A

Mini Project Report on

EcoWatt - Power Prediction System

Submitted in partial fulfillment of the requirements

for the degree of
BACHELOR OF ENGINEERING
IN
Computer Science & Engineering
Artificial Intelligence & Machine Learning
by
Soham Waradkar (23206002)
Jay Yadav(23206007)
Pranal Vernekar (23206008)
Manas Jagtap (23206011)

Under the guidance of

Prof. Sayali Badhan

Department of Computer Science & Engineering

(Artificial Intelligence & Machine Learning)
A. P. Shah Institute of Technology
G. B. Road, Kasarvadavali, Thane (W)-400615
University Of Mumbai
2024-2025
 A. P. SHAH INSTITUTE OF TECHNOLOGY

CERTIFICATE

This is to certify that the project entitled “EcoWatt - Power Prediction System” is a
bonafide work of Soham Waradkar (23206002), Jay Yadav(23206007), Pranal Vernekar
(23206008), Manas Jagtap (23206011) submitted to the University of Mumbai in partial
fulfillment of the requirement for the award of Bachelor of Engineering in Computer
Science & Engineering (Artificial Intelligence & Machine Learning).

________________ _______________
Prof. Sayali Badhan Dr. Jaya Gupta
Mini Project Guide Head of Department
 A. P. SHAH INSTITUTE OF TECHNOLOGY

Project Report Approval

This Mini project report entitled “EcoWatt - Power Prediction System” by

Soham Waradkar, Jay Yadav, Pranal Vernekar and Manas Jagtap is approved
for the degree of Bachelor of Engineering in Computer Science &Engineering,
(AI&ML) 2024-25.

External Examiner:

Internal Examiner:

Place: APSIT, Thane

Date:
 Declaration

We declare that this written submission represents my ideas in my own words and
where others' ideas or words have been included,have adequately cited and
referenced the original sources.We also declare that we have adhered to all
principles of academic honesty and integrity and have not misrepresented or
fabricated or falsified any idea/data/fact/source in my submission. We understand
that any violation of the above will be cause for disciplinary action by the Institute
and can also evoke penal action from the sources which have thus not been
properly cited or from whomsoever permission has not been taken when
needed.

Soham Waradkar Jay Yadav Pranal Vernekar Manas Jagtap

(23206002) (23206007) (23206008) (23206011)
 ABSTRACT

Predicting the power output of wind turbines is essential for improving the efficiency and
effectiveness of wind energy systems. This project explores a method for forecasting wind
turbine power using machine learning techniques combined with weather data. By analyzing
historical data on wind speed, direction, and other relevant factors, we developed a predictive
model that estimates turbine power output. We tested our model with data from several wind
farms and found that it performed better than traditional methods like linear regression. This
work highlights important features that influence power output and includes methods for
selecting the most relevant data. Our findings show that the model reduces prediction errors
and can be used in real-time to enhance grid management and energy distribution. This
project offers insights and tools for future work in wind energy forecasting, aiming to support
more efficient use of renewable energy.

Keywords: Prediction,Wind speed,Turbine,Power output

 Index

Index Page no.

Chapter-1 1
Introduction 2

Chapter-2 4
Literature Survey 5
2.1 History 5
2.1 Review 7

Chapter-3 11
Problem Statement 12

Chapter-4 13
Experimental Setup 14
4.1 Hardware setup 14
4.2 Software Setup 14

Chapter-5 15
Proposed system and Implementation 16
5.1 Block Diagram of proposed system 16
5.2 Description of Block diagram 17
5.3 Implementation 18
5.4 Advantages 23

Chapter-6 24
6.1 Conclusion 25
6.2 Future Scope 25

References 26
 CHAPTER 1
INTRODUCTION

1
 1. INTRODUCTION

As the global focus shifts toward sustainable energy solutions, wind power has become a
prominent player in the renewable energy sector. Wind turbines, which harness the kinetic
energy of the wind to generate electricity, are a key technology in this transition. However,
the inherently variable nature of wind makes it challenging to predict turbine output
accurately. The efficiency and reliability of wind power systems depend heavily on the
ability to forecast power generation to manage grid stability, optimize energy storage, and
ensure a consistent supply of electricity.

Accurate prediction of wind turbine power output involves understanding and modeling the
complex interplay between various meteorological factors, including wind speed, direction,
and atmospheric pressure. Traditional forecasting methods, such as linear regression, often
struggle to capture the dynamic and non-linear relationships between these factors and
turbine performance. As a result, there is a growing need for more sophisticated predictive
models that can handle the complexity of wind patterns and provide more reliable forecasts.

