- Sep 29, 2020 5:40 pm GMT
Big data analytics is big in more ways than one. Big data – information that is voluminous and diverse – and its analysis play central roles in TV shows, movies and books.
But big data analytics is also important in the real world. The technology powers drug discovery and energy production. Now a team from Idaho National Laboratory (INL) has looked at how big data analytics can be used to make smart electric utility grids even smarter, publishing their research findings in an IET Smart Grid paper.
“We wanted to explore what are the challenges and where are those challenges,” said Bishnu Bhattarai, the paper’s lead author, explaining the motivation for the research. “We wanted to say, ‘Here are some of the opportunities for utilities and system operators.’”
Those opportunities, the researchers concluded, include better energy management, as well as greater grid observability and reliability. Big data analytics also could improve security and enhance protection against both physical and cyberattacks.
INL authors on the paper include Yusheng Luo, Manish Mohanpurkar, Rob Hovsapian (now at the National Renewable Energy Laboratory) and Kurt Myers.
The paper’s list of co-authors includes experts from utilities and system operators as well as university and industry researchers. Input from specialists at a wide array of institutions was deliberate, ensuring that the paper included different perspectives, according to Bhattarai, now a senior research scientist at Pacific Northwest National Laboratory.
What is big data?
Big data consists of extremely large and diverse data sets, such as all biochemical compounds that might have therapeutic value in treating a disease. The promise is that computers and specialized software can analyze the data and reveal patterns and associations not readily visible. For drug discovery, this can mean uncovering a promising candidate.
The team of investigators looking at big data analytics for smart grids found that the data was big in some unexpected ways. One was the volume of information generated, with a typical distribution utility dealing with thousands of terabytes (1,000 gigabytes) of new data every year. Further research revealed other aspects of the data’s bulk.
“In the beginning, I was mostly thinking about data volume,” Bhattarai said. “I wasn’t thinking about the other dimensions of the big data. There are actually five dimensions.”
How can data have dimensions?
In addition to volume, the researchers identified four other characteristics: variety, velocity, value and veracity.
Variety, as the name implies, refers to the type of data, which can be from instruments or social media postings.
Velocity describes how fast the data arrives or is sampled, with some of it appearing continuously and other information only showing up every hour or day.
As for value, that ties into economics, with some data, for instance, related to revenue and other information not so clearly tied to money.
Finally, not all data is of the same accuracy and reliability. One example of data with possible veracity issues would be the weather forecast from a guy on the street versus a trained meteorologist. One may be more trustworthy than the other.
What are challenges to leveraging big data?
For such massive, motley data to be useful, the team identified several areas needing improvement. One area concerns standards, such as those that govern what form data must be in before it is shared, stored and analyzed.
Another key need is for interoperability, which would allow systems from different vendors to easily talk to one another. An easy and free-flowing exchange of information is required to leverage some existing big data analytics technologies.
A third hurdle lies in data privacy, which requires a regulatory framework for data sharing and storage. For example, traffic data can help predict utility demand by forecasting when people will get home. But that same information could provide insights into and information about individuals. Dealing with the data privacy issue will be necessary not only for smart grids but also for smart cities, Bhattarai said.
Can those puzzles be solved?
Having identified the big data challenges, the researchers then looked at possible solutions. They found that the tools and techniques developed by large social media, search and software corporations could be applicable to some of the challenges facing smart grids.
Yet a final solution will take some adapting and extensions of these tools and techniques, a process that is just getting underway at utilities and grid operators. Thus, while the initial steps have been taken, much more needs to be done, according to Bhattarai.
The IET Smart Grid paper has been well-received. Within its first year of being available online, ResearchGate ranked it among the top 1% of the site’s 2019 manuscripts and in the top 2% of all publications on the site. ResearchGate is a social networking site for scientists.
Yet Bhattarai notes that the publication is by no means a complete or final solution. Eventually, what it describes could make smart grids more intelligent and useful. For now, though, the paper appears to be doing what the co-authors hoped it would.
As Bhattarai said, “It’s intended to open up doors for new research.”