Gaining Traction

AI, IoT and ML solutions transforming the energy value chain

Power utilities are increasingly em­bracing new and emerging IT-based technological systems such as cloud computing, powerline communications, internet of things (IoT), digital substations, blockchain, remote mo­nitoring solutions, data solutions based on artificial intelligence and machine learning (AI/ML), etc. These technologies are revamping the way utilities operate and maintain their assets. Utility operators are now transitioning to decision-making based on insights from AI/ML systems that comprehend data aggregated by IoT devices across the whole utility.

IoT devices in generation, transmission and distribution

Embedding IoT devices in thermal po­wer plants can increase the efficiency of transporting and processing fuel (coal/ gas), as well as that of boilers, motors, turbines, flue gas desulphurisation systems, electrostatic precipitators, etc. For instance, in a conventional operational technology (OT) system with electronic communications, high and suboptimal temperatures in a boiler would be communicated to the system operator, who may ignore the signal or be late in recognising it, given that the information is not communicated in real time. In contrast, in a digitally connected OT system, a rise in boiler temperature would correspondingly trigger an alarm across the entire unit and solutions would be recommended to mitigate the issue in real time with limited delay. Furthermore, these systems would try to isolate the problem and possibly help in controlling the rise in temperature from affecting the rest of the plant.

Similarly, an IoT-based online monitoring system for transmission towers wo­uld be a more robust and reliable means of monitoring. These towers face damages due to several factors such as cyclones, thunderstorms, earthquakes, illegal constructions and theft, which might eventually lead to a tower collapsing. Tilt, vibration and weather sensors, and cameras installed on towers, can transfer accurate tower data for timely decision-making.

Substations at the intersection of transmission and distribution systems are pri­me candidates for digitalisation. Digital substations focus on digitalisation of both station-level and process-level operations by converting analog measurement data and binary status in­formation into digital data. Digitali­sa­tion provides a secure and reliable me­thod of data transmission, as well as significantly reduced investment and operating costs.

On the distribution side of the business, smart meters connected to the grid will help in enabling bidirectional energy fl­o­ws while reducing aggregate technical and commercial (AT&C) losses.

The abilities of smart meters include self-healing, providing adaptive power pricing, enabling real-time monitoring and remote control of systems. These fe­a­tures enhance energy efficiency, imp­ro­ve the reliability of the grid, increase its interoperability with other systems, and reduce outages by improving com­m­unications with it.

Power communication technologies

Wireless communication systems play a major role in enabling IoT, and hence, th­ere are a lot of considerations to be acc­ounted for before finalising a choice of communication technology. Wireless systems connect sensor devices to IoT gateways and perform end-to-end data communication between these IoT elements. Wireless systems are developed based on different wireless standards, and the use of one depends on the re­qu­i­rements of the application, be it communication ra­n­ge, bandwidth or power consumption. For example, re­ne­w­able sources of energy, including wind and solar power pla­n­ts, are mostly loca­ted in very remote are­as. Employing IoT systems on these sites requires the selecti­on of a suitable communication technology that can guarantee a continuous connection and support real-time data transfer in an energy-efficient manner.

Short distance power communication systems include the use of Wi-Fi for en­ergy metering and building energy management systems. However, due to the high power requirements of Wi-Fi, this technology is not the most ideal solution for the energy sector. Other communication technologies include low power wide area network (LPWAN) communication technologies such as narrowband IoT (NB-IoT) and Bluetooth low energy technologies. In April 2021, Tata Power Delhi Distribution Limited commissioned a smart meter system covering 230,000 households in Delhi. This system successfully deployed NB-IoT communications systems, and will be able to provide smooth flow of information in real time without any network interference or noise, as it uses a dedicated network to transmit data.

LPWAN solutions such as long range WAN, Sigfox and narrowband IoT are better suited for the energy sector, given that they can reliably send small packets of data continuously over long distances. These emerging LPWAN solutio­ns enable the establishment of reliable, low-cost, low-power, long-range, last-mile technologies for smart energy management.

