Alex West, Senior Principal Analyst
Whilst the use of cloud computing in some enterprise applications is near ubiquitous the uptake in manufacturing applications is still in its infancy and has been slowed by concerns around topics such as cybersecurity, vendor lock in and in-house skills. However, the rate of adoption is increasing both with a greater presence and focus from cloud providers as well a (gradual) transition to cloud native MES, PLM, CMMS and APM solutions from software vendors etc.
At the same time cloud providers are beginning to develop solutions that are more tailored to the needs of customers in different industry sectors, with a number having released (or with plan to release) a dedicated cloud solution for manufacturing. These packaged solutions provide a more focused set of tools relevant to the industry, along with access to an ecosystem of partners providing expertise and applications supporting the manufacturing sector.
A key question for the industry (and one that won’t be resolved in the short term) is the issue of how to work with multiple cloud providers. To date many companies will be partnering with different cloud vendors for different applications (e.g. ERP, CRM etc.). However, the ability of customers to share data between multiple clouds, or transition an existing project from one cloud to another can be cost prohibitive as a result of the egress charges.
As the way companies deploy compute to their manufacturing processes evolve, edge computing will be a central component of this. Manufacturers seeking to mature their digital transformation programs are collecting more data, that requires storage and also additional compute power. Whilst the cloud will be part of this, challenges such as low latency requirements, dealing with large volumes or frequent cycles of data etc. Such challenges will necessitate the deployment of edge solutions.
However, how companies look to deploy compute across their operations will vary from smart sensors through to on-premise data centers, with a range of products in between, including PLCs to gateways, also becoming more intelligent.
Whilst automation vendors are making more intelligent products, cloud vendors are building out their edge capabilities with both hardware and software solutions – both directly and through partnership.
As digital transformation projects come into focus, asset health analytics is at the forefront as low hanging fruit for companies to start projects with. Particularly in asset intensive industries.
Vendors from bearing manufacturers to cloud providers have entered the market with a range of solutions varying by:
Alongside this there have been numerous acquisitions and partnerships as companies look to consolidate the combination of data science expertise with industry/product knowledge to support the development and deployment of suitable algorithms. This isn’t straightforward, and whilst there has been pockets of success, currently less than a quarter of current asset health monitoring projects are utilizing AI, with the majority of applications still focused on more traditional condition monitoring solutions.
Vendors will continue to promote a variety AI based case studies at Hannover this year, but bringing these diverse expertise together can be a challenge for vendor and customer alike.
In recent years and amidst the backdrop of increasing demand for data analytics functionality at the edge and predictive maintenance from OT (Operational Technology) equipment, Industrial PCs have increasingly been adopted to bridge IT and OT environments.
So far, it has been difficult to collect data from PLCs across a range of vendors and the different network protocols that have historically been deployed.
Orchestration – Factory floors often deploy multiple PLCs from different suppliers. IPCs are expected to be used for orchestrating multiple PLCs with fully integrated control and to develop programs across heterogenous environments serving scalable tasks – not only to act as a controller but also edge computing device, data gateway, and AI enabler.
PLC virtualization technology – Although in the early phase of development, the transformation from existing hardware to virtualization will represent the next generation of factory automation. Several suppliers have already released hybrid-type controllers including software virtualized PLC-specific ASICs and firmware. Some products share data in the shared memory space between the computing and control functions, enabling data analytics and feed back to the control field without routing through a physical network.
The emergence of 3D vision has significantly impacted the outlook of investment into smarter factories and warehouses. Historically the biggest drawback for machine vision was the technological and cost barriers that prevented accurate 3D imaging in complex defect detection and volumetric sorting in logistics. Currently the highest adopters of 3D vision are applications in wood processing, integrated circuit inspection and logistics for track and trace applications.
