Short Decision Time Research Articles

A Novel Method of 3D Lyapunov Guidance Vector Field to Avoid Intercepting Satellite Based on Reinforcement Learning

This paper proposes a new 3D Lyapunov guidance vector field(3D-LGV) avoidance strategy based on reinforcement learning for the satellite evasion and interception problem. Combining it with the interfered fluid dynamical system (IFDS) enables the satellite to evade and smoothly enter orbit according to the state of the intercepting satellite in real time. 3D-LGV provides an initial flow field approaching an elliptical orbit, while IFDS provides a perturbed flow field based on the intercepting satellite position. The combined potential field of the initial flow field and the disturbed flow field is the planned velocity direction of the satellite. As a decision-making layer, the proximal policy optimization (PPO) dynamically adjusts the perturbed flow field in the IFDS to increase the avoidance success rate in different scenarios. The experimental results show that, compared with the particle swarm optimization with rolling horizon control algorithm, the algorithm proposed in this paper has a shorter decision time and a higher avoidance success rate. At the same time, Monte Carlo simulation shows that the evasion success rate of the proposed algorithm reaches 98%.

Deep Learning-Based High-Resolution Magnetic Resonance Angiography (MRA) Generation Model for 4D Time-Resolved Angiography with Interleaved Stochastic Trajectories (TWIST) MRA in Fast Stroke Imaging.

The purpose of this study is to improve the qualitative and quantitative image quality of the time-resolved angiography with interleaved stochastic trajectories technique (4D-TWIST-MRA) using deep neural network (DNN)-based MR image reconstruction software. A total of 520 consecutive patients underwent 4D-TWIST-MRA for ischemic stroke or intracranial vessel stenosis evaluation. Four-dimensional DNN-reconstructed MRA (4D-DNR) was generated using commercially available software (SwiftMR v.3.0.0.0, AIRS Medical, Seoul, Republic of Korea). Among those evaluated, 397 (76.3%) patients received concurrent time-of-flight MRA (TOF-MRA) to compare the signal-to-noise ratio (SNR), image quality, noise, sharpness, vascular conspicuity, and degree of venous contamination with a 5-point Likert scale. Two radiologists independently evaluated the detection rate of intracranial aneurysm in TOF-MRA, 4D-TWIST-MRA, and 4D-DNR in separate sessions. The other 123 (23.7%) patients received 4D-TWIST-MRA due to a suspicion of acute ischemic stroke. The confidence level and decision time for large vessel occlusion were evaluated in these patients. In qualitative analysis, 4D-DNR demonstrated better overall image quality, sharpness, vascular conspicuity, and noise reduction compared to 4D-TWIST-MRA. Moreover, 4D-DNR exhibited a higher SNR than 4D-TWIST-MRA. The venous contamination and aneurysm detection rates were not significantly different between the two MRA images. When compared to TOF-MRA, 4D-CE-MRA underestimated the aneurysm size (2.66 ± 0.51 vs. 1.75 ± 0.62, p = 0.029); however, 4D-DNR showed no significant difference in size compared to TOF-MRA (2.66 ± 0.51 vs. 2.10 ± 0.41, p = 0.327). In the diagnosis of large vessel occlusion, 4D-DNR showed a better confidence level and shorter decision time than 4D-TWIST-MRA. DNN reconstruction may improve the qualitative and quantitative image quality of 4D-TWIST-MRA, and also enhance diagnostic performance for intracranial aneurysm and large vessel occlusion.

The enigma of the ‘smoker's paradox’: Results from a single-center registry of patients with STEMI undergoing primary percutaneous coronary intervention

