app.py

import streamlit as st
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.ensemble import RandomForestClassifier
from prophet import Prophet

# Load dataset
@st.cache_data
def load_data(filename):
    return pd.read_csv(filename)

# Main application
st.sidebar.title("AI-Powered Industrial Management")
page = st.sidebar.radio("Choose a page", ["Predictive Maintenance for Wind Turbines", "Intelligent Energy Management Systems"])

# Predictive Maintenance for Wind Turbines
if page == "Predictive Maintenance for Wind Turbines":
    st.header("Predictive Maintenance for Wind Turbines")

    # Load and display dataset
    data = load_data('wind_turbine_data.csv')

    st.write("### Dataset Preview:")
    st.write(data.head())

    # Dataset Meta Information
    st.write("### Dataset Meta Information:")
    st.write("""
    - **Timestamp:** The time at which the data was recorded.
    - **Turbine_ID:** Unique identifier for each wind turbine.
    - **Wind_Speed (m/s):** The speed of the wind at the time of measurement.
    - **Ambient_Temperature (°C):** The outside temperature where the turbine is located.
    - **Power_Output (MW):** The power generated by the turbine at the time of measurement.
    - **Blade_Vibration (mm/s):** The vibration level of the turbine blades, a critical indicator of potential mechanical issues.
    - **Gearbox_Temperature (°C):** Temperature inside the turbine's gearbox.
    - **Oil_Level (%):** The level of lubrication oil in the turbine.
    - **Maintenance_History (0/1):** Whether the turbine underwent maintenance in the last 30 days (1) or not (0).
    - **Component_Failure (0/1):** Whether a component failure occurred (1) or not. This is the target variable.
    """)

    # Data Exploration
    st.write("### Data Exploration")

    # Select Graph Type
    graph_type = st.selectbox("Select a graph to visualize:", 
                              ["Wind Speed Distribution", 
                               "Ambient Temperature vs. Power Output", 
                               "Blade Vibration vs. Gearbox Temperature", 
                               "Oil Level Distribution"])

    # Plotting based on selection
    if graph_type == "Wind Speed Distribution":
        st.write("#### Wind Speed Distribution")
        plt.figure(figsize=(10, 6))
        plt.hist(data['Wind_Speed'], bins=20, color='skyblue', edgecolor='black')
        plt.xlabel('Wind Speed (m/s)')
        plt.ylabel('Frequency')
        plt.title('Distribution of Wind Speed')
        st.pyplot(plt)
        st.write("This graph shows the distribution of wind speeds recorded in the dataset. "
                 "It helps in understanding the common wind speeds at which the turbines operate, "
                 "which can be crucial for assessing turbine performance and predicting failures.")

    elif graph_type == "Ambient Temperature vs. Power Output":
        st.write("#### Ambient Temperature vs. Power Output")
        plt.figure(figsize=(10, 6))
        plt.scatter(data['Ambient_Temperature'], data['Power_Output'], color='green', alpha=0.6)
        plt.xlabel('Ambient Temperature (°C)')
        plt.ylabel('Power Output (MW)')
        plt.title('Ambient Temperature vs. Power Output')
        st.pyplot(plt)
        st.write("This scatter plot illustrates the relationship between ambient temperature and power output. "
                 "It shows how temperature variations can affect the power generated by the wind turbines, "
                 "which is important for understanding operational efficiency under different weather conditions.")

    elif graph_type == "Blade Vibration vs. Gearbox Temperature":
        st.write("#### Blade Vibration vs. Gearbox Temperature")
        plt.figure(figsize=(10, 6))
        plt.scatter(data['Blade_Vibration'], data['Gearbox_Temperature'], color='orange', alpha=0.6)
        plt.xlabel('Blade Vibration (mm/s)')
        plt.ylabel('Gearbox Temperature (°C)')
        plt.title('Blade Vibration vs. Gearbox Temperature')
        st.pyplot(plt)
        st.write("This scatter plot shows the correlation between blade vibration and gearbox temperature. "
                 "High levels of blade vibration can indicate mechanical issues, and this plot helps in identifying "
                 "potential stress on the gearbox due to vibrations.")

    elif graph_type == "Oil Level Distribution":
        st.write("#### Oil Level Distribution")
        plt.figure(figsize=(10, 6))
        plt.hist(data['Oil_Level'], bins=20, color='purple', edgecolor='black')
        plt.xlabel('Oil Level (%)')
        plt.ylabel('Frequency')
        plt.title('Distribution of Oil Level')
        st.pyplot(plt)
        st.write("This histogram displays the distribution of oil levels in the turbines. "
                 "Maintaining optimal oil levels is crucial for the smooth operation of turbine components. "
                 "This graph helps in understanding how often turbines operate with low or high oil levels, "
                 "which can impact their reliability and longevity.")

