Impact of Sensor Degradation on Measurement Uncertainty in Prognostics and Maintenance Decision-Making: State of the art and challenges

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Yuan et al. presented a data-driven method for predicting RUL of sensors considering impacts of sensor degradation on measurement uncertainty, employing a Weibull distribution to estimate failure probabilities over time based on historical sensor failure data.

“Yuan et al. (2024) presented a data-driven method for predicting the remaining useful life of sensors, considering the impacts of sensor degradation on measurement uncertainty, by employing statistical models such as the Weibull distribution to estimate failure probabilities over time based on historical sensor failure data.”

• cites Yuan et al., 2024 sourcePartly supported

  Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

  Experimental — accuracy not yet measured

Probably not a substantial claim. The paper’s conclusions may not depend on it.

After filtering, 85 papers were selected for the review.

“resulting in 85 papers being selected.”

• AssumptionRests on an assumption: The selected scientific databases (Web of Science, Scopus, IEEE Xplore, and ACM Digital Library) are sufficient for a literature search.
• AssumptionRests on an assumption: Article metadata (titles, abstracts, authors, and keywords) are meticulously curated by authors and editors to reflect the core content of the article.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

In the keyword occurrence network, sensor degradation is the most prominent and interconnected keyword.

“At the centre of the network,”

• AssumptionRests on an assumption: The selected scientific databases (Web of Science, Scopus, IEEE Xplore, and ACM Digital Library) are sufficient for a literature search.
• AssumptionRests on an assumption: Article metadata (titles, abstracts, authors, and keywords) are meticulously curated by authors and editors to reflect the core content of the article.

Probably not a substantial claim. The paper may give no reason why it holds.

No comprehensive review on the impacts of sensor degradation on prognostics and maintenance decision-making has been conducted to date.

“Despite the critical impact of sensor degradation on prognostics and maintenance decision-making, no comprehensive review on this topic has been conducted to date.”

• AssumptionRests on an assumption: The selected scientific databases (Web of Science, Scopus, IEEE Xplore, and ACM Digital Library) are sufficient for a literature search.
• AssumptionRests on an assumption: Article metadata (titles, abstracts, authors, and keywords) are meticulously curated by authors and editors to reflect the core content of the article.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The keyword-based search yielded 824 references from digital libraries.

“The keyword-based search yielded 824 references from digital libraries.”

• AssumptionRests on an assumption: The selected scientific databases (Web of Science, Scopus, IEEE Xplore, and ACM Digital Library) are sufficient for a literature search.
• AssumptionRests on an assumption: Article metadata (titles, abstracts, authors, and keywords) are meticulously curated by authors and editors to reflect the core content of the article.

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Operator perception of uncertain data, trust, and cognitive biases are underexplored.

“Data quality variations from degradation influence decisions; operator perception of uncertain data, trust, and cognitive biases underexplored.”

Probably not a substantial claim. The paper may give no reason why it holds.

There is insufficient integration of uncertainty management principles across the sensor lifecycle (design, manufacturing, deployment, maintenance).

“Insufficient integration of uncertainty management principles across sensor lifecycle (design, manufacturing, deployment, maintenance).”

Probably not a substantial claim. The paper may give no reason why it holds.

There is a gap in the research on quantifying how much sensor degradation can be tolerated while still ensuring reliable prognostics and maintenance decisions.

“For example, there is a gap in the research on quantifying how much sensor degradation can be tolerated while still ensuring reliable prognostics and maintenance decisions.”

Probably not a substantial claim.

Mitigation strategies such as state estimation, filtering techniques, and revised maintenance policies have been developed to account for sensor degradation and sustain system reliability.

“Furthermore, mitigation strategies such as state estimation, filtering techniques, and revised maintenance policies have been developed to account for sensor degradation and sustain system reliability.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Studies have highlighted a range of methods for addressing sensor degradation and measurement uncertainty in prognostics.

“Studies have highlighted a range of methods for addressing sensor degradation and measurement uncertainty in prognostics.”

Probably not a substantial claim.

Data quality variations from degradation influence maintenance decisions.

“Data quality variations from degradation influence decisions; operator perception of uncertain data, trust, and cognitive biases underexplored.”

Probably not a substantial claim. The paper may give no reason why it holds.

Condition-based maintenance lacks integrated models for sensor abnormal behaviours and failure patterns.

“CBM lacks integrated models for sensor abnormal behaviours/failure patterns.”

It may state two separate claims as one.

Sensor degradation significantly affects measurement reliability, potentially compromising the accuracy of fault detection, remaining useful life estimation, and maintenance optimization.

“The review highlighted that sensor degradation significantly affects measurement reliability, potentially compromising the accuracy of fault detection, remaining useful life estimation, and maintenance optimization.”

Probably not a substantial claim.

Maintenance for unknown system states is challenging.

“Policies often lack dynamic recalibration/adjustment; maintenance for unknown system states is challenging.”

Probably not a substantial claim. The paper may give no reason why it holds.

Maintenance policies often lack dynamic recalibration or adjustment.

“Policies often lack dynamic recalibration/adjustment; maintenance for unknown system states is challenging.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

There are limited self-healing mechanisms or advanced redundancy strategies.

“Limited self-healing mechanisms or advanced redundancy strategies.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Many of these approaches leverage optimization algorithms to enhance decision-making efficiency, ensuring that both sensor maintenance and overall system performance are effectively managed.

“Many of these approaches leverage optimization algorithms to enhance decision-making efficiency, ensuring that both sensor maintenance and overall system performance are effectively managed.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Theoretical and data-driven approaches, including statistical, probabilistic, and machine learning-based methods, offer deeper insights into degradation patterns and their impact on system performance.

“Additionally, theoretical and data-driven approaches, including statistical, probabilistic, and machine learningbased methods, offer deeper insights into degradation patterns and their impact on system performance.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

These approaches often utilize expert knowledge, physical degradation models, and data-driven or hybrid methods to enhance generalization across diverse applications.