BackgroundSmoker's paradox usually refers to the observation of a favorable outcome of smoking patients in acute myocardial infarction. MethodsFrom April 2006 to December 2018 a population of 2456 patients with ST segment elevation myocardial infarction (STEMI) treated with primary percutaneous coronary intervention (pPCI) were prospectively enrolled in the MATRIX registry. Ischemic time, clinical, demographics, angiographic data, and 1-year follow-up were collected. ResultsAmong 2546 patients admitted with STEMI, 1007 (41 %) were current smokers. Smokers were 10 years younger and had lower crude in-hospital and 1-year mortality (1.5 % vs 6 %, p < 0.0001 and 5 % vs 11 %, p < 0.0001), shorter ischemic time (203 [147–299] vs 220 [154–334] minutes, p = 0.002) and shorter decision time (60 [30–135] vs 70 [36–170] minutes, p = 0.0063). Smoking habit [OR:0.37(95 % CI:0.18–0.75)-p < 0.01], younger age [OR 1.06 (95%CI:1.04–1.09)-p < 0.001] and shorter ischemic time [OR:1.01(95%CI:1.01–1.02)-p < 0.05] were associated to lower in-hospital mortality. Only smoking habit [HR:0.65(95 % CI: 0.44–0.9)–p = 0.03] and younger age [HR:1.08 (95%CI:1.06–1.09)–p < 0.001] were also independently associated to lower all-cause death at 1-year follow-up. After propensity matching, age, cardiogenic shock and TIMI flow <3 were associated with in-hospital mortality, while smoking habit was still associated with reduced mortality. Smoking was also associated with reduced mortality at 1-year follow-up (HR 0.54, 95 % CI [0.37–0.78]; p < 0.001). ConclusionsSmoking patients show better outcome after PCI for STEMI at 1-year follow-up. Although “Smoking paradox” could be explained by younger age of patients, other factors may have a role in the explanation of the phenomenon.

THE ENIGMA OF THE ‘SMOKER‘S PARADOX‘ : RESULTS FROM A SINGLE–CENTER REGISTRY OF PATIENTS WITH STEMI UNDERGOING PRIMARY PERCUTANEOUS CORONARY INTERVENTION

Abstract Background Smoker’s paradox usually refers to the common observation of a favourable outcome of smoker patients in acute myocardial infarction. The interpretation of this phenomenon is still controversial and not fully explained. We aimed at assessing whether smoking could be associated with mortality in a large cohort of prospectively enrolled patients with ST segment elevation myocardial infarction (STEMI) referred to primary percutaneous coronary intervention (PCI). Methods From April 2006 to December 2018 a population of 2456 STEMI patients treated with primary PCI were prospectively enrolled in the MATRIX registry. Ischemic time, clinical, demographics, angiographic data and 1–year follow–up were collected. Results Among 2546 patients admitted with STEMI 1007 (41%) were current smokers. Smokers were 10 years younger and had lower crude in–hospital and 1–year mortality (1.5% vs 6%, p&amp;lt;0.0001 and 5% vs 11%, p&amp;lt;0.0001), shorter ischemic time (203 [147–299] vs 220 [154–334] minutes, p=0.002) and shorter decision time (60 [30–135] vs 70 [36–170] minutes, p=0.0063). After propensity–matching, age, cardiogenic shock and low TIMI score were associated with in–hospital mortality, while smoking habit was still associated with reduced mortality. Smoking was also associated with reduced death rate at 1–year follow–up (HR 0.54, 95% CI [0.37 – 0.78]; p&amp;lt;0.001). Conclusions Our data show a protective link between smoking status and mortality in patients with STEMI, the so called “smoker’s paradox”. This could be explained not only by the shorter ischemic time and younger age of the smoking population but also by a specific pathogenetic effect of smoking underlying the onset of STEMI and response to therapies.

A short decision time for transcatheter embolization can better associate mortality in patients with pelvic fracture: a retrospective study.

Early use of hemostasis strategies, transcatheter arterial embolization (TAE) is critical in cases of pelvic injury because of the risk of hemorrhagic shock and other fatal injuries. We investigated the influence of delays in TAE administration on mortality. Patients admitted to the Advanced Critical Care Center at Gifu University with pelvic injury between January 2008 and December 2019, and who underwent acute TAE, were retrospectively enrolled. The time from when the doctor decided to administer TAE to the start of TAE (needling time) was defined as "decision-TAE time." We included 158 patients, of whom 23 patients died. The median decision-TAE time was 59.5 min. Kaplan-Meier curves for overall survival were compared between patients with decision-TAE time above and below the median cutoff value; survival was significantly better for patients with values below the median cutoff value (p = 0.020). Multivariable Cox proportional hazards regression analysis revealed that the longer the decision-TAE time, the higher the risk of mortality (p = 0.031). TAE duration modified the association between decision-TAE time and overall survival (p = 0.109), as shorter TAE duration (procedure time) was associated with the best survival rate (p for interaction = 0.109). Decision-TAE time may play a key role in establishing resuscitation procedures in patients with pelvic fracture, and efforts to shorten this time should be pursued.

