On the scorching afternoon of 29th July 2012, circuit breakers on the 400 kV Agra-Gwalior-Bina line flung open under the intense load. With 1000 MW power being drawn from this single line whose maximum capacity was near 700 MW, the lines were snapped open, instantly breaking the circuit.
The load then burdened the Agra-Bareilly line, which ran in parallel. It also snapped. With each line getting disengaged, the power deficit snowballed out of control and toppled every line connected in parallel. Within minutes the entire Northern Region Grid, which innervates half of India, collapsed under this cascading catastrophe. 22 states from Assam to Rajasthan and Odisha to Kashmir were plunged into darkness for 2 days. This was the world’s largest wide area blackout, with more than 62 crore people’s lives rattled for days as the engineers scrimmaged to get the grid back on its feet. The disaster issued a poignant alarm over the dilapidated state of the Indian electric grid as it is today. With common citizens shrugging of the all-too-familiar outages as everyday occurrences, only the energy pundits know exactly about the smothering problems that our country’s power infrastructure faces.
India is attempting to do something no nation has ever done: build a modern industrialized economy, and bring light and power to its entire population, without dramatically increasing carbon emissions. Simply to keep up with rising demand for electricity, we must add around 15 gigawatts each year over the next 30 years. We get most of our electricity from aging, dirty coal-fired plants. Decrepit transmission grids running on obsolete hand-me-down technology from the West threaten the reliability of the entire energy infrastructure. At least 300 million of India’s 1.25 billion people live without electricity. The current Prime Minister’s 2014 manifesto has promised to increase India’s renewable-energy capacity to 175 gigawatts, including 100 gigawatts of solar, by 2022. In the very near future Indian engineers would have to orchestrate a pandemonium of micro-power generators spread across household rooftops and windy country sides along with Thorium and coal powered generators. SITARA is a step in that direction.
SITARA is an international collaboration to work out the gritty details of the architecture of the grid of the future. The 2-year project named SITARA: Smart Grid to Harness Satellite based Virtual Power Plants for Energy Sustainability, is one of 14 multilateral university partnerships which have been awarded grants by the Global Innovation Initiative in 2015: a joint effort of the UK and US to foster multilateral research collaboration with higher education institutions in Brazil, China, India and Indonesia. SITARA, with a grant of nearly £150,000, is one of eight grants awarded to UK-led partnerships by the British Council and the Department for Business, Innovation and Skills of the Government of UK.
The SITARA consortium includes the University of Bradford, North Carolina State University (US) and Indian Institute of Technology Madras (India). The project consortium will also collaborate with University of Hong Kong, which has a very strong smart grid research infrastructure, to validate some of the developed algorithms. Also the Media Lab Asia and the Centre of Development of Advanced Computing, both affiliated to the Ministry of Communications and Information Technology, Govt. of India would be involved in impact evaluation and dissemination activities.
Why do we need a smarter grid?
Enabling two way communication between the energy producer and consumer is the essence of the whole smart grid idea. This would currently allow for efficient usage of our limited energy resources. But the smart grid paradigm is more than teaching the old grid new tricks. It will become the enabling technology for large scale distributed grids to even exist in the future.
Power quality in distributed grids is a major concern. In a network where power influx can capriciously vary at a slightly strong gust or a cloudy dark sky, maintaining your wall socket at 220V 50Hz is going to be quite a challenge. Unlike a coal fired power plant where we can just yank up the flames at the peak demand, we need smarter solutions for a grid which depends mostly on renewables. Among other concerns, surges or transients are also a major one. These are brief overvoltage spikes or disturbances on a power waveform that can degrade or destroy electronic equipments. Transients can reach amplitudes of tens of thousands of volts even though they only last for microseconds.
Also in the future when anybody on the grid can supply electricity back to the it, new problems would arise. Like a serious issue, which is known as “islanding” in electrical engineering parlance. It’s a condition in which distributed generators like solar panels or wind turbines continue to generate power and feed the grid, even though the electrical power from the utility is turned off. Utility workers who may think that there is no power when the utility power is shut down, but the grid may still be powered by to the distributed generators.
