Abstract:In industry, forecast prediction and health management (PHM) is used to improve system reliability and efficiency. In PHM, remaining useful life (RUL) prediction plays a key role in preventing machine failures and reducing operating costs, especially for reliability requirements such as critical components in aviation as well as for costly equipment. With the development of deep learning techniques, many RUL prediction methods employ convolutional neural network (CNN) and long short-term memory (LSTM) networks and demonstrate superior performance. In this paper, a novel two-stream network based on a bidirectional long short-term memory neural network (BiLSTM) is proposed to establish a two-stage residual life prediction model for mechanical devices using CNN as the feature extractor and BiLSTM as the timing processor, and finally, a particle swarm optimization (PSO) algorithm is used to adjust and optimize the network structural parameters for the initial data. Under the condition of lack of professional knowledge, the adaptive extraction of the features of the data accumulated by the enterprise and the effective processing of a large amount of timing data are achieved. Comparing the prediction results with other models through examples, it shows that the model established in this paper significantly improves the accuracy and efficiency of equipment remaining life prediction.

What problem does this paper attempt to address?

The paper attempts to address the issue of how to improve system reliability and efficiency in the industrial sector, particularly for critical components (such as mechanical equipment in the aviation field) and high-cost equipment, by predicting the remaining useful life (RUL). Specifically, the paper aims to develop a hybrid deep learning model based on Particle Swarm Optimization (PSO), Convolutional Neural Network (CNN), and Bidirectional Long Short-Term Memory Network (BiLSTM) to achieve accurate prediction of the remaining useful life of mechanical equipment. ### Main Issues 1. **Improving System Reliability and Efficiency**: By predicting the remaining useful life of mechanical equipment, machine failures can be prevented, operating costs reduced, and normal operation and timely maintenance of equipment ensured. 2. **Addressing the Limitations of Traditional Methods**: - **Model-Driven Methods**: These rely on prior knowledge of mechanical systems, but as the complexity of equipment increases, obtaining sufficient prior knowledge becomes difficult. - **Data-Driven Methods**: Traditional data-driven methods are usually based on equipment operating time and empirical judgment, posing risks of "over-maintenance" and "under-maintenance," which may lead to equipment failure, production interruptions, or unnecessary maintenance cost increases. ### Solution The paper proposes a two-stage remaining life prediction model based on PSO-CNN-BiLSTM: 1. **Feature Extraction**: Using CNN to extract key features from sensor data and compress sequence length, improving deep learning performance and model training speed. 2. **Time Series Processing**: Using BiLSTM to mine the temporal characteristics of the data, achieving effective analysis and modeling of the equipment's full lifecycle data. 3. **Parameter Optimization**: Using the PSO algorithm to optimize network structure parameters, making the model better adapt to data characteristics and further improving prediction performance. ### Advantages - **Accuracy**: By combining CNN and BiLSTM, the model can deeply mine the spatiotemporal characteristics of the data, improving prediction accuracy. - **Flexibility**: Data-driven methods do not require in-depth understanding of the system's dynamic characteristics, making them suitable for complex mechanical systems. - **Optimization**: The PSO algorithm can effectively optimize the model's hyperparameters, accelerating the model's convergence process and improving prediction results. In summary, the paper proposes an innovative deep learning model to address key issues in the prediction of the remaining useful life of mechanical equipment, providing support for sustainable development in the industrial sector.

A Remaining Useful Life Prediction Method of Mechanical Equipment Based on Particle Swarm Optimization-Convolutional Neural Network-Bidirectional Long Short-Term Memory

Bi-LSTM-Based Two-Stream Network for Machine Remaining Useful Life Prediction

Rolling Bearing Residual Useful Life Prediction Model Based on the Particle Swarm Optimization-Optimized Fusion of Convolutional Neural Network and Bidirectional Long–Short-Term Memory–Multihead Self-Attention

Remaining Useful Life Prediction Method Based on Convolutional Neural Network and Long Short-Term Memory Neural Network

Convolutional neural network based on attention mechanism and Bi-LSTM for bearing remaining life prediction

Remaining Useful Life Prediction Based on an Integrated Neural Network

A Multi-Head Neural Network with Unsymmetrical Constraints for Remaining Useful Life Prediction

A Hybrid Method Based Deep Learning Approach for Predicting Residual Life of Machinery

Life Prediction for Machinery Components Based on CNN-BiLSTM Network and Attention Model

A Hybrid Prognostics Deep Learning Model for Remaining Useful Life Prediction

Remaining useful life prediction for equipment based on RF-BiLSTM

Memory-enhanced hybrid deep learning networks for remaining useful life prognostics of mechanical equipment

A Novel Combination Neural Network Based on ConvLSTM-Transformer for Bearing Remaining Useful Life Prediction

Long Short-Term Memory Neural Network with Transfer Learning and Ensemble Learning for Remaining Useful Life Prediction

Journal bearing seizure degradation assessment and remaining useful life prediction based on long short-term memory neural network

Remaining Useful Life Prediction Via Long‐short Time Memory Neural Network with Novel Partial Least Squares and Genetic Algorithm

Machine Remaining Useful Life Prediction Via an Attention-Based Deep Learning Approach

A Model Based on CNN-LSTM for the Remaining Life Prediction of Equipment in Aircraft Assembly Pulsation Production Line

A Two-Stage Data-Driven Approach to Remaining Useful Life Prediction Via Long Short-Term Memory Networks.

A dual path hybrid neural network framework for remaining useful life prediction of aero‐engine

A data-driven model for milling tool remaining useful life prediction with convolutional and stacked LSTM network