Volume 10, Issue 3, March – 2025 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar1046

Intelligent Power Management System

Manu C Madhu1; Jean Jacob2; Shiva Krishna M M3; Krishna Ramesh4
1
Student; 2Assistant Professor; 3Student; 4Student
1
Department of Computer Science and Engineering SNGCE Ernakulam, India
2
Department of Computer Science and Engineering SNGCE Ernakulam, India
3
Department of Computer Science and Engineering SNGCE Ernakulam, India
4
Department of Computer Science and Engineering SNGCE Ernakulam, India

Publication Date: 2025/03/28

Abstract: The paper deals with an Intelligent Power Manage- ment System, a System that works by SCADA model
supported by AI subsystems. The system will automate the operation of the power system.

Keywords: PMS(Power Management System), AI, ANN, ML, Rational.

How to Cite: Manu C Madhu; Jean Jacob; Shiva Krishna M M; Krishna Ramesh (2025). Intelligent Power Management System.
International Journal of Innovative Science and Research Technology, 10(3), 1372-1377.
https://doi.org/10.38124/ijisrt/25mar1046

I. INTRODUCTION generation and incurred losses, increased downtime,

increased maintenance costs. The final result of all this
Power Management System or PMS is a newer scaled- combined is higher power production costs. As technology
up version of Power Automation System designed to grows, their is higher demand for power as power demand
automate the Power Generation and supply of Energy increases the technology for power production is also being
throughout the Network of grids. The system collects data improved, but the cost for production and supply stays same
from users through specialized smart meters which can push or increased, this occurs as a result of the poor management
data to server though API interface and this data is used to strategies used by the authorities.
control and manage the grid. The system works similar to
SCADA but when SCADA is scaled up there will be no The newer technologies should kick into the
insight for the operators on how to control the system as the management phase of the Power Grids. The PMS is an
past data is not available. The present data will only be integral solution to the higher-level management of the
available after a delay in time in the future. Considering the power grids and the grid network. The idea of Incorporating
fact that the new system is designed to incorporate AI highly complex algorithms such as Random Forest Tree and
subsystems which train this available data of generators Isolation forest is to classify the generators and users and
and users from the past and predict the current statistics assign them rankings to choose the generator with the
which will help in the effective management of grids. An AI highest utility and also identify the abnormalities in the user
based real-time system can be implemented in PMS to and inform the authorities. By this process, we can minimize
handle real-time situations and avoid mishaps. The main waste and produce electricity with fewer CO2 emissions.
objective is to provide quality power supply along with
driving down the cost of energy production by choosing III. ARCHITECTURE
generator with higher rank and heuristics.
 The main Architecture of the System Consists of the
II. POWER MANAGEMENT SYSTEM the Server

The PMS is an intelligent system incorporated with  Power Generation management

mul- tiple AI as sub-modules to map the working of the  Usage Data Collection and Server Management
existing manually operated power grids. Considering  Power Load Management
existing power grids which are operated by mostly  User & Device Management Subsystems
unskilled labor force, the main issue lies within the  Assistive ML Modules
outcomes like unexpected power outages, excess power

IJISRT25MAR1046 www.ijisrt.com 1372

Volume 10, Issue 3, March – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar1046

Fig 1 Power Management System Architecture

 Power Generation Management mechanisms so that it is easier to implement above the

Power Generators are the source from which power is existing infrastructure. The ML is rational and can provide
generated and this power is transmitted through the power better output. The ML is equipped with Artificial Neural
infrastructure to reach the end users. The generators need Networks that can intelligently analyze the predefined data
constant management and control. The idea of this and rank the generators.
module is to develop a ML System that can automate the
power generation management. The older idea under The generator management is assisted in the PMS
SCADA was to use real-time available user data to manage with the help of Power Prediction Modules. Dynamic user
the power grids, but when scaled up to a larger system this data cannot be solely used for generator management. The
implementation fails as the real-time data is available with data needs high amount of processing. Since its time-
higher latency and a separate system needs to be used to consuming this cannot be used in real-time systems such as
collect and analyze the data to get high speed results. The Power Management Systems. The PMS has inbuilt modules
data analytics is harder and the collection requires a separate that train on the existing user data to predict the future
network so as to maintain reliability. These framework and power demands. The predicted power usage is used to
infrastructure would cost higher and is not easily increase the power production of the generators. This will be
implementable. assisting the future power demands and will be helping
create an Ideal System where demand and supply of power
An ML assisted module is used to rank the is always the same.
generators in real-time, the rank obtained, is based on
multiple parameters such as temperature, time, season, wind These two modules make sure that the power produced
, previous usage etc.., this is used to manage the power is greener as possible as well as the power is not wasted.
systems. There are two possible options to use the system These all are implemented as the sub-modules of the Main
one is to automate and control the generator automatically Power Management System Architecture.
and the other is to use them to assist the existing control

