With the dawn of Industry 4.0, manufacturing plants are applying artificial intelligence (AI) to operational data. Early adopters are drawn to the promise of digitalization — new revenue streams and improved operational efficiencies. Even those industrial facilities waiting on the sidelines are evaluating how to transform their operations, albeit incrementally.
There is one industrial sector that is less likely to celebrate the bullish research reports of Industry 4.0 — the mainstay of Industry 3.0: machinery OEMs. This is because of the perceived value shift from the equipment and systems within the industrial plant to the data that is generated from such machinery.
This article evaluates two models for Industry 4.0 — the digital twin and Hardware as a Service (HaaS). By comparing the two, this article explores how OEMs can play a leading role and secure their participation in this burgeoning market.
OEMs adjust from Industry 3.0 to Industry 4.0
The role of the OEM in Industry 3.0 was relatively simple. Equipment was sold to an industrial plant either directly or via a distributor. If the equipment was sold with a warranty, the OEM was responsible for repair or replacement during the warranty period. In many instances, the relationship between industrial plant and OEM was transactional and limited to the one-time sale of the original equipment or new parts or equipment after wear-out.
As plants migrate to Industry 4.0, OEMs will need to adjust. The competitive landscape is shifting, and technology is playing an increasingly important role. New agile competitors, fueled by venture capitalists, are entering the industrial arena. Traditional players such as GE and Siemens are reinventing themselves and spending on research and development (R&D) and innovation. If machinery will be purchased based on how it integrates into a company’s Industrial Internet of Things (IIoT) strategy, then OEMs will need to reconfigure their product offerings and incorporate new IIoT-based approaches.
The digital twin
Although GE did not invent IIoT or Industry 4.0, it has invested heavily in its Predix platform over the last couple of years and has played a major role in promoting the digital twin approach. Another digital twin proponent is Gartner, which included the digital twin in its annual lists of Top Strategic Technology Trends in 2017 and 2018. Gartner has predicted that by 2021 “half of large industrial companies will use digital twins, resulting in those organizations gaining a 10 percent improvement in effectiveness.”
The digital twin provides manufacturers with a virtual software clone of the physical machine. This clone is generated from the actual blueprints and requires significant man-hour investments of machine designers and software engineers. In real time, it mimics the underlying machine behavior, and allows simulated machine and process performance tracking. Some of the benefits to the industrial plant include asset maintenance and performance optimization. Secondary benefits include improved planning and decision-making.
There is no doubt that the digital twin provides tangible value to end customers. At the same time, industry giants have yet to make the compelling case that the digital twin can be applied to existing plant assets.
For example, when a digital twin is bundled with a jet engine, the customer gets immediate value from the insights contained within the system. Although in absolute terms the cost of the digital twin development is expensive, when compared with the overall cost of a new jet engine the cost is relatively minor.
The example of a jet engine is used because of its high price tag. However, if a digital twin is bundled with a new medium-sized water pump or electric generator, the relative cost of the digital twin model becomes prohibitive. Furthermore, when a digital twin is made for a piece of existing, old industrial machinery, it needs to be modeled on its exact physical blueprints, and it requires the original designer’s involvement. This is where the problem begins because this approach is not feasible and not scalable. Over time, blueprints are outdated and performance specifications are not always documented. People and knowledge move on and may not be available. To work accurately, there can be no inconsistencies between the digital twin and the performance requirements.
Another element to consider is the deployment cost. To create a virtual clone, plant technicians need to train the digital twin on its underlying behavior. This is an expensive, labor-intensive and time-
consuming exercise. As a result, the digital twin is only cost-effective for new and expensive equipment and less applicable to a plant’s existing asset base.
In some ways, the digital twin is an adversary to the OEM. To the extent that a large OEM can sell its machinery bundled with a digital twin, this creates a competitive advantage. However, 99 percent of OEMs lack this capability and the high R&D cost to develop digital twin capabilities is a significant barrier to entry.
