Training-Free Non-Intrusive Load Monitoring of
Electric Vehicle Charging with Low Sampling Rate
Zhilin Zhang
∗
, Jae Hyun Son, Ying Li, Mark Trayer, Zhouyue Pi
Samsung Research America – Dallas
1301 E. Lookout Drive, Richardson, TX 75082, USA
∗
Email: [email protected]
Dong Yoon Hwang, Joong Ki Moon
Smart Home Solution Lab
Samsung Electronics Inc.
Suwon, Kyeong-gi-do, Korea
Abstract—Non-intrusive load monitoring (NILM) is an im-
portant topic in smart-grid and smart-home. Many energy
disaggregation algorithms have been proposed to detect various
individual appliances from one aggregated signal observation.
However, few works studied the energy disaggregation of plug-
in electric vehicle (EV) charging in the residential environment
since EVs charging at home has emerged only recently. Recent
studies showed that EV charging has a large impact on smart-
grid especially in summer. Therefore, EV charging monitoring has
become a more important and urgent missing piece in energy
disaggregation. In this paper, we present a novel method to
disaggregate EV charging signals from aggregated real power
signals. The proposed method can effectively mitigate interference
coming from air-conditioner (AC), enabling accurate EV charging
detection and energy estimation under the presence of AC power
signals. Besides, the proposed algorithm requires no training,
demands a light computational load, delivers high estimation
accuracy, and works well for data recorded at the low sampling
rate 1/60 Hz. When the algorithm is tested on real-world data
recorded from 11 houses over about a whole year (total 125
months worth of data), the averaged error in estimating energy
consumption of EV charging is 15.7 kwh/month (while the
true averaged energy consumption of EV charging is 208.5
kwh/month), and the averaged normalized mean square error
in disaggregating EV charging load signals is 0.19.
Keywords—Non-intrusive load monitoring (NILM); Electric
Vehicle (EV); Smart Grid; Energy Disaggregation
I. INTRODUCTION
Non-intrusive load monitoring (NILM) or non-intrusive
appliance load monitoring (NIALM) is an important solution
to realize smart-grid and smart-home energy management
benefits. It aims to estimate operation status and energy
consumption of individual electronic appliances by monitoring
aggregated current/voltage/power signals in the main circuit
panel of a house or a building [1]–[3].
Electric vehicle (EV) charging is becoming an important
load element for smart grid analysis [4]–[6] although home
charging EVs recently entered the market. Due to the growing
number of the EV customers, a utility might start to experience
non-marginal impacts on parts of its distribution system.
Particularly, the gravity of this impact will depend on at what
time, for how long, with what utility rate, and in what season
these EVs are being charged [7]. Therefore, it is necessary to
solicit the importance and urgency of monitoring EV charging
load via energy disaggregation.
Another usage of monitoring EV charging load is to
provide house owners the monthly energy consumption of EV
0 500 1000
0
1
2
3
4
Power (kw)
Time Index (min.)
(a) EV Power Signal
200 400 600 800 1000 1200 1400
0
0.5
1
1.5
2
2.5
3
3.5
4
Power(kw)
Time Index(min.)
(b) AC Power Signal
Fig. 1. (a) An EV power signal. (b) An AC power signal exhibiting two
kinds of waveform patterns, i.e. spike trains and lumps.
charging. This monthly feedback information can help house
owners in bill-management and travel-management in the same
way as monthly gas bill and conventional monthly electricity
bill [8].
There are many algorithms available for the energy disag-
gregation of various residential appliances [1]–[3], [9]–[14],
such as hidden Markov model(HMM) algorithms [10], [11],
[13]. However, these algorithms were not specifically designed
for EV charging, and they require extensive training and a large
computational load. Therefore, for practical implementation
where simultaneously monitoring tens of thousands houses is
required, those algorithms may not be an attractive solution.
In this paper, a novel algorithm for energy disaggregation
of EV charging is presented. It has several desired advantages.
(1) It can mitigate the interference coming from air-conditioner
(AC) power signals. Thus, it could be very helpful for smart
grid load analysis and management during peak load time in
summer. (2) It does not require training, which is an highly
attractive feature toward practical implementation. (3) It de-
mands a light computational load, thus suitable for monitoring
tens of thousands residual houses in large scale. (4) It works
well for data sampled at 1/60 Hz, which aligns with the data
provision capability of many smart-meters. Experiments based
on real-world power data showed that it exhibits far better
performance than state-of-the-art algorithms.
II. C
HALLENGES
One big challenge of disaggregating an EV charging load
from aggregated power signals is mitigating interference from
AC. As shown in Fig.1(a), an EV charging load signal can be
characterized as a square wave of a high amplitude (higher
than 3 kW) and a long duration (longer than 30 minutes but
generally shorter than 200 minutes) [5]. AC power signals