Deep learning for predictive maintenance with Long Short Term Memory Networks

6월 21, 2017에 게시됨

Senior Data Scientist

Deep learning has proven to show superior performance in certain areas such as object recognition and image classification. It has also gained popularity in other domains such as finance where time-series data plays an important role. Similarly, in predictive maintenance, the data is collected over time to monitor the state of an asset with the goal of finding patterns to predict failures which can benefit from certain deep learning algorithms. Among the deep learning networks, Long Short Term Memory (LSTM) networks are especially appealing to the predictive maintenance domain since they are very good at learning from sequences. This fact lends itself to their applications using time series data by making it possible to look back for longer periods of time to detect failure patterns.

The traditional predictive maintenance machine learning models are based on feature engineering which is manual construction of right features using domain expertise and similar methods. This usually makes these models hard to reuse since feature engineering is specific to the problem scenario and the available data which varies from one business to the other. Perhaps the most attractive part of applying deep learning in the predictive maintenance domain is the fact that these networks can automatically extract the right features from the data, eliminating the need for manual feature engineering. However, determining the topology of deep learning networks such as deciding on the optimal number of layers, number of nodes and hyperparameters is also considered very labor intensive with no clear guidance.

In the notebook Deep Learning Basics for Predictive Maintenance, we build an LSTM network for the data set and scenario described at Predictive Maintenance Template to predict remaining useful life of aircraft engines using the Turbofan Engine Degradation Simulation Data Set. This notebook serves as a tutorial for beginners looking to apply deep learning in predictive maintenance domain and uses simulated aircraft sensor values to predict when an aircraft engine will fail in the future so that maintenance can be planned in advance.