Remaining Useful Life Prognosis for Turbofan Engine Using Explainable Deep Neural Networks with Dimensionality Reduction
This study prognoses the remaining useful life of a turbofan engine using a deep learning model, which is essential for the health management of an engine. The proposed deep learning model affords a significantly improved accuracy by organizing networks with a one-dimensional convolutional neural ne...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-11-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/20/22/6626 |
| _version_ | 1827701894295846912 |
|---|---|
| author | Chang Woo Hong Changmin Lee Kwangsuk Lee Min-Seung Ko Dae Eun Kim Kyeon Hur |
| author_facet | Chang Woo Hong Changmin Lee Kwangsuk Lee Min-Seung Ko Dae Eun Kim Kyeon Hur |
| author_sort | Chang Woo Hong |
| collection | DOAJ |
| description | This study prognoses the remaining useful life of a turbofan engine using a deep learning model, which is essential for the health management of an engine. The proposed deep learning model affords a significantly improved accuracy by organizing networks with a one-dimensional convolutional neural network, long short-term memory, and bidirectional long short-term memory. In particular, this paper investigates two practical and crucial issues in applying the deep learning model for system prognosis. The first is the requirement of numerous sensors for different components, i.e., the curse of dimensionality. Second, the deep neural network cannot identify the problematic component of the turbofan engine due to its “black box” property. This study thus employs dimensionality reduction and Shapley additive explanation (SHAP) techniques. Dimensionality reduction in the model reduces the complexity and prevents overfitting, while maintaining high accuracy. SHAP analyzes and visualizes the black box to identify the sensors. The experimental results demonstrate the high accuracy and efficiency of the proposed model with dimensionality reduction and show that SHAP enhances the explainability in a conventional deep learning model for system prognosis. |
| first_indexed | 2024-03-10T14:43:43Z |
| format | Article |
| id | doaj.art-4c28647bbb82449b8e0dc7b94ffa1733 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T14:43:43Z |
| publishDate | 2020-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-4c28647bbb82449b8e0dc7b94ffa17332023-11-20T21:31:50ZengMDPI AGSensors1424-82202020-11-012022662610.3390/s20226626Remaining Useful Life Prognosis for Turbofan Engine Using Explainable Deep Neural Networks with Dimensionality ReductionChang Woo Hong0Changmin Lee1Kwangsuk Lee2Min-Seung Ko3Dae Eun Kim4Kyeon Hur5School of Electrical & Electronic Engineering, Yonsei University, 50 Yonsei-Ro Seodamun-Gu, Seoul 03722, KoreaSchool of Electrical & Electronic Engineering, Yonsei University, 50 Yonsei-Ro Seodamun-Gu, Seoul 03722, KoreaSchool of Electrical & Electronic Engineering, Yonsei University, 50 Yonsei-Ro Seodamun-Gu, Seoul 03722, KoreaSchool of Electrical & Electronic Engineering, Yonsei University, 50 Yonsei-Ro Seodamun-Gu, Seoul 03722, KoreaSchool of Electrical & Electronic Engineering, Yonsei University, 50 Yonsei-Ro Seodamun-Gu, Seoul 03722, KoreaSchool of Electrical & Electronic Engineering, Yonsei University, 50 Yonsei-Ro Seodamun-Gu, Seoul 03722, KoreaThis study prognoses the remaining useful life of a turbofan engine using a deep learning model, which is essential for the health management of an engine. The proposed deep learning model affords a significantly improved accuracy by organizing networks with a one-dimensional convolutional neural network, long short-term memory, and bidirectional long short-term memory. In particular, this paper investigates two practical and crucial issues in applying the deep learning model for system prognosis. The first is the requirement of numerous sensors for different components, i.e., the curse of dimensionality. Second, the deep neural network cannot identify the problematic component of the turbofan engine due to its “black box” property. This study thus employs dimensionality reduction and Shapley additive explanation (SHAP) techniques. Dimensionality reduction in the model reduces the complexity and prevents overfitting, while maintaining high accuracy. SHAP analyzes and visualizes the black box to identify the sensors. The experimental results demonstrate the high accuracy and efficiency of the proposed model with dimensionality reduction and show that SHAP enhances the explainability in a conventional deep learning model for system prognosis.https://www.mdpi.com/1424-8220/20/22/6626deep neural networkdimensionality reductionexplainable artificial intelligencefeature selectionprognostics and health monitoringturbofan engine |
| spellingShingle | Chang Woo Hong Changmin Lee Kwangsuk Lee Min-Seung Ko Dae Eun Kim Kyeon Hur Remaining Useful Life Prognosis for Turbofan Engine Using Explainable Deep Neural Networks with Dimensionality Reduction Sensors deep neural network dimensionality reduction explainable artificial intelligence feature selection prognostics and health monitoring turbofan engine |
| title | Remaining Useful Life Prognosis for Turbofan Engine Using Explainable Deep Neural Networks with Dimensionality Reduction |
| title_full | Remaining Useful Life Prognosis for Turbofan Engine Using Explainable Deep Neural Networks with Dimensionality Reduction |
| title_fullStr | Remaining Useful Life Prognosis for Turbofan Engine Using Explainable Deep Neural Networks with Dimensionality Reduction |
| title_full_unstemmed | Remaining Useful Life Prognosis for Turbofan Engine Using Explainable Deep Neural Networks with Dimensionality Reduction |
| title_short | Remaining Useful Life Prognosis for Turbofan Engine Using Explainable Deep Neural Networks with Dimensionality Reduction |
| title_sort | remaining useful life prognosis for turbofan engine using explainable deep neural networks with dimensionality reduction |
| topic | deep neural network dimensionality reduction explainable artificial intelligence feature selection prognostics and health monitoring turbofan engine |
| url | https://www.mdpi.com/1424-8220/20/22/6626 |
| work_keys_str_mv | AT changwoohong remainingusefullifeprognosisforturbofanengineusingexplainabledeepneuralnetworkswithdimensionalityreduction AT changminlee remainingusefullifeprognosisforturbofanengineusingexplainabledeepneuralnetworkswithdimensionalityreduction AT kwangsuklee remainingusefullifeprognosisforturbofanengineusingexplainabledeepneuralnetworkswithdimensionalityreduction AT minseungko remainingusefullifeprognosisforturbofanengineusingexplainabledeepneuralnetworkswithdimensionalityreduction AT daeeunkim remainingusefullifeprognosisforturbofanengineusingexplainabledeepneuralnetworkswithdimensionalityreduction AT kyeonhur remainingusefullifeprognosisforturbofanengineusingexplainabledeepneuralnetworkswithdimensionalityreduction |