07 Nov DEXMA Data Scientists At NILM Energy Disaggregation Workshop
This week, the DEXMA data science team is participating in the 4th European Workshop on Non-Intrusive Load Monitoring (NILM) in London. The aim of this conference series is to serve as a meeting point for all European researchers that are working on the topic of energy disaggregation in both industry and academia to network, share research and exchange ideas.
What is NILM?
NILM stands for Non-Intrusive Load Monitoring, process for analysing changes in the voltage and current going into a building and deducing what appliances are used in the building as well as their individual energy consumption.
Also known as energy disaggregation, the NILM process aims to “unpack” or “sort” your total energy consumption into the energy readings of individual loads over a period of time, allowing end users to see how much power a particular appliance has used during that time.
How is energy disaggregation applied in the real world?
Energy disaggregation, or getting insights from individual appliances, can be very useful for different energy-related applications, such as energy monitoring or demand response. “Generally it’s quite easy to gather a huge amount of energy consumption data,” says DEXMA Data Scientist Carlos Juan Troyano. The problem is that most of that data is what data scientists like Juan call “dirty”: unlabelled or incomplete. “So the idea behind NILM is trying to “clean” this data in order to make it useful for energy-related applications with as little human intervention as possible,” says Juan.
NILM is widely recognised by the energy management industry a highly effective low-cost alternative to the traditional energy audit that requires attaching individual monitors on each load in order to identify energy savings opportunities.
Does DEXMA technology use NILM?
You bet! Our platform for frictionless energy savings, EnergyGrader, uses NILM technology to figure out which buildings in a large portfolio are eating up your energy spend and why.
EnergyGrader’s NILM system uses advanced machine learning algorithms to break load curves from fiscal metering down into individual consumptions for each recognised load type. This gives a load curve for each recognised appliance type or group of appliances, “as if” a meter had been used to monitor them.
EnergyGrader then benchmarks this load curve information against a database of more than 50,000 similar sites within the DEXMA database to approximate the energy consumption profile of the building under analysis.
To generate personalised energy savings recommendations, EnergyGrader uses a self-learning decision tree to define the best energy-saving options for each site. Each recommendation can be applied according to maximum possible investment, overall project budget, site activities, technology or location. Final recommendations can then be organised by payback period, or total expected annual savings.
The best part? No need for an on-site audit interruption or to stop production to carry out an in-depth study that can take up to an entire year to complete. EnergyGrader can identify how energy is being consumed and wasted using the magical combination of proactive analytics and NILM technology.
Curious to see some NILM magic for yourself?