Data-driven prognostics for lithium-ion battery based on Gaussian Process Regression

Liu, Datong; Pang, Jingyue; Zhou, Jianbao; Yu, Peng

doi:10.1109/phm.2012.6228848

Cited by 23 publications

(11 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Bolde et al [89] have proposed a novel method for the adaptation of parameters in an electrochemical model of a lithium-ion battery using an unscented Kalman filter (UKF) with battery current and voltage over randomized discharge profiles. Liu et al proposed a Gaussian process model [90] and an improved autoregressive (AR) time series model named Nonlinear Degradation Auto Regression (ND-AR) model [91] for RUL estimation of lithium-ion batteries. Olivares and Muñoz [92] have proposed a PF-based prognostic framework for estimating the state of health (SOH) and predicting the RUL of lithium-ion batteries.…”

Section: Fault Prognosismentioning

confidence: 99%

Fault Diagnosis And Prognosis Of Satellite Attitude Control System With Reaction Wheels and Control Moment Gyros

Rahimi¹

2021

Preprint

View full text Add to dashboard Cite

Condition-based maintenance (CBM) and prognostics and health management (PHM), as consisting parts of diagnosis, prognosis, and health monitoring (DPHM) framework, have developed over the past decades to remedy the limitations of the traditional maintenance practices for complex systems. In space, where mass and power budget are restricted, application of CBM and PHM has become more vital to the success of a mission. Reaction wheels (RW) and Control Moment Gyros (CMG), as the most commonly used actuators onboard satellites, are prone to faults and failures. The ability to detect faults, isolate their location and severity, and estimate the remaining useful life (RUL) of the faulty unit can enhance mission success rate and reduce maintenance and damage costs extensively. Therefore, in this thesis, a model-based DPHM framework is developed and evaluated. Firstly, a novel fault detection algorithm is proposed, using Unscented Kalman filters (UKF) in conjunction with residual and innovation sequences, for detecting agile faults in RW/CMG onboard satellites. Secondly, a novel fault isolation algorithm is proposed, using UKF, Bayes’ probability and interacting multiple models (IMM), to isolate the location of the fault and its severity. Finally, a new fault prognosis approach is proposed, using UKF and particle filters (PF) to estimate the RUL of a faulty unit. Extensive simulations were conducted for each phase of the DPHM to verify advantages of the proposed techniques over the available methods in the literature. Extensive simulations were conducted to evaluate the performance of the proposed methods in each module of the framework. Regarding the proposed fault detection scheme, results showed superior performance of the proposed adaptation technique compared to the original UKF and a previously developed AUKF. The proposed fault isolation scheme was able to successfully isolate the faulty unit at multiple levels of isolation including formation level, system level, and actuator level with over 99% success rate for formation level, over 99% success rate for the RW assembly and for up to 90% success rate for the CMG assembly in the system level. For the CMG assembly, due to direct estimation of the fault parameters, it was possible to determine the severity of the faults as well as their location. Finally, the proposed fault prognosis approach provided RUL estimates with errors as low as 1.5% compared to the actual remaining useful life. Overall, the proposed framework can be regarded as a promising tool for fault detection, isolation and identification, and prognosis of the complex nonlinear systems. Furthermore, the proposed framework can be extended to other complex systems in space including multi-agent formation systems and other areas where the model of the system under study is available.

show abstract

Section: Fault Prognosismentioning

confidence: 99%

Fault Diagnosis And Prognosis Of Satellite Attitude Control System With Reaction Wheels and Control Moment Gyros

Rahimi¹

2021

Preprint

View full text Add to dashboard Cite

show abstract

“…In light of their flexible, probabilistic non-parametric framework, GPs have found several applications in prognosis, e,g, nuclear component degradation [16], lithium-ion batteries [17], [19], [20] and bearings [18]. On the other hand, despite the increasing efforts in integrating DNNs with effective UQ techniques, very few of the methods mentioned in the previous subsection have been successfully transferred to prognosis tasks.…”

Section: Uq In Data-driven Prognosticsmentioning

confidence: 99%

“…The degradation trajectories are given in the form of multivariate time-series of sensor readings. Figure 3 shows the distribution of the flight envelopes for six training units (2,5,10,16,18,20) and three test units (11,14,15). It is worth noting that test unit 14 has an operation distribution that is significantly different from the training units.…”

Section: A Case Study Of Predicting the Rul Of Turbofan Enginesmentioning

confidence: 99%

“…GPs, in particular, provide a good adaptability to handle nonlinear, relatively complex regression problems and, compared to standard neural networks, they are based on a wellunderstood probabilistic formulation. Their flexibility and the availability of open-source software implementations [15] have led to a number of interesting applications in prognostics of engineered systems [16]- [20].…”

Section: Introductionmentioning

confidence: 99%

“…FIGURE 3: Approximate density distributions of the flight envelopes given by recordings of altitude, flight Mach number, throttle-resolver angle, and total temperature at the fan inlet for complete trajectories (from start until a failure occurs) of six training units(2,5,10,16,18,20) and three test units(11,14,15).…”

mentioning

confidence: 99%

See 2 more Smart Citations

Uncertainty-Aware Prognosis via Deep Gaussian Process

et al. 2021

View full text Add to dashboard Cite

The problem of Uncertainty Quantification (UQ) is of paramount importance when Machine Learning (ML) and Deep Learning (DL) models are deployed in the real world. In the context of safetycritical applications, such as the prediction of the Remaining Useful-Life (RUL) of infrastructure and industrial assets, the relevance of effective UQ approaches is even higher, given the potentially catastrophic consequences or substantial costs associated with maintenance decisions that are performed either too late or too early. Particularly in such safety critical application contexts, transparency and reliability are essential requirements and any ML-based solution not providing meaningful uncertainty estimates would not fully satisfy such desiderata. However, most ML and DL techniques used for RUL estimation are often not designed to perform UQ, thus limiting their effective applicability in real-world scenarios. To address this limitation, in this paper we investigate the performance of a recently proposed class of algorithms, Deep Gaussian Process (DGPs), for predicting the remaining useful lifetime (RUL). DGPs provide uncertainty estimates associated with their predictions, yet retaining the expressive power and learning capabilities of modern DL methods. DGPs are able to scale to very large datasets, which was a limitation of some of the previous algorithms but has increasingly become an essential requirement for many industrial applications and, in particular, for RUL prediction tasks for which UQ is of key importance. The main contribution of this paper is a thorough evaluation and comparison of several variants of DGPs applied to the problem of RUL prediction. The performance of the DGP algorithms is evaluated on the N-CMAPSS (New Commercial Modular Aero-Propulsion System Simulation) dataset from NASA for aircraft engines. The results demonstrate that DGPs are able to provide very accurate RUL predictions along with sensible uncertainty estimates, providing more reliable solutions for (safety-critical) real-life industrial applications.

show abstract

Analysis of Actuators Prognostic Health Monitoring in Spacecraft Attitude Control Systems

Borah

2020

Lecture Notes in Mechanical Engineering

View full text Add to dashboard Cite

Data-driven prognostics for lithium-ion battery based on Gaussian Process Regression

Cited by 23 publications

References 9 publications

Fault Diagnosis And Prognosis Of Satellite Attitude Control System With Reaction Wheels and Control Moment Gyros

Fault Diagnosis And Prognosis Of Satellite Attitude Control System With Reaction Wheels and Control Moment Gyros

Uncertainty-Aware Prognosis via Deep Gaussian Process

Analysis of Actuators Prognostic Health Monitoring in Spacecraft Attitude Control Systems

Contact Info

Product

Resources

About