Energy distribution systems, an inter-networked and intricately balanced collection of sources & load are experiencing a time of unprecedented change. Following decades of under investment, energy systems have become a priority area to address local issues of critical infrastructure rebuilding & resiliency as well as customer empowerment while also supporting the wider goals of climate change, energy sustainability and economic development.
Automation, which is an inherently data-driven effort, has become an integral part of the energy industry. Evolving from pneumatically-driven components designed to mimic manual human actions, initially, automation was implemented through relay logic or a series of standalone controllers, depending on the industry sector. Today, programmable controllers and protection devices are almost ubiquitous through the industry – used for data collection, breaker reclosing, metering points and everything up to full plant & system control. This has a number of results – the first is that it challenges staff to develop new skills. The second is that it allows for remote visibility and operations of facilities. The third is that it enables data collection, which sets the stage for welcoming the next industrial revolution of cyber-physical systems and AI. However, these essential systems also open new cybersecurity risks, create dependence on communications infrastructure and challenge traditional architectures.
Distributed Energy Resources (DER)
Power plants made their debut in the 18th century and were initially controlled either manually or through the use of mechanical automation. As technology advanced, the means to maintain the stability of the system and to protect the equipment also advanced. Automation has evolved through manual, pneumatic, electric and now programmable means of control. This advancement has been spurned by technology & the industrial revolutions and has generated positive changes within the industry. However, now we are embarking upon a different set of changes – the thought process underpinning the planning and distribution of energy is changing. Although we will likely still have large centralised harvesting centres close to the energy sources that capitalise on networks of transmission and distribution infrastructure to deliver the product to customers, there is a shift taking place towards the acceptance, use & deployment of Distributed Energy Resources (DER). This shift moves energy harvesting closer to the user, reduces the dependence on costly & vulnerable transmission systems and transforms distribution systems into bidirectional networks.
What data does
Programmable automation systems started out with limited I/O, only collecting data from sensors and transmitters connected via voltage and current signals to costly centralised processors for interpretation, calculation and storage. Modern automation systems also have the ability to be distributed, placing I/O, processing, and in many cases, intelligence in the field using both distributed I/O systems, networked processors and edge devices. IIoT-enabled devices and standardised communications protocols have not only streamlined the manner in which data is organised, formatted and collected, but have also enabled the collection of greater amounts of data. Currently, systems collect all of this data, serve it up for real-time and short-term trending functions and then send it off to longer-term storage. Having all of this data stored in a data historian, data warehouse or data lake allows visualisation and analysis. It empowers the organisation to identify patterns and correlations in the data which triggers action. This provides the foundation for data-hungry advances, such as Machine Learning, predictive control, mixed reality and AI applications. Enterprises, now, have the unprecedented ability to automate both industrial and business processes. Prosumers now have the ability to both produce and consume energy on the same network, all monitored, controlled and regulated by IIoT-enabled edge devices feeding optimised system models to determine needs, constraints and pricing. In addition to this, the increased level of data inflow also provides support to additional functions within the energy ecosystem.
A transactive energy approach
To use an example to illustrate, most energy companies are asset-intensive organisations. These assets require time and effort to ensure that they are in optimal working condition to deliver energy to consumers. Starting at the equipment level, Machine Learning algorithms, such as decision trees, fed from real operating data, have the ability to optimise the operation of the equipment and hence, the operation of the system overall. Additionally, similar algorithms can be used to extend preventative maintenance programmes into predictive maintenance programmes, which have the potential to cut costs, which is immediately reflected in the bottom line. If we collect all of the decision trees within the maintenance programmes for the equipment, then we have a random forest to deploy the asset management programme. Furthermore, ensuring that all assets are kept in the best operating condition possible ensures that availability is maximised. This sets the stage to develop and implement economic and control mechanisms that allow the dynamic balancing of supply and demand – a transactive market. A transactive energy approach provides greater efficiency of usage of grid assets, including generation and storage.
Additionally, it also provides greater resilience and reliability while engaging the consumer by providing them with a choice. Traditionally, the realm of large utilities and enterprises, now individuals would have the ability to transact on the market with the resources available to them. Automation provides the marginal pricing, time shifting and the ability to maximise the use of assets, not to mention the optimisation. The convergence of IT and OT systems allows transactions to be automated in the background, allowing energy to be generated and delivered in exchange for value. This continues until a sufficient supply is established to stabilise pricing and satisfy demand. Although prosumer ownership of renewables works well on the electrical side, it is not as common in other areas of energy. Consequently, we see Energy-as-a-Service models being established by manufacturers, consultants and utilities. This concept is already being used with Virtual Power Plants (VPP), whereby, the utility owns the generation and interconnection infrastructure placed in a distributed fashion across residential rooftops and simply rents the rooftop space from consumers. The utility uses the energy generated to offset purchased power, and the homeowner receives a steady income source from the utility.
The path forward
Where does this leave us? The energy industry will continue to evolve. It has always been dynamic, fearless and willing to be at the edge of technology. There are many advancements that have already started, such as VPPs, that will continue to be innovative within the industry. We have been seeing the convergence of IT and OT for some time now with the introduction of client/server technology, thin clients, virtualisation and ‘as a Service’ models into the automation world. These technologies both supply greater levels of data and also provide access to a wider range of data sources to the automation system, which, in turn, allows for more accurate predictions and better decision-making. This is complemented with OT advancements, such as two-wire Ethernet and modern standardised IP-based communications protocols. Data lakes & data warehouses are being implemented to capture and store the new levels of data being generated and cloud-based infrastructure will enable & simplify analysis and processing of the data. Companies look to Machine Learning & Machine Learning algorithms to identify patterns within their data and manage & understand the information in addition to making predictions regarding courses of action. Digital twins are being used to plan, design and test systems prior to implementation to reduce the time from concept to production. When combined with technology, such as Augmented Reality (AR), they can form a powerful diagnostic and troubleshooting tool that is safe and allows technical staff to diagnose & understand issues before arriving on-site, thus ensuring that they carry appropriate tooling and safety equipment with them so that outage time is reduced & safety is increased.
Entering an age of intelligent machines
Artificial Intelligence, Machine Learning and AR will continue to be applied to new situations and other areas of the energy value chain. While I foresee an industry that will evolve with and embrace the technology of today & tomorrow, the industry has to adapt quickly to new technologies and innovations while tempering the implementations with common sense so as to observe ethical data practices as we enter the age of intelligent machines.