Electrify Real-Time Analysis of Electricity Consumption and Bill Prediction
Electrify Real-Time Analysis of Electricity Consumption and Bill Prediction
Volume 8, Issue 3, March – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Electrify: Real-Time Analysis of Electricity
Consumption and Bill Prediction
Abhijith C Amanyujith Raj
TLY19CS001 TLY19CS012
Department of Computer Science Department of Computer Science
College of Engineering Thalassery College of Engineering Thalassery
Aiswariya R Fathimath Hanna Safeer
TLY19CS009 TLY19CS024
Department of Computer Science Department of Computer Science
College of Engineering Thalassery College of Engineering Thalassery
Prof. Nikhil Dharman M K
Department of Computer Science
College of Engineering Thalassery
Abstract:- In this scenario, it is essential to optimize energy usage in buildings, reducing utility costs. In this
power usage and promote efficient energy consumption. project, we focus on one specific method to facilitate it: the
The increasing electricity demand has put a strain on prediction of electricity consumption and real-time analysis
natural resources and the environment. To address this, of electricity consumption. In this project, we will make IoT
there is a need to optimize energy usage and reduce waste. Based electricity energy consumption reading application
The proposed system aims to help users achieve this by using ESP32.
providing real-time analysis of electricity consumption
and predicting monthly electricity bills using machine The purpose of this project is to develop a system for
learning algorithms. The system employs a current sensor real-time analysis of electricity consumption and bill
to track the energy consumed by each device in a house prediction. The project aims to address the problem of
and estimates the monthly electricity bill. The user can set inefficient electricity usage and provide a solution through the
a monthly desired bill, and as the energy consumption implementation of machine learning algorithms. The project
approaches the set limit, the user is notified through the will involve the development of a system that utilizes data
software application. The LSTM algorithm is used for from a current sensor and a Wi-Fi module and incorporates
predicting the monthly electricity bill based on the data various machine-learning techniques to analyze and predict
collected from the current sensor. The algorithm takes electricity consumption patterns. The resulting system will be
into account various factors such as the energy consumed tested and evaluated through a series of user studies to ensure
by different devices, time of day, and historical data to its accuracy and reliability.
provide accurate predictions. Firebase is used as a cloud
service for storing and processing data. It allows for II. LITERATURE SURVEY
efficient and secure storage of data and provides real-time
updates, ensuring that users always have access to the The paper [1] analyses the benefits of demand response
latest information. The proposed system offers numerous for utility companies and proposes prediction methods for
benefits, including improved energy efficiency and cost households to determine whether changing to a time-of-use
savings. (TOU) rate plan from a progressive rate plan can reduce their
electricity bill. The methods are developed using statistical
I. INTRODUCTION learning techniques and tested using hourly electricity usage
and bills obtained from ten apartment complexes through
The ever-increasing demand for power has driven many AMI, with one complex used for testing. The results show
natural resources to the verge of depletion. With the current high decision accuracy and low root-mean-square error for
technology of electricity meter, we need to go to the meter the saving prediction. The proposed prediction methods only
reading room and take down readings. Thus monitoring and rely on historical data, which may not accurately predict
keeping track records of your electricity consumption is a future electricity usage and bill savings.
difficult task. In such a case, optimal power-sharing and
efficient energy consumption are called into practice. So The paper [2] analyses the characteristics of electricity
balancing the power supply and demand is one of the most consumption data according to contract types and measures
fundamental problems for the operation and control of any the performance of deep learning models in predicting future
electric power grid. To automate this, we can use the Internet electricity consumption. The dataset includes hourly
of Things. IoT sensors can be used to monitor and optimize electricity consumption data over two years. The paper
IJISRT23MAR1320 www.ijisrt.com 1647
Volume 8, Issue 3, March – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
applies three deep learning models (LSTM, GRU, and CNN) Energy Consumption Monitoring
to the dataset. The results show that the LSTM model The energy consumption of each device is monitored
outperformed the other models for all contract types. The using the SCT-013 current sensor, which is connected to the
study only tests deep learning models and does not compare ESP32 microcontroller.
their performance to other predictive models.