    # Predictive Maintenance
    st.write("### Predict Equipment Failure")
    user_input = {
        'Wind_Speed': st.number_input('Wind Speed (m/s)'),
        'Ambient_Temperature': st.number_input('Ambient Temperature (°C)'),
        'Power_Output': st.number_input('Power Output (MW)'),
        'Blade_Vibration': st.number_input('Blade Vibration (mm/s)'),
        'Gearbox_Temperature': st.number_input('Gearbox Temperature (°C)'),
        'Oil_Level': st.number_input('Oil Level (%)'),
        'Maintenance_History': st.selectbox('Maintenance History (0/1)', [0, 1])
    }

    input_df = pd.DataFrame([user_input])

    # Load pre-trained model (assumed pre-trained, you should load it here)
    model = RandomForestClassifier()  # This is where you would load your trained model
    model.fit(data.drop(columns=['Component_Failure', 'Timestamp', 'Turbine_ID']), data['Component_Failure'])  # For demonstration only

    prediction = model.predict(input_df)
    st.write(f"Predicted Component Failure: {'Yes' if prediction[0] == 1 else 'No'}")

# Intelligent Energy Management Systems
elif page == "Intelligent Energy Management Systems":
    st.header("Intelligent Energy Management Systems")

    # Load the datasets
    energy_df = load_data('energy_consumption.csv')
    production_df = load_data('production_schedule.csv')
    pricing_df = load_data('energy_pricing.csv')

    # Function to display dataset metadata
    def display_metadata(df, title):
        st.subheader(f"{title} Metadata")
        st.write(f"Number of records: {df.shape[0]}")
        st.write(f"Number of columns: {df.shape[1]}")
        st.write("Columns:")
        for col in df.columns:
            st.write(f"- **{col}**: {df[col].dtype}")

    st.write("""
    This app demonstrates AI-powered systems that can dynamically manage and allocate energy resources in industrial settings, optimizing energy consumption in real time.
    """)

    # Display dataset metadata information
    st.write("### Dataset Metadata Information")

    st.write("""
    #### Energy Consumption Data:
    This dataset records the energy consumption of different machines over time, along with associated parameters like temperature, humidity, and operational status.
    """)
    display_metadata(energy_df, "Energy Consumption")

    st.write("""
    #### Production Schedule Data:
    This dataset contains the production schedules for different machines, including the start and end times of production, the type of product being produced, and the target output.
    """)
    display_metadata(production_df, "Production Schedule")

    st.write("""
    #### Energy Pricing Data:
    This dataset provides the energy pricing information over time, including indicators for peak hours, which can influence energy consumption optimization strategies.
    """)
    display_metadata(pricing_df, "Energy Pricing")

    # Visualization: Energy consumption over time for a selected machine
    st.header("Energy Consumption Analysis by Machine")

    selected_machine = st.selectbox("Select a Machine ID", energy_df['machine_id'].unique())

    machine_energy_df = energy_df[energy_df['machine_id'] == selected_machine]

    fig, ax = plt.subplots()
    ax.plot(pd.to_datetime(machine_energy_df['timestamp']), machine_energy_df['energy_consumed_kWh'], label='Energy Consumption (kWh)', color='blue')
    ax.set_xlabel('Timestamp')
    ax.set_ylabel('Energy Consumed (kWh)')
    ax.legend()

    st.pyplot(fig)

    st.write("""
    This graph shows the energy consumption for the selected machine over time.
    """)

    # AI-Powered Forecasting
    st.header("Energy Consumption Forecasting")

    # Prepare the data for Prophet
    forecast_data = machine_energy_df[['timestamp', 'energy_consumed_kWh']].rename(columns={'timestamp': 'ds', 'energy_consumed_kWh': 'y'})

    # Initialize Prophet model
    model = Prophet()
    model.fit(forecast_data)

    # Create future dataframe
    future = model.make_future_dataframe(periods=48, freq='H')  # Forecasting the next 48 hours

    # Forecast
    forecast = model.predict(future)

    # Plot the forecast
    st.subheader("Forecasting Results")
    fig1 = model.plot(forecast)
    st.pyplot(fig1)

    fig2 = model.plot_components(forecast)
    st.pyplot(fig2)

    st.write("""
    The above charts display the forecasted energy consumption for the selected machine over the next 48 hours, along with the trend and seasonality components.
    """)