“These approaches often utilize expert knowledge, physical degradation models, and data-driven or hybrid methods to enhance generalization across diverse applications.”

Probably not a substantial claim.

There is a lack of clear standards for managing sensor uncertainties.

“Lack of clear standards for managing sensor uncertainties; need to quantify tolerable degradation.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The contribution of journal articles has substantially increased in recent years.

“In addition, it is worth mentioning that the contribution of journal articles has substantially increased in recent years.”

It may state two separate claims as one.

The number of selected publications shows a trend of being low in the early years but shows a significant increasing trend from 2018 onwards.

“Figure1shows that the number of selected publications shows a trend of being low in the early years but shows a significant increasing trend from 2018 onwards.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Various modelling techniques have been employed to address sensor degradation challenges, ranging from empirical models based on experimental data to advanced filtering methods within state estimation frameworks.

“Various modelling techniques have been employed to address these challenges, ranging from empirical models based on experimental data to advanced filtering methods within state estimation frameworks.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Balancing maintenance costs, downtime, and degradation risk is complex.

“Balancing maintenance costs, downtime, and degradation risk is complex.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one. The quote only partly states this claim.

The review identified 824 papers across databases: Web of Science 245, Scopus 369, IEEE Xplore 147, and ACM 63.

“| Web Of Science | 245 |”

• Also stated in: § 8. CONCLUSION · plain

  “| Scopus | 369 |”

• Also stated in: § 8. CONCLUSION · plain

  “| IEEE Xplore | 147 |”

It may state two separate claims as one.

Uncertainties caused by sensor degradation and limitations in degradation mechanisms can lead to inaccurate determination of machinery condition, which in turn affects the reliability of maintenance decisions.

“However, the uncertainties caused by sensor degradation and limitations in degradation mechanisms can lead to inaccurate determination of machinery condition, which in turn affects the reliability of maintenance decisions (Li et al., 2021;Huynh, Barros and Bérenguer, 2012;Hegedus & Kosztyán, 2011).”

• cites Li et al., 2021Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Huynh, Barros and Bérenguer, 2012 sourceNot checked — no full text of the source
• cites Hegedus & Kosztyán, 2011 sourceNot checked — no full text of the source

It may state two separate claims as one.

Filtering methods encounter filtering error dynamic instability due to sensor degradation, indicating difficulty in deriving exact values of filtering error covariance (FEC) which tends to be increasing.

“Filtering methods encounter filtering error dynamic instability (He, Wang and Zhou, 2008), indicating difficulty in deriving exact values of filtering error covariance (FEC) which tends to be increasing (Huang & Shen, 2021).”

• cites He, Wang and Zhou, 2008Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Huang & Shen, 2021 sourceNot checked — no full text of the source

It may state two separate claims as one.

Sensor degradation is a well-known phenomenon that impacts the accuracy and reliability of measurements; as sensors degrade over time, the data they provide become less precise and more uncertain.

“Sensor degradation is a well-known phenomenon that impacts the accuracy and reliability of measurements, as sensors degrade over time, the data they provide become less precise and more uncertain”

• cites Abid, Sayed Mouchaweh and Cornez, 2019 sourceNot checked — no full text of the source
• cites Elattar, Elminir and Riad, 2016 sourceNot checked — no full text of the source
• cites Javed, Gouriveau and Zerhouni, 2017 sourceNot checked — no full text of the source

It may state two separate claims as one.

Sensor degradation results in distorted measurements, causing observations to deviate significantly from true values, which can mislead Condition-Based Maintenance (CBM) strategies.

“Liu et al. (2019) highlighted that sensor degradation results in distorted measurements, causing observations to deviate significantly from true values, which can mislead Condition-Based Maintenance (CBM) strategies.”

• cites Liu et al. (2019)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The review protocol followed in this paper adheres to the Kitchenham guidelines.

“The review protocol followed in this paper adheres to the Kitchenham guidelines (Keele, 2007), ensuring a thorough and systematic literature review.”

• cites Keele, 2007Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Sensor degradation causes increased measurement error and error rates over time.

“Many existing references in the literature note that sensor degradation causes increased measurement error (Mukhopadhyay, Liu, Bedford and Finkelstein, 2023;Ohsuga & Ohyama, 1988;Zhang, Si, Du and Hu, 2018;He, Sun, Xie and Kuo, 2022) and error rates over time (Zhang, Xin, Yin, Wang and Wang, 2016).”

• cites Mukhopadhyay, Liu, Bedford and Finkelstein, 2023Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Ohsuga & Ohyama, 1988 sourceNot checked — no full text of the source
• cites Zhang, Xin, Yin, Wang and Wang, 2016Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Accurate prognostics requires reliable measurement of the states of the system, which can be compromised by sensor degradation and the resulting measurement uncertainty.

“Accurate prognostics requires reliable measurement on the states of the system, which can be compromised by sensor degradation and the resulting measurement uncertainty”

• cites Zhao, Zhang, Liu and Qiu, 2019Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Sun, Zuo and Pecht, 2011 sourceNot checked — no full text of the source
• cites Tao, 2012 sourceNot checked — no full text of the source

Probably not a substantial claim.

Modelling of sensor degradation provides essential insights for advancing the acquisition of reliable measurements by translating understanding of degradation mechanisms into quantitative frameworks, enabling predictive analysis of sensor performance and reliability.

“Modelling of sensor degradation provides essential insights for advancing the acquisition of reliable measurements as it translates the understanding of degradation mechanisms into quantitative frameworks, thereby enabling predictive analysis of sensor performance and reliability.”

It may state two separate claims as one.

The quality of measurements provided by sensors is a critical factor impacting the performance of prognostics, maintenance decision-making and optimization, and it can be compromised by sensor degradation.