Shared decision-making for renal replacement treatment and illness perception in patients with advanced chronic kidney disease

BackgroundCurrent healthcare trends emphasize the use of shared decision-making (SDM) for renal replacement treatment (RRT) in patients with chronic kidney disease (CKD). This is crucial to understand the relationship between SDM and illness perception of CKD patients. Few studies have focused on SDM and illness perception status of CKD patients and the impact of illness perception on RRT after SDM.MethodsIn this cross-sectional study, we used a questionnaire with purposive sampling from March 2019 to February 2020 at the nephrology outpatient department of a medical center in southern Taiwan. The nephrology medical team in this study used the SHARE five-step model of SDM to communicate with the patients about RRT and Brief Illness Perception Questionnaire (BIPQ) was applied to evaluate illness perception of these patients at the beginning of SDM. According to the SDM decision time, the study participants were classified general and delayed SDM groups. The distribution between SDM groups was estimated using independent two sample t-test, chi-squared test or Fisher’s exact test. The correlation between illness perception and SDM decision time were illustrated and evaluated using Spearman’s correlation test. A p-value less than 0.05 is statistically significant.ResultsA total of 75 patients were enrolled in this study. The average time to make a dialysis decision after initiating SDM was 166.2 ± 178.1 days. 51 patients were classified as general group, and 24 patients were classified as delayed group. The median SDM decision time of delayed group were significantly longer than general group (56 vs. 361 days, P < 0.001). Our findings revealed that delayed group was significantly characterized with not created early surgical assess (delayed vs. general: 66.7% vs. 27.5%, p = 0.001) compared to general group. The average BIPQ score was 54.0 ± 8.1 in our study. We classified the patients into high and low illness perception group according to the median score of BIPQ. The total score of BIPQ in overall participants might increase by the SDM decision time (rho = 0.83, p = 0.830) and the linear regression line also showed consistent trends between BIPQ and SDM decision time in correspond cohorts. However, no statistically significant findings were found.ConclusionsThe patients with advanced chronic kidney disease took an average of five and a half months to make a RRT decision after undergoing SDM. Although there is no statistical significance, the trend of illness perception seems correlated with decision-making time. The stronger the illness perception, the longer the decision-making time. Furthermore, shorter decision times may be associated with earlier establishment of surgical access. We need more research exploring the relationship between illness perception and SDM for RRT in CKD patients.

Midfrontal theta as an index ofconflict strength inapproach-approach vs avoidance-avoidance conflicts.

The seminal theory of motivational conflicts distinguishes between approach-approach (AP-AP) conflicts, in which a decision is made between desirable alternatives, and avoidance-avoidance (AV-AV) conflicts, in which a decision is made between undesirable alternatives. The behavioral differences between AP-AP and AV-AV conflicts are well documented: abundant research showed that AV-AV conflicts are more difficult to resolve than AP-AP ones. However, there is little to no research looking into the neural underpinnings of the differences between the two conflict types. Here, we show that midfrontal theta, an established neural marker of conflict, distinguished between the two conflict types such that midfrontal theta power was higher in AV-AV conflicts than in AP-AP conflicts. We further demonstrate that higher midfrontal theta power was associated with shorter decision times on a single-trial basis, indicating that midfrontal theta played a role in promoting successful controlled behavior. Taken together, our results show that AP-AP and AV-AV conflicts are distinguishable on the neural level. The implications of these results go beyond motivational conflicts, as they establish midfrontal theta as a measure of the continuous degree of conflict in subjective decisions.

A DRL based approach for adaptive scheduling of one-of-a-kind production

Adaptive scheduling is an efficient strategy for one-of-a-kind production (OKP) widespread in heavy industries to address its challenges of the high degree of customization and frequent interference. However, the scheduling procedure meets the problems of complex constraints and short decision time. To address these issues, this study aims to develop a reinforcement learning-based algorithm to solve the adaptive scheduling problem of OKP. The objective is to minimize the makespan. Firstly, the OKP adaptive scheduling problem is modeled as a Markov decision process, and a reinforcement learning algorithm is used to train the scheduling agent offline. Then, the trained agent can make scheduling decisions adaptively according to the production state in a short time. To evaluate the effectiveness of the proposed algorithm, a large number of numerical experiments are performed on the benchmark datasets and a practical engineering case. The results show that the proposed algorithm is competitive in static testing. And it can also achieve the balance between scheduling performance and computation time during adaptive scheduling.

