Utilizing the Opportunities with IIoT Adoption
Based on surveys and conversations with our clients, we know IIoT is being adopted by equipment suppliers. As your equipment is replaced due to age, upgrades or other clear business benefits, IIoT capable devices will enter your plant. The equipment-related data will become increasingly available – IIoT success! Now what? Hopefully, you will use this information for algorithms that predict machine failure so problems are prevented before they occur i.e., predictive or prescriptive maintenance programs. The IIoT data is a resource to be utilized, and the associated analytics managed.
As I have written in the past, predictive maintenance (PdM) provides a key benefit that financially justifies most of today’s IIoT projects. But, research by ARC Advisory Group indicates that only 18% of plants now use data in the historian for PdM calculations. Currently, the historian is focused on process data (like temperature of the materials). With IIoT, this is expected to expand to include equipment data (like temperature of the bearings). Adding equipment data provides a significant increase in the opportunities for PdM calculations.
Growth in IIoT-based PdM Algorithms
Since the sensors and data acquisition come with an IIoT enabled equipment, the costs for a PdM implementation drop dramatically compared to previous methods. Now, the focus becomes an engineer’s time to investigate and write the algorithm. With the far lower implementation costs, we expect a proliferation of PdM implementations and associated algorithms based on data deposited in the historian or other time-series database.
Each PdM implementation will likely use equipment and process data in an algorithm. The algorithm can take many forms – algebra, Boolean logic, statistical process control (SPC), machine learning, or others. At least initially, the keep-it-simple approach should prevail with algebra and Boolean logic being predominate.
How Will You Manage Thousands of Algorithms?
With the relatively easy access to process and equipment data combined with the low cost of implementation, it makes economic sense to cover all critical and many non-critical machines – potentially thousands in larger plants like automotive, semiconductor, consumer products, pharmaceuticals, refinery, and bulk chemicals.
Now, let’s explore management and administration of a multitude of IIoT-based algorithms like:
- A narrowly focused control engineer changes a tag from T101 to T215 for the control system. How is this change handled among the algorithms including that tag? Is there a “where used” function with search and replace?
- An upgrade occurs and many points will change. Which algorithms need to be modified or removed? How is implementation scheduling managed?
- Can an algorithm become an object that is replicated across many equal/similar assets (by pointing it to the individual asset’s tags), and then modify/improve the math in one place for many assets?
- Governance: Who can edit which algorithms? Audit trail?
With predictive maintenance, technicians perform work just before the equipment incurs a problem i.e., when needed. Compared to preventive maintenance, a study by Shell shows that PdM reduces maintenance costs by half. The Plant Engineer’s Handbook has the following benefits for PdM:
- Maintenance costs – down by 50%
- Unexpected failures – reduced by 55%
- Repair and overhaul time – down by 60%
- Spare parts inventory – reduced by 30%
- 30% increase in machinery mean time between failures (MTBF)
- 30% increase in uptime
Also, the business impact of lower unplanned downtime has immediate benefits like increased capacity and revenue. In addition, a multitude of secondary benefits occur like lower inventory (less safety stock for unplanned events), and improved customer satisfaction (with more on-time shipments).
With the adoption of IIoT, users should consider how they will manage the related PdM algorithms in today’s dynamic manufacturing plants. Suppliers of IIoT software should look beyond the creation and implementation of algorithms, and consider how they will be supported and managed.
About ARC Advisory Group: Founded in 1986, ARC Advisory Group is a Boston based leading technology research and advisory firm for industry and infrastructure. For further information or to provide feedback on this article, please contact firstname.lastname@example.org