NILM – Focus on Non-Intrusive Monitoring
What is Non-Intrusive Monitoring?
The principle of non-intrusive monitoring, also known as NILM or NIALM (Non Intrusive (Appliance) Load Monitoring), involves measuring aggregated consumption data and using intelligent algorithms to identify electrical consumption by usage.
Studied in laboratories since the 1980s, these technologies have been commercialised by a few innovative companies globally since 2010. Among these companies, some (including Smart Impulse) have brought these technologies to maturity and industrialised their deployment. As a result, these solutions are now rapidly spreading due to their ease of implementation and the clarity of the data they provide for addressing energy management challenges.
Today there are two technological families for disaggregating the load curve: one is based on the analysis of events (switching on/off) on the load curve, while the other focuses on a detailed analysis of the shape of the electrical signal distorted by consuming equipment. These two technologies are presented in more detail in the paragraph: ‘The different NILM technologies‘.
What is NILM used for?
These technologies address many use cases, from identifying energy savings or consumption drifts to assisting with environmental certifications (HQE, BREEAM, LEED, etc.), as well as reporting, user awareness, or understanding the life of the building.
The Different NILM Technologies
The first approach is based on the detection of events (switching on/off) observed on the load curve at a frequency of about one second. This is referred to as low-frequency technology.
The second approach is based on an analysis of the shape of the electrical signal. It uses continuous measurement of the electrical current shape as input data to finely analyse the “distortions” caused by the consuming equipment on the network. This identification of distortions, or electrical signatures, is obtained through a very high-frequency measurement of the electrical signal. In this case, it is referred to as high-frequency technology.
- Low Frequency:
These technologies have the advantage of using simple input data that can be obtained with standard measurement components. However, they rely on event detection and are therefore limited when there are too many devices (and thus events) on an electrical network to separate their individual operations. Hence, these technologies are suitable for the residential sector, but not for large sites with numerous operating devices. Finally, since these technologies rely on variations to identify usages, they are not capable of identifying the nature of constant consumptions, which are found in the “base load” (night and weekend), where significant energy saving opportunities are often identified.
- High Frequency:
These technologies require measurement equipment capable of very precise acquisition of electrical signals. Since these technologies work based on electrical signatures by type of equipment (for example, different lighting or motor technologies) rather than by event related to an individual piece of equipment, there is no limitation on the size of the site being analysed, nor on characterising the “base load”.
And what about Smart Impulse in all this?
Smart Impulse is today a world leader in the state of the art of high frequency NIALM technologies, with 6 patents filed. Their solutions have been tested and validated by European energy service leaders by comparing data with sub-metering solutions. The results of this study show an average discrepancy per category of less than 5%. A summary of these results is available upon request from Smart Impulse.
To go further:
More articles are available on our blog:
Energy Performance: The Challenges of Implementing an Action Plan
How to save energy in commercial buildings in the UK?
Dassault Systèmes & Smart Impulse: From Energy Optimisation to ISO 50 001