Brain-Inspired Spike Echo State Network Dynamics for Aero-Engine Intelligent Fault Prediction

TL;DR

This work addresses aero-engine fault prediction by capturing rich spatiotemporal features in time-series data. It introduces Spike-ESN, a brain-inspired architecture with a Poisson-based spike input layer, a spike reservoir, and a ridge regression readout to learn temporal evolution from high-dimensional spike representations. Empirical results on real aero-engine data show Spike-ESN outperforming ESN, ARMA, CNN, LSTM, and Transformer in RMSE and MAPE, while offering fast training and potential for online fault warning. The approach enhances interpretability, memory, and robustness, enabling more reliable real-time monitoring and early fault detection for aero-engines.

Abstract

Aero-engine fault prediction aims to accurately predict the development trend of the future state of aero-engines, so as to diagnose faults in advance. Traditional aero-engine parameter prediction methods mainly use the nonlinear mapping relationship of time series data but generally ignore the adequate spatiotemporal features contained in aero-engine data. To this end, we propose a brain-inspired spike echo state network (Spike-ESN) model for aero-engine intelligent fault prediction, which is used to effectively capture the evolution process of aero-engine time series data in the framework of spatiotemporal dynamics. In the proposed approach, we design a spike input layer based on Poisson distribution inspired by the spike neural encoding mechanism of biological neurons, which can extract the useful temporal characteristics in aero-engine sequence data. Then, the temporal characteristics are input into a spike reservoir through the current calculation method of spike accumulation in neurons, which projects the data into a high-dimensional sparse space. In addition, we use the ridge regression method to read out the internal state of the spike reservoir. Finally, the experimental results of aero-engine states prediction demonstrate the superiority and potential of the proposed method.

Figures (13)

• Figure 1: Schematic diagram of data driven fault diagnosis of an aero-engine system.
• Figure 2: Schematic diagram of operating data of an aero-engine.
• Figure 3: Spike echo state network. It consists of a spike input layer, a spike reservoir and an output layer.
• Figure 4: Process of using Spike-Esn for aircraft fault early warning.
• Figure 5: The average spike interval of the exhaust temperature $T_6$ and the compressor opening angle $\alpha_2$ in the training dataset.

Theorems & Definitions (3)

• Remark 1
• Remark 2
• Remark 3