Probabilistic Neural Network Applications in Coal Mining Analytics

Probabilistic Neural Networks (PNNs) have emerged as a powerful classification tool within coal mining analytics, offering rapid decision support based on a Bayesian framework. By estimating class-conditional probability density functions through non-parametric methods, PNNs enable real-time recognition of operational states and material properties. In coal mining, applications range from identifying shearer cutting patterns and distinguishing coal-rock seams to predicting coal hardness and diagnosing equipment faults. Key advantages include straightforward network topology, fast training through direct use of training samples, and robustness to noise. Optimisation of smoothing parameters and feature extraction methods such as empirical mode decomposition further enhance accuracy. As coal operations increasingly adopt mechanised and automated systems, PNN-based analytics deliver crucial insights for improving safety, maximising resource recovery and reducing downtime.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

Recent studies have advanced PNN applications in coal mining through integration with signal processing and optimisation algorithms. One work employs a modified fruit fly optimisation algorithm to tune the PNN smoothing parameters for diagnosing shearer cutting states via vibration signals from the rocker transmission. Ensemble empirical mode decomposition isolates intrinsic mode functions, and Kullback–Leibler divergence selects state-sensitive components, yielding superior classification accuracy and demonstrating viability for on-site deployment. Another study proposes an online cutting pattern recogniser for shearer drums based on improved ensemble empirical mode decomposition and PNN. End-point continuation mitigates signal boundary effects, while correlation-based IMF selection reduces dimensionality. Energy and standard deviation features feed the PNN, achieving over 92% recognition accuracy in industrial trials. Both examples illustrate how coupling PNNs with adaptive signal decomposition and meta-heuristic optimisers can deliver rapid, reliable pattern identification under harsh mining conditions.

Probabilistic Neural Network Applications in Coal Mining Analytics publication trend

The graph below shows the total number of articles in probabilistic neural network applications in coal mining analytics across all publications each year (not limited to Nature Index journals).

Technical terms

Probabilistic Neural Network (PNN): A feed-forward network that performs classification by estimating class-conditional probability density functions and applying Bayes decision rule.

Smoothing parameter (σ): A bandwidth factor in PNN kernels controlling the degree of generalisation and overlap between class density estimates.

Ensemble Empirical Mode Decomposition (EEMD): A data-driven method that decomposes a non-stationary signal into intrinsic mode functions by averaging results over noise-assisted trials.

Intrinsic Mode Function (IMF): A component signal extracted by empirical mode decomposition, representing oscillatory modes at distinct frequency scales.

Fruit Fly Optimisation Algorithm (FOA): A swarm-intelligence method inspired by foraging behaviour, used to optimise model parameters through population-based search.

References

1. Cutting State Diagnosis for Shearer through the Vibration of Rocker Transmission Part with an Improved Probabilistic Neural Network. Sensors (2016).
2. A Cutting Pattern Recognition Method for Shearers Based on Improved Ensemble Empirical Mode Decomposition and a Probabilistic Neural Network. Sensors (2015).