P140 THE “SMOKER’S PARADOX”: A QUESTION OF SHORTER ISCHEMIC TIME AND YOUNGER AGE. RESULTS FROM A LARGE COHORT OF CONSECUTIVE STEMI PATIENTS TREATED WITH PRIMARY PCI

Abstract Background Smoker’s paradox usually refers to the common observation of a favourable outcome of smoker patients in acute myocardial infarction. The interpretation of this phenomenon is still controversial and not fully explained. We aimed at assessing whether smoking could be associated with mortality in a large cohort of prospectively enrolled patients with ST segment elevation myocardial infarction (STEMI) referred to primary percutaneous coronary intervention (PCI). Methods From April 2006 to December 2018 a population of 2456 STEMI patients treated with primary PCI were prospectively enrolled in the MATRIX registry. Ischemic time, clinical, demographics, angiographic data and 1–year follow–up were collected. Results Among 2546 patients admitted with STEMI 1007 (41%) were current smokers. Smokers were 10 years younger and had lower crude in–hospital and 1–year mortality (1.5% vs 6%, p&amp;lt;0.0001 and 5% vs 11%, p&amp;lt;0.0001), shorter ischemic time (203 [147–299] vs 220 [154–334] minutes, p=0.002) and shorter decision time (60 [30–135] vs 70 [36–170] minutes, p=0.0063). After adjustment for different risk factors using Cox regression model, age [adjHR: 1.11 (95%CI: 1.03 – 1.18) – p&amp;lt;0.01] and ischemic time [adjHR: 1.01 (95%CI: 1.01 – 1.02) – p&amp;lt;0.05] appeared to be the confounder factor of the relationship between smoking and in hospital mortality. Conclusions In our large prospectively–enrolled population the so–called “smoker’s paradox” is explained by the shorter ischemic time and younger age of the smoking population. Our data shows that smokers were admitted to the hospital for STEMI about 10 years before non–smokers, with a premature coronary atherothrombosis that could be the price to pay for smoking. Smokers had non–significant different risk of in–hospital mortality compared to non–smoking.

Metamodel-based metaheuristics in optimal responsive adaptation and recovery of traffic networks

ABSTRACT Different emerging threats highlighted the relevance of recovery and adaptation modelling in the functioning of societal systems. However, as modelling of systems becomes more complex, its effort increases challenging the practicality of the engineering analyses required for efficient recovery and adaptation. In the present work, metamodels are researched as a tool to enable these analyses in traffic networks. One of the main advantages of metamodeling is their synergy with the short decision times required in recovery and adaptation. A sequential global metamodeling technique is proposed and applied to three macroscopic day-to-day user-equilibrium models. Two reference contexts of application are researched: optimal recovery to a perturbation (with response times reduced by 98% with loss of accuracy lower than 1%) and adaptation under uncertainty with perturbation-dependent optimality. Results show that metamodeling-based metaheuristics enable fast resource-intensive engineering analyses of traffic recovery and adaptation, which may change the paradigm of decision-making in this field

The Limits of Defaults: The Influence of Decision Time on Default Effects

The stability of default effects to contextual features is critical to their use in policy. In this paper, decision time was investigated as a contextual factor that may pose limits on the efficacy of defaults. Consistent with the hypothesis that time constraints may increase reliance on contextual cues, four experiments, including a preregistered one of a nationally representative sample, and a meta-analysis that included four additional pilot experiments, indicated that short decision times increased the advantage of action defaults (i.e., the default option automatically endorsed the desired behavior) and that the default advantage was trivial or nonexistent when decision times were longer. These effects replicated for naturalistic as well as externally induced decision times and were present even when participants were unaware that time was limited. This research has critical implications for psychological science and allied disciplines concerned with policy in the domains of public health, finance and economics, marketing, and environmental sciences.

Bi-objective optimization for multi-task offloading in latency and radio resources constrained mobile edge computing networks

The Mobile Edge Computing (MEC) environment provides leading-edge services to smart mobile devices (SMDs). Besides, computation offloading is a promising service in 5G: it reduces battery drain and applications’ execution time. In this context, we consider a general system consisting of a multi-cell communication network where each base station (BS) is equipped with a MEC server to provide computation offloading services to nearby mobile users. In addition, each SMD handles multiple independent offloadable heavy tasks that are latency-sensitive. The purpose of this article is to jointly optimize tasks’ offloading decisions as well as the allocation of critical radio resources while minimizing the overall energy consumption. Therefore, we have formulated a bi-objective optimization problem that is NP-hard. Because of the short decision time constraint, the optimal solution implementation is infeasible. Accordingly, with the use of the weighted aggregation approach, we propose Intelligent Truncation based Hybrid Local Search (ITHLS) solution. In critical radio resources situations, our solution jointly minimizes the number of penalized SMDs and the overall consumed energy. Finally, simulation experiments were realized to study the ITHLS solution performance compared to some effective state of the art solutions, and the simulation results in terms of decision-making time, energy and number of truncated SMDs are very promising.

