Source code for RCAIDE.Library.Methods.Noise.Metrics.EPNL_noise_metric

# RCAIDE/Methods/Noise/Metrics/EPNL_noise_metric.py
# 
# 
# Created:  Jul 2023, M. Clarke  

# ----------------------------------------------------------------------------------------------------------------------
#  IMPORT
# ----------------------------------------------------------------------------------------------------------------------

# Python package imports   
import numpy as np  
    
# ----------------------------------------------------------------------------------------------------------------------  
#  EPNL_noise_metric
# ----------------------------------------------------------------------------------------------------------------------        


def EPNL_noise_metric(PNLT):
    """This method calculates the effective perceived noise level (EPNL) based on a
    time history Perceived Noise Level with Tone Correction (PNLT).
     
    Assumptions:
        N/A

    Source:
        N/A

    Inputs:
        PNLT - Perceived Noise Level with Tone Correction  [PNLdB]
     
     Outputs:
        EPNL - Effective Perceived Noise Level             [EPNdB]
     
    Properties Used:
        N/A  
    """           
    # Maximum PNLT on the time history data    
    PNLT_max = np.max(PNLT,axis=0)
    n_mic_x  = len(PNLT[0,:, 0])
    n_mic_y  = len(PNLT[0,0,:])
    
    # Calculates the number of discrete points on the trajectory
    nsteps   = len(PNLT)    
    EPNL     = np.zeros((n_mic_x,n_mic_y))
    for n_x in range(n_mic_x):
        for n_y in range(n_mic_y):
            # Finding the time duration for the noise history where PNL is higher than the maximum PNLT - 10 dB
            i = 0
            while PNLT[i][n_x][n_y]<=(PNLT_max[n_x][n_y]-10) and i<=nsteps:
                i = i+1
            t1 = i #t1 is the first time interval
            i  = i+1
        
            # Correction for PNLTM-10 when it falls outside the limit of the data
            if PNLT[nsteps-1][n_x][n_y]>=(PNLT_max[n_x][n_y]-10):
                t2=nsteps-2
            else:
                while i<=nsteps and PNLT[i][n_x][n_y]>=(PNLT_max[n_x][n_y]-10):
                    i = i+1
                t2 = i-1 #t2 is the last time interval 
            
            # Calculates the integral of the PNLT which between t1 and t2 points
            sumation = 0
            for i in range (t1-1,t2+1):
                sumation = 10**(PNLT[i][n_x][n_y]/10)+sumation
                
            # Duration Correction calculation
            duration_correction = 10*np.log10(sumation)-PNLT_max[n_x][n_y]-13
                        
            # Final EPNL calculation
            EPNL[n_x][n_y] = PNLT_max[n_x][n_y]+duration_correction
    
    return EPNL