Study on the P-S-N Curve of Sucker Rod Based on Three-Parameter Weibull Distribution

During the oil production process, sucker rods are subjected to cyclic alternating load. After a certain number of cycles, a sucker rod can experience fatigue failure. The number of cycles is called fatigue life (N), and the accurate relationship between maximum stress (S) and fatigue life (N) under a certain reliability (P), namely the P-S-N curve, is an important basis for the reliability analysis and fatigue life prediction of sucker rods. The Basquin model, based on log-normal distribution, is widely used for fitting the P-S-N curves of sucker rods. Due to the limitation of this model, it is difficult to extrapolate the conclusion obtained from a finite fatigue region to the high-cycle or ultra-high-cycle fatigue region, which makes it impossible to estimate the fatigue limit of the sucker rod. Compared to the log-normal distribution, Weibull distribution causes the sucker rod to have a minimum safety life, namely the safety life at 100% survival rate, which complies with the fatigue characteristics of the sucker rod and is more in line with the actual situation. In this study, the fatigue data for ultra-high-strength HL and HY grade sucker rods were obtained through experimental fatigue tests. A new fatigue life model was established and the P-S-N curves of two types of ultra-high strength sucker rods were obtained. For HL- and HY-type ultra-high strength sucker rods, the average error between the fitting result and fatigue test value is 1.25% and 4.39%, respectively. Compared to the S-N curve fitting result obtained from the Basquin model commonly used for sucker rods, the new model based on three-parameter Weibull distribution provides better fitting precision and can estimate fatigue limit more accurately, so this model is more suitable for estimating fatigue life and can better guide the design of ultra-high strength sucker rod strings.

The radicle emergence test and storage longevity of cucurbit rootstock seed lots

A study was carried out to test whether a single count of radicle emergence (RE test) would correlate with the storage potential of seed lots of cucurbit (C. maxima × C. moschata) rootstock cultivars. The RE test was performed by counting radicle emergence percentage between 26 and 100 hours of germination. Seed longevity was determined by storage at 75% relative humidity at 35°C over 100 days, with seed survival curves constructed based on normal germination percentages. Ki (estimated initial viability in probits), σ (standard deviation of the normal distribution of seed death in time) and p50 (time for viability to fall to 50%) were determined through probit analysis. Correlation analysis showed that RE counts particularly between 30 and 34 hours were highly (P < 0.01) correlated with normal germination percentage, Ki and p50. It can be concluded that RE tests can be used as an indicator of subsequent seed lot longevity in cucurbit rootstock seed cultivars.

Preliminary analysis on the noise characteristics of MWISP data

Noise is a significant part within a millimeter-wave molecular line datacube. Analyzing the noise improves our understanding of noise characteristics, and further contributes to scientific discoveries. We measure the noise level of a single datacube from MWISP and perform statistical analyses. We identified major factors which increase the noise level of a single datacube, including bad channels, edge effects, baseline distortion and line contamination. Cleaning algorithms are applied to remove or reduce these noise components. As a result, we obtained the cleaned datacube in which noise follows a positively skewed normal distribution. We further analyzed the noise structure distribution of a 3D mosaicked datacube in the range l = 40 ⋅ ° 7 to 43 ⋅ ° 3 and b = − 2 ⋅ ° 3 to 0 ⋅ ° 3 and found that noise in the final mosaicked datacube is mainly characterized by noise fluctuation among the cells.

Probabilities of Misclassification Linked with Inverse Hyperbolic Sine Normal (IHSN) Distribution

Probability of misclassification occurs when there is a choice of criteria that is not favourable for classification. The probabilities of misclassification associated with a family of Johnson’s system, the Inverse Hyperbolic Sine Normal distribution, was developed in this study. The distribution theory and rules, along with the formulation of the system, were generated. It was asserted that the estimation of the parameters of the system could be demystified if one or more variables under consideration are distributed normally.