The rapid rise of renewable energy is exposing the shortcomings in India’s current power transmission infrastructure.
As they increasingly produce and distribute this new kind of power, government agencies and companies—and their infrastructure—are grappling with the technical difficulties this transition ensues. These include frequent switching between conventional and renewable energy and forecasting demand and supply.
So a clutch of cleantech startups is tackling these problems using artificial intelligence, machine learning, and analytics. They’re either providing these services themselves or creating and selling software that firms and utilities can customise.
The share of renewable energy in India’s overall power capacity has shot up from around 13% in 2015 to around 20% this year. In fact, last year, the country added more renewable capacity than conventional power.
While this works well for India’s broader climate goals and financials—renewable energy is now cheaper than thermal power—this is a challenge for the country’s power-transmission infrastructure for multiple reasons.
For one, wind and solar power generation can often be erratic and limited in supply, given that they depend entirely on the weather and solar cycle. So transmission firms must find a balance between these and thermal power to ensure uninterrupted supply.
“Demand forecast, as well as procurement forecast, is very critical for the power sector both from a financial point of view and from a technical standpoint,” Umakant Panwar, a former power department secretary in the state of Uttarakhand, told Quartz. Utilities are penalised if they underprocure electricity and don’t meet the demand; if they overprocure, they lose money.
“With the ever-increasing mix of renewable energy into the grid, it is becoming less and less predictable,” Panwar said
Earlier, utilities would use simple computing techniques or do it manually. Now they’re adopting technology as it enables the utility to anticipate the power supply load for the day, he added.
“Even though rules on scheduling and forecasting got initiated in 2010, the real pace has started coming in the last couple of years,” said Vishal Pandya, director of REConnect Energy, a company that provides such forecasting and scheduling services to power producers and utilities. The company uses technology to help power producers forecast wind-power generation, and works with different state power utilities to schedule their power flows.
Utilities in Gujarat, Karnataka, West Bengal, and Uttar Pradesh are all working on such pilot projects with startups. “On the government utility side, there is much more seriousness and receptiveness to try out new things,” said Nitin Tanwar, CEO of Climate Connect Technologies, an Amsterdam-based startup that makes machine learning-based software that power producers and utilities can use to study their data. While some outsource the forecasting work to startups such as REConnect, power producers, and utilities are now beginning to subscribe to software from companies like Climate Connect.
Electricity producers, too, are keen since they’re now required to provide generation data to power utilities, failing which they could be penalised as per regulations being drawn up by several state authorities. States like Telangana, Andhra Pradesh, Rajasthan, Madhya Pradesh, and Karnataka already have such rules. Others such as Gujarat, Maharashtra, and Tamil Nadu are working on policies around this.
In any case, these cleantech startups are not just about forecasting.
Climate Connect, for instance, is set to launch a product that even helps in procurement. “The algorithm will tell them which power plant they should buy the power from because it is the cheapest and which area they should supply power to because they have the most revenue recovery from there. These decisions, which are now made by humans, may not be on a daily basis and may not be as efficient,” CEO Tanwar said.
Mumbai-based Quenext is looking to aggregate data on power plants and put it out for public use. Today, forecasting is done for mostly big plants. Quenext intends to it for even the smaller ones.
“We leverage satellites to identify solar panels, geotag it, overlay weather data on it, and start generating forecasts for every kilowatt. We are mapping each and every wind farm and each and every solar panel,” Awadhesh Kumar, CEO of Quenext, said.