Time and resource constrained offloading with multi-task in a mobile edge computing node

In recent years, the importance of the mobile edge computing (MEC) paradigm along with the 5G, the Internet of Things (IoT) and virtualization of network functions is well noticed. Besides, the implementation of computation-intensive applications at the mobile device level is limited by battery capacity, processing capabalities and execution time. To increase the batteries life and improve the quality of experience for computationally intensive and latency-sensitive applications, offloading some parts of these applications to the MEC is proposed. This paper presents a solution for a hard decision problem that jointly optimizes the processing time and computing resources in a mobile edge-computing node. Hence, we consider a mobile device with an offloadable list of heavy tasks and we jointly optimize the offloading decisions and the allocation of IT resources to reduce the latency of tasks’ processing. Thus, we developped a heuristic solution based on the simulated annealing algorithm, which can improve the offloading rate and reduce the total task latency while meeting short decision time. We performed a series of experiments to show its efficiency. Finally, the obtained results in terms of full-time treatrement are very encouraging. In addition, our solution makes offloading decisions within acceptable and achievable deadlines.

Frequency Disturbance Event Detection Based on Synchrophasors and Deep Learning

Power system frequency disturbances are caused by various generation and transmission events including generator trips, load disconnections, line trips, etc. Accurate detections of the events are crucial to bulk power system situation awareness and event investigation. This paper utilizes the recent advances of deep learning to build a convolutional neural network model to detect events in an accurate yet straightforward manner. In this paper, the rate of change of frequency and the relative angle shift are converted to images as the inputs of the proposed model. Finally, this paper uses two convolutional neural networks and classifier fusion to achieve the detection result. Compared with the conventional event detection algorithm and the frequency only deep learning model, the proposed model improves the detection accuracy by over 48%. As a promising tool for bulk power system situation awareness, the proposed model requires a short decision time, which is suitable for practical scenarios.

Psychological constraints on aggressive predation in economic contests.

When humans compete, they invest energy and effort to injure others and to protect against injury and exploitation. The psychology behind exploiting others and protecting against exploitation is still poorly understood and is addressed here in an incentivized economic contest game in which individuals invested in predatory attack and prey defense. Consistent with standard economic theory on production and predation, we find that individuals compete less intensely when they attack rather than defend and that attacks disproportionally often fail. We find, furthermore, 2 psychological mechanisms that restrain attack more than defense. First, individuals with stronger concern for others' welfare (Experiment 1a) and with stronger empathy (Experiment 1b) less frequently attack and when they attack, they do so less forcefully. Second, shorter decision times (Experiment 2a and Experiment 2b), along with cognitive taxation (Experiment 2b) associate with more forceful, but not with more frequent attack. Finally, investments in defense were neither predicted by other-concern and empathy, nor by decision time and cognitive taxation. Thus, individuals with stronger prosocial preferences and more deliberated decisions spent less energy on injuring others, and less often defeated their antagonists but ended-up personally wealthier. The waste of conflict can be reduced by strengthening prosocial preferences and cognitive resources available for deliberate decision-making. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Association between workers’ compensation claim processing times and work disability duration: Analysis of population level claims data

Delays in workers’ compensation claim processing (CP) times have been associated with reduced recovery and delayed return-to-work. This study aimed to (1) determine the injury, worker, and workplace factors associated with CP delays and (2) investigate whether CP delays are associated with longer disability duration after adjusting for these factors. Retrospective cohort analysis of Australian workers’ compensation claims was conducted from 1st July 2009 to 30th June 2016 for objective (1) and to 30th June 2014 for objective (2). CP times were derived by calculating differences in days between: injury and lodgement dates (lodgement); lodgement and decision dates (decision) and; injury and decision dates (total). All CP times were shorter for younger workers and those with fractures or traumatic injury, and longer for those with neurological or mental health conditions, and other diseases. Claims from self-insured employers had shorter decision times. With increasing lodgement, decision and total time there was significantly higher hazard of longer disability duration. Findings suggest the need for more efficient claims management to ensure fewer barriers to claim lodgement or approval. This in turn should reduce disability duration and ensure improved return-to-work outcomes.

