Industry 4.0 has been described as a game-changer that will transform both the global economy and our daily lives. For instance, Cisco has published a study that claims digital transformation in the Oil and Gas sector can lead to a 4% increase in supply that will stimulate economic growth as much as 0.8%.
Furthermore, industry analysts and technology vendors are amplifying the hype with research and articles with one message: Upgrade your IIoT infrastructure or risk falling behind competitors.
Full disclosure: Presenso is a pioneer in the field of AI and Machine Learning for Predictive Asset Maintenance. At the same time, we recognize that even if we are experiencing an industrial revolution, organizations need to justify incremental investment in technology from a Return on Investment (ROI) and Total Cost of Ownership (TCO) perspective.
As industrial plants struggle to optimize their existing O&M budgets, they deal with the following questions:
- What is the return on an investment in Machine Learning for Asset Maintenance?
- Does current plant staff have the required level of expertise to implement and use these solutions?
- What type of infrastructure is required to support Industrial Analytics?
We will address each of these topics separately.
What is the ROI for Industrial Analytics for Asset Maintenance?
Calculating the ROI on Machine Learning for Asset Maintenance requires the use of several assumptions which are not always easy to verify upfront. Let’s break down the “Return” and “Investment” components.
The “Return” Component of ROI
From a Return perspective, there are two potential areas of financial gain:
- Reduced Downtime and/or Increased throughput and yield: When machines shut down, there is a loss of production. On average, 17 days are lost on each industrial machine per year. Of course, this varies by industry and company. There is a direct cost associated with maintaining a machine or facility that is out of production and an opportunity cost for lost production. For instance, in the automotive industry, the estimated cost per minute of lost production is $22,000.
- Repair Cost: When machines break there is a cost to repair. In some cases, maintenance staff is on-hand and can expedite repairs. However, in most instances, it is hard to calculate the cost of random, unscheduled downtime and many companies do not track these costs carefully. As with downtime, the costs vary by industry and company. For instance, in the Wind Power Industry, repair costs are relatively high because wind farms are often situated in remote locations and there is a significant cost (crane and labor) associated with transporting equipment.
The “Investment” Component of ROI
Because of the number of new offerings from existing vendors and new market entrants, there is much confusion about the “Investment” component of ROI. For some companies, digitalization will require an investment in a new data infrastructure whereas other companies can leverage existing resources. IIoT has attracted some of the biggest industrial technology juggernauts including GE, Siemens, and SAP. Because no one player has gained market dominance, it has created a degree of uncertainty as to which of these entities will ultimately succeed.
Finally, when building the business case for Industrial Analytics one should be cautious about relying on consulting organizations that offer generic boilerplate guidance. There is a tendency to make recommendations without truly understanding both your industry and company. It is not unusual to read reports suggesting results of a 10% increase in production or 20% decreases in O&M expenditures. Much of the benchmark data on Industry 4.0 and Industrial Analytics are based on theoretical calculations that use industry averages.
What are the Organizational Requirements for Industry 4.0?
From an organization perspective, the key issues to consider are your organization’s existing expertise and the extent to which additional resources can be recruited.
In some cases, there are industry trends that will likely impact a company’s ability to migrate to Industry 4.0. Most noticeably, within the Oil and Gas industry, companies are facing an impending labor shortage with an aging workforce. According to Mercer Management’s Oil and Gas Talent Outlook 2016-2025, 20% of Geoscientists in Europe and 23% of Petroleum engineers will reach retirement age by 2020 in the US and Canada.
Why is this important? Implementing Industrial Analytics is disruptive and if an organization is undergoing labor difficulties, this issue needs to be recognized upfront.
A second consideration is the overall skills shortage of big data scientists and engineers. According to recent research, demand for data scientists is expected to jump by 28% in 2020. Many of the top data professionals are drawn to the financial services and high-tech industries where compensation levels are significantly higher than those in the industrial world.
The ability for an organization to recruit data professionals may limit the type of solution that can be implemented. For instance, the concept of building a Machine Learning Center of Excellence (CoE) may not be feasible for many industrial plants.
Similarly, companies should be skeptical about purchasing analytics tools that required internal Big Data expertise.
With solutions such as Presenso Machine Learning for Asset Maintenance, we start by providing Business Intelligence (BI) dashboards that visualize information about machine failure. Furthermore, there is no requirement to hire big data experts because the solution performs the analytics tasks without the need for human input.
Does your organization have the infrastructure to support Machine Learning for Asset Maintenance initiatives?
Machine Learning cannot be used for Industrial Analytics without access to accurate data. The quality and use of data vary by industry. For instance, according to an Accenture/IDC report, there are significant problems with generating and utilizing data in the oil and gas industry. Significantly, it takes more than three days or for 38% of companies in this sector to access onshore production data. This data is consistent with anecdotal evidence that suggests that less than 5% of sensor data in the oil and gas industry is analyzed.
That said, there is no consensus about the required infrastructure for Industrial Analytics. There are valid reasons to embrace Industry 4.0 that include using big data to improve manufacturing, inventory costs, logistics, and product quality. Industrial Analytics is not the only component of Industry 4.0 but for many organizations, it is the primary driver.
A manufacturing plant with a holistic definition of Industry 4.0 may require more of an investment in its data infrastructure especially if the same data needs to be used for different processes. However, if asset maintenance is the primary consideration, there are solution alternatives that do not require new infrastructure.
Conclusion: Caveat Emptor
This article summarizes some of the basic considerations for building a business case for Industrial Analytics. There are many variables and assumptions that need to be made and we caution against that use of over-hyped vendor sponsored statistics.
We are not suggesting that you delay implementing a Machine Learning for Asset Maintenance solution. Instead, we recommend a conservative, fact-based approach:
- When creating the business case for Industrial Analytics, test each solution provider by giving them with 2-3 years of historical data. Compare their predictions of machine failure relative to actual failures.
- Include 2 to 5 solution providers in the test.
- For final vendor selection and to forecast ROI and TCO, use data that is based on the pilot instead of relying on industry benchmarks.
We invite you to learn more about Presenso by signing up for a complimentary webcast briefing by clicking on this link.