“The quality of measurements provided by sensors is a critical factor impacting the performance of prognostics, maintenance decision-making and optimization, and it can be compromised by sensor degradation”

• cites Liu, Do, Iung and Xie, 2019Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Poor data quality can lead to incorrect assessments of equipment health, resulting in unnecessary maintenance or unexpected failures.

“Lukens, Rousis, Baer, Lujan and Smith (2022) echoed that poor data quality can lead to incorrect assessments of equipment health, resulting in unnecessary maintenance or unexpected failures.”

• cites Lukens, Rousis, Baer, Lujan and Smith (2022) sourceNot checked — no full text of the source

It may state two separate claims as one.

Sensor degradation weakens the correlation between sensor output and actual conditions, leading to reduced measurement reliability.

“Consequently, it weakens the correlation between sensor output and actual conditions (Li, Price, Stott and Marshall, 2007) leading to reduced measurement reliability (Wanga, Al Atata, Ghaffaria, Leea and Xib, 2008;Li & Dai, 2020).”

• cites Li, Price, Stott and Marshall, 2007Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Wanga, Al Atata, Ghaffaria, Leea and Xib, 2008Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Li & Dai, 2020 sourceNot checked — no full text of the source

It may state two separate claims as one.

Many maintenance models assume that sensor performance remains constant, which is often not the case in reality.

“Many maintenance models assume that sensor performance remains constant, which is often not the case in reality (Mukhopadhyay et al., 2023;van Oosterom, Maillart and Kharoufeh, 2017).”

• cites Mukhopadhyay et al., 2023Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites van Oosterom, Maillart and Kharoufeh, 2017Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Sensor degradation is a complex phenomenon driven by a variety of physical and chemical conditions, and is highly dependent on the sensor type and its operational environment.

“The degradation of sensors is a complex phenomenon driven by a variety of physical and chemical conditions, and are highly dependent on the sensor type and its operational environment.”

• cites Abdel-Jaber & Glisic, 2016 sourceNot checked — no full text of the source
• cites Bikmukhametov, Stanko and Jäschke, 2018 sourceNot checked — no full text of the source

It may state two separate claims as one.

Calibration methods suffer from a reduction in accuracy, possible errors in the calibration coefficients, and increased frequency of recalibration due to sensor degradation.

“Calibration methods suffer from a reduction in accuracy (Sun & Xiong, 2020;Yu & Wu, 2009), possible errors in the calibration coefficients (Kamei, Nakamura, Yamamoto, Nakamura, Tsuchida, Yamamoto and Wu, 2012), and increased frequency of recalibration (Aldrin, Medina, Allwine, QadeerAhmed, Fisher, Knopp and Lindgren, 2007).”

• cites Sun & Xiong, 2020Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Yu & Wu, 2009Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Kamei, Nakamura, Yamamoto, Nakamura, Tsuchida, Yamamoto and Wu, 2012 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Time-dependent degradation is typically modelled using discrete calibration-based methods or continuous normalization and deterministic models.

“Time-dependent degradation is typically modelled using either discrete approaches, such as calibration-based methods, or continuous approaches, including normalization techniques and deterministic models.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Normalized gain shows a decreasing trend.

“Another form, which is normalized gain, shows a decreasing trend (Jiang et al., 2006).”

• cites Jiang et al., 2006Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Deterministic models often struggle to capture the inherent randomness and variability of real-world degradation processes, and their accuracy is highly dependent on precise knowledge of physical parameters that can be difficult to obtain.

“However, these models often struggle to capture the inherent randomness and variability of real-world degradation processes, and their accuracy is highly dependent on precise knowledge of physical parameters that can be difficult to obtain.”

It may state two separate claims as one.

Strain-gauge acoustic emission transducers and ultrasonic eddy current sensors suffer from sensor bond degradation and can result in inconsistent sensitivity.

“Additionally, strain-gauge acoustic emission transducers and ultrasonic eddy current sensors suffer from sensor bond degradation processes and can result in inconsistent sensitivity (Aldrin et al., 2007).”

• cites Aldrin et al., 2007 sourceNot checked — no full text of the source

It may state two separate claims as one.

Using stochastic processes to model sensor degradation captures the probabilistic nature of degradation over time and allows a dynamic representation reflecting inherent randomness.

“Using stochastic processes to model sensor degradation offers the advantage of capturing the probabilistic nature of degradation over time, allowing for a dynamic representation that reflects the inherent randomness of the phenomena.”

It may state two separate claims as one.

Long-term use of antimony-based sensors leads to formation of an antimony oxide layer, resulting in decreased response sensitivity and longer response times.

“Long-term use of antimony-based sensors leads to the formation of an antimony oxide layer, resulting in decreased response sensitivity and longer response times (Liu et al., 2023).”

• cites Liu et al., 2023Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

As sensors age and accuracy diminishes, the complex nature of degraded measurements makes it difficult to accurately capture system states, directly affecting precision and credibility of measurements.

“As sensors age and their accuracy diminishes, the reliability of the data they produce is compromised as the complex nature of degraded measurements makes it difficult to accurately capture system states (Wen et al., 2023), directly affecting the precision and credibility of measurements.”

• cites Wen et al., 2023 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

A uniform distribution was used to model the sensor degradation process.

“A uniform distribution was also used to model sensor degradation process (Huang & Shen, 2021;Wen et al., 2023).”

• cites Huang & Shen, 2021 sourceNot checked — no full text of the source
• cites Wen et al., 2023 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Deterministic models offer high interpretability because they are based on well-understood physical laws and empirical evidence.

“The primary advantage of this approach lies in its high interpretability, as the models are based on well-understood physical laws and empirical evidence.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Sensitivity degradation is associated with optical sensors.

“It was mentioned to be associated with optical sensors (Kamei et al., 2012).”

• cites Kamei et al., 2012 sourceNot checked — no full text of the source

It may state two separate claims as one.

UWB sensors often encounter outliers and missing data.

“UWB sensors often encounter outliers and missing data (Chughtai et al., 2022).”

