创建光伏出力预测模型

pv/pv_train.py

@@ -0,0 +1,35 @@
+import os
+os.environ["CUDA_VISIBLE_DEVICES"] = "-1"
+
+import pandas as pd
+import numpy as np
+
+def time_series_to_supervised(data, n_in=10, n_out=1,dropnan=True):
+    """
+    :param data:作为列表或2D NumPy数组的观察序列。需要。
+    :param n_in:作为输入的滞后观察数（X）。值可以在[1..len（数据）]之间可选。默认为1。
+    :param n_out:作为输出的观测数量（y）。值可以在[0..len（数据）]之间。可选的。默认为1。
+    :param dropnan:Boolean是否删除具有NaN值的行。可选的。默认为True。
+    :return:
+    """
+    n_vars = 1 if type(data) is list else data.shape[1]
+    df = pd.DataFrame(data)
+    origNames = df.columns
+    cols, names = list(), list()
+    cols.append(df.shift(0))
+    names += [('%s' % origNames[j]) for j in range(n_vars)]
+    n_in = max(1, n_in)
+    for i in range(n_in-1, 0, -1):
+        time = '(t-%d)' % i
+        cols.append(df.shift(i))
+        names += [('%s%s' % (origNames[j], time)) for j in range(n_vars)]
+    n_out = max(n_out, 0)
+    for i in range(1, n_out+1):
+        time = '(t+%d)' % i
+        cols.append(df.shift(-i))
+        names += [('%s%s' % (origNames[j], time)) for j in range(n_vars)]
+    agg = pd.concat(cols, axis=1)
+    agg.columns = names
+    if dropnan:
+        agg.dropna(inplace=True)
+    return agg

pv/vmd_decom.py

@@ -0,0 +1,58 @@
+import numpy as np
+import pandas as pd
+import numpy as np
+from vmdpy import VMD
+from skopt import gp_minimize
+from skopt.space import Real, Integer, Categorical
+from skopt.utils import use_named_args
+
+
+# 加载数据
+def do_vmd(dataframe):
+    signal = dataframe['power'].values
+    # 定义VMD函数，用于优化过程
+    def vmd_objective(params):
+        alpha, tau, K, DC, init, tol = params
+        u, u_hat, omega = VMD(signal, alpha, tau, K, DC, init, tol)
+        # 定义目标函数，这里使用模态的平均带宽作为优化目标
+        # 你可以根据实际需求定义其他目标函数
+        objective = np.mean([np.diff(np.where(um != 0)) for um in u])
+        return -objective  # gp_minimize用于最小化，所以取负值
+
+    # 定义参数空间
+    space = [
+        Real(low=2, high=100, prior='log-uniform', name='alpha'),
+        Real(low=0, high=1, name='tau'),
+        Integer(low=2, high=10, name='K'),
+        Categorical(categories=[0, 1], name='DC'),
+        Categorical(categories=[0, 1], name='init'),
+        Real(low=1e-6, high=1e-2, prior='log-uniform', name='tol')
+    ]
+
+    # 使用贝叶斯优化
+    @use_named_args(space)
+    def objective_function(alpha, tau, K, DC, init, tol):
+        return vmd_objective((alpha, tau, K, DC, init, tol))
+
+    result = gp_minimize(objective_function, space, n_calls=50, random_state=0)
+
+    print('最优参数:', result.x)
+    print('最优目标函数值:', -result.fun)  # 取负值回到原始目标函数的值
+
+    # 使用找到的最优参数进行VMD分解
+    alpha_opt, tau_opt, K_opt, DC_opt, init_opt, tol_opt = result.x
+    u_opt, u_hat_opt, omega_opt = VMD(signal, alpha_opt, tau_opt, K_opt, DC_opt, init_opt, tol_opt)
+
+    # 保存或处理分解得到的模态函数
+    best_params = [alpha_opt, tau_opt, K_opt, DC_opt, init_opt, tol_opt]
+    vmd_rst = u_opt.T
+
+
+    vmd_train_data = pd.concat([data2vmd, pd.DataFrame.from_records(vmd_rst, index=data2vmd.index, columns=[f"vmd_{x}"for x in range(vmd_rst.shape[1])])], axis=1)
+    
+    return vmd_train_data
+
+if __name__ == '__main__':
+    data2vmd = pd.read_csv('./data/pv_data_hourly.csv', index_col=0)
+    train_data = do_vmd(data2vmd)
+    train_data.to_csv('./data/vmd_train.csv', index=False, encoding='utf-8-sig')

requirements.txt

@@ -0,0 +1,7 @@
+numpy==2.0.2
+pandas==2.2.3
+EMD_signal==1.6.0
+vmdpy==0.2
+PyEMD==1.0.0
+scikit-optimize==0.10.1
+scikit_learn==1.2.1