Predict Remaining Useful Life
Typically, you estimate the remaining useful life (RUL) of a system by developing a model that can perform the estimation based on the time evolution or statistical properties of condition indicator values. Predictions from such models are statistical estimates with associated uncertainty. They provide a probability distribution of the RUL of the test machine.
The model you use can be a dynamic model such as those you obtain using System Identification Toolbox™ commands. Predictive Maintenance Toolbox™ also includes some specialized models designed for computing RUL from different types of measured system data. For an overview of the types of models you can use, see Models for Predicting Remaining Useful Life.
Developing a model for RUL prediction is the next step in the algorithm-design process after identifying promising condition indicators. Because the model you develop uses the time evolution of condition indicator values to predict RUL, this step is often iterative with the step of identifying condition indicators.
|Exponential degradation model for estimating remaining useful life|
|Linear degradation model for estimating remaining useful life|
|Hashed-feature similarity model for estimating remaining useful life|
|Pairwise comparison-based similarity model for estimating remaining useful life|
|Residual comparison-based similarity model for estimating remaining useful life|
|Proportional hazard survival model for estimating remaining useful life|
|Probabilistic failure-time model for estimating remaining useful life|
|Estimate remaining useful life for a test component|
|Compare test data to historical data ensemble for similarity models|
|Estimate parameters of remaining useful life model using historical data|
|Plot survival function for covariate survival remaining useful life model|
|Reset remaining useful life degradation model|
|Update posterior parameter distribution of degradation remaining useful life model|
You can use recursive models, identified models, or state estimators to predict remaining useful life (RUL). There are also specialized models designed for computing RUL from system data.
- RUL Estimation Using Identified Models or State Estimators
- RUL Estimation Using RUL Estimator Models
Rank features to determine best indicators of system degradation and improve accuracy of remaining useful life (RUL) predictions.
This example shows how to segment data from a degrading system into frames, perform frame-based processing and feature extraction, and use prognostic ranking in Diagnostic Feature Designer.
Prediction Using RUL Models
As data arrives from a machine under test, you can update the RUL prediction with each new data point.
Build a complete Remaining Useful Life (RUL) estimation algorithm from preprocessing, selecting trendable features, constructing health indicator by sensor fusion, training similarity RUL estimators, and validating prognostics.
Build an exponential degradation model to predict the Remaining Useful Life (RUL) of a wind turbine bearing in real time. The exponential degradation model predicts the RUL based on its parameter priors and the latest measurements.
Prediction Using Identified Models or State Estimators
Estimate the states of a nonlinear system using an unscented Kalman filter in Simulink.
Extract features from vibration signals from a ball bearing, conduct health monitoring, and perform prognostics.
Prediction Using Artificial Intelligence
Predict the remaining cycle-life of a fast charging Li-ion battery using a supervised machine learning algorithm.
This example shows how to predict the RUL of engines using deep convolutional neural networks (CNN).