• cites Chughtai et al., 2022Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Antimony-based sensors exhibit performance characteristics causing potential drift.

“For instance, antimony-based sensors exhibit such performance characteristics, causing potential drift (Liu et al., 2023).”

• cites Liu et al., 2023Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Filtering technique-based models enhance adaptability and provide robust uncertainty quantification by leveraging inherent structures of filtering algorithms.

“This approach enhances adaptability and provides robust uncertainty quantification by leveraging inherent structures of filtering algorithms.”

It may state two separate claims as one.

When sensor deterioration occurs, estimating the exact gain reduction becomes challenging, complicating maintenance and prognostics.

“When sensor deterioration occurs, estimating the exact gain reduction becomes challenging, complicating maintenance and prognostics processes (He et al., 2008).”

• cites He et al., 2008Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

The normalization approach is fundamentally limited by its reliance on expert knowledge, which can be subjective, and the challenge of defining an accurate non-degraded baseline for comparison.

“However, the approach is fundamentally limited by its reliance on expert knowledge, which can be subjective, and the challenge of defining an accurate non-degraded baseline for comparison.”

It may state two separate claims as one.

Continuous glucose monitoring and energy expenditure sensors may experience missing signals, stuck signals, spikes, and other noise-related issues.

“Continuous glucose monitoring and energy expenditure sensors may experience missing signals, stuck signals, spikes, and other noise-related issues (Feng et al., 2018).”

• cites Feng et al., 2018 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Stochastic degradation is analysed using probability distributions or matrix transformations to capture inherent randomness.

“On the other hand, stochastic degradation is analysed using probability distributions or matrix transformations to capture the inherent randomness of the process.”

It may state two separate claims as one.

Ageing processes increase noise levels, affecting sensor accuracy.

“Ageing processes further increase noise levels, affecting sensor accuracy (van Oosterom et al., 2017).”

• cites van Oosterom et al., 2017Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Gain degradation is a common impact in sensor systems and can be caused by ageing, intermittent failure, and transmission congestion.

“Gain degradation is a common impact in sensor systems and can be caused by factors such as ageing, intermittent failure, and transmission congestion (Liu et al., 2016).”

• cites Liu et al., 2016Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Sensor degradation causes a shift in sensor measurements, altering the gradient of the load-response regression relationship and distorting the distribution of measurement data.

“Sensor degradation also causes a shift in sensor measurements; for instance, it can alter the gradient of the load-response regression relationship, which distorts the distribution of measurement data (Hickinbotham & Austin, 1999).”

• cites Hickinbotham & Austin, 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Measurement biases can be complex and might not follow a Gaussian distribution.

“Measurement biases can be complex, as reported in (Li & Ying, 2017) and might not follow a Gaussian distribution.”

• cites Li & Ying, 2017Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper may give no reason why it holds.

As sensors age and performance deteriorates, the reliability and accuracy of the data they produce are compromised, directly affecting the precision and credibility of measurements.

“As sensors age and their performance deteriorates, the reliability and accuracy of the data they produce are compromised, directly affecting the precision and credibility of measurements derived from them.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

The calibration approach heavily relies on the availability of stable and reliable reference data for frequent recalibration, which can be costly and challenging to obtain.

“however, can heavily relies on the availability of stable and reliable reference data for frequent recalibration, which can be costly challenging to obtain.”

It may state two separate claims as one.

Stochastic process modeling of sensor degradation has the limitation of significant mathematical complexity and the requirement for substantial historical data to select the appropriate process and estimate its parameters.

“The limitation of this approach lies in the significant mathematical complexity and the requirement for substantial historical data to accurately select the appropriate stochastic process and estimate its parameters.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Sensor data records may show long-term and erroneous drift due to a non-uniform degradation process.

“It was mentioned that sensor data records may show long-term and erroneous drift due to a non-uniform degradation process (Sun & Xiong, 2020).”

• cites Sun & Xiong, 2020Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Drift increases are associated with a gradual deterioration process.

“Additionally, drift increases were also mentioned as being associated with a gradual deterioration process (Murthy, 1982).”

• cites Murthy, 1982 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Degradation processes can be broadly categorized into time-dependent degradation, which may be gradual or abrupt, and stochastic degradation, which is probabilistic.

“Degradation processes can be broadly categorized into two main types: time-dependent degradation, which may occur either gradually (continuous) or abruptly (catastrophic), and stochastic degradation, which follows a probabilistic nature.”

It may state two separate claims as one.

Murthy et al. (1982) results showed that the optimal maintenance strategy often involves either full maintenance during a portion of the operational time or no maintenance, with maintenance improving the mean time to failure and reducing the variance of the signal estimation error.

“The results showed that the optimal maintenance strategy often involves either full maintenance during a portion of the operational time or no maintenance, with maintenance improving the mean time to failure and reducing the variance of the signal estimation error.”

• cites Murthy et al. (1982)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Liu et al. (2019) method achieves a near-optimal maintenance policy and a long-run cost rate lower than 36% compared to when sensor degradation is disregarded.

“The method achieves a near-optimal maintenance policy effectively addressing sensor degradation and achieving a long-run cost rate that is lower than 36% compared to when sensor degradation is disregarded.”

• cites Liu et al. (2019)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Liu et al. presented a physics-based approach using a Wiener process to model both system and sensor degradation of wastewater treatment plants and employed a Kalman filter to incorporate sensor degradation parameters into state estimation for RUL prediction.

“Liu et al. (2019) presented a physics-based approach that uses a Wiener process to model both system and sensor degradation of wastewater treatment plants and employed the Kalman filter to incorporate sensor degradation parameters into the system's state estimation for accurate RUL prediction through its cumulative distribution function.”

• cites Liu et al., 2019Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Disregarding sensor degradation significantly increases the risk of inaccurate system state estimation, leading to suboptimal maintenance decisions with ineffective actions and increased costs.

“Disregarding sensor degradation significantly increases the risk of inaccurate system state estimation, leading to suboptimal maintenance decisions with ineffective actions and increased costs.”