In this project, we explore the application of advanced machine learning techniques to

enhance the accuracy of wind power predictions. Our approach involves developing a model
that integrates historical wind data with real-time meteorological information. By analyzing
data from multiple wind farms, we aim to identify key variables that significantly influence
turbine output and use this information to train a machine learning model. This model is
designed to not only predict power output with greater precision but also adapt to changing
wind conditions in real-time.

To achieve these goals, we apply various machine learning algorithms, including regression
trees, neural networks, and ensemble methods, to determine which approaches offer the best
performance. Feature selection techniques are employed to isolate the most relevant
predictors from a broader set of data, thus improving the model’s efficiency and accuracy.
The effectiveness of our model is evaluated through rigorous testing, comparing its
predictions to actual power output data from the wind farms.

Our research also addresses the practical implications of implementing these predictive
models in operational settings. By providing more accurate forecasts, our approach can help
grid operators better manage energy resources, reduce the likelihood of power imbalances,

2
and make informed decisions about energy storage and distribution. Ultimately, this project
aims to contribute to the advancement of wind energy technology by offering a robust tool
for predicting turbine power output, thereby supporting the broader goal of optimizing
renewable energy systems.

3
 CHAPTER 2
LITERATURE SURVEY

4
 2. LITERATURE SURVEY
2.1 HISTORY

Wind energy has a long history, dating back to ancient times when windmills were used in
places like Persia and medieval Europe for tasks such as grinding grain and pumping water.
However, the development of modern wind energy technology, aimed at generating electricity,
really took off in the late 20th century. This was driven by the oil crises of the 1970s and 1980s,
which highlighted the need for alternative, sustainable energy sources and spurred
advancements in renewable energy technologies.

During the late 20th century, wind turbines began to evolve from simple designs into more
efficient and powerful systems. The early turbines were quite basic and had limited efficiency.
The 1990s saw major improvements as engineers developed larger turbines with better
materials and design features. This period also saw an increased emphasis on predicting how
much power these turbines could generate, which was crucial for integrating wind energy into
the electricity grid effectively.

Initially, predicting wind turbine power output relied on straightforward statistical methods.
These methods used historical data on wind speed and direction to estimate future power
production. Linear regression models were commonly used, but they had limitations,
particularly in capturing complex, non-linear relationships between wind conditions and power
output. As the size and number of wind farms increased, more sophisticated forecasting
methods became necessary. With the advent of the 21st century, numerical weather prediction
models started to play a key role. These models use complex mathematical equations to
simulate atmospheric conditions and provide more accurate forecasts by considering a wider
range of meteorological data and higher spatial resolutions. While earlier methods like time
series analysis were used, they were often limited in their ability to handle non-linear patterns.

Recently, machine learning has transformed wind power forecasting. Techniques like
regression trees, neural networks, and ensemble methods have greatly improved the accuracy of
predictions. Machine learning can analyze large amounts of historical and real-time data to
identify patterns that simpler models might miss. This has made it possible to provide more
accurate and timely predictions of wind turbine output.

5
Feature selection techniques have also been used to focus on the most important variables,
which helps improve the efficiency and accuracy of the models. Machine learning methods,
which were not available in earlier times, are now essential for making better predictions and
managing wind power systems effectively.

6
2.2 LITERATURE REVIEW

[1] “Wind Turbine Power Output Forecasting Using Artificial Intelligence”,

Tejas Bhardwaj et. al.(2022)

In this paper, wind turbine power output is predicted using Artificial intelligence (AI)
techniques. The AI techniques used for predictions are Machine Learning (ML) algorithm and
Deep Learning (DL). In ML, polynomial regression is used and Long Short Term
Memory(LSTM) was used in DL. This forecasting is long-term forecasting that uses three
years of data collected from NIWE (National Institute of Wind Energy) and the results can be
used directly for the planning of energy management. Various environmental factors were
taken into consideration for forecasting for better accuracy and results. The AI helps to predict
the wind turbine output with high accuracy by considering the linear and non-linear types of
dataset. This technique can also be used for the preventive maintenance of wind turbines and
before the installation of wind power plants in an unfamiliar place to determine the
corresponding wind potential.

