Predict Remaining Useful Life (RUL)

Predict RUL using specialized models designed for computing RUL from system data, state estimators, or identified models

One way of analyzing condition indicators is to use them in detecting faults, but you can also use a different type of condition-indicator analysis for predicting the RUL of a system. RUL of a machine is the expected life or usage time remaining before the machine requires repair or replacement.

Typically, you estimate the 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.

Apps

Health Indicator Designer

Interactively transform a set of features into a single composite health indicator that can be used to predict the remaining useful life (RUL) of a machine (Since R2024a)

Functions

expand all

Feature Selection

`monotonicity`	Quantify monotonic trend in condition indicators
`prognosability`	Measure of variability of condition indicators at failure
`trendability`	Measure of similarity between trajectories of condition indicators

RUL Prediction

RUL Models

`exponentialDegradationModel`	Exponential degradation model for estimating remaining useful life
`linearDegradationModel`	Linear degradation model for estimating remaining useful life
`hashSimilarityModel`	Hashed-feature similarity model for estimating remaining useful life
`pairwiseSimilarityModel`	Pairwise comparison-based similarity model for estimating remaining useful life
`residualSimilarityModel`	Residual comparison-based similarity model for estimating remaining useful life
`covariateSurvivalModel`	Proportional hazard survival model for estimating remaining useful life
`reliabilitySurvivalModel`	Probabilistic failure-time model for estimating remaining useful life

Training and Prediction

`predictRUL`	Estimate remaining useful life for a test component
`compare`	Compare test data to historical data ensemble for similarity models
`fit`	Estimate parameters of remaining useful life model using historical data
`plot`	Plot survival function for covariate survival remaining useful life model
`restart`	Reset remaining useful life degradation model
`update`	Update posterior parameter distribution of degradation remaining useful life model

Topics

RUL Basics

Models for Predicting Remaining Useful Life

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
Feature Selection for Remaining Useful Life Prediction
Rank features to determine best indicators of system degradation and improve accuracy of remaining useful life (RUL) predictions.
Perform Prognostic Feature Ranking for a Degrading System Using Diagnostic Feature Designer
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

Update RUL Prediction as Data Arrives
As data arrives from a machine under test, you can update the RUL prediction with each new data point.
Similarity-Based Remaining Useful Life Estimation
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.
Wind Turbine High-Speed Bearing Prognosis
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.
Live RUL Estimation of a Servo Gear Train Using ThingSpeak
This example shows how to estimate the Remaining Useful Life (RUL) of a servo motor gear train through real-time streaming of servo motor data from an Arduino-based data acquisition system to ThingSpeak™, and from ThingSpeak to an RUL estimation engine running in MATLAB®.
ThingSpeak Dashboard for Live RUL Estimation of a Servo Motor Gear Train
This example shows how to set up a ThingSpeak™ dashboard to estimate and visualize the Remaining Useful Life (RUL) of a servo motor gear train.

Prediction Using Identified Models or State Estimators

Nonlinear State Estimation of a Degrading Battery System
Estimate the states of a nonlinear system using an unscented Kalman filter in Simulink.
Condition Monitoring and Prognostics Using Vibration Signals
Extract features from vibration signals from a ball bearing, conduct health monitoring, and perform prognostics.

Prediction Using Artificial Intelligence

Battery Cycle Life Prediction from Initial Operation Data
Predict the remaining cycle-life of a fast charging Li-ion battery using a supervised machine learning algorithm.
Fault Detection and Remaining Useful Life Estimation Using Categorical Data
Use categorical data to improve the accuracy of machine fault predictions.
Remaining Useful Life Estimation Using Convolutional Neural Network
This example shows how to predict the RUL of engines using deep convolutional neural networks (CNN).
Battery Cycle Life Prediction Using Deep Learning
Predict the remaining cycle-life of a fast charging Li-ion battery by training a deep neural network.

Related Information

Predictive Maintenance, Part 3: Remaining Useful Life Estimation

Featured Examples

Similarity-Based Remaining Useful Life Estimation

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.

Open Live Script

Wind Turbine High-Speed Bearing Prognosis

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.

Open Live Script

Remaining Useful Life Estimation of a DC-Link Capacitor in a Power Converter

Design and Implement a remaining useful life (RUL) estimation algorithm for an aluminum electrolytic DC-link capacitor in an AC-DC-AC power converter system.

Open Live Script