Efficient Long-Term Degradation Profiling in Time Series for Complex Physical Systems

Abstract

The long term operation of physical systems inevitably leads to their wearing out, and may cause degradations in performance or the unexpected failure of the entire system. To reduce the possibility of such unanticipated failures, the system must be monitored for tell-tale symptoms of degradation that are suggestive of imminent failure. In this work, we introduce a novel time series analysis technique that allows the decomposition of the time series into trend and fluctuation components, providing the monitoring software with actionable information about the changes of the system's behavior over time. We analyze the underlying problem and formulate it to a Quadratic Programming (QP) problem that can be solved with existing QP-solvers. However, when the profiling resolution is high, as generally required by real-world applications, such a decomposition becomes intractable to general QP-solvers. To speed up the problem solving, we further transform the problem and present a novel QP formulation, Non-negative QP, for the problem and demonstrate a tractable solution that bypasses the use of slow general QP-solvers. We demonstrate our ideas on both synthetic and real datasets, showing that our method allows us to accurately extract the degradation phenomenon of time series. We further demonstrate the generality of our ideas by applying them beyond classic machine prognostics to problems in identifying the influence of news events on currency exchange rates and stock prices. We fully implement our profiling system and deploy it into several physical systems, such as chemical plants and nuclear power plants, and it greatly helps detect the degradation phenomenon, and diagnose the corresponding components.

Time Series Based Data Explorer and Stream Analysis for Anomaly Prediction

“…Namely, a complete (univariate or multivariate) time series is classified only by the time series prefix composed of the observation values of the first several time stamps. This method satisfies our requirements for early prediction of abnormalities (i.e., detecting abnormal signs as soon as possible) but cannot meet the requirements for the phase abnormal prediction (phase division and continuous monitoring of abnormal evolution paths); other related research results [3,27] also have similar problems. Illustratively, [3] ECG mining is applied to predict the mortality risk of patients but does not continuously monitor and stage patients' disease evolution process; [27] solve the system problem of evaluating the degree of aging, which considers the stages of the aging process to a certain extent, but for each stage, it fails to estimate the time interval from the occurrence of the abnormality.…”

“…Among the existing research results, a few studies [3, 5, 27-30, 42, 43] are related to the problem of abnormal prediction [3]. For instance, ECG mining is used to predict the mortality risk of patients [27], optimization methods are used to evaluate the aging degree of systems, and early time series classification is performed. However, the characteristics of anomaly prediction algorithms have not been discussed deeply.…”

Using spatio-temporal graph neural networks to estimate fleet-wide photovoltaic performance degradation patterns

“…(5) QP-Aging-Detect [15]: The QP-Aging-Detect is a mathematical model using Quadratic Programming (QP) to profile the long-term degradation in timeseries by decomposing them into the aging and fluctuation terms. For fair comparison, we remove the monotonic (strictly increasing or decreasing) constraint imposed on aging component.…”

“…Conventional methods [11][12][13][14] rely on domain-specific knowledge, requiring manual tuning, site specific meteorology data and ad-hoc decisions, which hinder large-scale PLR analysis. Diversified degradation patterns, including non-linear and non-monotonic behavior, cannot be accurately captured by methods assuming monotonic aging [15,16]. Traditional methods such as 6K, PVUSA, XbX, and XbX+UTC [11][12][13][14] are limited to standalone analysis of individual PV systems, require complete meteorology data and cannot leverage spatial coherence for large-scale fleet-level analysis [6].…”

Improving failures prediction by exploring weighted shape‐based time‐series clustering

“…These feature components are easier to understand and can be processed more conveniently. Thus, time‐series decomposition is usually applied as a preprocess technique, and various methods have addressed this problem, such as moving average, seasonal‐trend decomposition procedure based on loess, wavelets, SSA, and ensemble empirical mode decomposition (EEMD) . Among them, in this paper, SSA and EEMD are selected as the decomposition techniques to obtain the failures subseries from original data, considering the following reasons: (1) In contrast to other methods, both SSA and EEMD can produce a sufficient number of subseries to serve the clustering; (2) the 2 methods are both domain‐independent and nonparametric so that they can be applied in various failures time‐series data; (3) they are both well‐known and can be used as the representatives of time‐series decomposition techniques, hence the generalization of our clustering method can be validated.…”