IJISRT25MAR1046 www.ijisrt.com 1373

Volume 10, Issue 3, March – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar1046

Fig 2 Power Generation Management Sub-Module

 Usage Data Collection and Server Management The system will be having a server that is used to
The usage data collection is done through multiple process the incoming usage data from the users. It is a uni-
methods on the existing power systems. The PMS is channel communication and the data is saved on to the
developed with an inbuilt API access for the end-point edge database by the module. Separate module is responsible for
devices to ping into the server. These devices can send or communicating with the users so that the whole process is
receive data through the API through https requests which similar to transactions.
can be implemented over the existing Internet Architecture
rather than opting for a separate system. The server is  Power Load Management and Forecasting
maintained by Django framework which can handle basic Power Load Management is the essential task to
requests from all devices and work safely under the load. maintain the equilibrium between user demand and supply
the equilib- rium needs to be maintained so as to make the
grid effective. The main disadvantage in all the existing
system is the latency associated with the power usage data
collection, this affects the management is a negative way as
sudden spikes in usage disturbs the equilibrium and often
incur in losses or power outage. This has been one of the
major concern from the I-o-T integration as it was costly yet
still ineffective. The main idea is to incorporate an ML that
can predict the accurate power usage that can calculate the
spikes and manage the demand in a better possible way so
that it can result in reliable power supply.

The database holds the value of the usage data

when there is a spike (positive or negative), the load
management system checks for the load and calculates the
exact values passes it on to the server and modules so that
they can access it and along with this it broadcasts the
message across the power-grid so that the smart devices
can take the appropriate actions. This is important as real-
time mapping of input to output can be only done with the
help of Load Management.The system is essential as it can
Fig 3 Data to Server Uni-Channel Communication

IJISRT25MAR1046 www.ijisrt.com 1374

Volume 10, Issue 3, March – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar1046
also easily figure out overload, take the necessary actions IV. COMPARISON WITH
and inform the authorities regarding. The action part is EXISTING SYSTEM
specified by the Supply Management module.
There are existing power grid management systems
across the world but in most of the cases such as in rural
areas the grid is mostly operated manually this is not efficient
and causes accidents as well as power outages. This is
the main area of scope where the system can be
implemented. The older systems are manual time consuming
as well ass inefficient. Most of the people are yet not
familiar with electricity. In developing nations itself there
are places where power is only available for just one day per
week for which farmers have to wait to complete the process
of agriculture. The better power grid may not show its effect
directly in the light in a short period but it will reduce the
cost of operation across multiple sectors providing better
quality products at a lower rate and a better lifestyle to
majority of the people. Consider the example of agriculture,
if power is available continuously the crop will be better as
there is proper irrigation, pest control etc.. The farmers wont
have to wait for long hours the cost of production will also
decrease as power will be rather available at a lower cost
Fig 4 Load Forecasting this will increase the profit at the root level.

 Automated Error Triggering The profit is available to all markets and this also helps
The Errors are usual in the Power grid and this is not people to rely more on grids and thus indirectly reduce the
easily detectable at the management side, but since data power usage from Non-renewable energy thus reducing the
collection is implemented the usual change is predictable, carbon emissions. In case of Industries when there is
But when there is unusual changes the ML can recognize it shortage of power they rely on fossil fuel powered
this can be used to trigger automated error checking generators thus operation can produce more pollutants, If
mechanism that can run itself to find the point of issue and there is reliable power from Renewable sources they can
inform the authorities to correct it. This process only takes use it and run on a lower operation cost.
few minutes and is accurate that it can also inform the end
users about the same also it can anticipate the errors in case The system can be coupled and scaled across nation to
of Climate. The higher wind, thunders all of the natural develop a united power grid so that the intra-territorial
things can be predicted these all affect the power power exchanges can be easily monitored and the usage of
consumption directly but the linear connection cant be Non- renewable energy can be reduced.
mapped easily but ML can help with the same by mapping
the data and producing the anticipates error triggers so that
the users as well as the authorities stay alert so that errors
can be tolerated easily and in an effective manner rather than
going for outages

