CenterPoint Energy’s Forecasting Model Engine using HANA
CenterPoint Energy (CNP) serves energy needs of over five million residential and business customers throughout the
eastern half of the U.S of which 2.2 million have digital smart meters.
These digital smart meters feature two-way communications able to send and receive information to and from our consumers. CNP can read these meters remotely. The meters also measure usage in 15-minute intervals, which can give consumers valuable information about their consumption habits.
In the past, Load Studies were based on stratified random samples of residential, small commercial and all industrial customers.
Load Studies incorporating SMART METER data will significantly improve upon traditional Load Studies and results in better estimates of weather impacts, timely Load Studies and an improved understanding of changes in usage.
To support the load study process, Forecasting Model Engine platform was developed using SAP HANA technology. CNP in partnership with SAP data science team developed innovative applications, such as:
- GATE KEEPER WIZARD to verify the quality of interval and weather data
- Weather Response Function (WRF) WIZARD to validate estimates of weather impacts predicted by WRF’s against observed changes in hourly loads and
- Change Analysis dashboard to help understand period-over-period trends for various KPIs
Using HANA CNP is now able to correlate big data sets like 15 minutes interval data from smart meters ( ~ 5 billion records) with weather (1 year history), customer information (~2.3 million customers), and geographical location information and perform advanced analytics and predictive models more efficiently and accurately.
The data science team are building advanced analytics and predictive algorithms to achieve a better estimation accuracy the forecast functions has to be applied on group of customers with similar load profiles.
In future, CNP would be using SAP HANA for even more Smart Meter data innovations and insights.