• AssumptionRests on an assumption: Sensor degradation increases the probability of undetected failures and unnecessary repairs due to false alarms.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

A principal benefit of empirical models is practicality, as they can often be developed and implemented straightforwardly from available data.

“A principal benefit of this approach is its practicality, as these models can often be developed and implemented straightforwardly from available data.”

Probably not a substantial claim.

Zhang et al. (2023) developed maintenance decision-making in the context of SHM systems, specifically addressing how measurement uncertainty caused by sensor degradation impacts the reliability of maintenance strategies.

“Zhang et al. (2023) developed maintenance decision-making in the context of SHM systems, specifically addressing how measurement uncertainty caused by sensor degradation impacts the reliability of maintenance strategies.”

• cites Zhang et al. (2023) sourceNot checked — no full text of the source

It may state two separate claims as one.

Hachem et al. incorporated sensor degradation parameters into state estimation models for wastewater treatment using Gamma and Wiener stochastic processes with particle filters, and RUL was predicted by simulating the future degradation path; this resulted in a significant reduction in MSE and RMSE when sensor degradation was considered.

“Hachem et al. (2024) incorporated parameters about sensor degradation into state estimation models in the case of wastewater treatment using stochastic processes such as Gamma and Wiener, with state estimation carried out through particle filters, and RUL is predicted by simulating the future degradation path based on the current state estimation-based particle filter, specifically, for each”

• cites Hachem et al., 2024Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Sensor degradation impacts prognostics by reducing the availability of effective monitoring data and making it difficult to distinguish adjacent health states, leading to inaccurate predictions of health conditions and increased risk of failure in critical systems.

“Sensor degradation impacts prognostics by reducing the availability of effective monitoring data and making it difficult to distinguish adjacent health states, leading to inaccurate predictions of health conditions and increased risk of failure in critical systems (Yin et al., 2022).”

• cites Yin et al., 2022 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Predictive Maintenance leverages data through the prognostics process, enabling proactive predictions of future health based on current and historical data.

“Meanwhile, PdM leverages data through the prognostics process, enabling proactive predictions of future health based on current and historical data (Roehrich & Raffaele, 2023;Assagaf, Sukandi, Abdillah, Arifin and Ga, 2023).”

• cites Roehrich & Raffaele, 2023Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Assagaf, Sukandi, Abdillah, Arifin and Ga, 2023 sourceNot checked — no full text of the source

It may state two separate claims as one.

Investigating sensor degradation in prognostics is crucial, as accurate measurements are essential for predicting system health and sensor degradation is a significant source of measurement uncertainties that undermine the accuracy of prognostics.

“Investigating sensor degradation in prognostics is crucial, as accurate measurements are essential for predicting system health and sensor degradation is a significant source of measurement uncertainties that undermine the accuracy of the prognostics process (Mukhopadhyay et al., 2023;Zhang et al., 2018;Liu et al., 2019).”

• cites Mukhopadhyay et al., 2023Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Zhang et al., 2018 sourceNot checked — no full text of the source
• cites Liu et al., 2019Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Yoo et al. presented a hybrid prognostics method integrating physics-based models of sensor degradation accounting for bias, gain, and drift with linear discriminant analysis to estimate system resilience, resulting in a more precise resilience estimation revealing a 6% drop in resilience due to sensor degradation.

“Yoo et al. (2020) presented a hybrid prognostics method that integrates physics-based models of sensor degradation accounting for faults such as bias, gain, and drift, and with linear discriminant analysis to estimate system resilience demonstrated with the electro-hydrostatic actuator (EHA) system, resulting in a more precise resilience estimation that reveals a 6% drop in resilience due to”

• cites Yoo et al., 2020Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Development of integrated hybrid models can be more complex to properly fuse different modelling paradigms.

“Still, the development of such integrated models can be more complex to properly fuse the different modelling paradigms.”

It may state two separate claims as one.

Zhang et al. quantified the influence of sensor measurement errors on RUL prediction in a blast furnace by modelling system degradation with a Wiener process incorporating deteriorating sensor measurement errors quantified using Relative Entropy, and numerical results indicate controlling measurement error within specified permissible ranges significantly improves accuracy of lifetime estimates.

“Zhang et al. (2018) effectively quantified the influence of sensor measurement errors on RUL prediction in a blast furnace system by modelling system degradation with a Wiener process that incorporated deteriorating sensor measurement errors quantified using Relative Entropy, and estimated failure time using an Inverse Gaussian distribution, numerical results indicate that controlling the”

• cites Zhang et al., 2018 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The strength of the statistical/probabilistic approach lies in its versatility, offering a diverse toolkit to model a wide array of degradation effects.

“The strength of this approach lies in its remarkable versatility, offering a diverse toolkit of statistical methods to model a wide array of degradation effects.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Condition-Based Maintenance is a proactive approach that utilizes real-time monitoring of equipment conditions to assess and determine the necessity for maintenance.

“CBM is a proactive approach that utilizes real-time monitoring of equipment conditions to assess and determine the necessity for maintenance (Golmakani, 2022;Kroculick, 2014;de Meyer, Goosen, van Rensburg, du Plessis and van Laar, 2021).”

• cites Golmakani, 2022 sourceNot checked — no full text of the source
• cites Kroculick, 2014 sourceNot checked — no full text of the source
• cites de Meyer, Goosen, van Rensburg, du Plessis and van Laar, 2021Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Experience-based methods rely on historical data and expert knowledge to inform predictions and decision-making.

“Experience-based methods rely on historical data and expert knowledge to inform predictions and decision-making, drawing insights from observations and professional expertise.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.

The complexity of sensor degradation necessitates a multifaceted approach incorporating domain expertise, computational techniques, and real-world validation to ensure robustness and reliability.

