Energy consumption and price reduction
We’ve been researching how to monitor individual appliances consumption in order to decrease peak and energy prices for end-user. That’s what we’ve found.
The first method to separate the appliance’s consumption is obvious and that is Intrusive Load Monitoring (ILM): use a metering device for each individual appliance that you need.
Opposite, Non-Intrusive Load Monitoring (NILM) is the energy disaggregation technique that provides a method to separate the individual consumption for certain appliances, respecting consumers’ privacy and often using already-deployed smart meters [1].
ILM is more precise than NILM but it is more expensive. NILM, on the other hand, is much cheaper, often it only needs smart meter. Let’s focus on NILM.
Main stages in NILM are [1]:
- Data collection: electrical data, including current, voltage, and power data, are obtained from smart meters, acquisition boards or by using specific hardware;
- Event detection: an event is any change in the state of an appliance over time. An event implies variations in power and current, which can be detected in the electrical data previously collected by means of thresholds;
- Feature extraction: appliances provide load signature information or features that can be used to distinguish one from another;
- Load identification: using the features previously identified, a classification procedure takes place to determine which appliances are operating at a specified time or period, and/or their states
So, the input of NILM is consumption characteristics, e.g. current, voltage, and the output is which appliances are turned on/off at a given time and/or their state.
There are appliances that have different load profiles at a different state, that’s what state in fourth stage means. It adds complexity to the algorithm. Another problem is that the house can have multiple appliances of the same type. Also, there are appliances with low consumption, for example, LED, which are hard to disaggregate, and appliances with consumption not varying in a periodic fashion.
Thus, NILM is not quite an easy task. Heuristic algorithms, such as Genetic Algorithms (GA) or Particle Swarm Algorithms (PSA), along with machine learning techniques are often used to solve that.
One of the important applications of NILM is the Home Energy Management System (HEMS). HEMS is responsible for scheduling appliances to reduce energy bills (Real-Time Pricing should be present) while saving a user’s comfort. It is also responsible for renewable energy system management if any. HEMS not only saves user’s money but also reduces consumption peak, which is good for energy producers because they can use fewer power generators.
Next, there are two examples of HEMS.
In [2] the following scheme was used:
Energy Management Controller (EMC) uses the Constrained Swarm Intelligence-based Consumer-Centric DSM module and database of historical records to disaggregate energy and schedule appliances. Also, HEMS is taking into consideration alternative energy (solar power on the image).
A phenomenal reduction in peak power consumption is achieved by 13.97% in that scheme.
In [3] the following scheme was used:
Data Acquisition Device is present. It is used to get consumption data. ZigBee-based Plug-load Control Relays are used to switch appliances on and off. Also, there is a computer that runs NILM algorithm (upper) and Home Gateway (lower) which has Database and communicates with ZigBee-based Plug-load Control Relays. Homeowners can get access to data using an internet browser.
NILM in that scheme is pretty complex:
.
It has three stages: off-line load learning & modeling, on-line load monitoring, and day-ahead in-home load scheduling. Such techniques as Hard c-Means clustering, sequential forward selection (SFS), hard c-means-based k-nearest neighbor classifier (k-NNC) are used. (See [3] for details).
Also, we’ve found a company called Smappee that offers appliances submeter, which can be integrated with HEMS: https://www.smappee.com/uk/homepage
That company has three ways to record appliances consumption (https://www.smappee.com/uk/blog/smappee-complete-submetering/):
The first and second ways use ILM, while the third way, which is cheaper, uses NILM. In a third way you should install a clamp to your power cable. The company claims that its patented NILM has a 70% average accuracy.
We came to the conclusion that for our project we could use NILM. NILM is pretty complex but real task. It’s already used by companies, such as Smappee, but may be not very accurate. We are not experts in field, so we can try ready solutions. For that purpose NILM toolkit [4] would be handy, which gives opportunity to evaluate different NILM algorithms on open datasets, such as REDD, BLUED, PLAID, REFIT, TRACEBASE, WHITED, UK-DALE, DRED [1]. Core idea of our project is similar to HEMS, maybe we should try to schedule appliances too.
Students:
Iolanda De Almeida Oliveira, Pavel Ivliev, Michael Kämpf, Igor Kirianov
Supervising Professor:
Zarina Charlesworth
References
- A Ruano, A Hernandez, J Ureña, M Ruano, J Garcia NILM Techniques for Intelligent Home Energy Management and Ambient Assisted Living: A Review. Energies 12, 2203 (2019)
- YH Lin, YC Hu Residential consumer-centric demand-side management based on energy disaggregation-piloting constrained swarm intelligence: Towards edge computing. Sensors 2018, 18, 1365.
- YH Lin, MS Tsai An Advanced Home Energy Management System Facilitated by Nonintrusive Load Monitoring With Automated Multiobjective Power Scheduling. IEEE Trans. Smart Grid 2015, 6, 1839-1851.
- Nipun Batra, Jack Kelly, Oliver Parson, Haimonti Dutta, William Knottenbelt, Alex Rogers, Amarjeet Singh, Mani Srivastava. NILMTK: An Open Source Toolkit for Non-intrusive Load Monitoring. In: 5th International Conference on Future Energy Systems (ACM e-Energy), Cambridge, UK. 2014. DOI:10.1145/2602044.2602051 arXiv:1404.3878