Demand for residential energy insights and predictions based on the analysis of data coming out of smart meters is on the rise as utilities are catching up on their digital transformations and coping with the ongoing energy crisis. The analysis of energy consumption data is called energy disaggregation. While most energy disaggregation solutions available in the market today are aiming to help utilities achieve similar goals such as increased customer engagement, customer retention, and lowering churn rates, we see that accuracy, scalability, global reach, and data science make all the difference when it comes to picking the right solution over competing ones.
In this blog piece, we elaborate on 5 key reasons why NET2GRID differentiates from the competition.
A one-stop-shop for energy disaggregation software and hardware
Acing energy disaggregation accuracy on smart meter data is of paramount importance. When this is combined with in-house developed hardware which also unlocks the real-time data from smart meters worldwide, this makes for a winning combination.
Hardware expertise was the decisive factor for E.ON Germany to choose NET2GRID’s solution over other competing ones. NET2GRID was the only provider who could overcome the technical specificities in accessing digital meter data in the country. Multi-family homes and digital meters in the basement outside of the wifi range make the German conditions the most challenging in Europe for data acquisition. NET2GRID overcame the difficulties with its Smartbridge; we managed to unlock the real-time data from both the German ‘mME’ digital meter and 'iMSys' smart meters for consumers who owned one, and we achieved a 90% disaggregation accuracy and an 80% appliance events recognition coverage of residential energy consumption. The E.ON optical meter reader we designed is an affordable, white-label, country-agnostic, self-install hardware that captures both real-time and historical meter data, and it unlocks the meter PIN code with an IR sensor via an app or API.
Simultaneously, EDP chose to invest in NET2GRID because it is a one-stop shop for both energy disaggregation software and hardware. As Antonio Coutinho, CEO of EDP Ventures, has put it, ‘’NET2GRID’s hardware is competitive in pricing, which is relevant because if you lower the transaction costs, you do more transactions’’. At the same time, ‘’NET2GRID’s Non-Intrusive Load Monitoring is proven to be the most accurate compared to other competing solutions.’’
Watch as Antonio Coutinho elaborates on how does NET2GRID differentiates from the competition:
Investing in a solution provider with deep knowledge in offering full-stack IoT services, including software and hardware, is extremely important. Hardware is hard to develop; finding a vendor who has mastered both hardware and software development makes the entire deployment work seamless, cost-effective, and risk-free. Utilities can build a single customer journey for smart meter data and real-time data while migrating their traditional, commodity business to the new Energy-as-a-Service model.
Furthermore, busy innovation and data science departments within utilities don’t need to spend time and energy piecing together a solution from multiple providers. Instead, they can collaborate with one who orchestrates the whole integration process and who can ensure a smooth customer experience for both software and hardware.
Global reach
As every country is different, capturing meter data is not a simple task. Having a business partner who develops technology compatible with an extensive range of technical standards is extremely important. NET2GRID has developed various hardware products for most major markets which solve challenges such as Home Wi-Fi range or power availability close to the smart meter. NET2GRID's competitive hardware products can connect to a wide variety of meter models and numerous (wireless or wired) protocols like Wi-Fi, BLE, Zigbee and LoRa (point-to-point).
Our meter reader portfolio consists of a plethora of hardware models which fit specific meter hardware market conditions across the globe. Additionally, our algorithms have proven to be reliable and able to detect appliances across continents as our pilots from the US to Australia show.
Process smart meter data of all granularities
One main concern from utilities is that the energy insight services might not deliver accurate results due to the low granularity of available energy measurements in their country. When in-house data science teams develop their solutions, they usually focus their efforts on processing data coming only from the smart meters. The main reason is the lack of access to more granular smart meter data, like real-time, or sub-second data.
NET2GRID offers three types of AI-fueled energy services to support three use cases that have a radically different granularity of available consumption data. First, it applies general machine learning models that recognize activity categories in household energy consumption data, derived from smart meters, without additional hardware installed. Next, for real-time consumption measurements, a set of machine learning (ML) models are created from NET2GRID’s rich portfolio of ground truth data sources. These ml models are applied to each household, to create individual custom-made appliance models able to disaggregate energy consumption of individual appliances, achieving over 90 percent accuracy. The third algorithm uses pre-trained edge ML models to identify appliance events with 98 percent accuracy within 15 seconds. These models process 100 Hz data and run on the latest-generation smart meters, without cloud connectivity, to save cost, time, and energy.
The most accurate energy disaggregation in the market
NET2GRID’s energy disaggregation algorithms are trained based on real-time information and the largest labeled appliance data set in the world. As a result, we can disaggregate real-time data with over 90% accuracy.
That's a big differentiating factor from the competition, which mostly work with models trained based only on smart meter data. In that context, having a 15-minute data set of one million meters, or a 15-minute data set of five million meters doesn’t make the algorithms smarter, because the amount of detailed information contained in this data set is limited. What NET2GRID does is that it has acquired data and trained models based on one-second data which means a thousand times more data points and, thus, a thousand times more information. The output of those models is benchmarked against ground truth data numbering more than fifty thousand appliance events that NET2GRID has collected over the years from around the globe. The outcome of the one-second disaggregation is used as an input for the 15-minute data disaggregation models, boosting the training process to create much deeper models than the competition.
Proven technology resulting from 10 years of solid experience in the field
NET2GRID’s strong team is comprised of talented people with complementary skills coming from various backgrounds, including energy retail and tech. The company has devoted 10 years of solid effort in developing solutions that address challenging market needs and bring solid results. Satisfied customers include E.ON, ENI, EDF, and EDP while the company is backed by Shell, APG and EDP Ventures. NET2GRID has created a rich ecosystem of partnerships including AWS, SAP, Greenbird, Intellihub in Australia and ITRON in the USA.
Comments