Small Sample Reliability Research Articles

Small Sample Reliability Assessment With Online Time-Series Data Based on a Worm Wasserstein Generative Adversarial Network Learning Method

The scarcity of time-series data constrains the accuracy of online reliability assessment. Data expansion is the most intuitive way to address this problem. However, conventional small-sample reliability evaluation methods either depend on prior knowledge or are inadequate for time series. This article proposes a novel autoaugmentation network, the worm Wasserstein generative adversarial network, which generates synthetic time-series data that carry realistic intrinsic patterns with the original data and expands a small sample without prior knowledge or hypotheses for reliability evaluation. After verifying the augmentation ability and demonstrating the quality of the generated data by manual datasets, the proposed method is demonstrated with an experimental case: the online reliability assessment of lithium battery cells. Compared with conventional methods, the proposed method accomplished a breakthrough in the online reliability assessment for an extremely small sample of time-series data and provided credible results.

Hierarchical Bayesian support vector regression with model parameter calibration for reliability modeling and prediction

Support vector regression (SVR) has been widely used for reliability modeling and prediction in various engineering practices. In order to improve the accuracy and robustness of SVR models, this paper proposes a hierarchical Bayesian support vector regression (HBSVR) model, which can be used for dynamic high-dimensional reliability modeling with small data sets. First, a hierarchical Bayesian model is developed to obtain the posterior inference of HBSVR model parameters. Then, a step-size adaptive accelerated Markov Chain Monte Carlo (SAA-MCMC) method is proposed and combined with Sequential Minimal Optimization (SMO) for model parameter calibration. In addition, a HSBVR dynamic update algorithm is proposed to readjust the current parameters of HBSVR by using the updated dataset and SAA-MCMC to avoid repeated modeling during online prediction. Finally, an active learning algorithm is applied to guide the selection of new samples to improve the model accuracy continuously. Eight numerical examples and two real engineering cases are conducted. The results show that the proposed HBSVR outperforms other methods in terms of prediction accuracy and robustness in small sample reliability prediction tasks.

Reliability Analysis of Engine Shell Strength Under Rocket Sledge Vibration Environment

The strength reliability of a solid rocket motor shell is analyzed. The maximum entropy test method, which is based on information theory, improves the test entropy of a single test sample by reducing the load capacity of the engine shell design under the condition of keeping the load of the engine housing unchanged, so that the total sample size that is required for the test is greatly reduced. The results show that test method of maximum entropy can solve the problem of shell strength reliability assessment when the engine sample size is small. This method can also be applied to small sample reliability assessment in other fields.

1900P - Real-world prospective data of small biopsy samples for tumor PD-L1 expression in non-small cell lung cancer

Background: PD-1 antibody has been effective in patients with PD-L1-positive advanced NSCLC. However, surgically resected specimens or core-needle biopsy samples were used to estimate drug potency in past clinical trials. Methods: The aim is to prospectively investigate small sample reliability for NSCLC to determine the PD-L1 expression status. We prospectively enrolled patients who underwent diagnostic biopsy by any procedures (bronchoscopy, CT/US-guided core-needle) from 2017.3 to 2018.3. Pathologically confirmed NSCLC PD-L1 expression was evaluated in our institution using companion diagnostic PD-L1 IHC. We evaluated: 1) the total number of tumor cells and sample size; 2) compared PD-L1 expression for each procedure using tumor proportion score: TPS (50%≦, 1∼49%, <1%), 3) the concordance rate of PD-L1 expression status by biopsy and surgical materials. Results: 137 cases of PD-L1 expression were evaluated. 112 cases were sampled by bronchoscopy (98 TBBs using BF-1T260/P260F in 23/75 cases, 20 TBNAs), and 25 cases of CT or US-guided core-needle biopsy. The TPS (50%≦ / 1∼49% / <1% / undiagnosed) for total cases was 31/29/37/3% respectively which was similar to the past report. In histological subtype, TPS for adenocarcinoma (n = 82) were 28/27/43/2% and squamous cell carcinoma (n = 49) were 35/35/27/4% respectively. TPS for TBBs using BF-P260F (thin bronchoscopy) were 20/33/43/4% and TBBs using BF-1T260 (normal bronchoscopy) were 48/22/30/0%. TPS for TBNAs were 50/25/25/0% and CT or US-guided sample showed 37/21/37/5%. Four cases were not able to be diagnosed for TPS because we couldn’t obtain enough tumor cells (less than 100 tumor cells) for diagnosis of TPS. The concordance rate of PD-L1 expression status by biopsy and surgical materials was 83.3% for comparison of 24 cases. Conclusions: Utilizing smaller samples to evaluate PD-L1 expression, the frequencies of TPS were comparable to past clinical trials which used larger samples for evaluating TPS. In this study, the larger samples showed higher PD-L1 expression than smaller samples due to their containing more squamous cell carcinoma cases or advanced stage. The concordance rate of PD-L1 expression for surgically resected tissue and biopsy sample was relatively good in our institution. Clinical trial identification: UMIN000027030. Legal entity responsible for the study: Kei Morikawa. Funding: Has not received any funding. Disclosure: All authors have declared no conflicts of interest.