“The complexity of sensor degradation necessitates a multifaceted approach, incorporating domain expertise, computational techniques, and real-world validation to ensure robustness and reliability.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Machine learning models often act as black boxes with low interpretability, and their performance is critically dependent on the availability of large, representative training datasets.

“However, these models often act as”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Statistical and probabilistic models are founded on principles of statistics and probability, focusing on deterministic relationships while incorporating uncertainty.

“Statistical and probabilistic models are founded on the principles of statistics and probability, focusing on deterministic relationships while incorporating elements of uncertainty.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.

The strength of the hybrid philosophy is its ability to leverage first-principles models while using data-driven techniques to capture complex interactions or correct for unmodelled effects.

“The strength of hybrid philosophy is its ability to leverage first-principles models while using data-driven techniques to capture complex interactions or correct for unmodelled effects.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. The quote only partly states this claim.

These effects can be systematically analysed and modelled using a variety of techniques, ranging from empirical models to filtering methods, statistical, probabilistic, and machine learning approaches.

“These effects can be systematically analysed and modelled using a variety of techniques, ranging from empirical models based on observations and experimental data to more advanced filtering methods employed within state estimation frameworks.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Li and Ying (2017) proposed method significantly improves diagnostic accuracy, enabling the correct identification and quantification of components across five test cases.

“The proposed method significantly improves diagnostic accuracy, enabling the correct identification and quantification of components across five test cases.”

• cites Li and Ying (2017)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Murthy et al. (1982) developed maintenance for deteriorating sensors guided by degradation indicators, specifically signal intensity, noise intensity, and drift quantification, applicable broadly to industries relying on sensor-based systems.

“Murthy et al. (1982) developed maintenance for deteriorating sensors guided by degradation indicators, specifically, signal intensity, noise intensity, and drift quantification, which can be applied broadly to industries relying on sensor-based systems for real-time monitoring and control.”

• cites Murthy et al. (1982)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

A primary benefit of experience-based methods is high interpretability, as explicit encoding of expert rules allows clear insight into decision-making logic.

“A primary benefit of such methods is their high degree of interpretability, as the explicit encoding of expert rules allows for clear insight into the model's decisionmaking logic.”

It may state two separate claims as one.

Sensor degradation has discernible impacts on measurement quality, affecting both qualitative and quantitative aspects.

“In summary, it is clearly demonstrated that sensor degradation has discernible impacts on measurement quality, affecting both qualitative and quantitative aspects.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Empirical models are constructed using data observed from experiments or real-world measurements, providing data-driven insights into system behaviour.

“Empirical models are constructed using data observed from experiments or real-world measurements, providing datadriven insights into system behaviour.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

A major weakness of empirical models is their limited ability to extrapolate, as predictive accuracy can degrade when applied to varying conditions outside the original observed operating points.

“However, a major weakness of empirical models is their limited ability to extrapolate, as their predictive accuracy can degrade when applied to varying conditions outside of the original observed operating points.”

It may state two separate claims as one.

Yin et al. proposed a BRB model to encode expert knowledge about sensor degradation and failure mechanisms, represented by probabilities of failure at each time step, accommodating uncertainties and inaccuracies caused by sensor degradation, and using fuzzy logic under Membership Functions to predict future health state.

“Yin et al. (2022) proposed a Belief Rule Base (BRB) model to encode expert knowledge about the sensor's degradation and failure mechanisms, represented by probabilities of failure at each time step of data, which accommodate uncertainties and inaccuracies in the data caused by sensor degradation, and subsequently use fuzzy logic under Membership Functions (MF) to predict the sensor's future health”

• cites Yin et al., 2022 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The most significant advantage of the machine learning approach is its ability to automatically learn and represent highly complex and nonlinear degradation patterns directly from data.

“The most significant advantage of this approach is its ability to automatically learn and represent highly complex and nonlinear degradation patterns directly from data.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Hybrid methods integrate physics-based models with data-driven techniques, either combining physical models of prognostics with data-driven insights into sensor degradation or vice versa, enabling more accurate and robust system health predictions.

“Hybrid methods in prognostics and health management represent a powerful approach by integrating physics-based models with data-driven techniques, either by combining physical models of prognostics with data-driven insights into sensor degradation or vice versa, enabling more accurate and robust system health predictions.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The core strength of data-driven prognostics is the ability to build effective prognostic models directly from operational data even without detailed first-principles understanding; however, the quality and quantity of historical data govern predictive accuracy and generalizability.

“The core strength of this approach is its ability to build effective prognostic models directly from operational data, even in the absence of a detailed first-principles understanding of the system's failure physics; however, the quality and quantity of historical data directly govern the model's predictive accuracy and generalizability.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Li and Ying (2017) presented an enhancement to gas turbine reliability by addressing measurement inputs affected by sensor degradation.

“Li and Ying (2017) presented an enhancement to gas turbine reliability by addressing measurement inputs affected by sensor degradation.”

• cites Li and Ying (2017)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Sensor degradation leads to inaccuracies in lifetime estimation due to measurement errors, which can cause suboptimal maintenance decisions, increased costs, and elevated safety risks.

“Sensor degradation leads to inaccuracies in lifetime estimation due to measurement errors, which can cause suboptimal maintenance decisions, increased costs, and elevated safety risks.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Carratù et al. proposed a data-driven method for predicting RUL of MEMS sensors using a neural network trained to forecast RUL based on degradation trends captured by the Health Index.

“Carratù et al. (2024) proposed a data-driven method for predicting the RUL of MEMS sensors, that utilized a neural network trained to forecast RUL based on degradation trends captured by the Health Index.”

• cites Carratù et al., 2024 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

A notable challenge of the statistical/probabilistic approach is the selection and validation of the appropriate statistical model, as an improper choice can lead to non-descriptive results and/or inaccurate predictions.

“A notable challenge, however, lies in the selection and validation of the appropriate statistical model, as an improper choice can lead to non-descriptive results and/or inaccurate predictions.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Prognostics approaches can be classified into data-driven, physics-based, and hybrid methods.

