Smart Grid Stability and Charging Stations Demand Prediction System

Overview

This project consists of two main modules aimed at optimizing electric vehicle (EV) charging stations' power management:

1. Smart Grid Stability: Predicts whether a charging station will remain stable when multiple vehicles are using fast charging, and redirects vehicles to other stations if instability is predicted.
2. Identification of Peak Hours: Predicts upcoming peak hours to help power providers prepare for high demand and ensure sufficient power supply.

Modules

1. Smart Grid Stability

Objective

Predict the stability of charging stations under high load conditions, especially when multiple EVs are charging simultaneously. If instability is predicted, incoming EVs are redirected to the next available station.

How it Works

• Trained a Random Forest Classifier with input features such as charging power, power demand, and power loss.
• The model predicts whether the station is stable or unstable based on these features.
• If instability is detected, incoming EVs are redirected to other stations.

2. Identification of Peak Hours

Objective

Predict the upcoming peak hours to help power providers prepare in advance and ensure sufficient power supply during high demand periods.

How it Works

• Trained an XGBoost model using features like date, month, hour, and total kWh consumed at each hour.
• Users can input a date to predict the peak hours for the following day.
• Helps power providers plan their supply to meet demand during predicted peak hours.

Features

• Smart Grid Stability:
  • Predicts whether a charging station is stable or unstable.
  • Redirection of EVs to other stations in case of instability.
• Peak Hour Identification:
  • Predicts peak hours based on historical data.
  • Helps power providers plan for peak demand and avoid supply shortages.

Tech Stack

• Programming Language: Python
• Libraries:
  • Scikit-learn (for Random Forest)
  • XGBoost
  • Pandas (for data processing)
  • Numpy
  • Matplotlib/Seaborn (for visualization)
• Tools:
  • Jupyter Notebooks (for experimentation)

Installation

Clone the repository

git clone https://github.com/henry-jacq/SmartEV.git
cd SmartEV

Install the Dependencies

pip install -r requirements.txt

Usage

Smart Grid Stability

1. Input the charging power, power demand, and power loss data for a station.
2. The model predicts whether the station will remain stable or if redirection is required.
3. If the station is unstable, the system provides alternative stations for redirection.

Peak Hour Identification

1. Input the desired date to predict the peak hours for the next day.
2. The model predicts the times of day when the power demand will be highest.

Data

The models are trained using historical data from charging stations and power consumption logs. The dataset includes:

• Charging power
• Power demand
• Power loss
• Voltage levels
• Total kWh consumption grouped by date, month, and hour

Model Training

Smart Grid Stability Model

• Model: Random Forest Classifier
• Features: Charging power, power demand, power loss, voltage levels
• Output: Station stability (Stable/Unstable)

Peak Hour Identification Model

• Model: XGBoost
• Features: Date, month, hour, total kWh consumed
• Output: Predicted peak hours for the next day