P2.02-072 Reliability of Small Biopsy Samples for Tumor PD-L1 Expression in Non-Small-Cell Lung Cancer

Programmed death 1 immune checkpoint inhibitor antibody has been effective in patients with PD-L1 positive advanced NSCLC. However, surgically resected specimens or core-needle biopsy samples were used to estimate drug potency in past clinical trials. The aim of this study is to prospectively investigate small sample reliability for NSCLC to determine the PD-L1 expression status. We prospectively enrolled patients who underwent diagnostic biopsy by any procedures (under rigid bronchoscopy, flexible bronchoscopy, CT & US-guided core-needle, excisional method) from March 2017 to March 2018 (ongoing study). Pathologically confirmed non-small cell lung cancer (NSCLC), PD-L1 expression was evaluated in our institution using companion diagnostic PD-L1 immunohistochemistry：PD-L1 IHC 22C3 pharmDx (Daco) with autostainer Link 48, detecting a driver mutation in parallel. We evaluated: 1) the total number of tumor cells and sample size; 2) compared PD-L1 expression for each procedure using tumor proportion score: TPS (<1%, 1～49%, 50%≤), 3) the concordance rate of PD-L1 expression status by biopsy and surgical materials; and 4) the efficacy of administration for PD-1 immune checkpoint inhibitor antibody. During the first three months 44 cases of PD-L1 expression were evaluated. 33 cases were sampled by bronchoscopy (6 under rigid scope with BF 1T260, 17 TBBs using 1T260/P260F in 4/13 cases, 10 TBNAs), and 7 cases were CT-guided core-needle biopsy. The TPS (<1%, 1～49%, 50%≤) was 8/6/10 for TBB, 3/4/2 for TBNA, and 4/0/3 for CT-guided. Four cases harboring EGFR mutation showed a lower PD-L1 expression (<10%). Utilizing smaller samples to evaluate PD-L1 expression, and the frequency of TPS were comparable to past clinical trials using larger samples. Smaller samples might be an accurate alternative to assess PD-L1 expression. Current data will be updated at the conference.

Small sample reliability growth modeling using a grey systems model

ABSTRACTWhen performing system-level developmental testing, time and expenses generally warrant a small sample size for failure data. Upon failure discovery, redesigns and/or corrective actions can be implemented to improve system reliability. Current methods for estimating reliability growth, namely the Crow (AMSAA) growth model, stipulate that parameter estimates have a great level of uncertainty when dealing with small sample sizes. For purposes of handling limited failure data, we propose the use of a modified GM(1,1) model to predict system reliability growth parameters and investigate how parameter estimates are affected by systems whose failures follow a poly-Weibull distribution. Monte-Carlo simulation is used to map the response surface of system reliability, and results are used to compare the accuracy of the modified GM(1,1) model to that of the AMSAA growth model. It is shown that with small sample sizes and multiple failure modes, the modified GM(1,1) model is more accurate than the AMSAA model for prediction of growth model parameters.

Milling cutter condition reliability prediction based on state space model

Reliability analysis based on equipment's performance degradation characteristics is one of important research areas for reliability engineering. Many researcher work on multi-sample analysis, but it is limited for single equipment or small sample reliability prediction. Therefore, the method of reliability prediction based on state space model (SSM) is investigated in this research for small sample analysis. Firstly, signals about machine working conditions are collected based on-line monitoring technology. Secondly, wavelet packet energy parameters are determined based on the monitored signals. Frequency band energy is regarded as characteristic parameter. Then, the degradation characteristics of signal to noise ratio is improved by moving average filtering processing. In the end, SSM is established to predict degradation characteristics of probability density distribution, and the degree of reliability is determined. Milling cutter is used to demonstrate the rationality and effectiveness of this method. It can be concluded that this method is effective for milling cutter reliability estimation based on the data analysis. It also contributes to machine condition remaining useful life prediction.

Research on the Evaluation of Small Sample Reliability for CNC Grinding Machine Tools Based on Bayes Theory

Aiming at the limitation by the high cost of test and length cycle process of the CNC grinding machine reliability evaluation, this article adopted small sample technology which needs less data while getting higher evaluation accuracy, by using the theory of Bayes theorem and failure rate as of the random variables, the experiment data of tracking CNC grinding machine for a year is analyzed. Results indicate that results acquired by this mean are consistent with the actual; meanwhile, it shortens test cycle and reduces cost, which is a very effective way to analysis CNC grinding machine. Therefore, it’s very essential that the small sample technology research is applied in the CNC grinding machine evaluation.

Small Sample Reliability Analysis of Automatic Ultrasonic Testing of Austenitic Steel Defect

Reliability analysis of the procedure of ultrasonic technique before inspection has been proved to be essential to ensure the equipment operation safety. The conventional evaluation methods are based on statistical models and a remarkable sample size is needed to get a accurate result. In this work, the Bootstrap model for small sample data is established based on the re-sampling technique. The probability of detection (POD) curves in different conditions are calculated based on the Bootstrap model, and used for quantitative analysis the influence of inspection parameter on stainless steel auto ultrasonic testing result. The calculated result based on bootstrap model is compared with which based on model assisted POD method. The results indicate that Bootstrap model is an efficient tool for the reliability analysis of inspection with small sample data, and insure the inspection result.

Transformed Up-Down Methods in Psychoacoustics

During the past decade a number of variations in the simple up-down procedure have been used in psychoacoustic testing. A broad class of these methods is described with due emphasis on the related problems of parameter estimation and the efficient placing of observations. The advantages of up-down methods are many, including simplicity, high efficiency, robustness, small-sample reliability, and relative freedom from restrictive assumptions. Several applications of these procedures in psychoacoustics are described, including examples where conventional techniques are inapplicable.