAP, and Sin Min Yap, head of Strategy at ANSYS, tell me that the joint offering isn’t targeted at people who don’t understand simulations and their limitations. They want experts to create the simulations and to define what should happen next, given a simulation’s results (in other words, predictive maintenance, level 1). Asset managers can review the simulation’s outputs to identify specific issues to address or investigate further. And with both physics-based performance data and business-focused cost/benefit information, they’ll be able to make decisions to maximize uptime and production while minimizing lost profit. This, then, gets fed back into SAP to generate work orders, pull spare parts, order supplies and do all that other business stuff that has to happen in a real plant.
It’s early days —this ships in the Fall for early adopters and in Q1 2019 for general availability— but I can immediately see use cases. In offshore oil and gas, dangerous working conditions lead asset owners to move as much on-shore as possible. Simulating operations and maintenance from afar improves safety and means they may not have to support a crew on the platform. If a problem is found early enough, the asset operator can avoid the costs of an emergency response and plan in detail how to carry out a fix.
In non-hazardous industries, the motivations may not be as extreme but they are just as real. Keeping a production line going generates revenue; every second it’s offline costs real money in maintenance and lost profit. The latter are, unarguably, bad.
For the product maker, too, there’s an opportunity to see how customers use their products. Rather than using statistical modeling or theoretical simulation that indicate that “95% of units sold exhibiting this behavior fail within the next 1000 duty cycles”, the manufacturer can tell its customers exactly what will happen — and how to prevent it. For many, this could be a way of creating new lines of business and cementing relationships.
It’s an ambitious plan. If it works, ANSYS and SAP will create a mix of virtual, physical and commercial that will be hard to beat.
A quick note: don’t take this partnership with ANSYS as SAP losing focus on Fedem. Fedem is hiring, growing its capabilities and expanding its use cases. I hope to have more on that soon.
The image is a screen capture from the launch video, courtesy SAP.