Digital Power Plants - What can operators of renewable and thermal power plants learn from another to become AI-ready?

The power industry is facing the biggest transformation in its long and proud history. The secure digitalization of power generation assets – new and old, renewable and thermal – is right at the heart of an accelerated decarbonization of our energy system.

Research from 麦肯锡 suggests that via rigorous end-to-end digitalization/digitization (as well as data-orientation) thermal power plants can realize an additional EBITDA potential of 20 to 30%![1] Thermal plants face significant cost pressure because of renewables and rising emission costs.

These EBITDA improvements are also available for renewable generation (especially offshore and onshore wind). The strong growth of renewables, guaranteed offtake prices, focus on hardware engineering, as well as the relatively short history of large-scale renewable generation are root-causes for untapped profitability potential.

According to the research, even sophisticated generation players in thermal (and also renewables) have only implemented isolated digital use cases, lacking significant P&L impact or not improving the way of working rigorously. The business case for digitization also supports achieving decarbonization goals. Digitization can accelerate the reduction of greenhouse gas emissions of thermal plants.

What is a 'Digital Power Plant'?

Digital power plants are fully data-centric power generation facilities (renewable and thermal) that leverage advanced software, technologies, and AI with the following objectives:

(1) reduce operations & maintenance cost;

(2) improve the power output and increase asset flexibilities; and

(3) increase the HSSE (health, safety, security, environment) performance.

They integrate a broad spectrum of technologies: existing and new sensor technologies, OT and IIoT devices, AI, mobile and augmented reality devices/Metaverse, drones, robots, and more. The data and information generated is used to improve the planning, monitoring, and control of all aspects of power generation assets. A digital power plant is operational and workflow oriented. There are overlaps with the 'digital twin' concept, but this is a more holistic concept.

Digital Power Plant and the 'Control Room'

A key activity of any generation plant (renewable or thermal) is the control room. Operators use this 24/7 facility to ensure a smooth and safe performance of the plant. The control room aggregates all relevant OT and IT data and information and facilitates real-time monitoring and decision making. It should be a central starting point for an end-to-end digital power plant journey.

Control rooms are sensitive from a security perspective (both physical and digital security). This is an additional advantage because it also enables hardening the existing security mechanisms in place. This is required, because the status quo is always challenged by hackers and attackers and new threats are emerging.

New security technologies and data innovations enable utilizing operational systems (OT), control rooms and legacy (IT) infrastructure for an end-to-end digital transformation towards the 'Digital Power Plant'. This includes raising current data and cyber security levels. But how can this be implemented?

‘Virtual Operations Center’ for generation assets

Planning and implementing a ‘Virtual Operations Center’ for generation assets operationalizes the end-to-end digital transformation (both in renewables and thermal). It is a digital addition to the existing systems, control rooms, etc. and supports and implements the Digital Power Plant. A broad range of benefits is enabled:

1. improves the day-to-day control room work: remote steering, less paper work, optimized inspection rounds, better performance observation, consolidate multiple assets during nighttime;

2. data-driven maintenance planning: eased use to run AI for predictive maintenance, sequencing of maintenance work, spare-part and workflow management, and more;

3. improves overall plant management: real time asset views, financial impact, people management, consolidated views, and more;

4. create ‘AI readiness’: reduce the ‘time-to-AI’, access to data (availability) is key to digitize and introduce AI and data access outside the generation asset if often impossible (because of security concerns, air-gapping); solves this problem because it is based on virtualize data;

5. track decarbonization performance: real-time tracking of plant emissions (thermal) and full audit of data streams;

6. vendor agnostic and portfolio overview: data from heterogeneous hard- and software environments (multiple vendors) is released, and different plants/fleets can be aggregated for reporting and steering purposes.

How does the technology work?

The required software layer (secure and governed data abstraction layer) resides on top of all existing and new OT/IT data infrastructure. This enables generators to do anything with data coming from the generation asset plus data from additional internal and external data sources (e.g. energy market data, additional data from hardware vendors, maintenance data, AI and more) in a highly secure, governed and trusted way. This includes blending data from different sources, creating new virtual data sets which feed new AI algorithms or can be used for visualization and performance analysis, and more.

This approach overcomes the security concerns stemming from the OT/IT convergence (moving OT data into IT and sending back commands and data from the IT to OT). It does not require moving data around, building new IT infrastructure and copying large amounts of generation asset data into new data repositories.

What are the future challenges for plant operators?

Security challenges

Generation assets are an attractive target for hackers and malicious actors because they are critical infrastructure.? This has been a long-time, key risk for generation assets. One security strategy is air-gapping (isolating the OT and control room) and there are more. A recent conversation with the CEO of a very big generation company revealed that this has worked so far, but lately they had an attack. The hacker came via a hardware vendor who was connected to the control room. There is no doubt that additional security concepts and technologies need to be deployed in power generation.

Flexibility challenges

Another problem is the use of proprietary operational data collection, storage, and analysis systems in power generation. These systems can prevent Digital Power Plants to emerge, because they lock-up data and make it hard for generation companies to get access and develop additional use cases. Their cost and high vendor-lock potential are additional disadvantages. An additional challenge for the Digital Power Plant stems from heterogeneous generation equipment (different vendors) across larger asset portfolios. These come with proprietary data and security systems leading to the challenge how to holistically manage this portfolio and create synergies.

Solutions

The proposed concepts are a great opportunity to address data security and flexibility aspects and achieve improvements in overall data and cyber security. Managing and securing the vast amounts of real-time data generated by generation plants of all kinds as well as managing and integrating older legacy systems is a significant project. Air-gapping was already mentioned as an insufficient security strategy. Pursuing end-to-end digitization triggers reviewing all security related aspects of the operations and challenging their future-readiness. This can also include crypto-agility (what are the implications of the emergence of quantum computers for data and cyber security?).

New technologies and innovation

The security and flexibility challenges can be addressed with new technologies emerging on the horizon right now. These are based on ‘Zero Trust’ security models and approaches, reducing the potential attack surface (e.g. Explicit Private Networking (XPN)). Unlocking data for additional use cases is the core objective and added value of governed data virtualization. The combination of both delivers unique end-to-end data security and robustness from the generation sensor/device to the AI.

Challenges for generation companies that cannot be solved with technology are cultural aspects and access to a workforce that is familiar with modern OT and IT technologies. Power generation has a long tradition and strong engineering culture that is not software-led. Similar issues can be observed in other industries such as automotive, where software systems become a key element of the value proposition. The same will happen in generation over time.

?

[1] https://www.mckinsey.com/industries/electric-power-and-natural-gas/our-insights/the-digital-power-plant-of-the-future

?