Вы находитесь на странице: 1из 4




  
  

    
 

Electricity Bill Forecasting Application by


Home Energy Monitoring System
Charnon Chupong Boonyang Plangklang
Dept. of Electrical Engineering Dept. of Electrical Engineering
Faculty of Engineering, RMUTT Faculty of Engineering, RMUTT
Pathumthani, Thailand Pathumthani, Thailand
charnon.c@en.rmutt.ac.th boonyang.p@en.rmutt.ac.th

Abstract— Home energy monitoring system has importance management by experimentation [6] not just display the data.
rule in home energy management. Many reports show that it has Kathryn B., Riccardo R., and Ben A. suggest that, the home’s
effectiveness for reducing energy consumption in home. But in energy monitoring system should be tailor designed to the
medium term and long term of using home energy monitoring specified requirements of individual users and should be reduce
system there are some report show the rapidly dismiss of energy the user’s effort to dealing with system, perhaps by notify
saving effective because of user do not pay attention anymore. consumers to abnormal usage or specifying which appliances
For improve the traditional home energy monitoring system are on or providing some recommendation.[3] Ambient Energy
there are three concepts should be applied to the systems, 1) the Orb is a frosted glass ball that provides real-time data about
system must be a learning tool not just a monitoring tool, 2) the
energy consumption that apply eco-feedback concept by
system should be tailored made for individual users and 3) users
should use less effort to dealing with the system. This article
changing their color according to electricity demand and
applied these concepts to home energy monitoring system and pricing allow uses to know their energy consumption situation
create an application to forecasting the user’s electricity bill. The without any effort [7]. Somchai B. and Boonyang P. present
system has applications programming interface (API) that allow the non-intrusive appliances load monitoring (NILM) that
users to create applications upon their requirements. From API allow users to know the consumption of significant appliances
we have create an application to forecasting the user’s electricity without install metering device directly to it. [8]
bill that report to user via email daily, user have less effort to
In this article we present the Energy Monitoring System
receive and translate the information. And from that daily report
user can learn of how their behaviors or their measures effect the that connected to the internet via Wi-Fi and we provide
electricity cost. The accuracy of electricity bill forecasting Application Programing Interface (API) allow any user to
application was tested by comparing the forecast cost and actual create their own user interface, dashboard or application that
cost and found 96% of accuracy, the result is highly acceptable. specified to their individual requirement. And we use this API
to create Electricity Bill Forecasting Application that can
Keywords—home energy monitoring; electricity bill forecasting the electricity bill and update forecasting value to
forecasting; smart meter users via email every day. This application allows users to
deal with energy consumption information without any effort.
I. INTRODUCTION Users can know and learn about effect of their behavior to
In Thailand over last 5 years, residential sector has the their electricity bill with forecasting data that provided to them
highest growth rate of electrical energy consumption [1] even daily. The rest of this article is organized in 3 section, 2nd
higher than industrial sector. The high growth rate of electrical section is present detail about proposed system the 3rd section
energy consumption in residential sector may cause from the is experiment and result and 4th section is conclusion.
growth of real-estate business and inappropriate energy
management measures in residential sector.
From our review, home’s energy monitoring system has
importance rule in home energy management system. There are
some reports show the effective of home’s energy monitoring
can reduce energy consumption about 5-15% [2], [3]. But in
medium term and long term the effect of home’s energy
monitoring will diminish rapidly from weeks to four months
[4], [5] because users do not pay any attention anymore.
Researcher have acknowledged this problem and study, how to
improve the effect of home’s energy monitoring to be sustained
over time. Froehlich, J., Findlater, L. and Landay, J. suggest Fig.1 Ambient Energy Orb
that the home’s energy monitoring must be designed as a
learning tool which allows user to learning about energy
978-1-5090-4666-9/17/$31.00 ©2017 IEEE
 
  
  

    
 

II. PROPOSED SYSTEM B. Application Programing Interface (API)


Our Application Programing Interface (API) is a service
A. Overall System that allow users to get data from our database to use in their
Internet-Enabled Energy Monitoring System consist of applications as they require without direct access to the
some sub-system as Fig.2 database. For more flexibility to use in any platform we use
the concept of REST HTTP for our API, user just use GET
method with some query string to appropriate URI to receive
their required data. The data will return in array of object
format. Here is our API list and example of usage in Fig.4.

• now, get the most recent data from database.


• trend, get n most recent data that user can specified
the value of n.
• range, get data from date-time1 to date-time2 that
user can specify the value of date-time1 and date-time2.
• today get all today data
• 24hr get all data within last 24 hour
• cost/tariff112 calculate electricity cost with tariff
1.1.2 Home, Normal rate
• cost/tariff212 calculate electricity cost with tariff
Fig.2 System’s Block Diagram 2.1.2 Small Business, Normal Rate
• cost/fixed calculate electricity cost with some fixed
• Hardware – We use Multi-function energy meter with tariff that user can specified
Modbus RTU communication port connected to ESP8266 and
some ICs to convert Modus RTU protocol to Wi-Fi signal and
send data to server with sampling rate 1 minute.
• Server – We use cloud service to install database and
write service software that access and write database with data
send by hardware all service software written by Node JS
language. And we write Application Programing Interface
(API) service by Node JS language for allow user to get data
from database and create their own application.
• Database – We use MongoDB a non-SQL type
database because it’s fast and more flexible than traditional
SQL type database. Data store in database is an object one
object represent on sampling data with key, value pair.
• Client device – Any device with internet browser can
access data via the internet.

