The identification of suspicious microseismic events is the first crucial step in processing microseismic data. In this paper, we present an automatic classification method based on a deep learning approach for classifying microseismic records in underground mines. A total of 35 commonly used features in the time and frequency domains were extracted from waveforms. To examine the discriminative ability of these features, a genetic algorithm (GA)-optimized correlation-based feature selection (CFS) method was applied. As a result, 11 features were selected to represent microseismic records. By dividing each microseismic record into 50 frames, an 11×50 feature matrix was utilized as the input. A convolutional neural network (CNN) with 35 layers was trained on 20,000 samples recorded at the Huangtupo Copper and Zinc Mine. There are 5 types of events: microseismic events, blasting, ore extraction, mechanical noise, and electromagnetic interference. The event type was correctly determined by the trained CNN classifier 98.2% of the time, outperforming traditional machine learning methods.

Microseismic monitoring is an effective technique to ensure the safety of rock mass engineering. Moreover, P-wave arrival picking is crucial in the seismic/microseismic monitoring process. The existing methods of P-wave arrival picking are not fully qualified for practical application because they are mostly semiautomatic or need too much training data. To overcome the shortcoming of today's most elaborate methods, we leverage the recent advances in artificial intelligence and present PickCapsNet, a highly scalable capsule network for P-wave arrival picking from a single waveform without feature extraction. We apply the PickCapsNet to study the induced microseismic events in Dongguashan Copper Mine, China, and compare it with Akaike information criterion (AIC), short- and long-time average ratio (STA/LTA), and convolutional neural network (CNN). The differences between the PickCapsNet and manual picks have a mean value of 0.0023 s and a standard deviation of 0.0033 s; moreover, 97.46% of the picks are within 0.01 s of the manual pick. Furthermore, at different signal-to-noise ratios (SNRs), it has a higher accuracy and stability than other methods. These results indicate that the proposed method is of high picking precision and robustness.

The identification of suspicious microseismic events is the first crucial step in microseismic data processing. Existing automatic classification methods are based on the training of a large data set, which is challenging to apply in mines without a long-term manual data processing. In this paper, we present a method to automatically classify microseismic records with limited samples in underground mines based on capsule networks (CapsNet). We divide each microseismic record into 33 frames, then extract 21 commonly used features in time and frequency from each frame. Consequently, a 21 × 33 feature matrix is utilized as the input of CapsNet. On this basis, we use different sizes of training sets to train the classification models separately. The trained model is tested using the same test set containing 3,200 microseismic records and compared to convolutional neural networks (CNN) and traditional machine learning methods. Results show that the accuracy of our proposed method is 99.2% with limited training samples. It is superior to CNN and traditional machine learning methods in terms of Accuracy, Precision, Recall, F1-Measure, and reliability.

In order to mitigate economic and safety risks during mine life, a microseismic monitoring system is installed in a number of underground mines. The basic step for successfully analyzing those microseismic data is the correct detection of various event types, especially the rock mass rupture events. The visual scanning process is a time‐consuming task and requires experience. Therefore, here we present a new method for automatic classification of microseismic signals based on the Gaussian Mixture Model‐Hidden Markov Model (GMM‐HMM) by using only Mel‐frequency cepstral coefficient (MFCC) features extracted from the waveform. The detailed implementation of our proposed method is described. The performance of this method is tested by its application to microseismic events selected from the Dongguashan Copper Mine (China). A dataset that contains a representative set of different microseismic events including rock mass rupture, blasting vibration, mechanical drilling, and electromagnetic noise is collected for training and testing. The results show that our proposed method obtains an accuracy of 92.46%, which demonstrates the effectiveness of the method for automatic classification of microseismic data in underground mines.