Predictive maintenance using cox proportional hazard deep learning

Abstract

Predictive maintenance (PdM) has become prevalent in the industry in order to reduce maintenance cost and to achieve sustainable operational management. The core of PdM is to predict the next failure so corresponding maintenance can be scheduled before it happens. The purpose of this study is to establish a Time-Between-Failure (TBF) prediction model through a data-driven approach. For PdM, data sparsity is regarded as a critical issue which can jeopardize algorithm performance for the modelling based on maintenance data. Meanwhile, data censoring has imposed another challenge for handling maintenance data because the censored data is only partially labelled. Furthermore, data sparsity may hurt algorithm performance of the existing approaches when addressing the data censoring issue. In this study, a new approach called Cox proportional hazard deep learning (CoxPHDL) is proposed to tackle the aforementioned issues of data sparsity and data censoring that are common in the analysis of operational maintenance data. The idea is to offer an integrated solution by taking advantage of deep learning and reliability analysis. To start with, an autoencoder is adopted to convert the nominal data into a robust representation. Secondly, a Cox PHM is researched to estimate the TBF of the censored data. A long-short-term memory (LSTM) network is then established to train the TBF prediction model based on the pre-processed maintenance data. Experimental studies using a sizable real-world fleet maintenance data set provided by a UK fleet company have demonstrated the merits of the proposed approach where the algorithm performance based on the proposed LSTM network has been improved respectively in terms of MCC and RMSE.

An integrated deep learning-based approach for automobile maintenance prediction with GIS data

“…The data which contained these features were fed into an LSTM network for training. The algorithm performance in terms of model correlation coefficient and root mean square error is better than that of a fully-connected neural network [36]. However, when the number of ordinary numeric features increases, the algorithm performance of LSTM network could be compromised.…”

Kestirimci Bakım ve Kalan Yararlı Ömür Uygulama için Teknikler: Sistematik Haritalama Çalışması

“…As a result, the most optimum solution is found by comparing methods from given the best results. The area that appears on the graph as 'Other' contains performance metrics such as Model Correlation Coefficient [35], Software Product Quality Metrics (ISO / IEC 9126, 25041, 25051) [41], α−λ Metric [42] and Opinion of a Machining Expert [43]. To evaluate the performance of clustering algorithms, metrics that not sufficiently reliable are used such as homogeneity score, integrity score, V measurement, corrected Rand index, corrected mutual information, silhouette coefficient.…”

“…In[34], a total of 28 features; 11 time-domain features, 9 frequency-domain features, and 8 time-frequency domain features, have been extracted. Moreover[35] and[36] Auto Encoder based,[2] Fast Fourier Transform (FFT) and[37] and[38] K-means based feature extraction method is used.…”

The Digital Twin Landscape at the Crossroads of Predictive Maintenance, Machine Learning and Physics Based Modeling

“…Survival Models: When Data Encompasses Only the End Time of Failure From Similar Equipment Sometimes only the time point of failure is known, and survival models such as Kaplan-Meier, Cox proportional hazard (CPH) [28], or more advanced approaches using deep learning [11] can be utilized. Survival curves (probability of survival over time) for time-to-event (i.e., events are faults or failures in the PMx context) are generated from the failure history data.…”