工业大数据竞赛(基于深度学习的模型)

最新推荐文章于 2024-07-18 15:51:43 发布

wolfrosneT

最新推荐文章于 2024-07-18 15:51:43 发布

阅读量1.3k

点赞数 2

本文链接：https://blog.csdn.net/qq_40773753/article/details/94767344

版权

本次博客将分享一次工业大数据竞赛的经历，重点探讨如何运用深度学习技术进行模型构建和优化。通过对海量数据的预处理、特征工程和模型训练，最终实现对复杂工业问题的精准预测。

摘要由CSDN通过智能技术生成

#数据处理代码

import pandas as pd
import numpy as np

#提取M1数据
pinlv=[]
#M_number=['M1','M2','M3','M4','M5']
M_xilie=['a','b','c','d','e']
boxing=['压缩机非联端X','压缩机非联端Y','压缩机联端X','压缩机联端Y','压缩机轴位移A(波形)','压缩机轴位移B(波形)','压缩机轴位移C(波形)']

fenlei=pd.DataFrame({'压缩机非联端X':[[]],'压缩机非联端Y':[[]],'压缩机联端X':[[]],'压缩机联端Y':[[]],'压缩机轴位移A(波形)':[[]],\
                     '压缩机轴位移B(波形)':[[]],'压缩机轴位移C(波形)':[[]],'label':[[]]})

file_all_count=[]
k=0
import os
for h in range(5):
    for i in range(7):
        file_count=len(os.listdir('D:\训练数据\M1\M1%s\%s'%(M_xilie[h],boxing[i])))
        file_all_count.append(file_count)

for g in range(5):
    for h in range(7):
        for i in range(file_all_count[k]):
            df=pd.read_csv('D:\训练数据\M1\M1%s\%s\wave_%s.csv' %(M_xilie[g],boxing[h],(i+1)),skiprows=5)
            df=df.drop(['wave'],axis=1)    
            #df.loc[i+1]=list(data)
            listdata=list(df)
            listdata=[x[:4] for x in listdata]
            listdata=[float(x) for x in listdata]

            df2=pd.read_csv('D:\训练数据\M1\M1%s\%s\wave_%s.csv' %(M_xilie[g],boxing[h],(i+1)),nrows=1,header=None)
            freq=df2.loc[0,1]

            listdata2=[x*freq for x in listdata]

            pinlv=pinlv+listdata2    

        #import array
        #int16重要!
        #pinlv=np.array(pinlv,dtype=np.int16)
        #import wave
        #import numpy as np
        #import scipy.signal as signal
        #import scipy.io.wavfile

        #scipy.io.wavfile.write('D:\训练数据\M1\M1a\%s.wav' %(boxing[h]),44100,pinlv)   #7.2

        fenlei.loc[g,['%s'%(boxing[h])]]=[pinlv]
        pinlv=[]
        k=k+1
fenlei.to_csv('d:\训练数据\M1.csv')
import pandas as pd
import numpy as np

#提取M2数据
pinlv=[]
#M_number=['M1','M2','M3','M4','M5']
M_xilie=['a','b','c','d','e']
boxing=['压缩机非联端X','压缩机非联端Y','压缩机联端X','压缩机联端Y','压缩机轴位移A(波形)','压缩机轴位移B(波形)']

fenlei=pd.DataFrame({'压缩机非联端X':[[]],'压缩机非联端Y':[[]],'压缩机联端X':[[]],'压缩机联端Y':[[]],'压缩机轴位移A(波形)':[[]],\
                     '压缩机轴位移B(波形)':[[]],'压缩机轴位移C(波形)':[[]],'label':[[]]})

file_all_count=[]
k=0
import os
for h in range(5):
    for i in range(6):
        file_count=len(os.listdir('D:\训练数据\M2\M2%s\%s'%(M_xilie[h],boxing[i])))
        file_all_count.append(file_count)

for g in range(5):
    for h in range(6):
        for i in range(file_all_count[k]):
            df=pd.read_csv('D:\训练数据\M2\M2%s\%s\wave_%s.csv' %(M_xilie[g],boxing[h],(i+1)),skiprows=5)
            df=df.drop(['wave'],axis=1)    
            #df.loc[i+1]=list(data)
            listdata=list(df)
            listdata=[x[:4] for x in listdata]
            listdata=[float(x) for x in listdata]

            df2=pd.read_csv('D:\训练数据\M2\M2%s\%s\wave_%s.csv' %(M_xilie[g],boxing[h],(i+1)),nrows=1,header=None)
            freq=df2.loc[0,1]

            listdata2=[x*freq for x in listdata]

            pinlv=pinlv+listdata2    

        #import array
        #int16重要!
        #pinlv=np.array(pinlv,dtype=np.int16)
        #import wave
        #import numpy as np
        #import scipy.signal as signal
        #import scipy.io.wavfile

        #scipy.io.wavfile.write('D:\训练数据\M1\M1a\%s.wav' %(boxing[h]),44100,pinlv)   #7.2

        fenlei.loc[g,['%s'%(boxing[h])]]=[pinlv]
        pinlv=[]
        k=k+1
fenlei.to_csv('d:\训练数据\M2.csv')
import pandas as pd
import numpy as np

#提取M3数据
pinlv=[]
#M_number=['M1','M2','M3','M4','M5']
M_xilie=['a','b','c','d','e']
boxing=['压缩机非联端X','压缩机非联端Y','压缩机联端X','压缩机联端Y','压缩机轴位移A(波形)','压缩机轴位移B(波形)','压缩机轴位移C(波形)']

fenlei=pd.DataFrame({'压缩机非联端X':[[]],'压缩机非联端Y':[[]],'压缩机联端X':[[]],'压缩机联端Y':[[]],'压缩