Cloud computing and AI/ML

The improved efficiency of data management resulting from cloud computing allows manufacturers to create evolving digital profiles (digital twins) of physical objects or processes, which help in optimising business performan­ce by detecting physical issues in real time and by predicting outcomes more accurately.

On the demand side, integrating cloud-based technologies will also help utilities in predicting short-term demand. Short-term forecasts are used for operations, and are obtained based on historical loads and weather forecasts using different prediction methods, such as neural networks. Weather forecasts are also used to predict the transmission capacity of the grid. AI/ML-based forecasts of grid demand will increasingly ac­quire more importance given the rising penetration of variable renewable energy systems such as solar and wind power in the electricity mix. Studying the feasibility of integrating renewable en­ergy sources into the transmission network without destabilising grid freq­uency requires extensive high-fidelity simulations under various loads, generation profiles and weather conditions.

On the distribution side, increasing cloud computation capacity in combination with AI/ML will help discoms in predicting demand from consumers, in addition to aiding in the execution of time-of-day pricing systems. The discoms can also gamify the system and incentivise consumers to use electricity more efficiently by providing them with specific data pertaining to their consumption.

Other technologies

Robotics: States such as Rajasthan and Gujarat have a significant number of solar installations given that these regi­ons have a high solar irradiance. How­ever, these solar installations have to contend with high dust concentration, necessitating operators to employ maintenance workers to clean photovoltaic pa­­nels periodically so as to maximise per­­formance. In recent years, these op­erators have been deploying robotic cl­eaners that are enhanced with ML algorithms and AI techniques such as artificial neural networks and genetic algorithms. Auto­nomous robots empo­wered by AI will soon be more commonly dep­loyed across several areas of the power sector. In the coming years, generation companies with operations in harsh and inaccessible regions will prefer to employ robots for performing maintenance tasks.

Blockchain: Blockchain technology, with its emphasis on verifiable transpa­rency, will facilitate the adoption of green en­ergy trading, renewable energy certificates, carbon certificates, etc. The introduction of blockchain technology will help discoms or bulk buyers directly verify the source of the energy they purchase. Additionally, such verification will help in converting energy resources into digital assets, so that they can be traded on the blockchain. In 2021, Tata Power successfully tested blockchain-ba­sed solar energy trading among pro­su­mers, demonstrating a way for consu­mers with rooftop solar to tap into more revenue streams.

Issues and challenges

Reliability concerns: IoT devices need to be designed robustly so that they work in different environments. Power grids and networks often operate under harsh conditions such as high or low temperatures, high voltages, exposure to electromagnetic waves and exposure to water. Therefore, IoT devices must satisfy re­qu­irements such as reliability or compatibility under such conditions.

Communication technologies: A variety of communication technologies are available based on WiFi, Bluetooth, RF-net, cellular communications, etc. How­ever, there are still issues, as most of the­se technologies do not completely sa­t­is­fy the conditions of affordability, re­lia­bility, speed and interoperability.

Integration of IoT with subsystems: A big challenge involves the integration of IoT in energy sector subsystems, as th­ese subsystems employ various sensor and data communication technologies. Some solutions propose the designing of co-simulation models for energy systems to integrate IoT and minimise synchronisation delay errors among various subsystems.

Lack of financing: Most of these systems require substantial investments across several components in a coherent and composite manner in order to yield synergistic benefits. Additionally, the payback period for these technologies ranges from eight to ten years in many cases, therefore, making many fi­nancial institutions sceptical of exten­ding funding for them.

The way forward

Conventional energy systems are transitioning from a centralised and unified architecture to a more distributed and decentralised system, on account of increasing off-grid adoption of solar and wind. Additionally, high AT&C losses, as well as high operations and maintenance costs, reduce the profit margins of most power sector utilities and constrain them from expanding their operations. Therefore, integrating IT systems with OT systems will help them in shifting to the upcoming era of distributed generation, increased trading and bi-directional energy flows, while simultaneously helping them enhance their profitability.


Enter your email address