In addition to solving several such problems, a smart grids enables dynamic pricing of power even within a single day. This can greatly ease the loading on the grid. When the peak hours are charged more than the lean ones who wouldn’t schedule using their heavy appliances away from the peak hour?
The Virtual Power Plant
At the core of SITARA’s approach at this multidimensional problem, is the concept of a Virtual Power Plant. The aggregation of hundreds of homes with solar power and their battery storage will provide the utility with a cost-effective and innovative “virtual power plant”. So instead of installing a new centralized power plant, these hundreds of microgenerators can be dynamically orchestrated to behave as a single effective power plant. This can supplement the traditional energy delivery model thus improving the grid resilience, reliability and sustainability.
VPPs are supported by cloud-based software, monitoring and automation equipment and an optimisation engine. VPPs are able to extract flexibility out of small scale generators and customers, allowing them to respond to varying amounts of renewable supply on the system whilst compensating them for doing so.
IIT Madras and SITARA
SITARA is split in 5 work packages each focusing on ways to address separate aspects of the project. IIT Madras is responsible for the 2nd work package which focuses on developing predictive models for consumer behaviour. Prof Swarup and his team, in the Electrical Engineering department, have analyzed data sets in Virtual Power Plants consisting of renewable energy and storage to predict electrical behavior and cost behavior. The group has identified key variables which significantly factor in when considering the load behaviour and has developed a robust prediction algorithm based on artificial neural networks. Artificial neural networks roughly mimics the brain while it tries to ascertain the mathematical relationship between the final observed variable and the parameters. The parameters included were:
- Calendar variables: Like the day of the week, whether it is a weekday or holiday and hour of the day which influences energy forecasting significantly.
- Weather variables: Climatic conditions considered include temperature, humidity, wind chill index, illumination, rainfall, precipitation, cloud cover and some special events like typhoon or sleet occurrences.
- Holiday effects: Local events, including holidays and festivities, also affect the energy demand.
These events may lead to either an increase or decrease in demand. Influences of these events are usually local. Random events and disturbances such as abnormal consumption behavior, idiosyncratic and social habits of the individuals account for the randomness in the final data. This energy forecasting model was used to predict the consumption pattern in Sydney in the year 2010. The past hourly load pattern and temperatures data of Sydney was obtained for the years 2006 to 2009. This was used to “train” the neural network, that is to rigorously calculate the relationship between the energy load and different parameters.
The model was able to reliably forecast onto the test set, and thus could be used as the prediction engine in SITARA. Based on this analysis, Prof Swarup and his team have also proposed dynamic pricing schemes, i.e. pricing based on which cluster the consumer falls into, is it the heavy usage end or the nominal usage end, and which part of the day are they in.
Much integration of various work modules needs to be done. In the larger scheme of SITARA, IIT Madras’s contributions stand as an integral component. But many knobs and gears need to be turned before SITARA can truly mark an impact on the current grid scenario. In part, SITARA also aims to train Indian electrical engineers in aspects of the smart grid technology. Student exchange programs between collaborating universities have seen productive exchange of ideas and skills across international borders. The project has a major part of work coming ahead and we hope that we come one step closer to a better and smarter grid when it finds completion.
Meet the Prof
Prof K. Shanti Swarup received his M.Tech degree in Power Systems and Control Engineering from REC (NIT) Warangal. He obtained his PhD in Electrical Engineering from IISc Bangalore in the field of Power Systems and Artificial Intelligence. He has worked in Advanced R&D, Mitsubishi Electric Corporation, Japan as a consultant and software developer for intelligent power management systems. He has wide industrial and academic experience in India and abroad. His primary research interests at IIT Madras are in the areas of Power System Modelling, Operation and Control, and Distributed Artificial Intelligence.
Meet the Author
G Soorya is a third year chemical engineering student. He is an avid reader of popular science and is passionately curious about “networks, links and connections”. He is particularly interested in systems biology, circuit-design, and research-spun off entrepreneurship.