Industrial Big Data Analytics for Prediction of Remaining Useful Life Based on Deep Learning

Abstract:

Due to recent development of cyber-physical systems (CPSs), big data, cloud computing, and industrial wireless networks, a new era of industrial big data is introduced. Deep learning, which brought a revolutionary change in computer vision, natural language processing, and variety of other applications, has significant potential for solutions providing in sophisticated industrial applications. In this paper, a concept of device electrocardiogram (DECG) is presented and an algorithm based on deep denoising auto-encoder (DDA) and regression operation is proposed for prediction of the remaining useful life of industrial equipment. First, the concept of electrocardiogram is explained. Then, a problem statement based on manufacturing scenario is presented. Subsequently, the architecture of proposed algorithm called integrated deep denoising auto-encoder (IDDA) and algorithm workflow are provided. Moreover, DECG is compared with traditional factory information system, and the feasibility and effectiveness of proposed algorithm are validated experimentally. The proposed concept and algorithm combine typical industrial scenario and advance artificial intelligence, which has great potential to accelerate the implementation of Industry 4.0.

Existing System:

The experience-based models cannot deal with a large number of queries in expert systems, and they mostly rely on expert knowledge and engineering experience. The similar deficiency exits in physics-based models. The physics-based models require insight in system failure mechanisms, which are supposed to be converted into mathematical expressions.

Proposed System:

The device electrocardiogram (DECG) principle is introduced, and a new methodology based on deep learning and DECG is proposed for prediction of RUL of equipment as well as production line. DECG, which is similar to monitor the health of the human body, records devices’ cycle time with all its sub-processes. Due to much more data collected from DECG, it’s possible to introduce deep learning and fully enhance the performance of deep learning for specific application. Based on deep learning and a large number of run-to-failure samples, the proposed algorithm can provide an accurate prediction of device RUL.

Main Goal:

A concept of DECG in manufacturing environment, which provides a fine-grained status observation and reduces dependency on experts’ knowledge greatly, is proposed.

A RUL predicting methodology based on regression and deep denoising auto-encoders (DDA) is proposed to achieve an automatic feature engineering and a high-level features extraction.

The proposed algorithm and traditional factory information system are compared, and the experiment is performed in order to validate the feasibility and effectiveness of proposed algorithm.

Conclusions:

A new algorithm for RUL prediction based on DECG and deep learning is presented. Firstly, the concept of DECG was introduced. Then, the problem statement in manufacturing environment was explained. In addition, in order to reduce the impact of experts’ experience and human decision on prediction, a deep learning methodology, which embraces IDDA and regression operation, was used. The proposed algorithm was verified by experiments, wherein DECG was compared with FIS. The experimental result have proven DECG superiority over FIS in terms of response and reliability. Furthermore, the prediction accuracy of IDDA was validated by comparison with true RUL. The obtained results have shown a high effectiveness of proposed algorithm. Nevertheless, the comparison results have indicated superiority of proposed algorithm and its feasibility to accelerate the implementation of Industry 4.0.

REFERENCES

[1] J. Wan, S. Tang, Q. Hua, D. Li, C. Liu and J. Lloret, “Context-Aware Cloud Robotics for Material Handling in Cognitive Industrial Internet of Things,” IEEE Internet of Things Journal, to be published, DOI: 10.1109/JIOT.2017.2728722, 2017.

[2] E. Ahmed, I. Yaqoob, I. Hashem, I. Khan, A. Ahmed, M. Imran and A. Vasilakos, “The role of big data analytics in Internet of Things,” Computer Networks, vol. 129, pp. 459-471, 2017.

[3] J. Wan, D. Zhang, Y. Sun, K. Lin, C. Zou and H. Cai, “VCMIA: A Novel Architecture for Integrating Vehicular Cyber-Physical Systems and Mobile Cloud Computing,” Mobile Networks and Applications, vol. 19, no. 2, pp. 153-160, 2014.

[4] E. Ahmed, A. Ahmed, I. Yaqoob, J. Shuja, A. Gani, M. Imran, M. Shoaib, “Bringing computation closer towards user network: Is edge computing the solution?” IEEE Communications Magazine, vol. 55, no. 11, pp. 138-144, 2017.

[5] Y. Xu, Y. Sun, J. Wan, X. Liu and Z. Song, “Industrial Big Data for Fault Diagnosis: Taxonomy, Review, and Applications,” IEEE Access, to be published, DOI: 10.1109/ACCESS.2017.2731945, 2017.

[6] L. Linxia and F. Köttig, “Review of hybrid prognostics approaches for remaining useful life prediction of engineered systems, and an application to battery life prediction,” IEEE Transactions on Reliability, vol. 63, no. 1, pp. 191-207, 2014.

[7] V. Jindal and J. Aggarwal, “An approach towards designing of car troubleshooting expert system,” International Journal of Computer Applications, vol. 1, no. 23, pp. 107-114, 2010.

[8] F. Zhao, Z. Tian and Y. Zeng, “Uncertainty quantification in gear remaining useful life prediction through an integrated prognostics method,” IEEE Transactions on Reliability, vol. 62, no. 1, pp. 146-159, 2013.

[9] M. Daigle and K. Goebel, “Improving computational efficiency of prediction in model-based prognostics using the unscented transform,” Tech. Rep. DTIC Document, 2010.

[10] K. Medjaher, D. A. Tobon-Mejia and N. Zerhouni, “Remaining useful life estimation of critical components with application to bearings,” IEEE Transactions on Reliability, vol. 61, no. 2, pp. 292-302, 2012.

[11] Y. Peng and M. , “A hybrid approach of HMM and grey model for age-dependent health prediction of engineering assets,” Expert Systems with Applications, vol. 38, no. 10, pp. 12 946–12 953, 2011.

[12] N. Gebraeel, M. Lawley, R. Liu and V. Parmeshwaran, “Residual life predictions from vibration-based degradation signals: A neural network approach,” IEEE Transactions on industrial electronics, vol. 51, no. 3, pp. 694-700, 2004.

[13] Z. He, S. Wang, K. Wang, and K. Li, “Prognostic analysis based on hybrid prediction method for axial piston pump,” 2012 10th IEEE International Conference on Industrial Informatics (INDIN), pp. 688– 692, 2012.

[14] V. Cherkassky and Y. Ma, “Practical selection of SVM parameters and noise estimation for SVM regression,” Neural networks, vol. 17, no. 1, pp. 113-126, 2004.

[15] P. Rajesh, M. PandurangeRao, S. Sengottuvel, K. Gireesan and T. S. Radhakrishnan, “Effective Extraction of Visual Event-Related Pattern by Combining Template Matching With Ensemble Empirical Mode Decomposition,” IEEE Sensors Journal, vol. 17, no. 7, pp. 2146-2153, 2017.