Fig 5 Error-Triggering

IJISRT25MAR1046 www.ijisrt.com 1375

Volume 10, Issue 3, March – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar1046

Fig 6 Pie Chart showing of Distribution of Power Sources

The above figure shows that India is primarily computational power than SCADA.
dependent on the non-renewable sources the primary
dependency can on non-renewable sources can be brought VI. RESULTS
down to renewable sources reducing the carbon emissions.
The dependency is due to wastage of renewable energy The Predicted Outputs from the ML subsystems were
and the higher cost of implementation of the power plants useful in early management of the Power grids. These
based on renewable energy. The older system doesn’t supported the early detection of sudden spikes in the system
provide the scope for further improvements and with the and changing according to it. The early warning system
newer system the investments can be planned accordingly incorporated helped in the prediction of anticipated
and be made efficient. maintenance’s. These systems help in driving down the
production costs and the Error detection module works
V. COMPARISON WITH SCADA tirelessly to spot any errors in the system and alerts the
concerned authority. The main objective is to provide
SCADA stands for Supervisory Control and Data rational availability of power.
Acqui- sition, and it’s a software system that monitors and
controls industrial processes in real time. SCADA systems VII. CONCLUSION
are essential for managing and monitoring power plants,
especially wind and solar power plants. In conclusion the Smart Power Grid management
System equipped with multiple application specific AI and
Although SCADA is useful in Power Generation ML mod- ules can revolutionize the existing power grid
Management we cannot couple it to a power grid to manage by providing a better management option and also proving
the supply. The problem is that the power supply most rational output and from the power grid. The idea of
management needs an individual high speed network to equipping ML and AI will help in better management of the
handle the large load that might occur in peak time. This power systems. The better management results in the
might not be useful otherwise. reduced cost of operations as well as reliable power supply.
The system can ensure error and hassle free operation and
The Proposed model Helps to manage the grid by can also reduce pollution. The Power grids can also focus on
presump- tions and assumptions rather than real time data the power quality along with the quantity of supply. This
dependency this is a better option as real-time data will result in long term outcomes such as better lifestyle,less
availability is a large issue along with latency. Latency pollution, reduced cost of power, reduced price of goods and
issues need to be solved in the grid to ensure maximum services etc...// All these will be the long term results of
efficiency this is crucial. The ML can accurately predict the the system and can produce quick and anticipated actions.
power outcomes and provide management suggestions The manual operation needs time to adapt to a change but
the automation will help in the better adaptation with the
Although the Idea might seem similar. Both the situation.
systems differ the Predicted system can work with less

