One could argue that it is premature to consider the future of Maintenance 4.0 when many industrial plants are struggling with their initial deployments. In fact, research from McKinsey suggests that two-thirds of Proofs of Concept are stuck in pilot purgatory. In our opinion, an understanding of underlying trends can help inform current investment decisions. This article will explore Maintenance 4.0 concepts that are not yet mainstream but that should be considered part of the planning process.
1. Data Science without Data Scientists
Some pithy industry analysts like to compare Big Data with oil. We should keep in mind that by the time a gallon of gasoline is of value to the consumer, it has passed through downstream, midstream and upstream processes that are labor-intensive. This model cannot be scaled to support Maintenance 4.0. The industrial sector does not have access to a pool of Machine Learning experts to apply manual processes to extract, cleanse and analyze Big Data used for Industrial Analytics.
What does this mean for current Maintenance 4.0 investment plans? Future maintenance programs will need to be built based on the assumption that the science of Machine Learning will become automated. In fact, Automated Machine Learning (or AutoML) is a fast-growing area of innovation within the data science discipline.
The implication for industrial plants is that they will benefit from advances in Machine Learning without developing deep expertise in data science.
2. Emerging Roles for OEMs
In the early 1960’s Rolls Royce released its “Power by the Hour” sales model for jet engines. Almost six decades later, many OEMs are adopting a Hardware as a Service model whereby industrial equipment is leased to manufacturers. The OEMs continue to own the underlying asset and assume responsibility for maintenance activities.
In a research study conducted by Emory University and Presenso, 69% of O&M employees who responded to a survey indicated that in the next five years industrial plants will expect OEM vendors to integrate Predictive Maintenance into their equipment.
Industrial plants considering large scales investments in machinery should be aware of this trend. Furthermore, those Predictive Maintenance solutions that require the installation of hardware and software may become redundant if OEMs assume a greater maintenance role.
3. No-Fault Remediation
A future iteration of Maintenance 4.0 is the amalgamation of Industrial Analytics and Maintenance functions. When a Machine Learning algorithm identifies evolving failure, it will also provide prescriptive guidance on how to remediate. The default scenario is that the decision to performance a repair will be made by an O&M professional. However, we expect that a so-called “no-fault” class of repairs will be automated without human input. Of course, there will be rules in place so to limit these repairs to those that pose minimal risk.
Although Maintenance 4.0 is still in its nascency, planners will need to consider both current and next generation solutions. This additional level of complexity is not easy, but given the rapid pace of innovation, it is unavoidable.