Untitled

# Initialize variables to store the start and end times
start_time = None
end_time = None


#DfB has only values for cement type cement_type_cpiv_bool_pv and the mill in normal operation conditions

valid_indices = dfB[(~dfB['separator_pwr_kw_pv'].isna()) &
                    # (~dfB['mill_dp_mbar_pv'].isna()) &
                    (~dfB['separator_speed_rpm_pv'].isna())].index

valid_periods = []


# Iterate through each row of the dataset do define periods where we have continouys valid subsets of the dataset
for index, row in dfB.iterrows():
    if start_time is None:
        start_time = index
        flag=True
        
    if (index in valid_indices) and flag==True:

        duration = index - start_time

        if duration.total_seconds() == 60 * 60:
            end_time = index
            valid_periods.append((start_time, end_time))

    else:
        flag=False
        start_time = None
        
        
        
def check_cross_corr(column,df_temp,start_time,end_time,m_type,max_lag=30):
    df_temp = df_temp[start_time:end_time]

    speed_normalized = (df_temp[column] - np.mean(df_temp[column])) / np.std(df_temp[column])
    power_normalized = (df_temp['separator_pwr_kw_pv'] - np.mean(df_temp['separator_pwr_kw_pv'])) / np.std(df_temp['separator_pwr_kw_pv'])

    cross_corr = np.correlate(speed_normalized, power_normalized, mode='full')
    lags = np.arange(-max_lag, max_lag + 1)
    
    if m_type == 'MAX':
        max_corr_index = np.argmax(cross_corr[:len(lags)])  # Index of the maximum correlation
        max_corr_lag = lags[max_corr_index]  # Lag corresponding to the maximum correlation
        return start_time ,end_time ,np.max(cross_corr[:len(lags)]), max_corr_lag       
    else:
        min_corr_index = np.argmin(cross_corr[:len(lags)])  # Index of the minimum correlation
        min_corr_lag = lags[min_corr_index]  # Lag corresponding to the maximum correlation
        return start_time ,end_time ,np.min(cross_corr[:len(lags)]), min_corr_lag
        
        
        
        
lags = [] 
    
for period in valid_periods:
    result_tuple = check_cross_corr('separator_speed_rpm_pv',dfB,period[0],period[1],m_type = 'MAX', max_lag=60)
    lags.append(result_tuple[3])



    
plt.hist(lags, bins=50)
plt.show()

most_common_value = max(set(lags), key=lags.count)
print(most_common_value)