Data Preprocessing
The paper [3] proposes a machine learning-based The raw energy consumption data is preprocessed to
technique to anticipate household electricity consumption remove noise and prepare it for analysis. This involves
using weather data. It is based on various machine learning filtering, normalization, and feature extraction.
algorithms and found that the Random Tree algorithm
produced the best results, with a correlation coefficient of Real-Time Analysis
75.7%. A potential limitation of this paper is that it only The preprocessed data is then used for real-time analysis
focuses on the relationship between weather conditions and using machine learning algorithms. The LSTM algorithm is
electricity consumption. used to predict the energy consumption for the upcoming
period and estimate the monthly electricity bill.
In paper [4] reviews machine learning models for
predicting energy consumption based on a one-year dataset User Notification:
from a shoe store. Linear Regression and Support Vector As the energy consumption approaches the user's set
Regression achieved the best results with a success rate of monthly bill limit, the user is notified through the software
85.7%. The paper emphasizes the importance of predicting application, allowing them to adjust their energy consumption
energy consumption for energy supply companies to adapt to accordingly.
customer behavior and plan supply chain adaptations. The
limitation of this paper is the small size of the data set used Cloud Storage
for analysis. The preprocessed data and analysis results are stored in
Firebase cloud storage, allowing for easy access and retrieval.
III. EXISTING SYSTEMS
Application Development
The existing system for monitoring electricity The data and analysis results are presented to the user
consumption in households typically involves manually through a software application, allowing them to monitor
reading the meter and calculating the usage. This can be a their energy consumption and bill prediction in real time.
time-consuming process and can lead to inaccuracies. Smart
meters have been introduced as an alternative to manually Continuous Learning:
reading the meter, but they can be expensive to install and The machine learning algorithm is continually updated
may not be available in all areas. In addition, there are several and refined based on new data, allowing for improved
challenges in managing electricity consumption in accuracy in energy consumption and bill prediction over time.
households. Many households are not aware of their energy
consumption patterns and do not know how to reduce their V. CONCLUSION
energy usage.
The proposed system of real-time analysis of electricity
The existing system typically involves the use of traditional consumption and bill prediction using machine learning
electricity meters, which do not provide real-time data on algorithms has several advantages over the existing systems.
energy usage. Users are only able to view their electricity By continuously monitoring the energy consumption of each
consumption after the end of each billing cycle and are not device, the system enables efficient energy consumption and
able to make adjustments to their consumption in real time. optimal power-sharing, which can help in reducing energy
This can lead to wasteful energy consumption, which not only costs and conserving natural resources. The use of a current
increases the electricity bill but also hurts the environment. sensor and machine learning algorithms for bill prediction
makes the system accurate and reliable, providing users with
IV. PROPOSED SYSTEM an estimate of their monthly electricity bill.
The system aims to address the limitations of the The application developed as part of the system allows
existing system by introducing a smart metering solution that users to set a monthly desired bill and notifies them when
continuously monitors the energy consumption of each energy consumption approaches the set limit. This feature can
device in a household and provides real-time feedback to the help users in budgeting their electricity costs and reducing
user. This is achieved by using a current sensor to measure unnecessary energy consumption. In addition, the use of
the energy consumed by each device and then transmitting cloud services such as Firebase for backend and training data
this data wirelessly to a microcontroller (ESP32) for storage provides scalability and flexibility to the system. This
processing. The processed data is then sent to a cloud-based allows the system to handle large amounts of data and provide
backend (Firebase) for storage and analysis. real-time updates to users.
IJISRT23MAR1320 www.ijisrt.com 1648
Volume 8, Issue 3, March – 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
REFERENCES
[1]. Y. M. Chung, S. Kang, J. Jung, B. J. Chung, and D.S.
Kim, "Residential Electricity Rate Plans and Their
Selections Based on Statistical Learning," in IEEE
Access, vol. 10, pp. 74012-74022, 2022
[2]. C. -G. Lim and H. -J. Choi, "Deep Learning - Based
Analysis on Monthly Household Consumption for
Different Electricity Contracts," 2020 IEEE
International Conference on Big Data and Smart
Computing (BigComp), 2020
[3]. G. B. Brahim, "Weather Conditions Impact on
Electricity Consumption in Smart Homes: Machine
Learning Based Prediction Model," 2021 8th
International Conference on Electrical and Electronics
Engineering (ICEEE), 2021
[4]. A. González - Briones, G. Hernández, J. M. Corchado,
S. Omatu and M. S. Mohamad, "Machine Learning
Models for Electricity Consumption Forecasting: A
Review," 2019 2nd International Conference on
Computer Applications & Information Security
(ICCAIS), 2019
