rnn.py

# -*- coding: utf-8 -*-
"""rnn.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1ew-mmvutdj8KoySKwoBxFiXZq7_SH-9V
"""

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import confusion_matrix
from sklearn.utils import shuffle
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, SimpleRNN
from tensorflow.keras.utils import to_c
# Load data
scada_df = pd.read_csv('scada_data.csv')
scada_df['DateTime'] = pd.to_datetime(scada_df['DateTime'])

status_df = pd.read_csv('status_data.csv')
status_df['Time'] = pd.to_datetime(status_df['Time'])
status_df.rename(columns={'Time': 'DateTime'}, inplace=True)

fault_df = pd.read_csv('fault_data.csv')
fault_df['DateTime'] = pd.to_dat
# Merge data
df_combine = scada_df.merge(fault_df, on='Time', how='outer')
df_combine['Fault'] = df_combine['Fault'].fillna('NF')
df_combine = df_combine.drop(columns=['DateTime_x', 'Time', 'Error', 'WEC: ava. windspeed',
                                      'WEC: ava. available P from wind',
                                      'WEC: ava. available P technical reasons',
                                      'WEC: ava. Available P force majeure reasons',
                                     'WEC: max. windspeed', 'WEC: min. windspeed',
                                      'WEC: Operating Hours', 'WEC: Production kWh',
                                      'WEC: Production minutes', 'DateTime_y'])
etime(fault_df['DateTime'])          'WEC: ava. Available P force external reasons',


# Balancing classes
df_nf = df_combine[df_combine.Fault=='NF'].sample(300, random_state=42)
df_f = df_combine[df_combine.Fault!='NF']
df_combine = pd.concat((df_nf, df_f), axis=0).reset_index(drop=True)

# Prepare data for RNN
X = df_combine.drop(columns=['Fault']).values
y = pd.get_dummies(df_combine['Fault']).values  # One-hot encode fault categories

# Shuffle data
X, y = shuffle(X, y, random_state=42)

# Train-test split
x_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

# Normalize inputs
scaler = StandardScaler()
X_train = scaler.fit_transform(x_train)
X_test = scaler.transform(x_test)

# Reshape data for RNN input
X_train = X_train.reshape(X_train.shape[0], 1, X_train.shape[1])
X_test = X_test.reshape(X_test.shape[0], 1, X_test.shape[1])

# Build RNN model
model = Sequential()
model.add(SimpleRNN(64, input_shape=(X_train.shape[1], X_train.shape[2])))
model.add(Dense(32, activation='relu'))
model.add(Dense(y_train.shape[1], activation='softmax'))

# Compile model
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

# Train model
history = model.fit(X_train, y_train, epochs=50, batch_size=32, validation_data=(X_test, y_test), verbose=2)

# Evaluate model
accuracy = model.evaluate(X_test, y_test, verbose=0)[1]
print("Accuracy: {:.2f}%".format(accuracy * 100))

# Predictions on test data
y_pred = model.predict(X_test)
y_pred_classes = np.argmax(y_pred, axis=1)
y_true_classes = np.argmax(y_test, axis=1)
conf_matrix = confusion_matrix(y_true_classes, y_pred_classes)
# Generate confusion matrix
fault_categories = ['AF', 'EF', 'FF', 'GF', 'MF', 'NF']

# Plot confusion matrix
plt.figure(figsize=(10, 8))
sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues', xticklabels=fault_categories, yticklabels=fault_categories)
plt.xlabel('Predicted Label')
plt.ylabel('True Label')
plt.title('Confusion Matrix')
plt.show()