“Prognostics approaches can be classified into three categories: data-driven, physics-based, and hybrid methods (Guo, Li and Li, 2019).”

• cites Guo, Li and Li, 2019Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Machine learning methods leverage training processes to learn from data with complex patterns, enabling models that adapt and generalize to diverse scenarios.

“Machine learning methods leverage training processes to learn from data with complex patterns, enabling the development of models that can adapt and generalize to diverse scenarios.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Considering measurement uncertainty resulting from sensor degradation, the maintenance decision-making process has been adapted.

“Considering measurement uncertainty resulting from sensor degradation, the maintenance decision-making process has been adapted.”

It may state two separate claims as one. The quote only partly states this claim.

He et al. predicted RUL through reliability metrics by modelling degradation of both system and sensor using Wiener processes and dynamically updating through an ABC algorithm, incorporating measurement uncertainty from sensor degradation during sequential ADT; numerical studies on a gas turbine achieved significantly lower asymptotic variance in RUL prediction metrics compared to a benchmark that

“He et al. (2022) predicted the RUL through reliability metrics, including lifetime quantiles and system reliability, which is used to quantify how long a system will function before failure, by modelling the degradation of both the system and the sensor using Wiener processes and dynamically updating these models through an Approximate Bayesian Computation (ABC) algorithm, which incorporates”

• cites He et al., 2022Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

New policies define maintenance actions for both sensors and systems, supported by optimization algorithms.

“New policies define maintenance actions for both sensors and systems, supported by optimization algorithms.”

It may state two separate claims as one.

Distinguishing between sensor degradation and system degradation is critical, especially for applications such as SHM, yet advanced techniques to differentiate between these two phenomena are still lacking.

“Moreover, distinguishing between sensor degradation and system degradation is critical, especially for applications such as SHM, yet advanced techniques to differentiate between these two phenomena are still lacking (Mehdizadeh et al., 2012).”

• cites Mehdizadeh et al., 2012 sourceNot checked — no full text of the source

It may state two separate claims as one.

There is a recognized gap in the integration of physics-based and stochastic models for sensor degradation; combining them could provide more comprehensive and accurate degradation models.

“There is a recognized gap in the integration of physics-based and stochastic models for sensor degradation, while both approaches have been explored separately, combining them could provide more comprehensive and accurate degradation models (Hua et al., 2013).”

• cites Hua et al., 2013 sourceNot checked — no full text of the source

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.

Case studies on how degraded sensor data affect RUL predictions can yield valuable insights for enhancing algorithmic robustness.

“Case studies on how degraded sensor data affect RUL predictions can yield valuable insights for enhancing algorithmic robustness.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

The focus has been primarily on simple systems; however, prognostic frameworks must also account for more complex engineering systems, consisting of multiple sensors and components with interdependent degradation processes, which require further study.

“The focus has been primarily on simple systems; however, prognostic frameworks must also account for more complex engineering systems, consisting of multiple sensors and components with interdependent degradation processes, which require further study (Mukhopadhyay et al., 2023).”

• cites Mukhopadhyay et al., 2023Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Most current models rely on oversimplified assumptions and fail to capture the complexity of real-world dynamic operating conditions.

“Most current models rely on oversimplified assumptions and fail to capture the complexity of real-world dynamic operating conditions.”

Probably not a substantial claim. It may state two separate claims as one.

In brief, the performance of CBM and PdM fundamentally depends on measurement quality, which can be compromised by sensor degradation.

“In brief, the performance of CBM and PdM fundamentally depends on measurement quality, which can be compromised by sensor degradation.”

• AssumptionRests on an assumption: Sensor degradation increases the probability of undetected failures and unnecessary repairs due to false alarms.

Probably not a substantial claim. It may be too vague to check against evidence.

Incorporating measurement uncertainty into prognostics and maintenance frameworks has become a crucial strategy for improving system health predictions and optimizing maintenance decisions.

“Incorporating measurement uncertainty into prognostics and maintenance frameworks has become a crucial strategy for improving system health predictions and optimizing maintenance decisions.”

It may state two separate claims as one.

Most existing studies have largely overlooked how sensor deterioration affects RUL predictions, necessitating more focused research in this area.

“Most existing studies have largely overlooked how sensor deterioration affects RUL predictions, necessitating more focused research in this area (Mukhopadhyay et al., 2023;Hachem et al., 2024).”

• cites Mukhopadhyay et al., 2023Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Hachem et al., 2024Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Unlike traditional methods that rely on predefined degradation models, machine learning techniques can uncover complex, nonlinear relationships and patterns directly from the data that can be highly adaptable to various applications.

“Unlike traditional methods that rely on predefined degradation models, machine learning techniques can uncover complex, nonlinear relationships and patterns directly from the data that can be highly adaptable to various applications.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Data-driven approaches such as machine learning can support modelling degradation patterns by leveraging historical sensor data.

“Datadriven approaches, such as machine learning, can support modelling degradation patterns by leveraging historical sensor data.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.

Another prominent research gap is the need for more comprehensive indicators to capture measurement uncertainty caused by sensor degradation.

“Another prominent research gap is the need for more comprehensive indicators to capture measurement uncertainty caused by sensor degradation.”

It may state two separate claims as one.

Root-cause analysis of the nonlinear physical mechanisms of sensor degradation can provide more accurate modelling, but can be hard and costly due to the multi-factor complexity of degradation phenomena.

“Although root-cause analysis of the nonlinear physical mechanisms of sensor degradation can provide more accurate modelling; however, this can be hard and costly due to the multi-factor complexity of degradation phenomena.”

It may state two separate claims as one.

Dinh et al. (2024) proposed method achieved an optimal maintenance cost rate; compared to the conventional method, it reduced costs by approximately 12.5%.

“The proposed method achieved an optimal maintenance cost rate; compared to the conventional method, which had a higher cost rate, it reduced costs by approximately 12.5%.”

