The Future of Manufacturing Operations Management: At the cross-roads of MES and IoT platforms
Across industries, the acceleration of product innovation cycles and growing variability of market demands are challenging manufacturing plants and organizations that were mostly designed for efficiency. Add the general trend towards more personalized products, and you understand why a key driver of the “Industry 4.0” (I4.0) revolution is flexibility. Needless to say, productivity and quality imperatives will not go away; in fact, success today requires efficiency and flexibility. As they move from initial proofs-of-concepts to large scale implementations, manufacturers will require a new approach to manufacturing operations management (MOM), which is challenging the MES ecosystem as we’ve known it.
The future of MES
The accelerated renewal of products and their increased personalization require a much tighter integration between product lifecycle management (PLM) and manufacturing processes. A seamless flow of product definition data allows to speed up the introduction of new products or variants, saving weeks or months in the time-to-market of product or service innovations.
Adapting to fluctuating market demand requires flexible factories that can ramp up and down and change over near instantly, while ensuring end-to-end traceability for compliance or customer service reasons. For all these reasons, the role of MES as the “execution arm” of ERP is reinforced, and most Industry 4.0 leaders rightfully still consider it as the backbone of their manufacturing operations.
A new paradigm
While recognizing the importance of a well-designed MES core system, most manufacturers are confronted with similar issues. Most MES implementations started over a decade ago, when the IT/OT gap still was still wide and required lots of hard wiring. As a result, and to ensure uninterrupted operations, adding or evolving functionality is still tied to multi-years release/upgrade cycles. Secondly, the majority of MES solutions were architected at a time when the current standard patterns that make software solutions both flexible and scalable, in the cloud or on premise, still were in infancy.
When talking to production executives, they emphasize two main requirements for manufacturing operations management:
- To keep more complex and fast changing operations under control, they need end-to-end visibility, across the boundaries of systems. To make the right decisions, they not only need real time production data, but also in context of master data coming from the PLM or ERP.
- To cope with increasing pace of change, they do need a stable backbone, but also the ability to flexibly extend it with issue specific apps that will eventually enrich the core system.
Besides its traditional role, the MES then needs to provide flexible and powerful integration, analytics and apps development capabilities, all of which it clearly was not architected for.
First developed for connecting products and offer services extending those, IoT platforms – such as GE Predix, Siemens Mindsphere or PTC Thingworx – package these capabilities into an integrated solution, usually easily accessible in the cloud. Many manufacturers have then selected them to develop point solutions, exploring the potential of IoT and big data analytics to optimize their operations.
Small start-ups also saw the opportunity and used these platforms to offer packaged solutions in the areas of predictive maintenance, asset performance management and production optimization. Finally, the major cloud players have also started offering IoT specific frameworks that provide all the base functions, from shop floor connectivity to machine learning.
The majority of manufacturing clients we engage with are at the same juncture. Initial proof-of-concept pilots have demonstrated that meaningful savings could be achieved in reliability, yield or overall efficiency of assets and even complete production lines. Additionally, such pilot projects also demonstrated the flexibility of the IoT platforms, allowing them to dramatically shorten the time to develop point solutions to a variety of operational challenges. But, these mostly bottom-up pilots often led to the proliferation of incompatible technologies, sometime leading the CIOs to step in to stop runaway costs and security issues.
The question we get asked a lot is how to scale beyond those experiments, avoiding the technology proliferation but without stifling local process innovation.
Not just technology
The pilot projects have also confirmed that I4.0 is a much deeper transformation than deploying smarter automation. Optimizing operations on the fly, based on data, is a profound evolution of management culture. The production manager, whose value is in his intuitive understanding of operations, must accept that analytics will support and enhance his/her intuition and progressively automate it.
Similar to how I3.0 drive for automation required the welder to evolve towards robot programmer, using advanced technologies to make robots smarter and more autonomous, I4.0 means the same evolution for production management. Compared with spending 80% of their time running the factory, their role will require finetuning a system that will make it even smarter and eventually autonomous.
If you believe this is a trivial transformation, on your next factory visit, check whether the recently deployed production dashboard really reflects the reality of the shop floor and is used to drive production. But in a few rare cases, the supervisor will still quite convincingly explain why he still needs his own XL spreadsheet to drive operations and run the daily production meeting.
Digging deeper you’ll find that his latest requests for a couple more sensors and simple analytics to better understand a persistent issue took 6 months to be fulfilled and that automating the solution he finally worked out won’t be possible because of cost and the time it would take based on the existing IT/OT landscape.
The MES of the future
Across industries, CIOs are now considering a future where MES and IoT platforms would be deployed side-by-side and integrated over time into a flexible and scalable platform that can support their I4.0 strategies in the coming decade(s).
While the initial state of companies is quite different, the general direction is quite similar. First, declutter the MES of specific developments and deploy a streamlined core system, focused on the procedural side of operations, tightly integrated with the ERP, including standard/regulatory reporting requirements. Getting back to a standard “no mods” implementation limits costs while allowing to take advantage of the supplier’s innovation via regular upgrades. The general direction being towards more openness, based on cloud standards, MES systems will also become easier to integrate into modern architectures. Second, leverage the IoT platform to develop performance monitoring and optimization applications that require constant evolution and a much faster lifecycle. The platform is used to understand, solve and qualify precise issues, the qualified solution then feeding into the requirements for the next MES release. It obviously doesn’t hurt if the platform also allows to implement a temporary solution to start paying for the upgrade.
When properly supported by the right investments to develop the new skills mentioned above, the deployment of an IoT platform can then significantly accelerate the I4.0 transformation. The creation of a dedicated data team, comprising both forward looking process experts and data scientists who can go from plant to plant to support pilots and roll-outs, is one such area where most leaders have invested. Some of these data teams also include agile development teams who can churn successive releases of apps addressing high impact issues, in close collaboration with process engineers and production supervisors.
Developing architecture patterns that allow seamless integration between off-the-shelf MES implementations and platform based bespoke analytical applications is a major axis of co-innovation with leading clients. The good news is those are the same patterns we’ve successfully deployed at the front-end of businesses around CRM packages.
Multi-disciplinary teams of process engineers, data scientists and cloud architects are exploring the possibilities of getting to the state of autonomous operations one app at a time, leveraging the power of cloud technologies. In a not so distant future, the notion of Manufacturing App Store may very well be a reality.
By Pascal Brosset, CTO Digital Manufacturing, of Capgemini