SRI BALA JI CHOCKALINGAM ENGINEERING COLLEGE

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

SOLAR INTELLIGENCE PREDICTIVE MODELS FOR POWER

GENERATION AND RADIATION USING
MACHINE LEARNING
20

15
TEAM MEMBERS
ELAYARAJA.S (512721104005)
LOKESH.N (512721104014) 10

RANJITH.K (512721104022)
SABARI GIRI.V (512721104026)
5

GUIDED BY: PORKODI .S AP/CSE

ABSTRACT
Domain: Machine Learning (ML) is used in solar energy forecasting to predict power
generation and radiation levels with high accuracy. Advanced ML models like
Random Forest, ARD Regression, and Ridge Regression , Decision Tree Regression
analyze historical and real-time weather data to optimize energy efficiency and grid
stability
Existing System: Traditional solar forecasting relies on statistical models thatlack
real-time adaptability and struggle with weather variability.
Proposed System: Implements machine learning models like Random Forest,
ARD Regression, and Ridge Regression, Decision Tree Regression to enhance solar
power and radiation prediction using real-time and historical data.

How the Problem is Overcome: D Provides higher accuracy and

adaptability, ensuring better energy management and grid stability with
dynamic learning capabilities
OBJECTIVE
Develop machine learning models for accurate solar power and radiation forecasting.

Utilize historical and real-time meteorological data for improved prediction accuracy.

Enhance energy efficiency and grid stability through intelligent forecasting.

Optimize solar panel performance by predicting energy generation under varying

conditions.

Enable real-time monitoring and decision-making for better renewable energy utilization.
GOALS
Data Collection – Gather historical and real-time solar
data (irradiance, temperature, humidity, etc.).

Model Development – Train machine learning models (ARD Regression

,Decision Tree Regression,Ridge Regression,Random Forest Regression)
for accurate predictions.
Performance Optimization – Fine-tune models to improve forecast
accuracy and adaptability.

Deployment & Integration – Implement models in smart grids and

energy management systems.

Monitoring & Improvement – Continuously update models for better

efficiency and reliability.
FEASIBILITY STUDY
Existing System Feasibility:
Uses statistical models with limited real-time adaptability.
Requires costly satellite data and manual intervention.
Struggles with high variability in solar radiation predictions.
Slow adaptation to changing weather conditions affects
accuracy.

Proposed System Feasibility:

Uses ML models (ARD Regression ,Decision Tree Regression,Ridge Regression

Random Forest Regression) for real-time adaptation.
Reduces costs by automating solar power predictions.
Provides accurate forecasts, improving grid stability.
Enables faster and reliable solar energy forecasting.
SOFTWARE REQUIREMENTS
Operating System: Windows 10/11, Linux (Ubuntu), or
macOS
Programming Languages: Python (preferred), R, MATLAB
Libraries & Frameworks: , Pandas & NumPy (for data
processing) , Matplotlib & Seaborn (for data
visualization)

Database (if needed): MySQL, PostgreSQL, MongoDB

Cloud Platforms (Optional): Google Colab, AWS, Azure, or

Google Cloud for model training and deployment

IDE: Jupyter Notebook, PyCharm, VS Code

HARDWARE REQUIREMENTS
Processor: Intel Core i7/i9, AMD Ryzen 7/9, or higher (for training ML models)

GPU (Optional but Recommended): NVIDIA RTX 3060/RTX 4090 or equivalent for
deep learning acceleration

RAM: Minimum 16GB (32GB+ recommended for large datasets)

Storage: SSD (512GB or higher for faster data access)

Internet Connection: Required for cloud-based model training and data

retrieval
PROJECT PLAN
1. Data Collection & Preprocessing
Gather historical and real-time solar radiation, temperature, humidity,
and power generation data.
Clean, normalize, and prepare the data using Python libraries like Pandas
and NumPy.

2.Model Development & Training

Train machine learning models such as Random Forest, ARD Regression,
Ridge Regression, and Decision Tree Regression.
Optimize the models for better prediction accuracy and adaptability.
3.Deployment & Integration
Build a web application using Django (Python) for real-time solar power
prediction.
Deploy the system on cloud platforms (AWS, Google Cloud, or
PythonAnywhere) for accessibility.
LITERATURE SURVEY
1. K.C.Jayasankar, G.Anandhakumar and A.Kalaimurugan “Prediction of Solar Radiation using Deep LSTM-based
Machine Learning Algorithm” Journal of environmental nanotechnology 2024, Volume 13, Issue 3, Pages 01-
08,ISSN: 2279-0748
DESCRIPTION:

This study proposes a Deep LSTM-based machine learning model for predicting solar radiation, a crucial
factor for solar energy systems and weather forecasting. Traditional models struggle with complex
temporal patterns, but Deep LSTM (a type of RNN) effectively captures sequential dependencies. The
model is trained on historical meteorological data (temperature, humidity, wind velocity, and solar
radiation). Through careful feature selection and hyperparameter tuning, the Deep LSTM model
outperforms conventional methods, delivering higher accuracy and reliability in diverse weather
conditions.
Limitations:

High Data Requirement

Computationally Intensive
Overfitting Risk
Limited Generalization
LITERATURE SURVEY (CONTINUE)
2. Anas Al-lahham , Obaidah Theeb , Khaled Elalem , Tariq A. Alshawi and Saleh A. Alshebeili 1, "Sky Imager-Based
Forecast of Solar Irradiance Using Machine Learning" 2020, Volume 9, Issue 10, Article number 1700. The ISSN for
Electronics is 2079-9292.

