diff --git a/浙江电压等级电量/8_dy3.pth b/浙江电压等级电量/8_dy3.pth index ea83fd0..2621687 100644 Binary files a/浙江电压等级电量/8_dy3.pth and b/浙江电压等级电量/8_dy3.pth differ diff --git a/浙江电压等级电量/电压等级_输出为3.py b/浙江电压等级电量/电压等级_输出为3.py index 598cfbf..125b537 100644 --- a/浙江电压等级电量/电压等级_输出为3.py +++ b/浙江电压等级电量/电压等级_输出为3.py @@ -4,24 +4,166 @@ import numpy as np import matplotlib.pyplot as plt from torch import nn import os +from torch.utils.data import TensorDataset,DataLoader +import datetime torch.manual_seed(42) os.environ["KMP_DUPLICATE_LIB_OK"]="TRUE" # 解决OMP: Error #15: Initializing libiomp5md.dll, but found libiomp5md.dll already initialized. - +pd.set_option('display.width',None) class LSTM(nn.Module): def __init__(self,input_size,hidden_size,output_size,num_layers=2): super().__init__() - self.lstm = nn.LSTM(input_size,hidden_size,num_layers=2) - self.fc1 = nn.Linear(hidden_size,64) - self.ReLu = nn.ReLU() - self.fc2 = nn.Linear(64,output_size) - + self.lstm = nn.LSTM(input_size,hidden_size,num_layers) + self.fc1 = nn.Linear(hidden_size,output_size) + # self.ReLu = nn.ReLU() + # self.fc2 = nn.Linear(64,output_size) def forward(self,x): output,_ = self.lstm(x) s,b,h = output.shape output = output.reshape(-1,h) - output = self.ReLu(self.fc1(output)) - output = self.fc2(output) + output = self.fc1(output) + # output = self.ReLu(self.fc1(output)) + # output = self.fc2(output) return output + +def create_data(df_level,volt_level): + dataset_x = [] + dataset_y = [] + # 按月份分组 + grouped = df_level.groupby(df_level['stat_date'].dt.to_period('M')) + + # 遍历每个月的数据 + for name, group in grouped: + if len(group) == 31: + dataset_x.append(list(group[volt_level].values[1:28])) + dataset_y.append(list(group[volt_level].values[-3:])) + if len(group) == 30: + dataset_x.append(list(group[volt_level].values[:27])) + dataset_y.append(list(group[volt_level].values[-3:])) + if len(group) == 28: + fst = group[volt_level].values[0] + + dataset_x.append([fst,fst,fst]+list(group[volt_level].values[1:25])) + dataset_y.append(list(group[volt_level].values[-3:])) + else: + fst = group[volt_level].values[0] + if len([fst, fst]+list(group[volt_level].values[1:26])) != 27: + break + dataset_x.append([fst, fst]+list(group[volt_level].values[1:26])) + dataset_y.append(list(group[volt_level].values[-3:])) + + return np.array(dataset_x),np.array(dataset_y) + # 创建数据集 +file_dir = './浙江各地市分电压日电量数据' +print(os.listdir(file_dir)) +city1 = os.listdir(file_dir)[0] +df_city = pd.read_excel(os.path.join(file_dir,city1)).drop(columns='地市') +df_city = df_city[['stat_date','1-10kv','110kv(含66kv)','35kv']] +df_city[['1-10kv','110kv(含66kv)','35kv']] /= 10000 +df_city.stat_date = pd.to_datetime(df_city.stat_date) + +volt_level = '1-10kv' +df_level = df_city[['stat_date',volt_level]] +dataset_x,dataset_y = create_data(df_level,volt_level) + +for volt_level in df_city.columns[2:]: + df_level = df_city[['stat_date',volt_level]] + x,y = create_data(df_level,volt_level) + dataset_x = np.concatenate([dataset_x,x]) + dataset_y = np.concatenate([dataset_y,y]) + + +for