WTOAAM/dataAnalysisBusiness/algorithm/powerOscillationAnalyst.py

import os
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt  
import seaborn as sns  
from behavior.analyst import Analyst
from utils.directoryUtil import DirectoryUtil as dir
from algorithmContract.confBusiness import *

class PowerOscillationAnalyst(Analyst):
    """
    风电机组功率震荡分析
    """

    def typeAnalyst(self):
        return "power_diff"

    def turbineAnalysis(self,
                 dataFrame,
                 outputAnalysisDir,
                 outputFilePath,
                 confData: ConfBusiness,
                 turbineName):
        self.power_diff(dataFrame, outputFilePath,
                        confData.field_power,confData.field_rotor_speed)

    def power_diff(self, dataFrame, outputFilePath,  field_Active_Power, field_Rotor_Speed):
        # Floor the power column to the nearest 10
        dataFrame['power_col_floor'] = (
            dataFrame[field_Active_Power] // 10 * 10).astype('int32')

        # Group by the floored power column
        grouped = dataFrame.groupby('power_col_floor')

        # Calculate max, min, and diff of generator speed within each group
        agg_df = grouped[field_Rotor_Speed].agg(
            speed_max='max',
            speed_min='min'
        )
        agg_df['speed_diff'] = agg_df['speed_max'] - agg_df['speed_min']

        # Sort by the floored power column
        agg_df = agg_df.sort_index()

        # Write the result to a CSV file
        agg_df.to_csv(outputFilePath)

    def turbinesAnalysis(self,dataFrameMerge,outputAnalysisDir, confData: ConfBusiness):
        self.plot_power_oscillations(outputAnalysisDir,confData.farm_name)
  
    def plot_power_oscillations(self, csvFileDirOfCp, farm_name, encoding='utf-8'):
        """  
        Plot TSR trend from CSV files in a given input path and save the plots to an output path.  

        Parameters:  
        - csvFileDirOfCp: str, path to the directory containing input CSV files.
        - farm_name: str, name of the wind farm.
        - encoding: str, encoding of the input CSV files. Defaults to 'utf-8'.
        """        
        sns.set_palette('deep')
        field_Name_Turbine = "turbine_name"
        x_name = 'power_col_floor'  
        y_name = 'speed_diff'  

        # 初始化结果DataFrame  
        res = pd.DataFrame()  
        
        # 遍历输入路径下的所有文件  
        for root, dir_names, file_names in dir.list_directory(csvFileDirOfCp):  
            for file_name in file_names:  

                if not file_name.endswith(CSVSuffix):
                    continue

                file_path = os.path.join(root, file_name)  
                frame = pd.read_csv(file_path, encoding=encoding)  
                
                # 获取输出文件名前缀  
                turbine_name = file_name.split(CSVSuffix)[0]  
                # 添加设备名作为新列
                frame[field_Name_Turbine] = turbine_name
                 
                selected_data = frame.loc[:, [field_Name_Turbine, x_name, y_name]]  
                res = pd.concat([res, selected_data], axis=0)  
        
        # 重置索引  
        ress = res.reset_index(drop=True)  
        
        # 绘制所有设备的功率震荡图  
        fig, ax = plt.subplots(figsize=(16, 8))  
        ax = sns.lineplot(x=x_name, y=y_name, data=ress, hue=field_Name_Turbine)  
        ax.set_title('Power-Oscillation')  
        plt.legend(ncol=4)  
        plt.savefig(csvFileDirOfCp+ r'/{}-Power-Oscillation.png'.format(farm_name), bbox_inches='tight', dpi=300)  
        plt.close()  
        
        # 分组绘制每个设备的功率震荡图  
        grouped = ress.groupby(field_Name_Turbine)  
        for name, group in grouped:  
            color = ["lightgrey"] * len(ress[field_Name_Turbine].unique())  
            fig, ax = plt.subplots(figsize=(8, 8))  
            ax = sns.lineplot(x=x_name, y=y_name, data=ress, hue=field_Name_Turbine, palette=sns.color_palette(color), legend=False)  
            ax = sns.lineplot(x=x_name, y=y_name, data=group, color='darkblue', legend=False)  
            ax.set_title('turbine_name={}'.format(name))  
            plt.savefig(csvFileDirOfCp+ r'/{}.png'.format(name), bbox_inches='tight', dpi=120)  
            plt.close()