[2] “A Comparative Study of Machine Learning Techniques for Wind

Turbine Performance Prediction”, S. Muralidharan et. al.(2023)

This paper describes a comparative study of various machine learning techniques for wind
turbine performance prediction. The dataset used in this study is obtained from the National
Renewable Energy Laboratory (NREL) and contains meteorological data and power output
from a wind turbine. The machine learning techniques considered in this study include
artificial neural networks (ANN), decision trees (DT), and random forests (RF). The results
show that RF outperforms ANN and DT in terms of RMSE and MAE, while ANN
outperforms DT and RF in terms of R-squared. Overall, this research demonstrates the
effectiveness of machine learning techniques for wind turbine performance prediction and
provides insights on the advantages and disadvantages of certain machine learning approaches.
The findings of this research can be used to guide wind farm managers in selecting appropriate
machine learning techniques for wind turbine performance prediction.

7
[3] “Real-time power prediction approach for turbine using deep learning
techniques”, Lei Sun et. al.(2021)

This paper describes The accurate power forecasting is of great importance to the turbine
control and predictive maintenance. However, traditional physics models and statistical
models can no longer meet the needs of precision and flexibility when thermal power plants
frequently undertake more and more peak and frequency modulation tasks. In this study, the
recurrent neural network (RNN) and convolutional neural network (CNN) for power
prediction are proposed, and are applied to predict real-time power of turbine based on DCS
data (recorded for 719 days) from a power plant. In addition, the performances of two deep
learning models and five typical machine learning models are compared, including prediction
deviation, variance and time cost. It is found that deep learning models outperform other
shallow models and RNN model performs best in balancing the accuracy-efficient trade-off for
power prediction (the relative prediction error of 99.76% samples is less than 1% in all load
range for test 216 days). Moreover, the influence of training size and input time-steps on the
performance of RNN model is also explored. The model can achieve remarkable performance
by learning only 30% samples (about 216 days) with 3 input time-steps (about 60 s). Those
results of the proposed models based on deep-learning methods indicated that deep learning is
of great help to improve the accuracy of turbine power prediction. It is therefore convinced
that those models have a high potential for turbine control and predictable maintenance in
actual industrial scenarios.

[4] “Short-Term Power Prediction of Wind Turbine Applying Machine

Learning and Digital Filter”, Shujun Liu et. al.(2023)

As wind energy development increases, accurate wind energy forecasting helps to develop
sensible power generation plans and ensure a balance between supply and demand. Machine-
learning-based forecasting models possess exceptional predictive capabilities, and data
manipulation prior to model training is also a key focus of this research. This study trained a
deep Long Short-Term Memory (LSTM) neural network to learn the processing results of the
Savitzky-Golay filter, which can avoid over-fitting due to fluctuations and noise in
measurements, improving the generalization performance. The optimum data frame length to
match the second-order filter was determined by comparison. In a single-step prediction, the
method reduced the root-mean-square error by 3.8% compared to the model trained directly
with the measurements. The method also produced the smallest errors in all steps of the multi-

8
step advance prediction. The proposed method ensures the accuracy of the forecasting and, on
that basis, also improves the timeliness of the effective forecasts.

[5] “Integrated Energy Management and Forecasting Dataset”, K M

Karthick Raghunath(2023)

The Integrated Energy Management and Forecasting Dataset is a comprehensive data

collection specifically designed for advanced algorithmic modeling in energy management. It
combines two distinct yet complementary datasets - the Energy Forecasting Data and the
Energy Grid Status Data - each tailored for different but related purposes in the energy sector.
The integration of IoT (Internet of Things) technologies in the Integrated Energy Management
and Forecasting Dataset significantly enhances its functionality and applicability in real-world
scenarios. Sensors across various points in the energy grid and renewable energy sources
gather data like energy consumption, production, grid load, and environmental factors
(temperature, weather conditions). Sensors within the grid infrastructure also monitor load,
distribution, and transmission status. Sensors in solar panels, wind turbines, and traditional
power plants (like coal or gas-fired) provide detailed output data. IoT infrastructure ensures
that the data is not just accurate but also current, reflecting the latest status of the energy grid
and environmental conditions.

[6] “Prediction of Wind Power with Machine Learning Models”, Ömer Ali
Karaman(2023)

Wind power is a vital power grid component, and wind power forecasting represents a
challenging task. In this study, a series of multiobjective predictive models were created
utilising a range of cutting-edge machine learning (ML) methodologies, namely, artificial
neural networks (ANNs), recurrent neural networks (RNNs), convolutional neural networks,
and long short-term memory (LSTM) networks. In this study, two independent data sets were
combined and used to predict wind power. The first data set contained internal values such as
wind speed (m/s), wind direction (°), theoretical power (kW), and active power (kW). The
second data set was external values that contained the meteorological data set, which can
affect the wind power forecast. The k-nearest neighbours (kNN) algorithm completed the
missing data in the data set. The results showed that the LSTM, RNN, CNN, and ANN
algorithms were powerful in forecasting wind power. Furthermore, the performance of these
models was evaluated by incorporating statistical indicators of performance deviation to

