AHU Static Pressure Duration Comparison

This example demonstrates how to chart AHU static pressure duration for a time period by AHU. Load your data from BDXpy, select a timeframe and revise any variables or chart formatting.

Output

Show Code

import pandas as pd
import numpy as np
import plotly.graph_objects as go

# Parameters for dummy data
np.random.seed(0)  # For reproducibility
num_ahus = 5
ahus = [f'AHU_{i+1}' for i in range(num_ahus)]
date_range = pd.date_range(start='2023-10-01', end='2023-10-31', freq='H')
num_samples = len(date_range)

# Generate random static pressure data with different transformations for each AHU
data = {
    'datetime': np.tile(date_range, num_ahus),
    'AHU': np.repeat(ahus, num_samples),
    'static_pressure': []
}

# Apply different transformations to each AHU for variation
for i, ahu in enumerate(ahus):
    if i == 0:  # AHU 1 - Flat line with slight random variation
        pressures = np.random.uniform(1.5, 2.0, num_samples)
    elif i == 1:  # AHU 2 - Logarithmic increase
        pressures = np.logspace(0.1, 1, num_samples) / 10 + np.random.uniform(-0.1, 0.1, num_samples)
    elif i == 2:  # AHU 3 - Polynomial curve (quadratic)
        pressures = (np.linspace(1, 3, num_samples) ** 2) / 5 + np.random.uniform(-0.1, 0.1, num_samples)
    elif i == 3:  # AHU 4 - Polynomial curve (cubic)
        pressures = (np.linspace(1, 3, num_samples) ** 3) / 10 + np.random.uniform(-0.1, 0.1, num_samples)
    else:  # AHU 5 - Linear with slight random variation and top values increased
        pressures = np.linspace(1.0, 3.0, num_samples) + np.random.uniform(-0.1, 0.1, num_samples)
        top_5_percent = int(0.05 * num_samples)
        pressures[:top_5_percent] += np.linspace(0.5, 1.0, top_5_percent)  # Slight increase in top 5%

    data['static_pressure'].extend(pressures)

# Create the DataFrame
df = pd.DataFrame(data)

# Function to generate the duration curve with binned percent exceedance for AHU static pressure
def generate_binned_duration_curve(df, ahus, bin_width=5):
    duration_data = []
    for ahu in ahus:
        ahu_data = df[df['AHU'] == ahu].copy()
        ahu_data = ahu_data.sort_values(by='static_pressure', ascending=False).reset_index(drop=True)

        # Calculate percent exceedance and bin it
        ahu_data['percent_exceedance'] = np.floor(np.arange(1, len(ahu_data) + 1) / len(ahu_data) * 100 / bin_width) * bin_width
        # Group by binned percent_exceedance and calculate mean only on numeric columns
        binned_ahu_data = ahu_data.groupby('percent_exceedance', as_index=False)['static_pressure'].mean()
        binned_ahu_data['AHU'] = ahu  # Retain AHU identifier
        duration_data.append(binned_ahu_data)
    return pd.concat(duration_data, ignore_index=True)

# Generate binned duration curve data
duration_df = generate_binned_duration_curve(df, ahus, bin_width=5)

# Plot the duration curve as overlapping bar charts with wider bins
def plot_binned_duration_curve(duration_df, ahus):
    fig = go.Figure()
    for ahu in ahus:
        ahu_duration = duration_df[duration_df['AHU'] == ahu]
        fig.add_trace(go.Bar(
            x=ahu_duration['percent_exceedance'],
            y=ahu_duration['static_pressure'],
            name=ahu,
            opacity=0.6
        ))
    fig.update_layout(
        title="Binned Duration Curve for AHU Static Pressure (Overlapping Bar Charts)",
        xaxis_title="Percent Exceedance (%)",
        yaxis_title="Static Pressure (inches of water)",
        barmode='overlay',
        legend_title="AHU",
        template="plotly_white"
    )
    fig.show()

# Run the plot function to save the file as HTML
plot_binned_duration_curve(duration_df, ahus)