Fig.4 Example of output from calling API

C. Electricity Bill Forecasting Application


By API discuss above we create application that can
Fig.3 System’s Hardware install at home
forecast user electricity bill. The concept of this application is
using average daily energy consumption in last 7 days to
predict energy consumption until billing date and also
 
  
  

    
 
calculate electricity cost at due date as flow chart in Fig.5 and
then send forecasting value to user via email and popular
messaging application like LINE Application. This forecasting
process run daily repeat.

Fig.6 Email message and LINE Notification that send to user

III. EXPERIMENTS AND RESULTS

A. Data avalibility
We have tested the communication ability of system by
install it at home, let it measure electricity parameter and send
data via Wi-Fi to the internet every 1 minute. We have tested
Fig.5 Application’s flow chart the system for 40 days, if the system work perfectly it should
have 57,600 records in database during this test. After 40 days
we have checked data in database and found only 50,124
records available that mean 87% of data available or data loss
13%. The clause of this data loss is come from Wi-Fi Signal
lost due to electricity outage and hardware malfunction.
 
  
  

    
 
B. Accuracy of Forecasting designed (or have potential to tailored designed) to specified
We have tested the Electricity Bill Forecasting individual requirement of users, third the system should let
Application as discussed in section 2 by compare daily users to dealing with no or less effort. This article presents the
forecast electricity cost each 10 days with actual electricity home energy monitoring system and its application that
cost in electricity bill. The result show average accuracy of applied these three concepts. The proposed system has API
forecasting is 96%, maximum accuracy is 98% and minimum service that allow users to create any application upon their
accuracy is 94% as shown in Fig.7 and Fig.8. The clause of requirement as second concept. From proposed system and its
forecasting error is come from due date in application have set API, we create application that can forecast the electricity bill
to date 21st 0:00 AM of each month but actually due date is of user with accuracy 96% and send the result to them via
upon utility’s staff schedule some month is date 20th and some email every day that allow the use to have less effort to
month is date 21th. receive and translate information as third concept. Daily report
from proposed application let user to learning the effect of
their behavior or their measures to electricity bill quickly, as
first concept.
Future work, we will test the effect of energy saving from
this concepts of energy monitoring system in medium term
and long term with larger sampling size.

ACKNOWLEDGEMENT
This research was supported by Rajamankala University of
Technology Thanyaburi Research Fund. for Academic year
2017.

REFERENCES
Fig 7. Forecast in each date compare with actual
[1] Department of Development Alternative Energy and Efficiency,Ministry
of Energy of Thailand, “Energy Balance of Thailand 2015”, pp.51
[2] D. S. Parker, “Pilot Evaluation of Energy Savings from
Residential Energy Demand Feedback Devices,” Florida, 2008
[3] Kathryn Buchanan, Riccardo Russo and Ben Anderson,”Feeding back
about eco-feedback: How do consumers use and respond to energy
monitors?”,in Energy Policy, vol. 73,pp 138-146, 2014
[4] S. S. van Dam, C. a. Bakker, and J. D. M. van Hal, “Home
energy monitors: impact over the medium-term,” Build. Res.
Inf., vol. 38, no. 5, pp. 458–469, Oct. 2010
[5] M. a. Alahmad, P. G. Wheeler, A. Schwer, J. Eiden, and A.
Brumbaugh, “A Comparative Study of Three Feedback Devices
for Residential Real-Time Energy Monitoring,” IEEE Trans.
Ind. Electron., vol. 59, no. 4, pp. 2002–2013, Apr. 2012
[6] Froehlich, J. , Findlater, L. , Landay, J. , The design of eco-feedback
Fig.8 Accuracy of Forecasting technology. In Proceedings of the SIGCHI Conference on Human
Factors in Computing Systems. ACM, pp. 1999–2008, 2010
[7] Energy Device – Energy Orb., last accessed Jan. 6, 2017. [Online].
IV. CONCLUSION AND FUTURE WORK Available: http://www.ambientdevices.com/about/energy-devices
There are some reports show that the energy saving effect [8] S. Biansoongnern and B. Plangklang, "Nonintrusive load monitoring
(NILM) using an Artificial Neural Network in embedded system with
from home energy monitoring system is dismiss very fast in low sampling rate," 2016 13th International Conference on Electrical
medium term and long term running of system. And from our Engineering/Electronics, Computer, Telecommunications and
literature review we can conclude the way to improve the Information Technology (ECTI-CON), Chiang Mai, 2016, pp. 1-4.
energy monitoring system. The system should be applied three
concepts, first the system must be a leaning tool not just a
monitoring system, second the system should be tailored

Вам также может понравиться