9
demonstrate the efficacy of the employed methodology effectively. Moreover, the
performance of these models was evaluated by incorporating statistical indicators of
performance deviation, including the coefficient of determination (R2), root mean square error
(RMSE), mean absolute error (MAE), and mean square error (MSE) metrics to effectively
demonstrate the efficacy of the employed methodology. When the metrics are examined, it can
be said that ANN, RNN, CNN, and LSTM methods effectively forecast wind power. However,
it can be said that the LSTM model is more successful in estimating the wind power with an
R2 value of 0.9574, MAE of 0.0209, MSE of 0.0038, and RMSE of 0.0614.

10
 CHAPTER 3
PROBLEM STATEMENT

11
 3. Problem Statement

As the global shift toward renewable energy accelerates, wind power has emerged as a crucial
component of sustainable energy portfolios. However, the effectiveness of wind energy is
highly dependent on accurate forecasting of wind turbine power output, which remains a
significant challenge. Current forecasting methods often struggle to accurately predict short-
term power output due to the intricate and dynamic nature of wind patterns. Traditional
models fail to account for the complexity of wind behavior, leading to unreliable predictions.
Additionally, the high variability of wind conditions, characterized by rapid and unpredictable
changes, makes short-term forecasting even more difficult. This variability poses a significant
challenge for grid operators, who must balance supply and demand to ensure grid stability.

Another major challenge is data integration. Accurately forecasting wind power requires the
effective combination of diverse data sources, including historical wind data, real-time
weather forecasts, and turbine-specific parameters. However, integrating these data sources in
a meaningful way is a significant hurdle. Moreover, achieving a balance between prediction
accuracy and computational efficiency is crucial. High-precision models often demand
substantial computational resources, which can limit their practicality for real-time
applications.

This project aims to address these challenges by developing advanced machine learning
models designed specifically for wind power forecasting. By leveraging sophisticated
algorithms and integrating diverse data sources, the project seeks to improve the accuracy of
wind power predictions while optimizing computational efficiency. Enhanced forecasting
capabilities will lead to more reliable energy management, improved grid stability, and a
stronger contribution of wind energy to the renewable energy mix. This initiative aims to make
wind energy more predictable and efficient, supporting the transition to a sustainable energy
future and contributing to global efforts to combat climate change.

12
 CHAPTER 4
EXPERIMENTAL SETUP

13
 4. Experimental Setup

4.1 Hardware Setup

 Processor: Intel i5 8th gen Or Above
 Ram: 8GB Or Above
 Storage: 256 GB SSD Or Above
 OS: Windows 10 Or Above

4.2 Software Setup

 Visual Studio Code
 Node Package Manager (10.8.3)
 React JS (18.2.0)
 Firebase
 Python Flask (3.0.3)

14
 CHAPTER 5
PROPOSED SYSTEM
&
IMPLEMENTATION

15
 5.Proposed system & Implementation

5.1 Block diagram of proposed system

Fig. 5.1: Block Diagram For EcoWatt

16
5.2 Description of block diagram

 Landing Page: The entry point to the application that directs users to various sections
including About, Contacts, Login, and Services.

 About & Contacts: These sections provide general information about the application
and allow users to contact support for assistance.

 Login: Provides authentication for users, granting access to further application

functionalities like profiling, prediction, and reporting.

 Services: Offers a general overview of the features and benefits of using the
application.

 Home: After logging in, users land on the home page where they can access all core
functionalities.

 Profiling: Users can manage their profiles, including information about datasets they
have uploaded for wind energy predictions.

 Prediction: This feature allows users to generate power predictions based on wind
energy datasets by analyzing historical data and weather conditions.

 Reports: Enables users to generate detailed reports on power predictions. Users can
specify starting and ending dates for generating these reports, ensuring the data
reflects the desired time range.

 Upload: Users can upload new datasets related to wind energy. These datasets are
used to enhance prediction accuracy and generate reliable energy forecasts.

 Settings: Users can manage their preferences and configurations within the
application.