DESCRIPTION:

This study introduces a sky imager-based machine learning model for solar irradiance forecasting. The model
extracts features from Total Sky Imager (TSI-880) and All Sky Imager (ASI-16) datasets, containing 350,000+
images spanning 16 years. Using Random Forest (RF) and K-Nearest Neighbors (KNN), the model predicts
Global Horizontal Irradiance (GHI) up to 4 hours ahead with lower computational complexity than deep
learning approaches. The results show competitive accuracy while maintaining efficiency, making it a viable
solution for real-time solar forecasting.

Limitations:

Weather Dependence
Computational Constraints
Data-Specific Performance
Prediction Horizon Limits
SYSTEM ARCHITECTURE
SYSTEM ARCHITECTURE(CONT.)
Step-by-Step Description
Data Collection: Gather solar radiation and power generation data from
Kaggle or Real Time Data.
Data Preprocessing: Clean, normalize, and prepare data using NumPy and
Pandas.
Data Visualization: Analyze trends using Matplotlib and Seaborn.
Model Training: Train and compare ARD Regression, Decision Tree, Ridge
Regression, and Random Forest models.
Model Selection & Saving: Choose the best model and save it as a .pkl file.
Web Application Development: Use Django (Python) for backend and HTML,
CSS, JavaScript for frontend.
Database Management: Store user inputs and predictions in SQLite3.
Model Deployment & Integration: Connect the trained model with the web
app for real-time predictions.
User Interaction: Enable login, registration, data input, and prediction display
via web pages.
System Deployment: Deploy on cloud platforms like AWS or PythonAnywhere
for accessibility
MODULES
DATA PRE-PROCESSING:

Validation techniques in machine learning are used to get the error rate of the Machine
Learning (ML) model, which can be considered as close to the true error rate of the dataset.
If the data volume is large enough to be representative of the population, you may not need the
validation techniques.
However, in real-world scenarios, to work with samples of data that may not be a true
representative of the population of given dataset.
To finding the missing value, duplicate value and description of data type whether it is float
variable or integer.
The sample of data used to provide an unbiased evaluation of a model fit on the training
dataset while tuning model hyper parameters.
Exploration data analysis of visualization (continue)

Data visualization is a key part of understanding and analyzing data in statistics and machine
learning.
It helps in identifying patterns, trends, and anomalies, making data easier to interpret. Charts
and plots, such as histograms, scatter plots, and box plots, provide a clear view of
distributions and relationships.
Visualization is crucial for detecting outliers, missing values, and correlations between
variables.Tools like Matplotlib and Seaborn in Python allow for effective data representation,
making it easier to draw insights and communicate findings to stakeholders. How to chart time
series data with line plots and categorical quantities with bar charts. How to summarize data
distributions with histograms and box plots.
Model Validation & Performance Evaluation Module (continue)

After training, models are evaluated in the Model Validation & Performance Evaluation Module
to ensure they provide accurate and reliable predictions.

Validation techniques like K-Fold Cross-Validation are used, and performance metrics such as
Mean Squared Error (MSE), Root Mean Square Error (RMSE), and R² Score help determine which
model performs best. The model with the highest accuracy and lowest error is selected for
deployment.
Web Application & Deployment Module (continue)
The Web Application Module provides a user-friendly interface for solar power forecasting. It
allows users to input weather parameters (such as temperature, humidity, and radiation levels)
and receive real-time predictions. The frontend is developed using HTML, CSS, and JavaScript,
while the backend is built with Django or Flask to process user inputs and display predictions.
CONCLUSION AND FUTURE ENHANCEMENT
Machine learning models like Random Forest, ARD Regression, Ridge Regression, and Decision
Tree Regression were effectively used for solar power and radiation forecasting.
These models demonstrated improved prediction accuracy compared to traditional statistical
methods.
Historical and real-time weather data were successfully utilized to enhance forecast reliability.

FUTURE ENHANCEMENT
Use advanced models like Random Forest, ARD Regression, Ridge Regression, and Decision Tree
Regression to capture more complex patterns and improve prediction accuracy.
Expand data sources by including satellite imagery and real-time IoT sensor inputs for more
robust forecasting.
Create a mobile version of the web app to provide users with on-the-go access to solar power
predictions.
REFERENCES
https://www.sciencedirect.com/science/article/pii/S2666603023000064

https://scijournals.onlinelibrary.wiley.com/doi/10.1002/ese3.1234

https://link.springer.com/article/10.3103/S0003701X22010170

https://www.sciencedirect.com/science/article/pii/S2352484723011228

https://www.sciencedirect.com/science/article/pii/S2214157X24009559

https://www.sciencedirect.com/science/article/pii/S036054422301201X

https://www.mdpi.com/2071-1050/15/11/8927

https://ietresearch.onlinelibrary.wiley.com/doi/epdf/10.1049/rpg2.12408
SAMPLE OUTPUT
SAMPLE OUTPUT
SAMPLE OUTPUT
SAMPLE OUTPUT
SAMPLE OUTPUT
SAMPLE OUTPUT
SAMPLE OUTPUT
SAMPLE OUTPUT
SAMPLE OUTPUT
SAMPLE OUTPUT
SAMPLE OUTPUT
THANK YOU!