• cites Dinh et al. (2024) sourceNot checked — no full text of the source

It may state two separate claims as one.

By simulating the effects of sensor degradation on system behaviour, digital twins can provide better predictions of RUL and identify critical components requiring attention.

“By simulating the effects of sensor degradation on system behaviour, digital twins can provide better predictions of RUL and identify critical components requiring attention.”

Probably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.

Despite these advancements, significant gaps remain.

“Despite these advancements, significant gaps remain.”

It may state two separate claims as one.

Less attention has been given to fusing degradation data from multiple sensors, and data fusion algorithms can be implemented for multi-sensor degradation systems.

“Furthermore, less attention has been given to fusing degradation data from multiple sensors, multiple sensor data, and data fusion algorithms can be implemented for multi-sensor degradation systems (Arosh et al., 2015).”

• cites Arosh et al., 2015 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Various approaches have been developed to mitigate sensor degradation by incorporating it into maintenance frameworks through additional indicators, state estimation, and filtering.

“Various approaches have been developed to mitigate this issue, incorporating sensor degradation into maintenance frameworks through additional indicators, state estimation, and filtering.”

It may state two separate claims as one.

Dinh et al. (2024) further analysis showed that the maintenance cost rate is sensitive to the sensor replacement threshold: lower thresholds increase costs due to frequent sensor replacements, while higher thresholds reduce replacements but raise costs from system failures or poor maintenance timing caused by degraded sensor accuracy.

“Further analysis showed that the maintenance cost rate is sensitive to the sensor replacement threshold: compared to the optimal value, lower thresholds increase costs due to frequent sensor replacements, while higher thresholds reduce replacements but raise costs from system failures or poor maintenance timing caused by degraded sensor accuracy.”

• cites Dinh et al. (2024) sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Monitoring an industrial robot's sensors over several years can reveal patterns of degradation that inform maintenance schedules and predictive models.

“For instance, monitoring an industrial robot's sensors over several years can reveal patterns of degradation that inform maintenance schedules and predictive models.”

Probably not a substantial claim. It may be too vague to check against evidence.

Integration of sensor degradation modelling into analysis of measurement uncertainty takes a pivotal role in the domain of estimation and filtering.

“Integration of sensor degradation modelling into analysis of measurement uncertainty takes a pivotal role in the domain of estimation and filtering.”

Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

Physics-based models excel at describing degradation through well-defined mechanisms, while stochastic models effectively capture the probabilistic and uncertain nature of degradation processes.

“While physics-based models excel at describing degradation through well-defined mechanisms, stochastic models effectively capture the probabilistic and uncertain nature of degradation processes.”

Probably not a substantial claim. The paper’s conclusions may not depend on it.

Long-term monitoring projects provide a unique opportunity to study degradation as it unfolds in real time, offering insights that static datasets cannot capture.

“These projects provide a unique opportunity to study degradation as they unfold in real time, offering insights that static datasets cannot capture.”

Probably not a substantial claim. The authors may not put it forward as true. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.

Cross-domain case studies comparing sensor degradation and prognostics in different industries can help identify universal principles and best practices that apply across domains.

“Another opportunity lies in the development of cross-domain case studies that compare sensor degradation and prognostics in different industries.This approach can help identify universal principles and best practices that apply across domains.”

It may state two separate claims as one.

Although sensor measurement degradation can be conveniently detected, challenges persist in quantifying the consequent uncertainty, for example, the difficulty in estimating the detected gain degradation.

“Although sensor measurement degradation can be conveniently detected, challenges persist in quantifying the consequent uncertainty, for example, the difficulty in estimating the detected gain degradation (He et al., 2008).”

• cites He et al., 2008Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim. The paper’s conclusions may not depend on it.

The literature witnessed a growing interest related to the relationship between sensor degradation and measurement uncertainty and prognostics and maintenance.

“This can indicate that the literature witnessed a growing interest related to the relationship between sensor degradation and measurement uncertainty and prognostics and maintenance.”

It may state two separate claims as one.

Extensive monitoring data and expert knowledge are often required, but effective sensor health data can often be scarce.

“Extensive monitoring data and expert knowledge are often required, but effective sensor health data can often be scarce (Yin et al., 2022).”

• cites Yin et al., 2022 sourceNot checked — no full text of the source

It may state two separate claims as one.

There is a lack of research on the simultaneous degradation of both the monitored system and the sensor, which may require decoupling their dynamics for effective degradation detection.

“In addition, there is a lack of research on the simultaneous degradation of both the monitored system and the sensor, which may require decoupling their dynamics for effective degradation detection (Jiang et al., 2006).”

• cites Jiang et al., 2006Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

It may state two separate claims as one.

Current studies frequently overlook shocks or assume their effects are constant, failing to capture the stochastic nature of these events.

“For instance, current studies frequently overlook shocks or assume their effects are constant (Hachem et al., 2021), failing to capture the stochastic nature of these events.”

• cites Hachem et al., 2021 sourceNot checked — no full text of the source

Probably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.

Physics-Informed Machine Learning (PIML) can be a standout approach to modelling sensor degradation by integrating physical terms with data-driven adaptability, ensuring predictions remain consistent with real-world physical behaviours, especially when training data may be sparse or noisy, improving model accuracy and explainability.

“Physics-Informed Machine Learning (PIML) can be a standout approach to modelling sensor degradation, by integrating physical terms with the data-driven adaptability power of machine learning, which ensures predictions remain consistent with real-world physical behaviours, especially in scenarios where training data may be sparse or noisy, that eventually improve model accuracy and explainability.”

It may state two separate claims as one.

Currently, measurement errors are often treated as time-independent random variables; however, in reality, sensor errors tend to change over time due to degradation.

“Currently, measurement errors are often treated as time-independent random variables; however, in reality, sensor errors tend to change over time due to degradation (Zhang et al., 2018).”

• cites Zhang et al., 2018 sourceNot checked — no full text of the source