Solving Practical Problems in Smart Grids - Novel Smart Meter Design, Implementation and Algorithms
By Rohit Gupta
Wednesday, 13 Mar
EEG 301, GG Bldng.
Smart meters form the key infrastructure important for the growth of smart
grids that can enable the realization of tasks like demand response,
efficient power quality management and sharing of resources like battery
and solar PV. A large number of meters installed are designed with the
purpose of collecting energy consumption data only. These meters do not
support the rapid change in the smart grid paradigm. Smart meters if used
efficiently can provide a lot of insight into the home energy consumption
and help in improving the overall energy scenario.
This work demonstrates how smart meters can be used judiciously to
understand energy usage. This work is based on research and practical
experiments conducted in SEIL and IIT Bombay campus for the purpose of
efficient smart meter design and infrastructure utilization. The
contributions of this work include:
- Collection of accurate and fine grain voltage and current signatures for
energy disaggregation: A large number of disaggregation techniques use
high frequency data to identify the appliance state. These signatures are
not provided by any commercial or research smart meter. We have designed
and constructed a smart meter that can provide such high-frequency
- Provide the algorithms to correct the data collection errors: With large
scale installations it is highly probable that there will be data
collection errors. We have provided algorithms for correction of such
errors through the data attributes.
- Efficient data collection from a massive number of smart meters: Current
metering infrastructure is designed for billing purpose where the meter
reading is communicated only after the 15-minute interval. Such
low-resolution data does not provide much insight into the energy
disaggregation. Collecting data every second is not efficient and provides
huge data that requires high communication bandwidth, storage space and
processing capacity. The approach provided by us can reduce the data by
huge quantity without losing on the quality of the analysis.
- Optimally install smart sensors to reduce infrastructure cost: It is not
possible to identify the stats of appliance in large buildings and
offices. To get a complete insight, more than one smart meter installed at
the mains is required for this purpose. We demonstrate how a minimum
number of meters can be installed to get the state of the appliances in a
About the Speaker
Rohit Gupta is currently pursuing his Ph.D. degree from the department of
Computer Science and Engineering at IIT Bombay under the guidance of Prof.
Krithi Ramamritham. His research interests are in the areas of databases,
IoT, big data, distributed systems, and cloud computing. Prior to Ph.D.,
he worked for Oracle for 12 years.