excel in os.listdir(file_dir)[1:]: + df_city = pd.read_excel(os.path.join(file_dir, excel)).drop(columns='地市') + df_city = df_city[['stat_date', '1-10kv', '110kv(含66kv)', '35kv']] + df_city[['1-10kv', '110kv(含66kv)', '35kv']] /= 10000 + df_city.stat_date = pd.to_datetime(df_city.stat_date) + for volt_level in df_city.columns[1:]: + df_level = df_city[['stat_date', volt_level]] + x, y = create_data(df_level, volt_level) + dataset_x = np.concatenate([dataset_x, x]) + dataset_y = np.concatenate([dataset_y, y]) + + +print(dataset_x.shape,dataset_y.shape) + +device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') +# 标准化到0~1 +max_value = np.max(dataset_x) +min_value = np.min(dataset_x) +dataset_x = (dataset_x - min_value) / (max_value - min_value) +dataset_y = (dataset_y - min_value) / (max_value - min_value) + +print(max_value,min_value) + +# 划分训练集和测试集 +train_size = int(len(dataset_x)*0.8) +train_x = dataset_x[:train_size] +train_y = dataset_y[:train_size] +eval_x = dataset_x[train_size:] +eval_y = dataset_y[train_size:] + +# # 将数据改变形状，RNN 读入的数据维度是 (seq_size, batch_size, feature_size) +train_x = train_x.reshape(-1, 1, 27) +train_y = train_y.reshape(-1, 1, 3) +eval_x = eval_x.reshape(-1, 1, 27) +eval_y = eval_y.reshape(-1, 1, 27) + +# # 转为pytorch的tensor对象 +train_x = torch.from_numpy(train_x).to(device).type(torch.float32) +train_y = torch.from_numpy(train_y).to(device).type(torch.float32) +eval_x = torch.from_numpy(eval_x).to(device).type(torch.float32) + +train_ds = TensorDataset(train_x,train_y) +train_dl = DataLoader(train_ds,batch_size=32,drop_last=True) +eval_ds = TensorDataset(eval_x,eval_y) +eval_dl = DataLoader(eval_ds,batch_size=64,drop_last=True) + + +model = LSTM(27, 16, output_size=3, num_layers=2).to(device) # 导入模型并设置模型的参数输入输出层、隐藏层等 + +train_loss = [] +loss_function = nn.MSELoss() +optimizer = torch.optim.Adam(model.parameters(), lr=0.001) +min_loss = 1 +for i in range(200): + for j,(x,y) in enumerate(train_dl): + x,y = x.to(device),y.to(device) + out = model(x) + loss = loss_function(out, y) + loss.backward() + optimizer.step() + optimizer.zero_grad() + train_loss.append(loss.item()) + # if (i+1) % 100 == 0: + # print(f'epoch {i+1}/1500 loss:{round(loss.item(),5)}') + if (j + 1) % 50 == 0: + print(f'epoch {i+1}/200 step {j+1}/{len(train_dl)} loss:{loss}' ) + model.eval() + for k,(x,y) in enumerate(eval_dl): + pred = model(eval_x) + + +# 保存模型 +torch.save(model.state_dict(),'dy3.pth') + +# for test +model = model.eval() + +dataset_x = dataset_x.reshape(-1, 1, 27) # (seq_size, batch_size, feature_size) +dataset_x = torch.from_numpy(dataset_x).to(device).type(torch.float32) + +pred_test = model(dataset_x) # 全量训练集 +# 模型输出 (seq_size, batch_size, output_size) +pred_test = pred_test.view(-1).cpu().detach().numpy() + +plt.plot(pred_test.reshape(-1), 'r', label='prediction') +plt.plot(dataset_y.reshape(-1), 'b', label='real') +plt.plot((train_size*3, train_size*3), (0, 1), 'g--') # 分割线 左边是训练数据 右边是测试数据的输出 +plt.legend(loc='best') +plt.show() + + +