Enablers and Applications in Industrial Autonomy
On the journey to industrial autonomy, each industry operates with its own unique set of accelerators and brakes. One important takeaway from the interviews conducted with industry experts and technology thought leaders was that to maximize the potential benefits, all related stakeholders will need to come together with a unified vision of what they hope to achieve for their particular industry.
AI and Robotic technologies developments at the heart of industrial autonomy
- Augmenting or replacing human involvement from certain tasks, both manual and decision making -
It was clear from speaking with industry experts that autonomy levels can vary widely, not only across industries but also within individual organizations. And while there are applications that have already achieved the level of Autonomous Orchestration, with autonomous systems responsible for decision making, even most new projects are just moving to Automated or Semi-Autonomous levels.
This is because movement to greater levels of autonomy is influenced by the type of application. In more complex multi-system applications, or in applications possessing greater levels of uncertainty and unpredictability, human operators will be present.
Moreover, while it is an exciting time for new technologies, various challenges exist and must be overcome for industrial autonomy implementations to be successful.
Overall, the most common technologies being labeled as “game-changers” by industry leaders are artificial intelligence (AI) and robotics. For instance, AI comes into play in the handling of large volumes of data, and in supporting predictive analytics that can be applied to applications such as asset health monitoring, process optimization, and quality-management. However, the benefits of analyzing data without domain knowledge is limited. The value comes when these technologies are combined with expertise that can identify both the business pain points alongside understanding the context of the data feeding the algorithms.
On the other hand, oil and gas professionals discussed how robotic solutions are increasingly being utilized in both different formats and environments, from mobile robots such as drones being used to support aerial inspection, for example in the monitoring for methane emissions, to remote operated vehicles with robotic arms that can be utilized in subsea maintenance of pipelines.
In addition, “fixed” land-based robots have been introduced in applications on the drill floor to automate pipe-handling and tool operations.
The what, where, when, and how to introduce industrial autonomy solutions is also greatly influenced by the facility and its readiness level. Those with brownfield sites commented that introducing advanced technologies presents big hurdles, as many of these sites were not designed with autonomy in mind, nor to limit involvement of human operators. But these sites were also highlighted as where maximum benefits can be achieved since they use more energy, deal with greater inefficiencies, and require more maintenance. In contrast, greenfield sites are increasingly being designed to include autonomous operations and elements, such as robotics.
Challenges to be faced on the journey
- Hurdles to be cleared on the road to industrial autonomy should be carefully considered and navigated -
Despite the clear benefits, companies hoping to introduce autonomous solutions must evaluate and overcome certain challenges. These include:
Cost: While autonomous technologies can significantly reduce operating costs in the long run, even industry leaders acknowledged that the upfront costs are often considerable, especially in upgrading brownfield sites. In some areas, interviewees had found that sites with access to low-cost labor, and where the overriding consideration is cost, will prefer to hire people instead of introducing autonomous solutions.
Regulations: Working with new technology can introduce a new set of risks.
For example, AI can be used to improve processes, by adapting to varying conditions (for example in pharmaceutical production it could be the changes in humidity, ingredient quality, ambient temperature etc.). However, in such heavily regulated industries, the question of how to validate a self-learning algorithm is not yet clear. As the algorithm modifies itself, it has changed and is no longer covered by previous validation.
Technology limitations: As systems grow increasingly complex, autonomous features and capabilities become more challenged. Whilst analytics solutions can derive meaning from large volumes of data, current solutions are narrow and deep, but less flexible in resolving non-standard issues. As such, interviewees had found that partnering these technologies with people can bring together the best of both worlds. However, even here considerations need to be made in how the responsibility for different decision-making processes are shared alongside legal and ethical implications.
Trust and ethics: One point mentioned in multiple interviews was that the successful introduction of new technologies requires the acceptance and trust of workers. For their part, organizations must consider how industrial autonomy impacts jobs and how the existing workforce can be supported and upskilled.
Identify and accelerate
- Leveraging the value to business from autonomy to support growth and competitiveness -
Whilst these challenges can be significant, and industrial autonomy may not be suitable for all applications, the consensus from interviews was that benefits to be obtained once corresponding systems and solutions are in place are too great to ignore. These include improved productivity, the ability for autonomy solutions to fill the gap in skills and experience created by a retiring workforce, and a reduction in costs as a result of more efficient manufacturing and efficient autonomy solutions in place.
However, for projects to be successful, it’s vital that applications are identified, and a positive cost-benefit analysis is clear. With that in place investment decisions must be made both for the immediate and long-term, to begin or accelerate the journey to industrial autonomy.