IJISRT25MAR1046 www.ijisrt.com 1376

Volume 10, Issue 3, March – 2025 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/25mar1046
REFERENCES Asia Conference on Power and Electrical En-
gineering (ACPEE), Chengdu, China, 2020, pp. 934-
[1]. R. Al-Roomi and M. E. El-Hawary, ”Fast AI-Based 939, doi: 10.1109/ACPEE48638.2020.9136300.
Power Flow Analysis for High-Dimensional Electric keywords: Decision support systems; Electrical
Networks,” 2020 IEEE Elec- tric Power and Energy engineering; Power supplies; Decision making;
Conference (EPEC), Edmonton, AB, Canada, 2020, Power quality; Companies; Power markets; premium
pp. 1-6, doi: 10.1109/EPEC48502.2020.9320057. power supply; decision support system; user demand;
keywords: Power systems; Mathematical model; voltage sag,
Transfer functions; Energy manage- ment; [7]. K. Shu, X. Ai, J. Wen, J. Fang, Z. Chen and C. Luo,
Generators; Artificial neural networks; Conferences; ”Optimal Energy Management for the Integrated
power operation; power flow analysis; load flow Power and Gas Systems via Real-time Pricing,”
analysis; Newton-Raphson; artificial neural 2018 IEEE Power Energy Society General Meeting
networks,. (PESGM), Portland, OR, USA, 2018, pp. 1-5, doi:
[2]. M. A. Rajabinezhad, a. G. Baayeh and j. M. 10.1109/PESGM.2018.8585873. keywords:
Guerrero, ”fuzzy- based power management Mathematical model; Fluid flow; Natural gas; Power
and power quality improvement in microgrid system dynamics; Energy management; Real-time
using battery energy storage system,” 2020 systems; Power system transients; integrated energy
10th smart grid conference (sgc), kashan, iran, system; bilevel programming; equilibrium
2020, pp. 1- 6, doi: constraints,
10.1109/sgc52076.2020.9335758. Keywords: [8]. W. Wanlu, Q. Lijuan, H. Shuai, L. Xiqiao and Z. Bo,
reactive power; power system management; ”Functional Design and System Development of
microgrids; power system harmonics; control Power Sales Management and Control Platform in
systems; batteries; load modeling; microgrid; energy Electric Power Market,” 2018 IEEE 4th Information
storage system; renewable energy resources; power Technology and Mechatronics Engineering
quality; energy management, Conference (ITOEC), Chongqing, China, 2018, pp.
[3]. T. E. Rao, S. Elango and G. G. Swamy, ”Power 1275-1279, doi: 10.1109/ITOEC.2018.8740450.
Management Strategy Between PV-Wind-fuel keywords: Companies; Power grids; Power markets;
Hybrid System,” 2021 7th Interna- tional Conference Power system management; Contracts; electric power
on Electrical Energy Systems (ICEES), Chennai, market; power sales management and control
India, 2021, pp. 101-107, doi: platform; power sale business,
10.1109/ICEES51510.2021.9383706. keywords: [9]. N. Yan, S. Li, T. Yan and S. h. Ma, ”Study on the
Photovoltaic systems; Wind energy; Power system Whole Life Cycle Energy Management Method of
manage- ment; Fuel cells; Hybrid power systems; Energy Storage System with Risk Correction
Mathematical model; Matlab; PV System; WECS; Control,” 2020 IEEE 4th Conference on Energy
Fuel Cell system; MPPT Technique; Hybrid System; Internet and Energy System Integration (EI2),
and Power Management, Wuhan, China, 2020, pp. 2450-2454, doi:
[4]. M. Hanan, W. Yousaf, X. Ai, E. Asghar, M. Y. 10.1109/EI250167.2020.9346933. keywords: Power
Javed and S. Salman, ”Multi-operating Modes system dynamics; Stability analysis; Regulation;
Based Energy Management Strategy of Virtual Safety; Energy management; Power generation;
Power Plant,” 2018 2nd IEEE Conference on Energy Energy storage; risk defense; dynamic safety margin;
In- ternet and Energy System Integration (EI2), energy storage systems; energy management,
Beijing, China, 2018, pp. 1-6, doi: [10]. Krishnan, Gokul M, Mohammed Soby, Ebin.
10.1109/EI2.2018.8582406. keywords: Batteries; (2017). Smart Grid Implementation in India with
Energy management; Wind turbines; State of charge; HVDC Transmission and Mi- croGrids. International
Mathematical model; Energy Management; Micro- Journal of Engineering Research and. V6.
grid; VPP; Simulink, 10.17577/IJERTV6IS020223.
[5]. G. Q. Tang, ”Smart grid management visualization: [11]. Rajeev Kumar, M. L. Dewal and Kalpana Saini,
Smart Power Management System,” 2011 8th ”Utility of SCADA in power generation and
International Conference Expo on Emerging distribution system,” 2010 3rd Interna- tional
Technologies for a Smarter World, Hauppauge, Conference on Computer Science and Information
NY, USA, 2011, pp. 1-6, doi: Tech- nology, Chengdu, China, 2010, pp. 648-652,
10.1109/CEWIT.2011.6135870. keywords: doi: 10.1109/ICC- SIT.2010.5564689. keywords:
Maintenance engineering; Real time systems; Smart Switches; Software; controllers; data acquisition;
grids; Power system faults; Monitoring; Power supervisory; atomic,
system reliability; Smart grid; management; powe [12]. Manu C Madhu, Jean Jacob, Krishna Ramesh, Shiva
system; modeling; real time; analysis; data Krishna M M Department of Computer Science
collection; visualisation; aumotmation; control; and Engineering SNGCE Ernakulam, India on
maintenance; Neural Networks, IJSREM Volume 8 December 2025 Intelligent
[6]. H. Zhang, H. Zhao, H. Li, Y. Chen, J. Ai Power Management System www.doi.org/10.55041/
and Q. Wang, ”A Decision Support System of IJSREM40128
Premium Power Supply Invest- ment,” 2020 5th

IJISRT25MAR1046 www.ijisrt.com 1377