HaaS as an IIoT entry point for OEMs
One topic that is sometimes overlooked is that industrial plants lack the internal competencies to manage multiple IIoT systems. McKinsey has forecast a shortage of up to 190,000 people with deep analytical skills in the U.S. in 2018. For the industrial sector where pay scales tend to lag behind financial services and high-tech industries, the ability to recruit big data scientists and engineers is further limited.
Therefore, although the digital twin is seen by many as a new shiny object, its mass market implementation and deployment is impractical. Enter HaaS. OEMs can now achieve not only the one-time hardware sale itself, but also bundle it into a service platform.
They can transform their product into a service platform offering by operationalizing the data generated from the already-embedded machine sensors. As long as there is internet connectivity, the data can be analyzed remotely and effectively in a cloud-based location. Scalable and automatic machine learning algorithms can be applied to the sensor data to detect anomalous behavior patterns, which indicate asset degradation or evolving failure.
Another benefit of this approach for the OEM is that once its equipment is operating in the field and the data is uploaded to the cloud platform, it can keep in touch with its equipment and can learn and track its ongoing field performance to improve future models.
The key selling point to the digital twin is that it predicts failure, thereby reducing downtime and maintenance costs. If the primary driver for a plant to consider the digital twin is for asset maintenance, then this same service can be provided by the OEMs themselves at substantially reduced costs.
In both greenfield and brownfield scenarios, HaaS provides compelling value that is cost-effective and scalable.
In a greenfield scenario, where new equipment is sold to the industrial plant, it is likely that the customer would prefer a bundled solution — a predictive maintenance package based on the actual hardware sold. The competitive advantage relative to the digital twin is lower cost, scalability and the limited input required from production plant technicians.
Perhaps the more interesting and more immediate opportunity for OEMs is their existing customer base — the brownfield scenario. If OEMs can develop a HaaS offering and refit assets that have already been installed with predictive maintenance capabilities, it provides an upsell opportunity to extend the OEMs footprint within its customer base.
Another aspect to consider is the business model that applies to both the greenfield and brownfield scenarios. With a HaaS model, OEMs can adopt the software business model where solutions are sold on a subscription basis with limited upfront investment on the part of the industrial plant. These multiyear agreements secure long-term and predictable revenue streams for the OEM. Depending on the taxation jurisdiction, the industrial plant can benefit from using operational versus capital expenditures.
IIoT predictive maintenance: A nascent market
According to ARC Research, the predictive maintenance category is growing at a compound rate of 39 percent and is expected to reach $11 billion in worldwide sales by 2022. As industry leaders such as IBM, SAP, Cisco, GE and Siemens look to this market for incremental revenue, OEMs have an opportunity to use their existing hardware business to gain market entry.
Most OEMs do not have competencies in AI and machine learning. The obvious question for OEMs is how can they enter the HaaS market without significant R&D investments? With advances within unsupervised machine learning, algorithms can be more easily applied to industrial scenarios and applications. The result is that a new breed of solution providers is emerging. OEMs can find ways to access the machine-learning discipline via partnerships and licensing agreements.
It has been estimated that $20 billion a year is lost globally in the process industries alone.
The promise of predictive maintenance — higher uptime and lower maintenance expenditures — is now firmly on the radar of the CEOs and CFOs within the industrial domain.
Hardware OEMs should use the disruptive change brought about by Industry 4.0 to reposition their products as services and benefit and reorient their business models.
Dr. David Almagor is the chairman of Presenso and a serial entrepreneur. He has more than 30 years of experience in managing complex R&D and business entities, and taking them from startup inception to business success. He was previously the founder and CEO of Panoramic Power that was acquired by Centrica PLC. Almagor is author of more than 40 publications and co-author of five patents. He holds a doctorate and a master’s degree in electrical engineering from the University of California, San Diego, and a Bachelor of Science in electrical engineering from the Technion, Israel Institute of Technology.