17
5.3 Implementation

Fig. 5.3.1 Landing

Fig. 5.3.2 About Us

18
Fig. 5.3.3 Contacts

Fig. 5.3.4 Services

19
Fig 5.3.5 Register

Fig 5.3.6 Login

20
Fig. 5.3.7 Profiling

Fig. 5.4.8 Reports

21
5.3.9 Prediction

5.3.10 Settings

22
5.4 Advantages

Enhanced Accuracy: ML models can analyze complex patterns and relationships in

historical wind data, leading to more accurate predictions of wind power output compared
to traditional methods.
Real-Time Forecasting: By integrating real-time weather data and turbine-specific
parameters, ML algorithms can provide timely and dynamic forecasts, which are crucial
for operational decisions and grid management.
Adaptability: ML models can continuously learn from new data, allowing them to adapt
to changing wind patterns and improve predictions over time, enhancing their reliability
and precision.
Optimized Resource Allocation: Accurate wind power forecasts enable better planning
and resource management, including more efficient scheduling of maintenance and better
utilization of wind energy in the grid.
Improved Grid Stability: Reliable predictions help in maintaining grid stability by
allowing for better integration of wind power into the energy mix, reducing the risk of
power imbalances and outages.
Cost Efficiency: By minimizing errors in power forecasts, ML models can help reduce
operational costs and increase the financial viability of wind energy projects.
Advanced Data Integration: ML can effectively combine various data sources, including
historical data, weather forecasts, and turbine performance metrics, to provide a
comprehensive view of wind power potential.
Scalability: ML algorithms can be scaled to handle large datasets from multiple wind
farms, making them suitable for both small and large-scale wind energy operations.
Predictive Maintenance: Accurate forecasts allow for predictive maintenance of turbines,
identifying potential issues before they lead to failures, thus extending the lifespan of
equipment and reducing downtime.

23
CHAPTER 6
CONCLUSION

24
 6.Conclusion

Conclusion:
Accurate wind power forecasting is crucial for the effective management of wind energy and
its integration into the power grid. The development of advanced machine learning models
holds the potential to significantly enhance the accuracy of wind turbine power output
predictions. By leveraging sophisticated algorithms and integrating diverse data sources such
as historical wind data, real-time meteorological forecasts, and turbine-specific parameters this
project aims to address the limitations of traditional forecasting methods.

The anticipated improvements in forecasting accuracy will enable better decision-making for
energy management, leading to optimized turbine operation and grid stability. Enhanced
predictions will facilitate more efficient scheduling of maintenance activities, reduce
operational uncertainties, and help in balancing supply and demand more effectively. As a
result, wind energy can be utilized more efficiently, contributing to a more stable and reliable
power system. This advancement will support the broader adoption of renewable energy and
aid in the transition towards a sustainable energy future.

Future Scope:
To build on the improvements made in wind power forecasting with machine learning, future
research should focus on real-time applications. Testing the developed models in live
operational settings will be essential to assess their performance and make necessary
refinements. Additionally, incorporating additional data sources, such as satellite imagery and
advanced remote sensing, could further enhance the accuracy of predictions by providing
more detailed information on wind conditions.

Evaluating the long-term effects of enhanced forecasting on grid stability and energy
efficiency will also be crucial. This involves understanding how improved predictions impact
operational strategies and maintenance planning. Furthermore, exploring advancements in
machine learning techniques and adapting models for different wind farm types and
geographic regions will ensure broader applicability and effectiveness. These steps will help
refine wind power forecasting and support more efficient use of wind energy.

25
 References

[1] “Wind Turbine Power Output Forecasting Using Artificial Intelligence”,Tejas Bhardwaj,
Sumit Mehenge & B. Sri Revathi, International Virtual Conference on Power Engineering
Computing and Control: Developments in Electric Vehicles and Energy Sector for
Sustainable Future (PECCON), 2022

[2] “A Comparative Study of Machine Learning Techniques for Wind Turbine Performance
Prediction”, S. Muralidharan, S.Parthasarathy, Deepa A & Jermin Jersha, International
Conference on Smart Engineering for Renewable Energy Technologies (ICSERET), 2023

[3] “Real-time power prediction approach for turbine using deep learning techniques”, Lei
Sun, Tianyuan Liu, Yonghui Xie, Di Zhang, Xinlei Xia, Elsevier BV, 2021

[4] “Short-Term Power Prediction of Wind Turbine Applying Machine Learning and Digital
Filter”, Shujun Liu, Yaocong Zhang, Xiaoze Du, Tong Xu & Jiangbo Wu, Applied
Sciences, 2023

[5] “Integrated Energy Management and Forecasting Dataset”, K M Karthick Raghunath,

International Conference on Sustainable Computing and Smart Systems (ICSCSS), 2024

[6] “Prediction of Wind Power with Machine Learning Models”, Ömer Ali Karaman, Applied
Sciences, 2023

26