Using four hazardous dispersion models, the ability to predict Lower Flammability Distance (LFD), the distance from a release through which the released gas is present in concentration below flammable range, has been evaluated using a set of large‐scale propane release field experiment data. The four models evaluated were the DEGADIS model, the SLAB model, the OME Simple Gas model and the OME Heavy Gas model (1983). The observed LFDs were compared with the predicted LFDs for three atmospheric conditions. The complete data set was broken into two parts: (1) the cyclone type releases (representing instantaneous release situations) and (2) the nozzle type releases (representing horizontal jet releases). A procedure based on the USEPA guidelines on air quality models was followed to evaluate the models. The complex models (DEGADIS and SLAB models) appear unreasonably weak in predicting LFD's compared with the relatively simple Gaussian model used in the OME Simple Gas model for instantaneous releases. The OME simple gas model is significantly better than the DEGADIS and SLAB models for cyclone type (instantaneous) releases under unstable and neutral atmospheric conditions; and the DEGASIS model is significantly better than the SLAB model for nozzle type (horizontal jet) releases under stable atmospheric conditions. Since the size of the data set is not large enough (n = 20 for cyclone type release and n = 18 for nozzle type release), it is difficult to show 95% significant differences between models.

No abstract

No abstract

Although a superhydrophobic surface could realize rapid rebounding (i.e., short contact time) of an orthogonal impacting droplet, the rebounding along the original impacting route may limit its engineering application; in contrast, the directional transportation seems to be more promising. Here, we achieve directional transportation of a droplet impacting on a wettability-controlled surface. When the droplet eccentrically impacts on the boundary between the superhydrophobic part and the hydrophilic part, it undergoes spreading, retracting, departure, throwing and breaking up stages, and finally bounces off directionally. The directional transportation distance could even reach more than ten times of the droplet size, considered the adhesion length (i.e., covering length on the hydrophilic part by the droplet at the maximum spreading) is optimized. However, there is a critical adhesion length, above which the directional transportation does not occur. To be more generalized, the adhesion length is de-dimensionalized by the maximum spreading radius, and the results show that as the dimensionless adhesion length increases, the transportation distance first increases and then decreases to zero. Under the present impacting conditions, the optimal dimensionless adhesion length corresponding to the maximum transportation distance is near 0.4, and the critical dimensionless adhesion length is about 0.7. These results provide fundamental understanding of droplet directional transportation, and could be useful for related engineering applications.

No abstract

To eliminate the limitations of the peak‐over‐threshold (POT) extrapolation method, the paper proposes an improved POT extrapolation method to reconstruct the load sequence and frequency of the part. A threshold model is established based on a generalized Pareto distribution (GPD). Assuming a road surface's irregularity obeys a normal distribution, the load between the upper and lower thresholds is reconstructed to obtain more dynamic road load data. Rain‐flow counting method is used to calculate the peak‐to‐valley value and frequency of the load signal, and the characteristics of the load mean and amplitude before and after the improvement of the POT extrapolation method were discussed. The effectiveness and conservativeness of the enhanced method are verified by predicting the fatigue life of a swing support rod.

No abstract

No abstract