Energy‐efficient and delay‐aware multitask offloading for mobile edge computing networks

AbstractMobile edge computing (MEC) is a recent technology that intends to free mobile devices from computationally intensive workloads by offloading them to a nearby resource‐rich edge architecture. It helps to reduce network traffic bottlenecks and offers new opportunities regarding data and processing privacy. Moreover, MEC‐based applications can achieve lower latency level compared to cloud‐based ones. However, in a multitask multidevice context, the decision of the part to offload becomes critical. Actually, it must consider the available communication resources, the resulting delays that have to be met during the offloading process, and particularly, both local and remote energy consumption. In this paper, we consider a multitask multidevice scenario where smart mobile devices retain a list of heavy offloadable tasks that are delay constrained. Therefore, we formulated the corresponding optimization problem, and we derive an equivalent multiple‐choice knapsack problem formulation. Because of the short decision time constraint and the NP‐hardness of the obtained problem, the optimal solution implementation is infeasible. Hence, we propose a solution that provides, in pseudopolynomial time, the optimal or near‐optimal solutions depending on the problem's settings. In order to evaluate our solution, we carried out a set of simulation experiments to evaluate and compare the performances of the different components of this solution. Finally, the obtained results in terms of execution's time as well as energy consumption are satisfactory and very encouraging.

Sample position inside abrasive fluidized bed to obtain morphology uniformity

Abrasive fluidized bed (AFB) machining of aluminum samples was studied with a specific focus on surface roughness formation, according to the sample position inside AFB operating in a minimum fluidization regime. The experimentation was aimed at defining optimum process conditions in terms of morphology uniformity and developing a mathematical model which relates the sample position to the final roughness. Measurements performed by profilometry have shown that a uniform average roughness lower than 1 μm can be obtained, starting from 3.5 μm, when the process parameters and handling are optimized. The analysis of variance (ANOVA) was performed to support the experimental and modeling strategy. This study led to the definition of a rapid and efficient procedure for finding the optimal positioning of samples in terms of finishing and surface uniformity. The proposed procedure does not require complex systems for component movement, and it is based on the study of just two reference positions of samples. The advantage of the procedure developed and validated in this study consists in its ease of implementation and its applicability in industrial strategies, where short decision times are essential.

Last-minute hotel-booking behavior: The impact of time on decision-making

Abstract Last-minute bookings are becoming more prevalent in the hotel industry. Consequently, this study aims to identify distinctive booking attributes associated with last-minute booking, and to examine how temporal distance between the time of booking and the date of arrival affects customers' price perception and booking behavior. An online survey of 357 individuals was conducted, and a series of one-way ANCOVA (analysis of covariance) were performed to test the hypotheses. The results indicated that customers who have the shortest temporal distance (0–2 days before the date of stay) exhibited a relatively smaller evoked set size, lower involvement, and shorter decision time than any other temporal distance group. Moreover, the group with the shortest temporal distance (0–2 days before the date of stay) diverged noticeably from the rest of the temporal distance groups on booking attributes. This research is one of the pioneering studies specifically focusing on last-minute booking behavior. The findings contribute to the understanding of the impact of time on customers' hotel-booking decisions. The current study offers new insights into the set of salient and non-static booking attributes that drive hotel choices over time as the date of stay draws near. This study offers pointed suggestions on how to drive online hotel bookings more effectively depending on customers' time of booking.

A new ensemble-based classifier for IGBT open-circuit fault diagnosis in three-phase PWM converter

Three-phase pulse width modulation converters using insulated gate bipolar transistors (IGBTs) have been widely used in industrial application. However, faults in IGBTs can severely affect the operation and safety of the power electronics equipment and loads. For ensuring system reliability, it is necessary to accurately detect IGBT faults accurately as soon as their occurrences. This paper proposes a diagnosis method based on data-driven theory. A novel randomized learning technology, namely extreme learning machine (ELM) is adopted into historical data learning. Ensemble classifier structure is used to improve diagnostic accuracy. Finally, time window is defined to illustrate the relevance between diagnostic accuracy and data sampling time. By this mean, an appropriate time window is achieved to guarantee a high accuracy with relatively short decision time. Compared to other traditional methods, ELM has a better classification performance. Simulation tests validate the proposed ELM ensemble diagnostic performance.